The impact of the progranulin-sortilin axis on breast cancer stem cell activity and patient outcome Karoline Berger Department of Laboratory Medicine Institute of Biomedicine Sahlgrenska Academy, University of Gothenburg Gothenburg 2021 Cover illustration “Lung metastasis in mice” by Karoline Berger. The impact of the progranulin-sortilin axis on breast cancer stem cell activity and patient outcome © Karoline Berger 2021 Karoline.berger@gu.se ISBN 978-91-8009-390-3 (PRINT) ISBN 978-91-8009-391-0 (PDF) Printed in Borås, Sweden 2021 Printed by Stema Specialtryck AB To my family “If we knew what it was we were doing, it would not be called research, would it?” - Albert Einstein Trycksak 3041 0234 SV AN ENMÄRKET Trycksak 3041 0234 SV AN ENMÄRKET Cover illustration “Lung metastasis in mice” by Karoline Berger. The impact of the progranulin-sortilin axis on breast cancer stem cell activity and patient outcome © Karoline Berger 2021 Karoline.berger@gu.se ISBN 978-91-8009-390-3 (PRINT) ISBN 978-91-8009-391-0 (PDF) Printed in Borås, Sweden 2021 Printed by Stema Specialtryck AB To my family “If we knew what it was we were doing, it would not be called research, would it?” - Albert Einstein I A bstract Breast cancer is the most common cancer in women worldwide. Still today, despite current breast cancer therapies , many patients experience treatment resistance and relapse, which are believed to be due to failure in targeting treatment- resistant cancer stem cells. Cytokines and growth factors secreted by various cell types present in the tumor microenvironment have the potential to affect this challenging cell subpopulation . This thesis focuses on a complex cellular communication system based on hypoxia - induced secretion, where we identified the growth factor progranulin as one of the key mediators driving cancer ste m cell propagation. In this thesis , we demonstrate that progranulin mediates cancer stem cell propagation in various breast cancer cell lines. By chemically degrading and modulating sortilin expression, or using a s mall sortilin binding molecule, AF38 469 , we could reduce the progranulin - induced cancer stem cell propagating effect in vitro, suggesting that the progranulin - induced cancer progression is dependent on sortilin. Importantly , using breast cancer xe nograft models, we were able to confirm th e progranulin - mediated cancer stem cell propagating effect in vivo. Strikingly , p rogranulin induced a significant increase in lung metastasis, which could be reduced by oral administration of AF3 8469. Moreover, when investigating the mechanism s behind sortilin - driven progranulin- induced cancer stem cell activation, we found that progranulin induced secretion of the inflammatory cytokine interleukin - 6 and could demonstrate a crosstalk between progranulin and interleukin - 6 protein expression . Similar to progranulin, interleukin - 6 affected breast cancer stem cell expansion via sortilin, altogether suggesting that sortilin is a highly relevant biological target in breast cancer. Furthermore , in a tissue microarray of breast cancer patients, high co - expression of progranulin and sortilin defined a novel and highly malignant subgroup of breast cancer, suggesting that these proteins can be used as prognostic biomarkers. Combined, results presented in this thesis propose that targeting the progranulin- sortilin communication axis represents a potential novel breast cancer therapeutic approach, inhibiting tumor progressio n driven by secretion and microenvironmental influences. Accordingly, we are currently in the proses of developing sortilin- targeting drugs for the treatment of breast cancers with high expression of progranulin and sortilin . Keyw ord s: Breast cancer, biomarker, cancer stem cells, microenvironment, progranulin, sortilin, interleukin - 6, targeted therapy, prog nostic ISBN 97 8- 91- 8 009- 3 90- 3 ( PRI N T) ISBN 97 8 - 91- 8 009- 3 91- 0 ( PD F) I A bstract Breast cancer is the most common cancer in women worldwide. Still today, despite current breast cancer therapies , many patients experience treatment resistance and relapse, which are believed to be due to failure in targeting treatment- resistant cancer stem cells. Cytokines and growth factors secreted by various cell types present in the tumor microenvironment have the potential to affect this challenging cell subpopulation . This thesis focuses on a complex cellular communication system based on hypoxia - induced secretion, where we identified the growth factor progranulin as one of the key mediators driving cancer ste m cell propagation. In this thesis , we demonstrate that progranulin mediates cancer stem cell propagation in various breast cancer cell lines. By chemically degrading and modulating sortilin expression, or using a s mall sortilin binding molecule, AF38 469 , we could reduce the progranulin - induced cancer stem cell propagating effect in vitro, suggesting that the progranulin - induced cancer progression is dependent on sortilin. Importantly , using breast cancer xe nograft models, we were able to confirm th e progranulin - mediated cancer stem cell propagating effect in vivo. Strikingly , p rogranulin induced a significant increase in lung metastasis, which could be reduced by oral administration of AF3 8469. Moreover, when investigating the mechanism s behind sortilin - driven progranulin- induced cancer stem cell activation, we found that progranulin induced secretion of the inflammatory cytokine interleukin - 6 and could demonstrate a crosstalk between progranulin and interleukin - 6 protein expression . Similar to progranulin, interleukin - 6 affected breast cancer stem cell expansion via sortilin, altogether suggesting that sortilin is a highly relevant biological target in breast cancer. Furthermore , in a tissue microarray of breast cancer patients, high co - expression of progranulin and sortilin defined a novel and highly malignant subgroup of breast cancer, suggesting that these proteins can be used as prognostic biomarkers. Combined, results presented in this thesis propose that targeting the progranulin- sortilin communication axis represents a potential novel breast cancer therapeutic approach, inhibiting tumor progressio n driven by secretion and microenvironmental influences. Accordingly, we are currently in the proses of developing sortilin- targeting drugs for the treatment of breast cancers with high expression of progranulin and sortilin . Keyw ord s: Breast cancer, biomarker, cancer stem cells, microenvironment, progranulin, sortilin, interleukin - 6, targeted therapy, prog nostic ISBN 97 8- 91- 8 009- 3 90- 3 ( PRI N T) ISBN 97 8 - 91- 8 009- 3 91- 0 ( PD F) II S ammanfattning på svenska Bröstcancer är den vanligaste cancerformen bland kvinnor i Sverige , med över 8000 nydiagnostiserade fall varje år. Trots den ökade överlevnaden de senaste åren får många patienter återfall , och fortfarande avlider ungefär 1300 kvinnor i Sverige varje år. Bröstcancer är en mycket heterogen och komplex sjukdom , och kännetecknas av stora variationer i tumörerna mellan olika patienter. Tumörerna påverkas i stor grad av om givningen kring tumören, den så - kallade mikromiljön, bestående av proteinfiber - nätverk , immunceller, cytokiner och andra celler som kan påverka hur tumören s varar på olika behandlingar. Här finns också en liten population av celler som kallas cancerstamcel ler som är mer aggressiva och behandlingsr esistenta än andra cancerceller och kan i högre grad ge upphov till spridning och återfall. Det är därför viktigt att rikta behandlingen mot dessa celler. Syftet med detta arbete är att bättre förstå hur cancercell er påverkas av omgivningen , och vi har valt att fokusera på ett antal proteiner som just påverkar mängden cancerstamceller i bröstcancer. I en screen av utsöndrade proteiner från cancerceller där vi undersökte hur hypoxi (lågt syretryck) påverkar olika tum öregenskaper, identifierade vi progranulin som ett viktigt protein som både utsöndrades av cancerceller under hypoxi, men som också utsöndras i höga nivåer i hormonreceptornegativa bröstcancerceller. Vi har vidare studerat hur progranulin påverkar receptor n sortilin och vad detta får för effekter på cancercellerna. Våra resultat visar att progranulin , via sortilin, ökar mängden cancerstamceller i bröstcancer. Dessutom har vi detaljstuderat hur progranulin och sortilin uttrycks på proteinnivå i bröstcancer o ch funnit att ungefär 20% av alla premenopausala bröstcancerfall uttrycker höga nivåer av både progranulin och sortilin , vilket också visade sig vara starkt kopplat till dålig prognos för patienterna. Dessa data tyder på att signalering via sortilin kan dr iva elakartade egenskaper . Genom att blockera eller bryta ner sortilin i cancerceller kan vi minska progran ulins effekt på cancerstamcellerna. Vi har även sett att progranulin - behandling ger en ökad mängd lungmetastaser i möss, något som också kan hämmas vid användning av en liten molekyl som binder till sortilin . Vi har ytterligare definierat hur olika klyvda peptid- delar av progranulin påverkar cancerstamceller via sortilin, såväl som undersökt samspelet mellan progranulin, sortilin och den inflammatoriska cytokinen IL- 6 , som även den visade sig kunna bidra till att påverka cancerstamcells - egenskaper i brös tcancer via receptorn sortilin. Sammanfattningsvis har vi genom detta arbete identifierat ett nytt sätt att angripa elakartade egenskaper i cancer. Blockerin g av sortilin kan vara ett effektivt sätt att inhibera tumörprogression som drivs av sekretion i tumörens mikromiljö. II I L ist of papers This thesis is based on the following studies, referred to in the text by their Roman numerals: I. R host S, Hughes É, Harrison H, Rafnsdóttir S, Jacobsson H, Gregersson P, Magnusson M, Fitzpatrick P, Andersson D, Berger K, Ståhlberg A and Landberg G. S ortilin inh ibition limits secretion- ind uced p rogranulin- d ep end ent breast cancer p rogression and cancer stem cell ex p ansion. Breast Cancer Res. 201 8 Nov 20;20( 1) :1 37. II. Berger K, Rhost S, Rafnsdóttir S, Hughes É, Magnusson Y, Ekholm M, Stål O, Rydén L and Landberg G . T umor co- ex p ression of p rogranulin and sortilin as a p rognostic biomark er in breast cancer. BMC Cancer. 2021 Feb 22;21( 1) :1 85 . III. Berger K* , P ersson E *, Gregersson P, Jonasson E, Ståhlberg A, Landberg G and Rhost S. I nterleuk in- 6 ind uces stem cell p rop agation th rough liaison w ith th e sortilin- p rogranulin ax is in breast cancer. *Authors contributed equally. ( Manuscript ) IV. Berger K* , Rhost S*, Hughes É, Gregersson P and Landberg G. G ranulin p ep tid e d omains ind uce breast cancer stem cell p rop agation vi sortilin. *Authors contributed equally. (Manuscript) V. Berger K, Pauwels E, Parkin son G, Landberg G, Le T, Demillo V.G, Lumangtad L.A , Jones D.E, Islam M.A, Olsen R , Kapri T, Intasiri A , Vermeire K, Rhost S and Bell T.W . R ed uction of P rogranulin- I nd uced Breast Cancer S tem Cell P rop agation by S ortilin- T argeting Cyclotriaz ad isulf onamid e ( CA D A ) Comp ound s. ( Manuscript under revision ) II S ammanfattning på svenska Bröstcancer är den vanligaste cancerformen bland kvinnor i Sverige , med över 8000 nydiagnostiserade fall varje år. Trots den ökade överlevnaden de senaste åren får många patienter återfall , och fortfarande avlider ungefär 1300 kvinnor i Sverige varje år. Bröstcancer är en mycket heterogen och komplex sjukdom , och kännetecknas av stora variationer i tumörerna mellan olika patienter. Tumörerna påverkas i stor grad av om givningen kring tumören, den så - kallade mikromiljön, bestående av proteinfiber - nätverk , immunceller, cytokiner och andra celler som kan påverka hur tumören s varar på olika behandlingar. Här finns också en liten population av celler som kallas cancerstamcel ler som är mer aggressiva och behandlingsr esistenta än andra cancerceller och kan i högre grad ge upphov till spridning och återfall. Det är därför viktigt att rikta behandlingen mot dessa celler. Syftet med detta arbete är att bättre förstå hur cancercell er påverkas av omgivningen , och vi har valt att fokusera på ett antal proteiner som just påverkar mängden cancerstamceller i bröstcancer. I en screen av utsöndrade proteiner från cancerceller där vi undersökte hur hypoxi (lågt syretryck) påverkar olika tum öregenskaper, identifierade vi progranulin som ett viktigt protein som både utsöndrades av cancerceller under hypoxi, men som också utsöndras i höga nivåer i hormonreceptornegativa bröstcancerceller. Vi har vidare studerat hur progranulin påverkar receptor n sortilin och vad detta får för effekter på cancercellerna. Våra resultat visar att progranulin , via sortilin, ökar mängden cancerstamceller i bröstcancer. Dessutom har vi detaljstuderat hur progranulin och sortilin uttrycks på proteinnivå i bröstcancer o ch funnit att ungefär 20% av alla premenopausala bröstcancerfall uttrycker höga nivåer av både progranulin och sortilin , vilket också visade sig vara starkt kopplat till dålig prognos för patienterna. Dessa data tyder på att signalering via sortilin kan dr iva elakartade egenskaper . Genom att blockera eller bryta ner sortilin i cancerceller kan vi minska progran ulins effekt på cancerstamcellerna. Vi har även sett att progranulin - behandling ger en ökad mängd lungmetastaser i möss, något som också kan hämmas vid användning av en liten molekyl som binder till sortilin . Vi har ytterligare definierat hur olika klyvda peptid- delar av progranulin påverkar cancerstamceller via sortilin, såväl som undersökt samspelet mellan progranulin, sortilin och den inflammatoriska cytokinen IL- 6 , som även den visade sig kunna bidra till att påverka cancerstamcells - egenskaper i brös tcancer via receptorn sortilin. Sammanfattningsvis har vi genom detta arbete identifierat ett nytt sätt att angripa elakartade egenskaper i cancer. Blockerin g av sortilin kan vara ett effektivt sätt att inhibera tumörprogression som drivs av sekretion i tumörens mikromiljö. II I L ist of papers This thesis is based on the following studies, referred to in the text by their Roman numerals: I. R host S, Hughes É, Harrison H, Rafnsdóttir S, Jacobsson H, Gregersson P, Magnusson M, Fitzpatrick P, Andersson D, Berger K, Ståhlberg A and Landberg G. S ortilin inh ibition limits secretion- ind uced p rogranulin- d ep end ent breast cancer p rogression and cancer stem cell ex p ansion. Breast Cancer Res. 201 8 Nov 20;20( 1) :1 37. II. Berger K, Rhost S, Rafnsdóttir S, Hughes É, Magnusson Y, Ekholm M, Stål O, Rydén L and Landberg G . T umor co- ex p ression of p rogranulin and sortilin as a p rognostic biomark er in breast cancer. BMC Cancer. 2021 Feb 22;21( 1) :1 85 . III. Berger K* , P ersson E *, Gregersson P, Jonasson E, Ståhlberg A, Landberg G and Rhost S. I nterleuk in- 6 ind uces stem cell p rop agation th rough liaison w ith th e sortilin- p rogranulin ax is in breast cancer. *Authors contributed equally. ( Manuscript ) IV. Berger K* , Rhost S*, Hughes É, Gregersson P and Landberg G. G ranulin p ep tid e d omains ind uce breast cancer stem cell p rop agation vi sortilin. *Authors contributed equally. (Manuscript) V. Berger K, Pauwels E, Parkin son G, Landberg G, Le T, Demillo V.G, Lumangtad L.A , Jones D.E, Islam M.A, Olsen R , Kapri T, Intasiri A , Vermeire K, Rhost S and Bell T.W . R ed uction of P rogranulin- I nd uced Breast Cancer S tem Cell P rop agation by S ortilin- T argeting Cyclotriaz ad isulf onamid e ( CA D A ) Comp ound s. ( Manuscript under revision ) IV V Table of contents ABSTRACT .................................................................................................................. I SAMMANFATTNING PÅ SV ENSKA .......................................................................... II LIST OF PAPERS ........................................................................................................ III ABBREVIATIONS ...................................................................................................... VII INTRODUCTION ...................................................................................................... 1 BREAST CANCE R ........................................................................................................ 1 N ORMAL BREAST DEVELOP MENT ................................................................................ 1 BREAST CANCE R DEVELOP MENT AND PROGRESSION .................................................... 2 BREAST CANCE R CLASSIF ICATION AND SUBTYPES ........................................................ 4 BREAST CANCER TREATME NT ..................................................................................... 5 Endocrine therapy ............................................................................................... 6 Monoclonal antibody-targeted therapy ............................................................. 7 Chemotherapy .................................................................................................... 7 T H E TUMOR MICROE NVIRO NMENT AND BREAST CAN CER HE TE ROGENEITY .................. 8 The clonal evolution theory and the hierarchal CSC model ................................ 9 Progranulin ....................................................................................................... 11 IL-6 .................................................................................................................... 12 IL-8 .................................................................................................................... 13 Hypoxia ............................................................................................................. 14 AIMS .......................................................................................................................... 15 METHODS ................................................................................................................ 17 T UMOR MODEL SYSTEMS .......................................................................................... 17 In vitro models .................................................................................................. 17 In vivo animal models ....................................................................................... 17 In vivo-like 3D models ....................................................................................... 18 E XPE RIMENTAL METHODS ........................................................................................ 19 Functional cancer stem cell enriching assays ................................................... 19 Gene expression analysis (qPCR) ...................................................................... 19 Protein analysis ................................................................................................. 19 P ATIENT DATA AND ANAL YSIS ................................................................................. 21 RESULTS AND DISCUSSI ON .................................................................................. 23 P APER I ................................................................................................................... 23 P APER II .................................................................................................................. 26 P APER III ................................................................................................................. 29 IV V Table of contents ABSTRACT .................................................................................................................. I SAMMANFATTNING PÅ SV ENSKA .......................................................................... II LIST OF PAPERS ........................................................................................................ III ABBREVIATIONS ...................................................................................................... VII INTRODUCTION ...................................................................................................... 1 BREAST CANCE R ........................................................................................................ 1 N ORMAL BREAST DEVELOP MENT ................................................................................ 1 BREAST CANCE R DEVELOP MENT AND PROGRESSION .................................................... 2 BREAST CANCE R CLASSIF ICATION AND SUBTYPES ........................................................ 4 BREAST CANCER TREATME NT ..................................................................................... 5 Endocrine therapy ............................................................................................... 6 Monoclonal antibody-targeted therapy ............................................................. 7 Chemotherapy .................................................................................................... 7 T H E TUMOR MICROE NVIRO NMENT AND BREAST CAN CER HE TE ROGENEITY .................. 8 The clonal evolution theory and the hierarchal CSC model ................................ 9 Progranulin ....................................................................................................... 11 IL-6 .................................................................................................................... 12 IL-8 .................................................................................................................... 13 Hypoxia ............................................................................................................. 14 AIMS .......................................................................................................................... 15 METHODS ................................................................................................................ 17 T UMOR MODEL SYSTEMS .......................................................................................... 17 In vitro models .................................................................................................. 17 In vivo animal models ....................................................................................... 17 In vivo-like 3D models ....................................................................................... 18 E XPE RIMENTAL METHODS ........................................................................................ 19 Functional cancer stem cell enriching assays ................................................... 19 Gene expression analysis (qPCR) ...................................................................... 19 Protein analysis ................................................................................................. 19 P ATIENT DATA AND ANAL YSIS ................................................................................. 21 RESULTS AND DISCUSSI ON .................................................................................. 23 P APER I ................................................................................................................... 23 P APER II .................................................................................................................. 26 P APER III ................................................................................................................. 29 V I P APER IV ................................................................................................................. 32 P APER V .................................................................................................................. 34 C ON CLUSIONS ....................................................................................................... 37 FUTURE PERSPECTIVES ........................................................................................... 39 ACKNOWLEDGEMENTS ........................................................................................ 43 RE F E RE N C ES ............................................................................................................. 45 VII A bbreviations 5 - FU : fluorouracil BCSS: breast cancer - specific survival BRCA: breast cancer CADA: cyclotriazadisulfonamide CD4: cluster of differentiation 4 cDNA: complementary DNA ctDNA: circulating tumor DNA CLL: chronic lymphocytic leukemia CPH: cox proportional hazards CSC: cancer stem cells CXCL: chemokine (C - X - C motif) ligand CXCR: chemokine (C- X - C motif) receptor ECM: extracellular matrix ELISA : enzyme - linked immunosorbent assay EMT : epithelial mesenchymal transition EphA2: Ephrin type - A receptor 2 ERBB2: receptor tyrosine - protein kinase 2 ERK1/2: extracellular signal- regulated kinase 1 and 2 ERα : estrogen receptor alpha ERβ : estrogen receptor beta FAK: focal adhesion kinase FASL: Fas ligand FPA : fluorescence polarization binding assay gp130: glycoprotein 130 GRN : progranulin gene HER2: human epidermal receptor 2 HIF: hypoxia inducible factor HIV: human immunodeficiency virus HR: hazard ratio HRE: hypoxic response element HRP: horseradish peroxidase IHC: immunohistochemistry V I P APER IV ................................................................................................................. 32 P APER V .................................................................................................................. 34 C ON CLUSIONS ....................................................................................................... 37 FUTURE PERSPECTIVES ........................................................................................... 39 ACKNOWLEDGEMENTS ........................................................................................ 43 RE F E RE N C ES ............................................................................................................. 45 VII A bbreviations 5 - FU : fluorouracil BCSS: breast cancer - specific survival BRCA: breast cancer CADA: cyclotriazadisulfonamide CD4: cluster of differentiation 4 cDNA: complementary DNA ctDNA: circulating tumor DNA CLL: chronic lymphocytic leukemia CPH: cox proportional hazards CSC: cancer stem cells CXCL: chemokine (C - X - C motif) ligand CXCR: chemokine (C- X - C motif) receptor ECM: extracellular matrix ELISA : enzyme - linked immunosorbent assay EMT : epithelial mesenchymal transition EphA2: Ephrin type - A receptor 2 ERBB2: receptor tyrosine - protein kinase 2 ERK1/2: extracellular signal- regulated kinase 1 and 2 ERα : estrogen receptor alpha ERβ : estrogen receptor beta FAK: focal adhesion kinase FASL: Fas ligand FPA : fluorescence polarization binding assay gp130: glycoprotein 130 GRN : progranulin gene HER2: human epidermal receptor 2 HIF: hypoxia inducible factor HIV: human immunodeficiency virus HR: hazard ratio HRE: hypoxic response element HRP: horseradish peroxidase IHC: immunohistochemistry V III IL: interleukin IL - 6R: interleukin - 6 receptor IL - 8RA/B: interleukin - 8 receptor A or B (Also called CXCR1 and 2) IVIS: In Vivo Imaging Software JAK: Janus tyrosine family kinase LDL: low - density lipoprotein M APK: mitogen - activated protein kinase MPE P: 1 - [ 2 - ( 2 - tert- butyl- 5 - methylphenoxy)- ethyl] - 3 - methylpiperidine mRNA: messenger RNA MYC: cellular myelomatosis NGS: next generation sequencing NOD/SCID: non - obese diabetic/severe combined immunodeficiency PDX: pati ent derived xenografts PEA: proximity extension assay PI3 - K: phosphoinositide 3 - kinase PR: progesterone receptor qPCR: quantitative polymerase chain reaction RB: retinoblastoma protein SERD : selective estrogen receptor degrader SER M: selective estrogen receptor modulator sIL- 6R : soluble interleukin - 6 receptor siRNA: small interfering RNA SLPI: secretory leukocyte protease inhibitor SORT1 : sortilin gene sSortilin : soluble sortilin STA T: the signaling transducer and activator of transcription TIL: tumor infiltrating lymphocyte TMA : tissue microarray TNF: tumor necrosis factor TNFR 1/2: tumor necrosis factor receptors 1 and 2 TNM: tumor node metastasis TP53: tumor protein p53 Vps10p: vacuolar protein sorting 10 protein 1 Introduction B reast cancer Breast cancer is the most common cancer among the female population and the leading cause of death to cancer in women worldwide [1] . In Sweden, 30% of all cancers in women are breast cancer and the numbers are increasing every year [2] . In 2019, almost 8300 women were diagnosed with breast cancer in Sweden [3] . Nevertheless, earlier detection and improved treatment strategies have led to a relatively good prognosis, with 5 - year and 10- year survival rates of 92% and 86.1% (Sweden, 201 6) [2] . There are a number of risk factors associated with breast cancer [4, 5] . Some are linked to hormones, like early menstruation, late menopause, having your first child late in life or having no children at all. Other risk factors are ass ociated with age, gender, genetic predisposition, family history of breast cancer, radiation exposure, obesity, diet, exercise- level and alcohol consumption [5] . Normal breast development From newborn until menopause, the female breast undergoes cycles of development and differentiation [6] . Through distinct developmental stages, the mammary glands develop from a few ste m cells and give rise to phenotypically and functionally different cell types in the breast (Figure 1) [7] . In an infant, the breast tissue consists only of a tiny duct composed of epithelial cells, which is similar in both genders until puberty. During puberty, hormones like estrogen allows the breast duct to grow rapidly and divide into primary and secondary ducts. The lobules develop and grow into tree- like structures. Throughout pregnancy, the breasts fully develop. Estrogen promotes further proliferation and differentiation of the ductal trees, and progesterone induces additional growth of the lobules. Additional increase in growth hormones and prolact in promotes the complete development of the mammary glands. V III IL: interleukin IL - 6R: interleukin - 6 receptor IL - 8RA/B: interleukin - 8 receptor A or B (Also called CXCR1 and 2) IVIS: In Vivo Imaging Software JAK: Janus tyrosine family kinase LDL: low - density lipoprotein M APK: mitogen - activated protein kinase MPE P: 1 - [ 2 - ( 2 - tert- butyl- 5 - methylphenoxy)- ethyl] - 3 - methylpiperidine mRNA: messenger RNA MYC: cellular myelomatosis NGS: next generation sequencing NOD/SCID: non - obese diabetic/severe combined immunodeficiency PDX: pati ent derived xenografts PEA: proximity extension assay PI3 - K: phosphoinositide 3 - kinase PR: progesterone receptor qPCR: quantitative polymerase chain reaction RB: retinoblastoma protein SERD : selective estrogen receptor degrader SER M: selective estrogen receptor modulator sIL- 6R : soluble interleukin - 6 receptor siRNA: small interfering RNA SLPI: secretory leukocyte protease inhibitor SORT1 : sortilin gene sSortilin : soluble sortilin STA T: the signaling transducer and activator of transcription TIL: tumor infiltrating lymphocyte TMA : tissue microarray TNF: tumor necrosis factor TNFR 1/2: tumor necrosis factor receptors 1 and 2 TNM: tumor node metastasis TP53: tumor protein p53 Vps10p: vacuolar protein sorting 10 protein 1 Introduction B reast cancer Breast cancer is the most common cancer among the female population and the leading cause of death to cancer in women worldwide [1] . In Sweden, 30% of all cancers in women are breast cancer and the numbers are increasing every year [2] . In 2019, almost 8300 women were diagnosed with breast cancer in Sweden [3] . Nevertheless, earlier detection and improved treatment strategies have led to a relatively good prognosis, with 5 - year and 10- year survival rates of 92% and 86.1% (Sweden, 201 6) [2] . There are a number of risk factors associated with breast cancer [4, 5] . Some are linked to hormones, like early menstruation, late menopause, having your first child late in life or having no children at all. Other risk factors are ass ociated with age, gender, genetic predisposition, family history of breast cancer, radiation exposure, obesity, diet, exercise- level and alcohol consumption [5] . Normal breast development From newborn until menopause, the female breast undergoes cycles of development and differentiation [6] . Through distinct developmental stages, the mammary glands develop from a few ste m cells and give rise to phenotypically and functionally different cell types in the breast (Figure 1) [7] . In an infant, the breast tissue consists only of a tiny duct composed of epithelial cells, which is similar in both genders until puberty. During puberty, hormones like estrogen allows the breast duct to grow rapidly and divide into primary and secondary ducts. The lobules develop and grow into tree- like structures. Throughout pregnancy, the breasts fully develop. Estrogen promotes further proliferation and differentiation of the ductal trees, and progesterone induces additional growth of the lobules. Additional increase in growth hormones and prolact in promotes the complete development of the mammary glands. 2 Figure 1. Normal breast tissue. The normal breast consists of milk - producing glands called lobules, which are connected by small tubes, called ducts. Lobules and ducts are surrounded by the breast stroma, consisting of fat cells, immune cells and connective tissue, as well as blood and lymph vessels. Image modified from https://www .teresewinslow.com . Breast cancer development and progression Cancer occur when a cell transforms to a malignant state through a series of processes characterized as the “Hallmarks of Cancer”, described by Hanahan and Weinberg in 2000 [8] . The six original hallmarks, shown in Figure 2, include: (1) cells becoming self - suffi cient in growth signals, leading to sustainable proliferation, (2) loss of sensitivity to growth - inhibition signals, (3) apoptotic resistance, (4) unlimited replication, (5) angiogenesis, providing the tumor with new blood vessels, and (6) activating invasion and metastasis, where the tumor becomes able to invade surrounding tissues. After new evidence linking inflammation and cancer, it became evident that additional hallmarks should be added [9] . They include: (7) avoiding immune surveillance and destruction, as well as (8) reprogramming of cellular metabolism, resulting from the cancer cells providing the surrounding microenvironment with various growth signals driving tumor progression [10] . Additionally, (9) genomic instability and mutations, along with (1 0) tumor - promoting inflammation were added. These are so- called enabling characteristics, leading to the acquirement of all the listed hallmarks. 3 Figure 2. The Hallmarks of Cancer. Th e origin of cancer are described through processes including the six original hallmarks, as well as two assisting factors and two emerging hallmarks. Image created with BioRender.com. In breast cancer, the tumor progresses from a normal cell to hyperplasia, where the cells appear normal, but divide and proliferate uncontrollably (Figure 3) [11- 13] . Tumor progression can occur through a gradual accumulation of genetic and epigenetic changes, from deletions that alter tumor suppressor genes, such as tumor protein p53 (T P5 3) or retinoblastoma protein (RB), to amplifications or activation of oncogenes, such as receptor tyrosine - protein kinase 2 (ERBB2) or MYC (cellular myelomatosis). Most of the mutations are sporadic, but some can be inherited, such as the germline mutations breast cancer 1 (BRCA1) and BRCA2. Through further alterations, the cells proliferative capacity increases and gradua lly become more abnormal in shape and orientation (atypical hyperplasia). In carcinoma in situ (ductal or lobular), the cells appear abnormal and grow uncontrollably, but have not yet broken through the tumor boundary. In time, more genetic alterations and mutations occur. Some might inactivate DNA repair genes leading to more genetic 2 Figure 1. Normal breast tissue. The normal breast consists of milk - producing glands called lobules, which are connected by small tubes, called ducts. Lobules and ducts are surrounded by the breast stroma, consisting of fat cells, immune cells and connective tissue, as well as blood and lymph vessels. Image modified from https://www .teresewinslow.com . Breast cancer development and progression Cancer occur when a cell transforms to a malignant state through a series of processes characterized as the “Hallmarks of Cancer”, described by Hanahan and Weinberg in 2000 [8] . The six original hallmarks, shown in Figure 2, include: (1) cells becoming self - suffi cient in growth signals, leading to sustainable proliferation, (2) loss of sensitivity to growth - inhibition signals, (3) apoptotic resistance, (4) unlimited replication, (5) angiogenesis, providing the tumor with new blood vessels, and (6) activating invasion and metastasis, where the tumor becomes able to invade surrounding tissues. After new evidence linking inflammation and cancer, it became evident that additional hallmarks should be added [9] . They include: (7) avoiding immune surveillance and destruction, as well as (8) reprogramming of cellular metabolism, resulting from the cancer cells providing the surrounding microenvironment with various growth signals driving tumor progression [10] . Additionally, (9) genomic instability and mutations, along with (1 0) tumor - promoting inflammation were added. These are so- called enabling characteristics, leading to the acquirement of all the listed hallmarks. 3 Figure 2. The Hallmarks of Cancer. Th e origin of cancer are described through processes including the six original hallmarks, as well as two assisting factors and two emerging hallmarks. Image created with BioRender.com. In breast cancer, the tumor progresses from a normal cell to hyperplasia, where the cells appear normal, but divide and proliferate uncontrollably (Figure 3) [11- 13] . Tumor progression can occur through a gradual accumulation of genetic and epigenetic changes, from deletions that alter tumor suppressor genes, such as tumor protein p53 (T P5 3) or retinoblastoma protein (RB), to amplifications or activation of oncogenes, such as receptor tyrosine - protein kinase 2 (ERBB2) or MYC (cellular myelomatosis). Most of the mutations are sporadic, but some can be inherited, such as the germline mutations breast cancer 1 (BRCA1) and BRCA2. Through further alterations, the cells proliferative capacity increases and gradua lly become more abnormal in shape and orientation (atypical hyperplasia). In carcinoma in situ (ductal or lobular), the cells appear abnormal and grow uncontrollably, but have not yet broken through the tumor boundary. In time, more genetic alterations and mutations occur. Some might inactivate DNA repair genes leading to more genetic 4 instability, g iving rise to invasive lobular/ductal carcinoma. In invasive carcinoma, some tumor cells have gained a more motile phenotype by remodeling of the extracellular matrix (ECM). The cells are then able to break through the membrane boundary and started invading nearby tissues [12, 13] . Eventually, cells can enter the blood or lymphatic circulation where they can form colonies (metastasize) at distant sites . Figure 3. Breast cancer progression. The tumor progresses from normal cells going through alterations leading to elevated proliferation of the cells (hyperplasia). Through additional increases in proliferation, as well as downregulation of apoptosis and a more abnormal morphology, the tumor gradually continue to progress leading to atypical hyperplasia. Proceeding further, we have carcinoma in situ , where the cells are still growing inside the boundary of the tumor. Upregulation of markers related to angiogenesis, epithelial mesenchymal transition and extra cellular matrix - remodeling direct the tumor into an invasive carcinoma. Image adapted from [11] and created with BioRender.com. Breast cancer classification and subtypes Breast cancer is a highly heterogeneous disease that can be divided into various subtypes. The subtypes are defined using different classification systems that are based on histology or the molecular basis of the tumor [14] . These subtypes are used as both prognostic markers (predicting the likely outcom e of the disease) and treatment - predictive indicators (probability to benefit from a certain treatment) to help decide treatment strategies for the patients [15- 17] . Histological classification primarily divides breast can cer into ductal or lobular carcinoma in situ, the precursor of breast cancer, or invasive carcinoma, where ductal carcinoma is the most common [17] . Another classification of breast cancer utilizes prognostic markers to characterize the tumor stage or tumor node metastasis (TNM) , depending on tumor size, lymph node status and metastatic spread, o r the tumor grade [18] . Tumor grade is divided into hig h or low grade based on how well differentiated the cells are and how fast the cells grow [19] . A high - grade 5 tumor is defined as poorly differentiated and have a high expressi on of the cellular proliferation marker Ki67. Classification by immunohistochemistry (IHC) is carried out through examining protein expression levels, especially the status of the commonly used biomarkers estrogen receptor alpha (ER α , the progesterone receptor (PR) and human epidermal receptor 2 (HER2). Molecular subtype classification was first proposed in year 200 0 by Perou and Sørlie, and are based on gene and protein expression data and epithelial cell origin (PAN 50: a 50 gene expression signature) [15, 16 , 20] . The molecul ar subtypes are summarized in Table 1 and LQFOXGHWKH(5α- positive luminal tumors, which are further separated into luminal A and luminal B tumors, the HER2 - enriched and basal - like (most often triple QHJDWLYHIRU(5α35DQG+(5 EUHDVWFDQFHU Table 1. Molecular subtypes of breast cancer. Based on data from [5, 15, 21 - 2 5 ] . Subtype Hormone status Grade/outcome Proliferation (Ki67) Luminal A (50 - 6 0% ) High ER and PR expression, low HER2 Low grade, good outcome Low proliferation Luminal B (10 - 2 0% ) ER and PR expression (lower than Luminal A), variable HER2 expression Higher grade, poor survival compared to luminal A Higher expression of proliferative genes compared to Luminal A HER2 (E RBB2 overexpressing) (15 - 2 0% ) Low or no ER and PR expression High grade, often aggressive, intermediate prognosis (but respond well to HER2 - targeted therapies) High proliferation Basal- like (10 - 1 5% ) Often triple negative (no ER, PR or HER2 ) High grade, poor outcome (only 20% respond to chemotherapy) High proliferation Breast cancer treatment Standard treatments for breast cancer patients are usually breast surgery, either breast conserving or removal of the entire breast, which in some cases also are followed by radiotherapy. More specific, systemi c treatments are also given, such as chemotherapy, endocrine treatment or targeted therapy ( e.g. small molecular inhibitors and antibody - based therapy) [16, 26] . These treatments can be given as neoadjuvant treatment, prior to surgery, to sh rink the tumor or as a response - indi cator. Alternatively, they 4 instability, g iving rise to invasive lobular/ductal carcinoma. In invasive carcinoma, some tumor cells have gained a more motile phenotype by remodeling of the extracellular matrix (ECM). The cells are then able to break through the membrane boundary and started invading nearby tissues [12, 13] . Eventually, cells can enter the blood or lymphatic circulation where they can form colonies (metastasize) at distant sites . Figure 3. Breast cancer progression. The tumor progresses from normal cells going through alterations leading to elevated proliferation of the cells (hyperplasia). Through additional increases in proliferation, as well as downregulation of apoptosis and a more abnormal morphology, the tumor gradually continue to progress leading to atypical hyperplasia. Proceeding further, we have carcinoma in situ , where the cells are still growing inside the boundary of the tumor. Upregulation of markers related to angiogenesis, epithelial mesenchymal transition and extra cellular matrix - remodeling direct the tumor into an invasive carcinoma. Image adapted from [11] and created with BioRender.com. Breast cancer classification and subtypes Breast cancer is a highly heterogeneous disease that can be divided into various subtypes. The subtypes are defined using different classification systems that are based on histology or the molecular basis of the tumor [14] . These subtypes are used as both prognostic markers (predicting the likely outcom e of the disease) and treatment - predictive indicators (probability to benefit from a certain treatment) to help decide treatment strategies for the patients [15- 17] . Histological classification primarily divides breast can cer into ductal or lobular carcinoma in situ, the precursor of breast cancer, or invasive carcinoma, where ductal carcinoma is the most common [17] . Another classification of breast cancer utilizes prognostic markers to characterize the tumor stage or tumor node metastasis (TNM) , depending on tumor size, lymph node status and metastatic spread, o r the tumor grade [18] . Tumor grade is divided into hig h or low grade based on how well differentiated the cells are and how fast the cells grow [19] . A high - grade 5 tumor is defined as poorly differentiated and have a high expressi on of the cellular proliferation marker Ki67. Classification by immunohistochemistry (IHC) is carried out through examining protein expression levels, especially the status of the commonly used biomarkers estrogen receptor alpha (ER α , the progesterone receptor (PR) and human epidermal receptor 2 (HER2). Molecular subtype classification was first proposed in year 200 0 by Perou and Sørlie, and are based on gene and protein expression data and epithelial cell origin (PAN 50: a 50 gene expression signature) [15, 16 , 20] . The molecul ar subtypes are summarized in Table 1 and LQFOXGHWKH(5α- positive luminal tumors, which are further separated into luminal A and luminal B tumors, the HER2 - enriched and basal - like (most often triple QHJDWLYHIRU(5α35DQG+(5 EUHDVWFDQFHU Table 1. Molecular subtypes of breast cancer. Based on data from [5, 15, 21 - 2 5 ] . Subtype Hormone status Grade/outcome Proliferation (Ki67) Luminal A (50 - 6 0% ) High ER and PR expression, low HER2 Low grade, good outcome Low proliferation Luminal B (10 - 2 0% ) ER and PR expression (lower than Luminal A), variable HER2 expression Higher grade, poor survival compared to luminal A Higher expression of proliferative genes compared to Luminal A HER2 (E RBB2 overexpressing) (15 - 2 0% ) Low or no ER and PR expression High grade, often aggressive, intermediate prognosis (but respond well to HER2 - targeted therapies) High proliferation Basal- like (10 - 1 5% ) Often triple negative (no ER, PR or HER2 ) High grade, poor outcome (only 20% respond to chemotherapy) High proliferation Breast cancer treatment Standard treatments for breast cancer patients are usually breast surgery, either breast conserving or removal of the entire breast, which in some cases also are followed by radiotherapy. More specific, systemi c treatments are also given, such as chemotherapy, endocrine treatment or targeted therapy ( e.g. small molecular inhibitors and antibody - based therapy) [16, 26] . These treatments can be given as neoadjuvant treatment, prior to surgery, to sh rink the tumor or as a response - indi cator. Alternatively, they 6 are given after surgery, alone or in combinations, depending on tumor burden or subtype, to prevent recurrence and prolong survival. Patient prognosis and treatment options differ widely depending on the molecular subtype of the tumor, having to take into account the presence or absence of hormone receptors, grade, lymph node sta tus and gene expression of specific markers [16, 27] . Individualization of therapy have become more common, directing the treatment towards the biology of the tumor and to reduce adverse side - effects and prevent resistance, by rewiring of signaling pathways, which usually arises from conventional therapy where all patients receive the same treatment [5, 19] . Endocrine therapy $SSUR[LPDWHO\  RI DOO WXPRUV H[SUHVV (5α DQG DUH WUHDWHG ZLWK endocrine adjuvant therapy, such as tamoxifen or aromatase inhibitors [28] . In the breast, ER α is the predominant estrogen receptor and is used in clinical setting for treatment decisions, as ERβOHYHOVKDYHEHHQVKRZQ to vary when it comes to treatment response [29] . The hormone estrogen binds to its main receptors ER αDQG E RβDQGLQGXFHWKHH[SUHVVLRQRI35 expression, activated by progester one. Tamoxifen is a selective estrogen receptor modulator (SER M) , functioning by competing with estrogen for the binding to the ER , blocking the effect of estrogen (estrogen antagonist). Tamoxifen is effective against metastatic breast cancer and are often given as adjuvant therapy up to 5 - 10 years after surgery, depending on the risk of relapse. Tamoxifen also function as an ER agonist in the bone, by increasing bone mineralization through decreasing low - density lipoprotein (LDL) cholesterol levels [28] . However, despite tamoxifen’s success in improving the survival of the hormone receptor positive patien t group, many patients experience tumor relapse or therapy resistance [26] . Another way of blocking the estrogen receptor is by the use of selective estrogen receptor degrader (SERD ) , such as fulvestrant [19] . Fulvestrant compete for estrogen receptor as estrogen antagonist, but its binding causes targeting of the ER for destruction by the immune system. Aromatase inhibitors like anastrozole or letrozole are blocking estrogen production by inhibiting the enzyme aromatase that normally converts testosterone to estradiol [26, 28] . 7 Monoclonal antibody- targeted therapy HE R 2 positive tumors are tumors with an ERBB2 amplification or HER2 overexpression, which represent 15 - 20% of all breast cancers. These patients usually respond well to monoclonal antibody treatment targeting the receptor HER2, such as Trastuzumab (Hercept in) [19, 30] . ERBB2 is a proto - oncogene that when amplified, it causes the activation of the HER 2 pathway. This activates proliferation, cell survival, metastasis through various pathways [5] . Chemotherapy There are several types of chemotherapy drugs availa ble, such as microtubule stabilizers (Docetaxel, Pacliaxel), anti- metabolic factors (5- FU , fluorouracil), DNA intercalators (cisplatin, doxorubicin) and CDK4 and CDK6 inhibitors (Palbociclib) [1 9] . The different types can be used alone or in combinations with each other, and also together with other types of therapy, such as endocrine therapy or HER2 targeted therapy to impr ove effectiveness and patient survival. In addition, 15 % of all breast cancers are basal- like tumors. The majority of basal- OLNHWXPRUVDUHWULSOHQHJDWLYHODFNLQJ(5α35DQG+(57KHVH tumors are difficult to treat, as they lack effective drug targets [23] . The standard treatment for this subtype is still chemotherapy, but patients often have a shortened disease- free and overall survival rate, and an increased risk of developing distant metastases compared to other forms of the disease. Metastases accounts for more than 90% of cancer - related deaths [3 1] . Therefore, there is a great need to develop treatments for triple negative breast cancers and identify subgroups that respond to specific therapeutic agents [32, 33 ] . In conclusion, there is an increas ing need to identify biomarkers and key mediators involved in breast cancer progression , to be able to distinguish subgroups of breast cancer patients which will benefit from specific treatments [34] . 6 are given after surgery, alone or in combinations, depending on tumor burden or subtype, to prevent recurrence and prolong survival. Patient prognosis and treatment options differ widely depending on the molecular subtype of the tumor, having to take into account the presence or absence of hormone receptors, grade, lymph node sta tus and gene expression of specific markers [16, 27] . Individualization of therapy have become more common, directing the treatment towards the biology of the tumor and to reduce adverse side - effects and prevent resistance, by rewiring of signaling pathways, which usually arises from conventional therapy where all patients receive the same treatment [5, 19] . Endocrine therapy $SSUR[LPDWHO\  RI DOO WXPRUV H[SUHVV (5α DQG DUH WUHDWHG ZLWK endocrine adjuvant therapy, such as tamoxifen or aromatase inhibitors [28] . In the breast, ER α is the predominant estrogen receptor and is used in clinical setting for treatment decisions, as ERβOHYHOVKDYHEHHQVKRZQ to vary when it comes to treatment response [29] . The hormone estrogen binds to its main receptors ER αDQG E RβDQGLQGXFHWKHH[SUHVVLRQRI35 expression, activated by progester one. Tamoxifen is a selective estrogen receptor modulator (SER M) , functioning by competing with estrogen for the binding to the ER , blocking the effect of estrogen (estrogen antagonist). Tamoxifen is effective against metastatic breast cancer and are often given as adjuvant therapy up to 5 - 10 years after surgery, depending on the risk of relapse. Tamoxifen also function as an ER agonist in the bone, by increasing bone mineralization through decreasing low - density lipoprotein (LDL) cholesterol levels [28] . However, despite tamoxifen’s success in improving the survival of the hormone receptor positive patien t group, many patients experience tumor relapse or therapy resistance [26] . Another way of blocking the estrogen receptor is by the use of selective estrogen receptor degrader (SERD ) , such as fulvestrant [19] . Fulvestrant compete for estrogen receptor as estrogen antagonist, but its binding causes targeting of the ER for destruction by the immune system. Aromatase inhibitors like anastrozole or letrozole are blocking estrogen production by inhibiting the enzyme aromatase that normally converts testosterone to estradiol [26, 28] . 7 Monoclonal antibody- targeted therapy HE R 2 positive tumors are tumors with an ERBB2 amplification or HER2 overexpression, which represent 15 - 20% of all breast cancers. These patients usually respond well to monoclonal antibody treatment targeting the receptor HER2, such as Trastuzumab (Hercept in) [19, 30] . ERBB2 is a proto - oncogene that when amplified, it causes the activation of the HER 2 pathway. This activates proliferation, cell survival, metastasis through various pathways [5] . Chemotherapy There are several types of chemotherapy drugs availa ble, such as microtubule stabilizers (Docetaxel, Pacliaxel), anti- metabolic factors (5- FU , fluorouracil), DNA intercalators (cisplatin, doxorubicin) and CDK4 and CDK6 inhibitors (Palbociclib) [1 9] . The different types can be used alone or in combinations with each other, and also together with other types of therapy, such as endocrine therapy or HER2 targeted therapy to impr ove effectiveness and patient survival. In addition, 15 % of all breast cancers are basal- like tumors. The majority of basal- OLNHWXPRUVDUHWULSOHQHJDWLYHODFNLQJ(5α35DQG+(57KHVH tumors are difficult to treat, as they lack effective drug targets [23] . The standard treatment for this subtype is still chemotherapy, but patients often have a shortened disease- free and overall survival rate, and an increased risk of developing distant metastases compared to other forms of the disease. Metastases accounts for more than 90% of cancer - related deaths [3 1] . Therefore, there is a great need to develop treatments for triple negative breast cancers and identify subgroups that respond to specific therapeutic agents [32, 33 ] . In conclusion, there is an increas ing need to identify biomarkers and key mediators involved in breast cancer progression , to be able to distinguish subgroups of breast cancer patients which will benefit from specific treatments [34] . 8 The tumor microenvironment and breast cancer heterogeneity As mentioned earlier, breast cancer is a very heterogeneous disease, where functional and phenotypical differences can be seen both between tumors and within the tumor itself [21, 35] . These var iations are due to several factors, including various genetic and epigenetic alterations, the surrounding tumor microenvironment, as well as the presence of a small subpopulation of tumor cells, called cancer stem cells (CSCs), which potentially could lead to therapy resistance, cancer progression, metastasis or even tumor relapse [36, 37] . Moreover, cells such as fibroblasts and immune cells that are present in the tumor microenvironment communicate with each other and the surrounding tumor niche, secreting a range of different cytokines and growth factors, promoting tumor growth and metasta sis (Figure 4) [38, 39] . Figure 4. Components of the tumor microenvironment. The tumor microenvironment consists of various cell types ( e . g. fibroblasts and immune cells) and extracellular matrix components ( e . g. collagens, laminins and fibronectin). The cells in the tumor microenvironment communicate through paracrine signaling mediated by various cytokines and growth factors, which modifies cellular processes leading to tumor progression. Image based on [38] and created with BioR ender.com. 9 Further , low oxygen levels (hypoxia) play a major role in cancer progression, as it has been linked to poor overall patient survival and a more malignant tumor phenotype [40] . Areas of hypoxia are present in many solid tumors and potentially in the metastatic sites, suggest ing that hypoxia influence cancer cells including the CSC s. The hypothesis of the origin of CSCs starts with a single stem cell that have acquired different mutations, or from cells that are more differentiated , and acquire stem cell traits during tumor progression [17, 41] . CSCs h ave characteristics in common with normal stem cells, including the ability to self - renew, give rise to progenitor cells or cells that are more differentiated, and share common signaling pathways, such as the Notch, Wnt and Hedgehog pathways [35, 36 , 42 , 43] . Furthermore, CSC are characterized as possessing anchor - independent growth abilities, being low proliferative, acquire tumor - initiating capacities and metastatic potential and express some common surface markers, including the CD44 + /CD24 -/low and ALDH high phenotypes (although no universal maker for CSCs have been identified) [44- 47] . Importantly, CSCs have been described as resistant to radio - and chemotherapy, and may therefore contribute to cancer relapse [38, 42] . These CSCs have therefore been proposed as a promising target for treating breast cancer [43] . CSCs can be identified by different methods, including functional assay such as tumor sphere formation and in vivo xenograft assays, as well as specific cell surface markers and Hoechst staining/side - population sorting (drug resistance in CSCs) [4 2, 48] . The clonal evolution theory and the hierarchal CSC model There are two current models explaining the cause of cancer and intra - tumor heterogeneity (Figure 5) [49, 50] . The classical hypothesis, the clonal evolution theory, is based on clonal expansion, where a single cell has gone through random genetic mutations providing it with growth advantages [50, 51] . Through clonal expansion of that cell or clones with a more dominant and aggressive phenotype, it will eventually give rise t o a tumor. A heterogeneous cell population then appears when some of the cells acquire properties that are advantageous on their own, possibly due to influences from the tumor microenvironment. 8 The tumor microenvironment and breast cancer heterogeneity As mentioned earlier, breast cancer is a very heterogeneous disease, where functional and phenotypical differences can be seen both between tumors and within the tumor itself [21, 35] . These var iations are due to several factors, including various genetic and epigenetic alterations, the surrounding tumor microenvironment, as well as the presence of a small subpopulation of tumor cells, called cancer stem cells (CSCs), which potentially could lead to therapy resistance, cancer progression, metastasis or even tumor relapse [36, 37] . Moreover, cells such as fibroblasts and immune cells that are present in the tumor microenvironment communicate with each other and the surrounding tumor niche, secreting a range of different cytokines and growth factors, promoting tumor growth and metasta sis (Figure 4) [38, 39] . Figure 4. Components of the tumor microenvironment. The tumor microenvironment consists of various cell types ( e . g. fibroblasts and immune cells) and extracellular matrix components ( e . g. collagens, laminins and fibronectin). The cells in the tumor microenvironment communicate through paracrine signaling mediated by various cytokines and growth factors, which modifies cellular processes leading to tumor progression. Image based on [38] and created with BioR ender.com. 9 Further , low oxygen levels (hypoxia) play a major role in cancer progression, as it has been linked to poor overall patient survival and a more malignant tumor phenotype [40] . Areas of hypoxia are present in many solid tumors and potentially in the metastatic sites, suggest ing that hypoxia influence cancer cells including the CSC s. The hypothesis of the origin of CSCs starts with a single stem cell that have acquired different mutations, or from cells that are more differentiated , and acquire stem cell traits during tumor progression [17, 41] . CSCs h ave characteristics in common with normal stem cells, including the ability to self - renew, give rise to progenitor cells or cells that are more differentiated, and share common signaling pathways, such as the Notch, Wnt and Hedgehog pathways [35, 36 , 42 , 43] . Furthermore, CSC are characterized as possessing anchor - independent growth abilities, being low proliferative, acquire tumor - initiating capacities and metastatic potential and express some common surface markers, including the CD44 + /CD24 -/low and ALDH high phenotypes (although no universal maker for CSCs have been identified) [44- 47] . Importantly, CSCs have been described as resistant to radio - and chemotherapy, and may therefore contribute to cancer relapse [38, 42] . These CSCs have therefore been proposed as a promising target for treating breast cancer [43] . CSCs can be identified by different methods, including functional assay such as tumor sphere formation and in vivo xenograft assays, as well as specific cell surface markers and Hoechst staining/side - population sorting (drug resistance in CSCs) [4 2, 48] . The clonal evolution theory and the hierarchal CSC model There are two current models explaining the cause of cancer and intra - tumor heterogeneity (Figure 5) [49, 50] . The classical hypothesis, the clonal evolution theory, is based on clonal expansion, where a single cell has gone through random genetic mutations providing it with growth advantages [50, 51] . Through clonal expansion of that cell or clones with a more dominant and aggressive phenotype, it will eventually give rise t o a tumor. A heterogeneous cell population then appears when some of the cells acquire properties that are advantageous on their own, possibly due to influences from the tumor microenvironment. 10 Figure 5. Tumor heterogeneity models. Clonal evolution theo ry vs cancer stem cell model. Adapted from [52, 53] . Image created with BioRender.com. CSC: cancer stem cell. The CSC hypothesis, on the other hand, explains tumor heterogeneity through a hierarchical cellular organization [37] . In this organization, a small proportion of the cancer cells have the ability to sustain tumor growth and generate heterogeneity through differentiation. This provides the cells with stem cell properties, such as tumor - initiating capacity and metastatic potential [46] . It has been shown in immune - deficient mice that only a small subset of the cancer cells are able to proliferate and cause tumor growth, regenerating the original tumor [46 , 54] . In order for t he CSCs to maintain the CSC pool, they undergo symmetric division, but also asymmetric division to generate cells with low tumorigenic potential. Moreover, some researchers have proposed a plastic CSC model, explaining how some non - CSCs can retrieve their CSC phenotype through dedifferentiation [46] . Other researchers explain this through the induction of epithelial mesenchymal transition (EMT) [4 6, 55] . EMT is normally seen in embryonic development, where epithelial cells gain properties, making them more mesenchymal and fibroblast - like. In cancer, this enables the cells to leave the primary tumor, enter the circulation and metastasize at distant sites [56] . CSCs may be created by a single stem cell that have acquired different mutations, or from cells that are more differentiated and acquire stem cell traits during tumor progre ssion. The CSCs have been described as resistant to radiotherapy and chemotherapy, and may thus contribute to cancer relapse after surgery and treatment [42, 50] . The tumor heterogeneity and progression might be explained by a combination of the two theories, and also in combination with 11 microenvironmental influences, as none of them explains how the tumor formation is initiated [50] . Progranulin Progranulin, also known as PC cell- derived growth factor or granulin - epithelin precursor, is a cysteine - rich, heavily glycosylated, autocrine growth factor involved in various biological processes, such as wound healing, tumorigen esis, inflammation, as well as various neurological diseases [57- 5 9] . This 88- kDa glycoprotein can be cleaved by neutrophil elastase, proteases and different matrix metallopeptidases to produce different biologically active peptide domains of about 6 kDa (Figure 6) [60 , 61] . These granulins are named from para- granulin, which is a half - length domain, to granulin G, F, B, A, C, D and E, ordered from the N- terminus of progranulin to the C - terminus [58] . The interest in progranulin has emerged over the last few years, with publications demonstrating an overexpression of progranulin in a range of different cancer types, as well as being associated with poor prognosis and survival, suggesting that progranulin may be a relevant predictive and prognostic biomarker in various types of cancer [58 ] . Figure 6. Progranulin and its associated receptors. Progranulin and its respective domains are shown as circles. Cleavage of progranulin by proteolytic processing ( e . g. by MMP - 9, MMP - 14, ADAMTS - 7, NE or PR3 ) produces biologically active granulins that are thought to be involved in inflammation. The full - length progranulin protein has been shown to have a n anti- inflammatory effect by binding to tumor necrosis factor receptors (TNFR1 /2 ), and has also been shown to bind to Ephrin type - A receptor 2 (EphA2) and sortilin. Illustration created in BioRender.com, adapted from [58] . 10 Figure 5. Tumor heterogeneity models. Clonal evolution theo ry vs cancer stem cell model. Adapted from [52, 53] . Image created with BioRender.com. CSC: cancer stem cell. The CSC hypothesis, on the other hand, explains tumor heterogeneity through a hierarchical cellular organization [37] . In this organization, a small proportion of the cancer cells have the ability to sustain tumor growth and generate heterogeneity through differentiation. This provides the cells with stem cell properties, such as tumor - initiating capacity and metastatic potential [46] . It has been shown in immune - deficient mice that only a small subset of the cancer cells are able to proliferate and cause tumor growth, regenerating the original tumor [46 , 54] . In order for t he CSCs to maintain the CSC pool, they undergo symmetric division, but also asymmetric division to generate cells with low tumorigenic potential. Moreover, some researchers have proposed a plastic CSC model, explaining how some non - CSCs can retrieve their CSC phenotype through dedifferentiation [46] . Other researchers explain this through the induction of epithelial mesenchymal transition (EMT) [4 6, 55] . EMT is normally seen in embryonic development, where epithelial cells gain properties, making them more mesenchymal and fibroblast - like. In cancer, this enables the cells to leave the primary tumor, enter the circulation and metastasize at distant sites [56] . CSCs may be created by a single stem cell that have acquired different mutations, or from cells that are more differentiated and acquire stem cell traits during tumor progre ssion. The CSCs have been described as resistant to radiotherapy and chemotherapy, and may thus contribute to cancer relapse after surgery and treatment [42, 50] . The tumor heterogeneity and progression might be explained by a combination of the two theories, and also in combination with 11 microenvironmental influences, as none of them explains how the tumor formation is initiated [50] . Progranulin Progranulin, also known as PC cell- derived growth factor or granulin - epithelin precursor, is a cysteine - rich, heavily glycosylated, autocrine growth factor involved in various biological processes, such as wound healing, tumorigen esis, inflammation, as well as various neurological diseases [57- 5 9] . This 88- kDa glycoprotein can be cleaved by neutrophil elastase, proteases and different matrix metallopeptidases to produce different biologically active peptide domains of about 6 kDa (Figure 6) [60 , 61] . These granulins are named from para- granulin, which is a half - length domain, to granulin G, F, B, A, C, D and E, ordered from the N- terminus of progranulin to the C - terminus [58] . The interest in progranulin has emerged over the last few years, with publications demonstrating an overexpression of progranulin in a range of different cancer types, as well as being associated with poor prognosis and survival, suggesting that progranulin may be a relevant predictive and prognostic biomarker in various types of cancer [58 ] . Figure 6. Progranulin and its associated receptors. Progranulin and its respective domains are shown as circles. Cleavage of progranulin by proteolytic processing ( e . g. by MMP - 9, MMP - 14, ADAMTS - 7, NE or PR3 ) produces biologically active granulins that are thought to be involved in inflammation. The full - length progranulin protein has been shown to have a n anti- inflammatory effect by binding to tumor necrosis factor receptors (TNFR1 /2 ), and has also been shown to bind to Ephrin type - A receptor 2 (EphA2) and sortilin. Illustration created in BioRender.com, adapted from [58] . 12 Progranulin binds to the sortilin receptor, also called neurotensin receptor - 3, to mediate progranulin internalization. Sortilin is a member of the vacuolar protein sorting 10 protein (Vps10p) domain receptor family, mostly expressed in neurons to regulate neuronal function and viability, as well as in other tissues and cell types, such as B - lymphocytes, metabolic tissues and solid tumors [62, 63] . Sortilin has multiple roles in cellular transport and signaling, both intracellularly and as a cell surface receptor, involved in targeting and sorting proteins to different fates [63, 64] . Consequently, sortilin is involv ed in tumorigenesis, cardiovascular and metabolic diseases and neurological disorders, such as dementia, and have been proposed as a potential drug target for these diseases [62, 65] . Furthermore, progranulin binds to tumor necrosis factor (TNF) receptor 1 and 2, highlighting its importance in the immune response [57, 66] . More recently, progranulin has also been shown to bind to the newly identified receptor Ephrin type - A receptor 2 (Eph A2) [67] . IL- 6 Cytokines and chemokines are secreted by various cells in the tumor microenvironment, creating a link between inflammation and cancer [68, 69] . This type of inflammation is disrupting the balance in the tumor microenvironment, between cytokines, chemokines, transcriptional factors and reactive oxygen species, leading to tumor growth and cancer progression [70] . IL - 6 is an inflammatory cytokine produced and secreted by numerous cells and is invo lved in the regulation of B - and T - cell activation, growth and differentiation, as well as in recruiting neut rophils [68, 71, 72] . In addition, IL - 6 is thought to be involved in tumorigenesis and resis tance to cancer therapy by regulating signaling pathways important for tumor development and progression [73, 74] . High IL - 6 levels in serum and tissue have been detected in various types of cancer, including breast cancer, and are correlated with poor prognosis, advanced disease, metastasis and worse response to therapy [68, 72, 74] . Moreover, IL - 6 as well as its receptor have been shown to play a role in proliferation and expansion of the CSC pool in several types of cancer, as well as to induce EMT, cell migration and invasion, leading to metastasis formation [73, 74] . IL - 6 binds to the IL - 6 receptor (IL - 6R) and forms a complex that associates with a signal transducing receptor glycoprotein 130 (gp130, expressed on most cells) on the receiving cell, called classical signaling (Figure 7, left) [71, 75] . However, binding can also occur through the soluble form of the IL - 6 receptor (sIL - 6R) , so called trans - signaling, where gp130 is activated 13 without a membrane bound IL - 6 receptor (Figure 7, right) [71, 75] . Trans - signaling is thought to activate more pro - inflammatory pathways and to be more important in cancer [74] . Signaling pathways activated by IL- 6 are the Janus tyrosine family kinase (JAK) and the signaling transducer and activator of transcription (ST AT) pathway (JAK - STA T pathway), the extracellular signal- regulated kinase 1 and 2 (ERK1/2) , mitogen - activated protein kinase (MAPK) pathway (ERK1/2 - MAPK pathway), as well as the phosphoinositide 3 - kinase (PI3- K) pathway [70 , 71 , 76] . Fig. 7. IL -6 signaling . In the classical signaling pathway, IL - 6 binds to IL - 6R and recruits pg130 for cell signal initiation, causing an anti- inflammatory response. Trans- signaling occurs when IL - 6 binds to the soluble form of the IL - 6 receptor (sIL- 6R), which is cleaved by me talloproteases (ADAM10 or 17) or created from alternative splicing of the mRNA. This binding complex then interacts with gp130 and activates pro - inflammatory signaling pathways. Image adapted from [74] and created with BioRender.com. IL- 8 An additional pro - inflammatory cytokine, IL - 8, also called chemokine (C - X - C motif) ligand 8 (CXCL8), recruits inflammatory neutrophils and is involved in the promotion of angiogenesis by synthesis of matrix metalloproteases [68, 77] . IL - 8 is produced by many different cell types, including various cancer cells. [68, 77] . High levels of IL - 8 are associated with tumor size, stage, drug resistance, angiogenesis and infiltration in several cancer types, incl uding metastatic breast cancer [68, 77] . IL - 8 binds to the receptors IL - 8R A (CXCR1) and IL - 8RB (CXCR2) and stimulates pathways, including PI3- K/Akt and MAPK/ERK [68, 77] . 12 Progranulin binds to the sortilin receptor, also called neurotensin receptor - 3, to mediate progranulin internalization. Sortilin is a member of the vacuolar protein sorting 10 protein (Vps10p) domain receptor family, mostly expressed in neurons to regulate neuronal function and viability, as well as in other tissues and cell types, such as B - lymphocytes, metabolic tissues and solid tumors [62, 63] . Sortilin has multiple roles in cellular transport and signaling, both intracellularly and as a cell surface receptor, involved in targeting and sorting proteins to different fates [63, 64] . Consequently, sortilin is involv ed in tumorigenesis, cardiovascular and metabolic diseases and neurological disorders, such as dementia, and have been proposed as a potential drug target for these diseases [62, 65] . Furthermore, progranulin binds to tumor necrosis factor (TNF) receptor 1 and 2, highlighting its importance in the immune response [57, 66] . More recently, progranulin has also been shown to bind to the newly identified receptor Ephrin type - A receptor 2 (Eph A2) [67] . IL- 6 Cytokines and chemokines are secreted by various cells in the tumor microenvironment, creating a link between inflammation and cancer [68, 69] . This type of inflammation is disrupting the balance in the tumor microenvironment, between cytokines, chemokines, transcriptional factors and reactive oxygen species, leading to tumor growth and cancer progression [70] . IL - 6 is an inflammatory cytokine produced and secreted by numerous cells and is invo lved in the regulation of B - and T - cell activation, growth and differentiation, as well as in recruiting neut rophils [68, 71, 72] . In addition, IL - 6 is thought to be involved in tumorigenesis and resis tance to cancer therapy by regulating signaling pathways important for tumor development and progression [73, 74] . High IL - 6 levels in serum and tissue have been detected in various types of cancer, including breast cancer, and are correlated with poor prognosis, advanced disease, metastasis and worse response to therapy [68, 72, 74] . Moreover, IL - 6 as well as its receptor have been shown to play a role in proliferation and expansion of the CSC pool in several types of cancer, as well as to induce EMT, cell migration and invasion, leading to metastasis formation [73, 74] . IL - 6 binds to the IL - 6 receptor (IL - 6R) and forms a complex that associates with a signal transducing receptor glycoprotein 130 (gp130, expressed on most cells) on the receiving cell, called classical signaling (Figure 7, left) [71, 75] . However, binding can also occur through the soluble form of the IL - 6 receptor (sIL - 6R) , so called trans - signaling, where gp130 is activated 13 without a membrane bound IL - 6 receptor (Figure 7, right) [71, 75] . Trans - signaling is thought to activate more pro - inflammatory pathways and to be more important in cancer [74] . Signaling pathways activated by IL- 6 are the Janus tyrosine family kinase (JAK) and the signaling transducer and activator of transcription (ST AT) pathway (JAK - STA T pathway), the extracellular signal- regulated kinase 1 and 2 (ERK1/2) , mitogen - activated protein kinase (MAPK) pathway (ERK1/2 - MAPK pathway), as well as the phosphoinositide 3 - kinase (PI3- K) pathway [70 , 71 , 76] . Fig. 7. IL -6 signaling . In the classical signaling pathway, IL - 6 binds to IL - 6R and recruits pg130 for cell signal initiation, causing an anti- inflammatory response. Trans- signaling occurs when IL - 6 binds to the soluble form of the IL - 6 receptor (sIL- 6R), which is cleaved by me talloproteases (ADAM10 or 17) or created from alternative splicing of the mRNA. This binding complex then interacts with gp130 and activates pro - inflammatory signaling pathways. Image adapted from [74] and created with BioRender.com. IL- 8 An additional pro - inflammatory cytokine, IL - 8, also called chemokine (C - X - C motif) ligand 8 (CXCL8), recruits inflammatory neutrophils and is involved in the promotion of angiogenesis by synthesis of matrix metalloproteases [68, 77] . IL - 8 is produced by many different cell types, including various cancer cells. [68, 77] . High levels of IL - 8 are associated with tumor size, stage, drug resistance, angiogenesis and infiltration in several cancer types, incl uding metastatic breast cancer [68, 77] . IL - 8 binds to the receptors IL - 8R A (CXCR1) and IL - 8RB (CXCR2) and stimulates pathways, including PI3- K/Akt and MAPK/ERK [68, 77] . 14 Hypox ia Solid tumors, such as most breast cancers, often contain regions with low oxygen levels (hypoxia), due to insufficient vascularization, and is a part of the intricate tumor microenvironment [39, 78, 79 ] . Patients with these tumors often have a poor prognosis, showing more malignant and treatment- resistant properties due to the tumor cells adapting to the lower levels of oxygen, controll ed by hypoxia inducible factors (HIFs) [31, 79, 80] . HIFs, often called master transcriptional regulators of the hypoxic response, are involved in numerous processes, such as inducing proliferation, EMT, metastasis, survival, angiogenesis, invasion and metastasis, pH regulation, changes in metabolism and g lucose uptake, as well as the maintenance of stem cells [39 , 81] . In addition, HI Fs have been proposed as a potential therapy target for cancer [78 , 79] . HIF1 is a heterodimer consisting of an α- subunit that is only expressed at low oxygen concentrations (oxygen sensitive), and a β - subunit, which is constitutively active [39] . The transcription factor HIF1 α is universally expressed in t he body, while the other isoforms, HIF2 α and HIF 3 α, are only expressed in some tissues and at varying oxygen levels [79] . Under normoxic conditions (21 % O 2 ) , HI F1 has almost no activity, due to low levels of HI F1 α. When oxygen is present, oxygen - sensitizing enzymes (hydroxylases) facilitates HIF1 α hydroxylation on the oxygen - dependent - degradation domain on HIF1 α, leading to ubiquitination and degradation [79] . During hypoxia, HIF 1 α is stabilized and translocated to the nucleus, where it forms a dimer with HI F1β . The HI F1 complex then bind to different hypoxic response elements (HREs), and i nteracts with various co - activators to regulate transcrip tion of different target genes. 15 Aims The overall aim within the project is to elucidate the role of the microenvironmentally induced autocrine growth factor progranulin and its associated receptor sortilin in breast CSC propagation and prognosis prediction. Limiting breast cancer progression by targeting sortilin could be a potential therapeutic strategy for the treatment of patients with breast tumors having elevated progranulin and sortilin expression. The more specific aims for each paper are: P ap er I : To determine the influence of progranulin and its receptor sortilin on breast cancer propagation and CSC expansion using functional assays and in vivo mouse models. P ap er I I : To identify the prognostic value of progranulin and sortilin tumor expression in premenopausal breast cancer patients using a unique randomized tissue microarray cohort with long follow - up. P ap er I I I : To identify and characterize progranulin - induced secreted compounds that affect CSC activity and determine the clinical relevance of these factors, using functional assays, protein expression analysis, as well as an in vivo- like culturing system. P ap er I V : To delineate the role of cleaved progranulin peptides on CSC expansion and their reliance on sortilin receptor binding. P ap er V : To determine the sortilin down - modulating effect of various CADA molecules and identify the most potent candidate inhibiting progranulin - induced CSC propagation that can be selected for further studies in vivo. 14 Hypox ia Solid tumors, such as most breast cancers, often contain regions with low oxygen levels (hypoxia), due to insufficient vascularization, and is a part of the intricate tumor microenvironment [39, 78, 79 ] . Patients with these tumors often have a poor prognosis, showing more malignant and treatment- resistant properties due to the tumor cells adapting to the lower levels of oxygen, controll ed by hypoxia inducible factors (HIFs) [31, 79, 80] . HIFs, often called master transcriptional regulators of the hypoxic response, are involved in numerous processes, such as inducing proliferation, EMT, metastasis, survival, angiogenesis, invasion and metastasis, pH regulation, changes in metabolism and g lucose uptake, as well as the maintenance of stem cells [39 , 81] . In addition, HI Fs have been proposed as a potential therapy target for cancer [78 , 79] . HIF1 is a heterodimer consisting of an α- subunit that is only expressed at low oxygen concentrations (oxygen sensitive), and a β - subunit, which is constitutively active [39] . The transcription factor HIF1 α is universally expressed in t he body, while the other isoforms, HIF2 α and HIF 3 α, are only expressed in some tissues and at varying oxygen levels [79] . Under normoxic conditions (21 % O 2 ) , HI F1 has almost no activity, due to low levels of HI F1 α. When oxygen is present, oxygen - sensitizing enzymes (hydroxylases) facilitates HIF1 α hydroxylation on the oxygen - dependent - degradation domain on HIF1 α, leading to ubiquitination and degradation [79] . During hypoxia, HIF 1 α is stabilized and translocated to the nucleus, where it forms a dimer with HI F1β . The HI F1 complex then bind to different hypoxic response elements (HREs), and i nteracts with various co - activators to regulate transcrip tion of different target genes. 15 Aims The overall aim within the project is to elucidate the role of the microenvironmentally induced autocrine growth factor progranulin and its associated receptor sortilin in breast CSC propagation and prognosis prediction. Limiting breast cancer progression by targeting sortilin could be a potential therapeutic strategy for the treatment of patients with breast tumors having elevated progranulin and sortilin expression. The more specific aims for each paper are: P ap er I : To determine the influence of progranulin and its receptor sortilin on breast cancer propagation and CSC expansion using functional assays and in vivo mouse models. P ap er I I : To identify the prognostic value of progranulin and sortilin tumor expression in premenopausal breast cancer patients using a unique randomized tissue microarray cohort with long follow - up. P ap er I I I : To identify and characterize progranulin - induced secreted compounds that affect CSC activity and determine the clinical relevance of these factors, using functional assays, protein expression analysis, as well as an in vivo- like culturing system. P ap er I V : To delineate the role of cleaved progranulin peptides on CSC expansion and their reliance on sortilin receptor binding. P ap er V : To determine the sortilin down - modulating effect of various CADA molecules and identify the most potent candidate inhibiting progranulin - induced CSC propagation that can be selected for further studies in vivo. 16 17 Methods Tumor model systems In vitro models In this thesis, we have used various established cancer cell lines, including the (5αSRVLWLYHFHOOOLQHV0&)DQG7'DQGWKH(5αQHJDWLYHFHOOOLQHV CAL - 12 0, MDA - MB - 231 and MDA - MB - 468, as well as the non- malignant breast epithelial cell line MCF10a as a control. Established cell lines are commonly used for studying disease mechanism s, as they are inexpensive , easy to grow and maintain in cell culture and experiments can be performed with high throughput [82, 83] . In addition, cell lines allow the identificatio n of active substances and make it easy to control external factors, permitting higher reproducibility [82] . One of the drawbacks of using cell lines for studying tumor development and progression is that these models do not fully resemble real life situations, as they usually do not include surrounding factors ( e.g. environmental stimuli influencing tumor growth). Moreover, growing cel ls on plastic can also induce a selective pressure on the cells, altering gene expression and cell phenotype. For drug development, it is therefore important to add optimal model systems in the validation phase to remove less suitable candidates [82 , 83] . In vi vo animal models It is difficult to mimic real life situations in vitro using only molecular techniques and cell culturing. For that reason, we included mouse models in our study . When working with animals, you always have to keep in mind the principles of the 3Rs: Replacement , Reduction and Refinement , concerning the optimization of animal welfare and minimizing the use of animals in research. Ethics permissions for the use of animal models were obtained by the Research Animal Ethics Committee in Gothenburg. Mice are suitable models for research on various human diseases due to their similar anatomy and physiology, where 99% of the genes in mice have a human homolog [84] . Mice are relatively small, easy to house and have a short reproduction time. Nevertheless, they have a higher metabolic rate than h umans have, and might require a higher drug concentration to get the same biological effect. For the animal studies performed in this thesis, we used luciferase - tagged breast cancer cells that were injected subcutaneously into NOD/SCID (non - obese diabetic/severe combined immunodeficiency) gamma mice. These 16 17 Methods Tumor model systems In vitro models In this thesis, we have used various established cancer cell lines, including the (5αSRVLWLYHFHOOOLQHV0&)DQG7'DQGWKH(5αQHJDWLYHFHOOOLQHV CAL - 12 0, MDA - MB - 231 and MDA - MB - 468, as well as the non- malignant breast epithelial cell line MCF10a as a control. Established cell lines are commonly used for studying disease mechanism s, as they are inexpensive , easy to grow and maintain in cell culture and experiments can be performed with high throughput [82, 83] . In addition, cell lines allow the identificatio n of active substances and make it easy to control external factors, permitting higher reproducibility [82] . One of the drawbacks of using cell lines for studying tumor development and progression is that these models do not fully resemble real life situations, as they usually do not include surrounding factors ( e.g. environmental stimuli influencing tumor growth). Moreover, growing cel ls on plastic can also induce a selective pressure on the cells, altering gene expression and cell phenotype. For drug development, it is therefore important to add optimal model systems in the validation phase to remove less suitable candidates [82 , 83] . In vi vo animal models It is difficult to mimic real life situations in vitro using only molecular techniques and cell culturing. For that reason, we included mouse models in our study . When working with animals, you always have to keep in mind the principles of the 3Rs: Replacement , Reduction and Refinement , concerning the optimization of animal welfare and minimizing the use of animals in research. Ethics permissions for the use of animal models were obtained by the Research Animal Ethics Committee in Gothenburg. Mice are suitable models for research on various human diseases due to their similar anatomy and physiology, where 99% of the genes in mice have a human homolog [84] . Mice are relatively small, easy to house and have a short reproduction time. Nevertheless, they have a higher metabolic rate than h umans have, and might require a higher drug concentration to get the same biological effect. For the animal studies performed in this thesis, we used luciferase - tagged breast cancer cells that were injected subcutaneously into NOD/SCID (non - obese diabetic/severe combined immunodeficiency) gamma mice. These 18 mice are immunocompromised, meaning that they lack functional T - and B- lymphocytes, as well as having reduced NK cell and macrophage functions [85] . Assessments of tumor growth and metastasis were performed using an In Vivo Imaging Software (IVIS) whole body imager based on luciferase expression from stably transfected cell lines. The use of cell line derived xenografts or patient derived xenografts (PD X) , where you graft cell lines or tumor cells into immunodeficient mice, are commonly used for studying cancer [86, 87] . These models more accurately mimic human tumors and correlate better with treatment response in patients than in vitro models [88] . However, these models can be very time consuming to dev elop, and since these mice models do not have functional immune cells, it is not possible to study immune responses [89] . Recently, humanization of mouse models, using tumor infiltrating lymphocytes (TILs) and inflammatory cytokines ( e.g. IL - 2) have been developed, as well as the use of genetically engineered mouse models with intact immune systems, where you can study disease progression more accurately [90, 91] . In vi vo - like 3D models To try to minimize the use of animals when performing experiments, various in vivo- like 3D models have been developed, generating similar cell responses and even reflect clinical features of the original tumors [92, 93] . These models make it possible to answer important scientific questions related to human health and biology without the use of time - consuming and costly animal experiments. Matrigel is often used to create a 3D model, as it provides the cells with a 3D environment and ECM components important for cell adhesion and signaling. However, matrigel is a hydrogel with an undefined concentration and content, derived from mouse sarcoma basement membran es, making it hard to reproduce experiments [94] . More recently, various in vitro 3D tumor models have been developed, using tissues from different organs and tumors, such as scaffolds derived from cells, tissues or even bioprinted [ 95] . These models support the growth of cell cultures on 3D structures to study the interaction with the tumor microenvironment. They preserve cell interactions with the ECM in the tumor microenvironment and are implied to generate more robust data for predicting patient outcome and treatment responses [95] . In addition, they can be used for drug screening and provide information about the role of ECM remodeling and malignant properties of the cells affected by the microenvironment during cancer progression [96- 98] . In our research group, we have developed in vivo- like 3D model systems based on cell - free 19 patient - derived scaffolds (PDSs ) for breast cancer, and more recently a lso for colorectal cancers that can be used as drug - testing platforms [93, 99 , 100] . Experimental methods Functional cancer stem cell enriching assays To study CSC characteristics in cells, we performed in vitro mammosphere formation assays, to study non - adherent conditions of cells, originating from the neurosphere assay developed in the early 1990s [10 1] . Here, researchers were able to culture and identify cells with stem cell properties from the adult brain, where only cells surviving anoikis resistance were able to form spheres [47, 101 , 102] . These cells have the ability to differentiate and self- renew [101] . The mammosphere formation assay is a relatively time- consuming and slightly complicated assay that involves specific culturing requirements and training to perform and analyze correctly. However, compared to tumor - initiating studies in mice, this assay is an effective way to assess CSC activity in cell lines and tumors. Gene expression analysis (qPCR) To measure gene expression i n cells or tissues, various methods can be used, including real - time quantitative polymerase chain reaction (qPCR), microarray analysis, hybridization - based assays and various sequencing techniques . qPCR is the most common technique for gene expression measurement, studying biomarkers and validating microarray data [10 3] . In Paper I, real - time qPCR was performed on single cells or on bulk - level with various gene - specific primers for key regulators or markers for differentiation, proliferation and pluripotency, to define the existence of subpopulations of cells. RNA from cells were extracted, followed by reverse transcription of RNA to complementary DNA (cDNA). Single cell analysis requires cell sorting and pre amplification [10 4] . qPCR on cDNA samples were run with different gene - specific primers using a thermal amplification cycle program with SYBR green detection system, followed by gene expression data analysis on mRNA levels using GenEx TM ( MultiID) . Protein analysis In this thesis, we performed western blots for protein expression of various markers, validation and pathway analysis. Western blots are routinely used in various research fields and has multiple applications, including 18 mice are immunocompromised, meaning that they lack functional T - and B- lymphocytes, as well as having reduced NK cell and macrophage functions [85] . Assessments of tumor growth and metastasis were performed using an In Vivo Imaging Software (IVIS) whole body imager based on luciferase expression from stably transfected cell lines. The use of cell line derived xenografts or patient derived xenografts (PD X) , where you graft cell lines or tumor cells into immunodeficient mice, are commonly used for studying cancer [86, 87] . These models more accurately mimic human tumors and correlate better with treatment response in patients than in vitro models [88] . However, these models can be very time consuming to dev elop, and since these mice models do not have functional immune cells, it is not possible to study immune responses [89] . Recently, humanization of mouse models, using tumor infiltrating lymphocytes (TILs) and inflammatory cytokines ( e.g. IL - 2) have been developed, as well as the use of genetically engineered mouse models with intact immune systems, where you can study disease progression more accurately [90, 91] . In vi vo - like 3D models To try to minimize the use of animals when performing experiments, various in vivo- like 3D models have been developed, generating similar cell responses and even reflect clinical features of the original tumors [92, 93] . These models make it possible to answer important scientific questions related to human health and biology without the use of time - consuming and costly animal experiments. Matrigel is often used to create a 3D model, as it provides the cells with a 3D environment and ECM components important for cell adhesion and signaling. However, matrigel is a hydrogel with an undefined concentration and content, derived from mouse sarcoma basement membran es, making it hard to reproduce experiments [94] . More recently, various in vitro 3D tumor models have been developed, using tissues from different organs and tumors, such as scaffolds derived from cells, tissues or even bioprinted [ 95] . These models support the growth of cell cultures on 3D structures to study the interaction with the tumor microenvironment. They preserve cell interactions with the ECM in the tumor microenvironment and are implied to generate more robust data for predicting patient outcome and treatment responses [95] . In addition, they can be used for drug screening and provide information about the role of ECM remodeling and malignant properties of the cells affected by the microenvironment during cancer progression [96- 98] . In our research group, we have developed in vivo- like 3D model systems based on cell - free 19 patient - derived scaffolds (PDSs ) for breast cancer, and more recently a lso for colorectal cancers that can be used as drug - testing platforms [93, 99 , 100] . Experimental methods Functional cancer stem cell enriching assays To study CSC characteristics in cells, we performed in vitro mammosphere formation assays, to study non - adherent conditions of cells, originating from the neurosphere assay developed in the early 1990s [10 1] . Here, researchers were able to culture and identify cells with stem cell properties from the adult brain, where only cells surviving anoikis resistance were able to form spheres [47, 101 , 102] . These cells have the ability to differentiate and self- renew [101] . The mammosphere formation assay is a relatively time- consuming and slightly complicated assay that involves specific culturing requirements and training to perform and analyze correctly. However, compared to tumor - initiating studies in mice, this assay is an effective way to assess CSC activity in cell lines and tumors. Gene expression analysis (qPCR) To measure gene expression i n cells or tissues, various methods can be used, including real - time quantitative polymerase chain reaction (qPCR), microarray analysis, hybridization - based assays and various sequencing techniques . qPCR is the most common technique for gene expression measurement, studying biomarkers and validating microarray data [10 3] . In Paper I, real - time qPCR was performed on single cells or on bulk - level with various gene - specific primers for key regulators or markers for differentiation, proliferation and pluripotency, to define the existence of subpopulations of cells. RNA from cells were extracted, followed by reverse transcription of RNA to complementary DNA (cDNA). Single cell analysis requires cell sorting and pre amplification [10 4] . qPCR on cDNA samples were run with different gene - specific primers using a thermal amplification cycle program with SYBR green detection system, followed by gene expression data analysis on mRNA levels using GenEx TM ( MultiID) . Protein analysis In this thesis, we performed western blots for protein expression of various markers, validation and pathway analysis. Western blots are routinely used in various research fields and has multiple applications, including 20 protein abundance, protein- protein interaction, cellular location, kinase activity, post- translational modification (phosphorylation, glycosylation) [105] . It is a method used to separate proteins based on t heir molecular mass, using an electrical field. Proteins are then transferred onto a nitrocellulose membrane and proteins can be detected by the use of specific primary antibodies (binding to antigens or proteins directly). Primary antibody incubation is f ollowed by staining with a secondary antibody labelled with e.g. the enzyme horseradish peroxidase (HR P) , allowing signal amplification and detection by chemiluminescence [105] . This allows the detection of relative protein concentration at relatively high sensitivity and specificity. However, western blots are relatively time - consuming, as you generally study one protein at a time, and you need to know which protein you are looking for and have adequate antibodies. Other assays t o study protein expression are ELISAs (enzyme- linked immunosorbent assays), where you also look at only one protein at a time. ELISAs require a smaller volume and can be more sensitive than western blots. However, it is not possible to see differences in s ize of the proteins, e.g. if there are isoforms present. Protein arrays, on the other hand, allow you to study multiple target proteins in a single sample. Various protein targets or pathways can be identified and later verified using western blot. Mass sp ectrometry is another way of studying protein expression at high throughput. Mass spectrometry is expensive and requires a lot of equipment and large sample volumes, but can be used to confirm antibody specificity and determine protein interaction after immune - precipitation. In addition, it is semi - quantitative and can uncover post - translational modifications and detect isoforms. Proximity extension assays (PE As) are commonly used for cell media and different body fluids. The PEA is based on a set of DNA oligo- labeled antibodies (probes) that when in proximity to their target proteins, they will bind and hybridize. Using a DNA polymerase, the probes are extended and you can then perform pre - amplification of the probes and quantify the leve ls using a qPCR detection system. This offers a high throughput and requires only a small sample volume. On the other hand, you need to use defined panels and can only get relative protein values. 21 Patient data and analysis Tissue microarrays (TMAs) are often used to study patient data, where you can analyze protein expression and potential biomarkers in relation to well- defined patient subgroups. TMAs are a collection of multiple tumor tissue cores, allowing the study of tumors from many different pa tients at the same time. IHC staining can be used to assess hormone status and proliferation status of patients, as well as detection of a biomarker of interest. These biomarkers can potentially be used as prognostic markers or treatment- predictive indicators for various patients. In Paper II, we used a TMA from a randomized clinical trial including 444 premenopausal breast cancer patients. These patients were diagnosed between 1984- 1 99 1 and received either two years of adjuvant tamoxifen treatment (n=21 2) or no systemic treatment (n=23 2). All patients were followed up for up to 30 years. This study was approved by the Ethics Committees of Linköping and Lund Universities in Sweden. Data for patient follow up were taken from the Swedish Causes of Death regist er. The TMAs were stained for progranulin and sortilin tissue expression to study their relation to other clinical markers and patient outcome. The survival analysis includes time to an event data, e.g. time from initial breast cancer diagnosis to disease - specific death. In this thesis, we study breast cancer- specific survival (BCSS). Kaplan - Meier survival curves are commonly used to estimate the survival probability of an individual over time, providing a summary of the data, as well as an estimation of the median survival time [106] . Hazard ratio (HR) measures the relative survival between two groups (the risk of surviving) [106] . A HR equal to one means that there is no difference in survival between the two groups, while HR above one means an increas ed mortality (less likely to survive at an indicated time). Moreover, performing multivariable analyses are central, as they take into consideration other factors (covariates) that might affect patient prognosis (survival) [10 7] . The Cox proportional hazards (CPH) model is a regression model analyzing survival time data in relation to the effects of a set of covariates . 20 protein abundance, protein- protein interaction, cellular location, kinase activity, post- translational modification (phosphorylation, glycosylation) [105] . It is a method used to separate proteins based on t heir molecular mass, using an electrical field. Proteins are then transferred onto a nitrocellulose membrane and proteins can be detected by the use of specific primary antibodies (binding to antigens or proteins directly). Primary antibody incubation is f ollowed by staining with a secondary antibody labelled with e.g. the enzyme horseradish peroxidase (HR P) , allowing signal amplification and detection by chemiluminescence [105] . This allows the detection of relative protein concentration at relatively high sensitivity and specificity. However, western blots are relatively time - consuming, as you generally study one protein at a time, and you need to know which protein you are looking for and have adequate antibodies. Other assays t o study protein expression are ELISAs (enzyme- linked immunosorbent assays), where you also look at only one protein at a time. ELISAs require a smaller volume and can be more sensitive than western blots. However, it is not possible to see differences in s ize of the proteins, e.g. if there are isoforms present. Protein arrays, on the other hand, allow you to study multiple target proteins in a single sample. Various protein targets or pathways can be identified and later verified using western blot. Mass sp ectrometry is another way of studying protein expression at high throughput. Mass spectrometry is expensive and requires a lot of equipment and large sample volumes, but can be used to confirm antibody specificity and determine protein interaction after immune - precipitation. In addition, it is semi - quantitative and can uncover post - translational modifications and detect isoforms. Proximity extension assays (PE As) are commonly used for cell media and different body fluids. The PEA is based on a set of DNA oligo- labeled antibodies (probes) that when in proximity to their target proteins, they will bind and hybridize. Using a DNA polymerase, the probes are extended and you can then perform pre - amplification of the probes and quantify the leve ls using a qPCR detection system. This offers a high throughput and requires only a small sample volume. On the other hand, you need to use defined panels and can only get relative protein values. 21 Patient data and analysis Tissue microarrays (TMAs) are often used to study patient data, where you can analyze protein expression and potential biomarkers in relation to well- defined patient subgroups. TMAs are a collection of multiple tumor tissue cores, allowing the study of tumors from many different pa tients at the same time. IHC staining can be used to assess hormone status and proliferation status of patients, as well as detection of a biomarker of interest. These biomarkers can potentially be used as prognostic markers or treatment- predictive indicators for various patients. In Paper II, we used a TMA from a randomized clinical trial including 444 premenopausal breast cancer patients. These patients were diagnosed between 1984- 1 99 1 and received either two years of adjuvant tamoxifen treatment (n=21 2) or no systemic treatment (n=23 2). All patients were followed up for up to 30 years. This study was approved by the Ethics Committees of Linköping and Lund Universities in Sweden. Data for patient follow up were taken from the Swedish Causes of Death regist er. The TMAs were stained for progranulin and sortilin tissue expression to study their relation to other clinical markers and patient outcome. The survival analysis includes time to an event data, e.g. time from initial breast cancer diagnosis to disease - specific death. In this thesis, we study breast cancer- specific survival (BCSS). Kaplan - Meier survival curves are commonly used to estimate the survival probability of an individual over time, providing a summary of the data, as well as an estimation of the median survival time [106] . Hazard ratio (HR) measures the relative survival between two groups (the risk of surviving) [106] . A HR equal to one means that there is no difference in survival between the two groups, while HR above one means an increas ed mortality (less likely to survive at an indicated time). Moreover, performing multivariable analyses are central, as they take into consideration other factors (covariates) that might affect patient prognosis (survival) [10 7] . The Cox proportional hazards (CPH) model is a regression model analyzing survival time data in relation to the effects of a set of covariates . 22 23 R esults and discussion Paper I - Sortilin inhibition limits secretion- induced progranulin- dependent breast cancer progression and cancer stem cell expansion Cell- to- cell communication and signaling through secretion of various cytokines and growth factors in the tumor microenvironment can drive tumor progression and influence treatment responses. Similarly, hypoxia is common in solid tumors like breast cancer and influence cancer cells and their secretion. The main hypothesis within this project is that hypoxia induces secretion of proteins and growth factors in the tumor cells, spreading a CSC propagating signal to its surrounding microenvironment. CSCs are th ought to be important drivers of tumor progression and treatment resistance [38 , 108] . Published data from our research group demonstrates that hypoxia increases the amount of &6&VLQ(5αSRVLWLYH breast cancer whereas it in contrast decreased the CSC - IUDFWLRQ LQ(5α negative disease [41, 10 9] . For that reason, targeting this aggressive subpopulation of cancer cells could be an appealing therapy to improve patient outcome. In this paper, we aimed to elucidate the role of th e secreted factor progranulin and its receptor sortilin in breast cancer propagation. Different in vitro and in vivo relevant conditions were used to validate breast CSC expansion mediated by progranulin, through its receptor sortilin. A hypoxic tumor micr oenvironment induces secretion of components stimulating cancer stem cell activation In order to study the influence of the microenvironment, and more specifically how a hypoxic environment induce secretion and affect breast CSC activity, we treated different breast cancer cell lines with conditioned PHGLD IURP (5α SRVLWLYH EUHDVW FDQFHU FHOOV RU SULPDU\ EUHDVW FDQFHU explants cultured in hypoxia. We then examined the mammosphere forming potential of the cell lines treated with the hypoxic conditioned media. Results showed an increase in the mammosphere forming capacity RIERWK(5αSRVLWLYHDQG(5αnegative cell lines with hypoxic conditioned media compared to cells treated with normoxic conditioned media, suggesting that the hypoxic - induced secreted m icroenvironment from the (5αpositive conditioned media induced CSC propagation in various breast FDQFHUFHOOOLQHV,QFRQWUDVWWKH(5αQHJDWLYHK\SR[LFFRQGLWLRQPHGLDKDV 22 23 R esults and discussion Paper I - Sortilin inhibition limits secretion- induced progranulin- dependent breast cancer progression and cancer stem cell expansion Cell- to- cell communication and signaling through secretion of various cytokines and growth factors in the tumor microenvironment can drive tumor progression and influence treatment responses. Similarly, hypoxia is common in solid tumors like breast cancer and influence cancer cells and their secretion. The main hypothesis within this project is that hypoxia induces secretion of proteins and growth factors in the tumor cells, spreading a CSC propagating signal to its surrounding microenvironment. CSCs are th ought to be important drivers of tumor progression and treatment resistance [38 , 108] . Published data from our research group demonstrates that hypoxia increases the amount of &6&VLQ(5αSRVLWLYH breast cancer whereas it in contrast decreased the CSC - IUDFWLRQ LQ(5α negative disease [41, 10 9] . For that reason, targeting this aggressive subpopulation of cancer cells could be an appealing therapy to improve patient outcome. In this paper, we aimed to elucidate the role of th e secreted factor progranulin and its receptor sortilin in breast cancer propagation. Different in vitro and in vivo relevant conditions were used to validate breast CSC expansion mediated by progranulin, through its receptor sortilin. A hypoxic tumor micr oenvironment induces secretion of components stimulating cancer stem cell activation In order to study the influence of the microenvironment, and more specifically how a hypoxic environment induce secretion and affect breast CSC activity, we treated different breast cancer cell lines with conditioned PHGLD IURP (5α SRVLWLYH EUHDVW FDQFHU FHOOV RU SULPDU\ EUHDVW FDQFHU explants cultured in hypoxia. We then examined the mammosphere forming potential of the cell lines treated with the hypoxic conditioned media. Results showed an increase in the mammosphere forming capacity RIERWK(5αSRVLWLYHDQG(5αnegative cell lines with hypoxic conditioned media compared to cells treated with normoxic conditioned media, suggesting that the hypoxic - induced secreted m icroenvironment from the (5αpositive conditioned media induced CSC propagation in various breast FDQFHUFHOOOLQHV,QFRQWUDVWWKH(5αQHJDWLYHK\SR[LFFRQGLWLRQPHGLDKDV 24 been shown to decrease the mammosphere forming capacity of the cells, independent on hormone status [41, 10 9] . Progranulin influences cancer stem cell propagation both in vitro and in vivo by inducing mammosphere formation and metastasis formation in mice To define secreted factors within the conditioned media resulting in CSC spreading, D F\WRNLQH VFUHHQXVLQJ FRQGLWLRQHGPHGLD IURPK\SR[LF(5α positive breast cancer cells (T47D and MCF7) was performed. From this screen a total of 507 cytokines and proteins were examined and we identified progranulin as being significantly upregulated in the hypoxic (5αSRVLWLYHFRQGLWLRQHGPHGLD,QGHHGSURJUDQXOLQWXUQHGRXWWREHRQH of the main mediators of CSC activation, shown by an incre ased mammosphere forming capacity in both progranulin - treated MCF7 and MDA - MB - 231 cells. This implies that progranulin is driving CSC SURSDJDWLRQLQERWK(5αQHJDWLYHDQGSRVLWLYHEUHDVWFDQFHU Next, we assessed if progranulin could influence breast cance r growth and progression in vivo by performing repetitive injections of progranulin in tumor- bearing mice using a luciferase - tagged MDA - MB 231 (or T47D ) breast cancer cell line xenograft. After 3 weeks of treatment during xenograft growth, there were no significant difference in the tumor burden between progranulin - injected mice and the vehicle control. However, a significant increase in lung metastasis in mice subjected to progranulin injections were observed, demonstrating that progranulin induce tumor progression and mediate a more metastasizing cellular subtype. In addition, when studying tumor initiation, cells pretreated with progranulin were injected into mice in a serial dilution f ormat and required a lower concentration of cells for tumor initiation, suggesting a higher CSC frequency in these cells. Progranulin is acting through the receptor sortilin Previously, researchers have observed an increase in extracellular levels of prog ranulin in the brain after inhibiting sortilin , suggesting that sortilin regulates brain progranulin levels and can be used to treat dementia caused by progranulin haploinsufficiency [11 0, 11 1] . Sortilin has been reported to be overexpressed in various cancer types, including melanoma, chronic lymphoc ytic leukemia (CLL), breast and ovarian carcinomas [62] . Sortilin overexpression is linked to proliferation, migration and invasion in these cancer types, and anti - sortilin antibodies have been proposed as a 25 therapy option inducing apoptosis in CLL [62] . In prostate cancer, reports have shown that progranulin binds to sortilin, leading to progranulin internalization and degradation, suppressing progranulin - induced proliferation, migration and ancho r- independent growth [112] . However, in ovarian carcinoma, sortilin knockdown induces apoptosis and reduces proliferation [113] , implying t hat most studies regarding sortilin in cancer link high levels of sortilin with bad prognosis. To illuminate the role of sortilin in progranulin - induced breast cancer spreading we assessed several ways of targeting sortilin with the aim to inhibit the prog ranulin induced CSC propagation in vitro, using different breast cancer cell lines. By targeting sortilin using small interfering RNA (siRNA) for sortilin, a sortilin degrader (1- [ 2 - ( 2 - tert- butyl- 5 - methylphenoxy)- ethyl] - 3 - methyl - piperidine, termed MPE P) or by pharmacological inhibition of sortilin using a small sortilin- binding compound (AF38469 [ 114] ) , we were able to inhibit the progranulin mediated mammosphere forming capacity of MDA - MB 231 cells. Similar results could be observed using MCF7, T47D and CAL - 12 0 cell lin es. To test if sortilin inhibition had the same effect on the progranulin - induced cancer progression in vivo, we used a more potent part of the progranulin protein, named granulin A, as it had a more pronounced effect in vivo compared to full - length progranulin. Importantly, results show ed that the granulin A - induced increase in lung metastasis could be blocked by AF38 46 9. Combined, these results demonstrate that sortilin is a functional receptor of progranulin and is responsi ble for driving progranulin - induced breast CSC propagation in vitro, leading to tumor progression and metastasis, as shown in vivo. 24 been shown to decrease the mammosphere forming capacity of the cells, independent on hormone status [41, 10 9] . Progranulin influences cancer stem cell propagation both in vitro and in vivo by inducing mammosphere formation and metastasis formation in mice To define secreted factors within the conditioned media resulting in CSC spreading, D F\WRNLQH VFUHHQXVLQJ FRQGLWLRQHGPHGLD IURPK\SR[LF(5α positive breast cancer cells (T47D and MCF7) was performed. From this screen a total of 507 cytokines and proteins were examined and we identified progranulin as being significantly upregulated in the hypoxic (5αSRVLWLYHFRQGLWLRQHGPHGLD,QGHHGSURJUDQXOLQWXUQHGRXWWREHRQH of the main mediators of CSC activation, shown by an incre ased mammosphere forming capacity in both progranulin - treated MCF7 and MDA - MB - 231 cells. This implies that progranulin is driving CSC SURSDJDWLRQLQERWK(5αQHJDWLYHDQGSRVLWLYHEUHDVWFDQFHU Next, we assessed if progranulin could influence breast cance r growth and progression in vivo by performing repetitive injections of progranulin in tumor- bearing mice using a luciferase - tagged MDA - MB 231 (or T47D ) breast cancer cell line xenograft. After 3 weeks of treatment during xenograft growth, there were no significant difference in the tumor burden between progranulin - injected mice and the vehicle control. However, a significant increase in lung metastasis in mice subjected to progranulin injections were observed, demonstrating that progranulin induce tumor progression and mediate a more metastasizing cellular subtype. In addition, when studying tumor initiation, cells pretreated with progranulin were injected into mice in a serial dilution f ormat and required a lower concentration of cells for tumor initiation, suggesting a higher CSC frequency in these cells. Progranulin is acting through the receptor sortilin Previously, researchers have observed an increase in extracellular levels of prog ranulin in the brain after inhibiting sortilin , suggesting that sortilin regulates brain progranulin levels and can be used to treat dementia caused by progranulin haploinsufficiency [11 0, 11 1] . Sortilin has been reported to be overexpressed in various cancer types, including melanoma, chronic lymphoc ytic leukemia (CLL), breast and ovarian carcinomas [62] . Sortilin overexpression is linked to proliferation, migration and invasion in these cancer types, and anti - sortilin antibodies have been proposed as a 25 therapy option inducing apoptosis in CLL [62] . In prostate cancer, reports have shown that progranulin binds to sortilin, leading to progranulin internalization and degradation, suppressing progranulin - induced proliferation, migration and ancho r- independent growth [112] . However, in ovarian carcinoma, sortilin knockdown induces apoptosis and reduces proliferation [113] , implying t hat most studies regarding sortilin in cancer link high levels of sortilin with bad prognosis. To illuminate the role of sortilin in progranulin - induced breast cancer spreading we assessed several ways of targeting sortilin with the aim to inhibit the prog ranulin induced CSC propagation in vitro, using different breast cancer cell lines. By targeting sortilin using small interfering RNA (siRNA) for sortilin, a sortilin degrader (1- [ 2 - ( 2 - tert- butyl- 5 - methylphenoxy)- ethyl] - 3 - methyl - piperidine, termed MPE P) or by pharmacological inhibition of sortilin using a small sortilin- binding compound (AF38469 [ 114] ) , we were able to inhibit the progranulin mediated mammosphere forming capacity of MDA - MB 231 cells. Similar results could be observed using MCF7, T47D and CAL - 12 0 cell lin es. To test if sortilin inhibition had the same effect on the progranulin - induced cancer progression in vivo, we used a more potent part of the progranulin protein, named granulin A, as it had a more pronounced effect in vivo compared to full - length progranulin. Importantly, results show ed that the granulin A - induced increase in lung metastasis could be blocked by AF38 46 9. Combined, these results demonstrate that sortilin is a functional receptor of progranulin and is responsi ble for driving progranulin - induced breast CSC propagation in vitro, leading to tumor progression and metastasis, as shown in vivo. 26 Paper I I - T umor co- expression of progranulin and sortilin as a prognostic biomarker in breast cancer Each year, almost 1400 women in Sweden dies from breast cancer [2] . The survival rate from breast cancer is relatively high. Nonetheless, many patients experience therapy re sistance and tumor relapse. As breast cancer is the most common cancer in women, with more than 8000 diagnosed in Sweden every year, early detection and identification of robust clinical markers are central to increase treatment efficiency and patient surv ival. In Paper I, we established the role of progranulin and sortilin on breast CSC propagation, tumor progression and metastasis [115] . In paper II, we determined the clinical impact of progranulin and sortilin tumor expression in breast cancer and investigated if these markers can be used as prognostic or treatment - predictive biomarkers for breast cancer patients. To evaluate whether progranulin and sortilin tumor expression could be used as biomarkers in breast cancer, we analyzed protein expression data from a TMA consisting of 560 premenopausal breast cancer patients with long follow - up time. These patients were randomized and given either two years of tamoxifen treatment or no adjuvant therapy. The samples were stained for progranulin and sortilin tumor tissue expression using IHC, then scored for high or low expression and evaluated in relation to various clinical markers and patient outcomes. In this breast cancer cohort, more WKDQRIWKHSDWLHQWVZHUHSRVLWLYHIRU(5αZKLFKLVUHSUHVHQWDWLYHIRU the whole population of breast cancer patients [28] . Progranulin and sortilin scoring and association with clinical parameters Progranulin and sortilin tissue expression were evaluated by IHC, using specific antibodies for progranulin and sortilin. Tissue microarrays from 444 patients with good enough material and clinical data were successfully stained and selected for further analysis. The staining for each patient were then given a s core 1 - 4, where l - 2 were seen as a low expression of the markers and 3- 4 as high expression of the markers. In this cohort, 50 % of the patients were categorized as having high sortilin tissue expression, and the other 50% as having low expression. For prog ranulin, 66% were categorized in the high expression group, and 34% in the low expression group. 27 Interestingly, there was a positive correlation with the expression of progranulin and sortilin, where patients with high expression of progranulin also tended to have a higher expression of sortilin. In addition, high progranulin expression significantly correlated with high grade, the SUROLIHUDWLRQPDUNHU.LDQGWKHK\SR[LFPDUNHU+,)αDOODVVRFLDWHG with CSC characteristics, tumor aggressiveness, resistan ce and formation of the pre - metastatic niche in solid cancers [19, 116 , 11 7] . This indicates that tumors with high expression of progranulin tend to be more aggressive. )XUWKHU(5αSRVLWLYH WXPRUVKDG ORZHU H[SUHVVLRQ RI SURJUDQXOLQ WKDQ (5αQHJDWLYHWXPRUV7KLVFRUUHODWHVZLWKUHVXOWVIURPRWKHUUHVHDUFKHUV [ 118] , as well as from Paper I where secretion of progranulin were higher in MDA - MB - 231 than in MCF7 cells. In contrast, sortilin expression was KLJKHU LQ (5α SRVLWLYH WXPRUV DQG KLJK VRUWLOLQ H[SUHVVLRQ FRUUHODWHG negatively with age. Patients with high co - expression of progranulin and sortilin have worse breast cancer - specific survival In addition to the clinical relevance for progranulin as a prognostic and predictive biomarkers in various types of cancer [58, 11 9- 121] , sortilin has been shown to be overexpressed various cancer types, including prostate, ovarian and breast cancers [62] . As recent work in our lab (Paper I), as well as that other researchers have reported sortilin as a functional receptor for progranulin and its link to more clinically aggressive properties [112, 11 5, 122] , we set out to determine if a combination of both progranulin and sortilin tumor expression were related to BCSS in our patient cohort . Dual tissue expression of progranulin and sortilin could be scored in 395 of the 444 tumors, where 20% of these were double high, expressing both high progranulin and hig h sortilin levels. When comparing the double high group against tumors with variable expression levels of both progranulin and sortilin, we observed a significantly decreased BCSS in the double high group compared to the mixed expression group for all pati ents in the cohort. In addition, this effect was also observed when looking at patients not receiving initial tamoxifen treatment, providing prognostic information independent of treatment interference. Only 14% of the patients had high levels of progranul in and low levels of sortilin, showing that most of the patients with high progranulin levels also had high sortilin levels. When analyzing the groups with various expression levels of sortilin and progranulin separately, the difference in BCSS between dou ble high and high single progranulin expression was not 26 Paper I I - T umor co- expression of progranulin and sortilin as a prognostic biomarker in breast cancer Each year, almost 1400 women in Sweden dies from breast cancer [2] . The survival rate from breast cancer is relatively high. Nonetheless, many patients experience therapy re sistance and tumor relapse. As breast cancer is the most common cancer in women, with more than 8000 diagnosed in Sweden every year, early detection and identification of robust clinical markers are central to increase treatment efficiency and patient surv ival. In Paper I, we established the role of progranulin and sortilin on breast CSC propagation, tumor progression and metastasis [115] . In paper II, we determined the clinical impact of progranulin and sortilin tumor expression in breast cancer and investigated if these markers can be used as prognostic or treatment - predictive biomarkers for breast cancer patients. To evaluate whether progranulin and sortilin tumor expression could be used as biomarkers in breast cancer, we analyzed protein expression data from a TMA consisting of 560 premenopausal breast cancer patients with long follow - up time. These patients were randomized and given either two years of tamoxifen treatment or no adjuvant therapy. The samples were stained for progranulin and sortilin tumor tissue expression using IHC, then scored for high or low expression and evaluated in relation to various clinical markers and patient outcomes. In this breast cancer cohort, more WKDQRIWKHSDWLHQWVZHUHSRVLWLYHIRU(5αZKLFKLVUHSUHVHQWDWLYHIRU the whole population of breast cancer patients [28] . Progranulin and sortilin scoring and association with clinical parameters Progranulin and sortilin tissue expression were evaluated by IHC, using specific antibodies for progranulin and sortilin. Tissue microarrays from 444 patients with good enough material and clinical data were successfully stained and selected for further analysis. The staining for each patient were then given a s core 1 - 4, where l - 2 were seen as a low expression of the markers and 3- 4 as high expression of the markers. In this cohort, 50 % of the patients were categorized as having high sortilin tissue expression, and the other 50% as having low expression. For prog ranulin, 66% were categorized in the high expression group, and 34% in the low expression group. 27 Interestingly, there was a positive correlation with the expression of progranulin and sortilin, where patients with high expression of progranulin also tended to have a higher expression of sortilin. In addition, high progranulin expression significantly correlated with high grade, the SUROLIHUDWLRQPDUNHU.LDQGWKHK\SR[LFPDUNHU+,)αDOODVVRFLDWHG with CSC characteristics, tumor aggressiveness, resistan ce and formation of the pre - metastatic niche in solid cancers [19, 116 , 11 7] . This indicates that tumors with high expression of progranulin tend to be more aggressive. )XUWKHU(5αSRVLWLYH WXPRUVKDG ORZHU H[SUHVVLRQ RI SURJUDQXOLQ WKDQ (5αQHJDWLYHWXPRUV7KLVFRUUHODWHVZLWKUHVXOWVIURPRWKHUUHVHDUFKHUV [ 118] , as well as from Paper I where secretion of progranulin were higher in MDA - MB - 231 than in MCF7 cells. In contrast, sortilin expression was KLJKHU LQ (5α SRVLWLYH WXPRUV DQG KLJK VRUWLOLQ H[SUHVVLRQ FRUUHODWHG negatively with age. Patients with high co - expression of progranulin and sortilin have worse breast cancer - specific survival In addition to the clinical relevance for progranulin as a prognostic and predictive biomarkers in various types of cancer [58, 11 9- 121] , sortilin has been shown to be overexpressed various cancer types, including prostate, ovarian and breast cancers [62] . As recent work in our lab (Paper I), as well as that other researchers have reported sortilin as a functional receptor for progranulin and its link to more clinically aggressive properties [112, 11 5, 122] , we set out to determine if a combination of both progranulin and sortilin tumor expression were related to BCSS in our patient cohort . Dual tissue expression of progranulin and sortilin could be scored in 395 of the 444 tumors, where 20% of these were double high, expressing both high progranulin and hig h sortilin levels. When comparing the double high group against tumors with variable expression levels of both progranulin and sortilin, we observed a significantly decreased BCSS in the double high group compared to the mixed expression group for all pati ents in the cohort. In addition, this effect was also observed when looking at patients not receiving initial tamoxifen treatment, providing prognostic information independent of treatment interference. Only 14% of the patients had high levels of progranul in and low levels of sortilin, showing that most of the patients with high progranulin levels also had high sortilin levels. When analyzing the groups with various expression levels of sortilin and progranulin separately, the difference in BCSS between dou ble high and high single progranulin expression was not 28 significant. However, when using multivariate analysis and studying the high progranulin group alone, we detected that high sortilin expression in this group was related to lower BCSS. This implies th at sortilin adds prognostic information on survival when combined with progranulin and that there is a link between progranulin and sortilin expression. We therefore hypothesize that targeting the progranulin - sortilin axis may be a therapeutic option for t his patient group with high co - expression of progranulin and sortilin. To further strengthen our results, we performed multivariate CPH regression analysis adjusting for known prognostic factors, such as histological grade, tumor size, age, lymph node staWXV (5α VWDWXV DQG treatment, in addition to progranulin and sortilin score. In these analyses, we identified co - expression progranulin and sortilin, together with high tumor grade and lymph node positivity as independent prognostic covariates, associated with high risk factors giving a lower BCSS. )XUWKHUPRUH ZKHQ DQDO\]LQJ WKH (5α SRVLWLYH EUHDVW FDQFHU SDWLHQWV separately to study the tamoxifen treatment effect, our results indicates that progranulin expression is not associated with tamoxifen resistan ce. These results are in contrast to previous reports suggesting a link between SURJUDQXOLQ H[SUHVVLRQDQG WDPR[LIHQ UHVLVWDQFH LQ(5αSRVLWLYH EUHDVW cancer patients [123] . However, patients in this cohort only received tamoxifen treatment for two years, opposed to the now recommended five to ten years [124, 125] , implying that further studies are needed in order to determine the role of progranulin in tamoxifen treatment resistance. In addition, any following treatment decisions after the initial trial start are not taken into consideration, which may have an impact on the results. To further validate and strengthen our results, an external cohort should be added to make sure the model can be applied to other datasets before the use in clinical practice. Off note, this cohort only included premenopausal breast cancer patients and therefore may not be representative of the whole population. Collectively, these results suggest that co- expression of progranulin and sortilin defines a highly malignant subgroup of breast cancer patients, and that targeting progranulin through its receptor sortilin could be a potential novel breast cancer therapeutic approach in addition to conventional treatment strategies. 29 Paper I II - Interleukin- 6 induces stem cell propagation through liaison with the sortilin- progranulin axis in breast cancer The tumor microenvironment plays an important role in tumor progression and therapy resistance, where cell - to- cell communication is affected by secretion of signaling molecules and interactions with the surrounding tumor stroma [38, 126] . To further study the effect of progranulin on cancer cells, we aimed to elucidate the role of progranulin - induced secretion of cytokines and growth factors on the CSC population in breast cancer. Progranulin- induced secretion in breast cancer cells Our research group is focusing on the crosstalk between the tumor microenvironment and the cancer cells, including the impact of cell secretion. In this study , we intended to explore how progranulin affects secretion from our breast cancer cell lines. In addition, we examined if the enhancing effects on the CSC population caused by progranulin - induced secretion could be prevented by the use of sort ilin modulators. This was accomplished by analyzing conditioned media from breast cancer cell lines treated with progranulin, the sortilin inhibitor AF38 469, or a combination of both, using a proximity extension assay ( PEA ) performed by OLINK at SciLifeLab in Uppsala. Panels containing cardiovascular and immuno - oncology markers were chosen for this high throughput multiplex assay due to their relevance and involvement in tumor biology. This allowed us to study almost 200 different proteins. The secretion sc reen revealed distinct secretion profiles between the treatments in MCF7 and MDA - MB - 231 cells. In particular, progranulin treatment increased the secretion of IL - 6 and IL - 8, as well as other cytokines involved in inflammation, metastasis and CSC formation, including tumor necrosis factor (TN F) , CXCL1, Fas ligand (FASL) [116, 127- 13 3] . Secretion profiles and cytokines induced by the other treatment combinations should also be studied in more detail. Crosstalk between progranulin and IL- 6 expression in breast cancer cells Further, to validate if progranulin - induced secretion of IL - 6 and IL - 8 also affected the internal expression of these interleukins. Western blot analysis on cell lysates confirmed that progranulin treatment induced protein expression of both IL - 6 and IL - 8 in a dose - dependent manner. In 28 significant. However, when using multivariate analysis and studying the high progranulin group alone, we detected that high sortilin expression in this group was related to lower BCSS. This implies th at sortilin adds prognostic information on survival when combined with progranulin and that there is a link between progranulin and sortilin expression. We therefore hypothesize that targeting the progranulin - sortilin axis may be a therapeutic option for t his patient group with high co - expression of progranulin and sortilin. To further strengthen our results, we performed multivariate CPH regression analysis adjusting for known prognostic factors, such as histological grade, tumor size, age, lymph node staWXV (5α VWDWXV DQG treatment, in addition to progranulin and sortilin score. In these analyses, we identified co - expression progranulin and sortilin, together with high tumor grade and lymph node positivity as independent prognostic covariates, associated with high risk factors giving a lower BCSS. )XUWKHUPRUH ZKHQ DQDO\]LQJ WKH (5α SRVLWLYH EUHDVW FDQFHU SDWLHQWV separately to study the tamoxifen treatment effect, our results indicates that progranulin expression is not associated with tamoxifen resistan ce. These results are in contrast to previous reports suggesting a link between SURJUDQXOLQ H[SUHVVLRQDQG WDPR[LIHQ UHVLVWDQFH LQ(5αSRVLWLYH EUHDVW cancer patients [123] . However, patients in this cohort only received tamoxifen treatment for two years, opposed to the now recommended five to ten years [124, 125] , implying that further studies are needed in order to determine the role of progranulin in tamoxifen treatment resistance. In addition, any following treatment decisions after the initial trial start are not taken into consideration, which may have an impact on the results. To further validate and strengthen our results, an external cohort should be added to make sure the model can be applied to other datasets before the use in clinical practice. Off note, this cohort only included premenopausal breast cancer patients and therefore may not be representative of the whole population. Collectively, these results suggest that co- expression of progranulin and sortilin defines a highly malignant subgroup of breast cancer patients, and that targeting progranulin through its receptor sortilin could be a potential novel breast cancer therapeutic approach in addition to conventional treatment strategies. 29 Paper I II - Interleukin- 6 induces stem cell propagation through liaison with the sortilin- progranulin axis in breast cancer The tumor microenvironment plays an important role in tumor progression and therapy resistance, where cell - to- cell communication is affected by secretion of signaling molecules and interactions with the surrounding tumor stroma [38, 126] . To further study the effect of progranulin on cancer cells, we aimed to elucidate the role of progranulin - induced secretion of cytokines and growth factors on the CSC population in breast cancer. Progranulin- induced secretion in breast cancer cells Our research group is focusing on the crosstalk between the tumor microenvironment and the cancer cells, including the impact of cell secretion. In this study , we intended to explore how progranulin affects secretion from our breast cancer cell lines. In addition, we examined if the enhancing effects on the CSC population caused by progranulin - induced secretion could be prevented by the use of sort ilin modulators. This was accomplished by analyzing conditioned media from breast cancer cell lines treated with progranulin, the sortilin inhibitor AF38 469, or a combination of both, using a proximity extension assay ( PEA ) performed by OLINK at SciLifeLab in Uppsala. Panels containing cardiovascular and immuno - oncology markers were chosen for this high throughput multiplex assay due to their relevance and involvement in tumor biology. This allowed us to study almost 200 different proteins. The secretion sc reen revealed distinct secretion profiles between the treatments in MCF7 and MDA - MB - 231 cells. In particular, progranulin treatment increased the secretion of IL - 6 and IL - 8, as well as other cytokines involved in inflammation, metastasis and CSC formation, including tumor necrosis factor (TN F) , CXCL1, Fas ligand (FASL) [116, 127- 13 3] . Secretion profiles and cytokines induced by the other treatment combinations should also be studied in more detail. Crosstalk between progranulin and IL- 6 expression in breast cancer cells Further, to validate if progranulin - induced secretion of IL - 6 and IL - 8 also affected the internal expression of these interleukins. Western blot analysis on cell lysates confirmed that progranulin treatment induced protein expression of both IL - 6 and IL - 8 in a dose - dependent manner. In 30 line with this, other researchers have observed that treatment with IL - 6 in liver and bile duct cancer increased the expression of progranulin in vitro, as well as that progranulin - induced production of IL - 6 in adipose tissue [134- 13 6] . However, to our knowledge, no feedback loop have been documented in the same system. Importantly , treating bre ast cancer cell lines with increasing concentrations of IL - 6 or IL - 8 elevated the levels of progranulin in the cells in a dose - dependent manner, suggesting a crosstalk between the expression of progranulin and IL - 6 , and IL - 8. IL- 6- induced CSC propagation and its dependence on sortilin IL - 6 and IL - 8 have both been associated with the induction of CSC characteristics in cancer [71, 13 7] . Therefore, we wanted to explore if IL - 6 and IL - 8 affect the amount of CSCs in MCF7 and MDA - MB - 23 1 breast cancer cells. When treating the cells with recombinant IL - 6 or IL - 8, we observed an increase in mammosphere formation, similar to what we could see with progranulin. This indicates that both IL - 6 and IL - 8 mediate breast CSC expansion in breast cancer. Interestingly, sortilin has also been suggested as a high - affinity receptor for IL - 6 in immune cells [64, 138] . Using a FPA , we were able to show that IL - 6 outcompeted the fluorescently labeled neurotensin bound to the soluble sortilin receptor , confirming this interaction. However, IL- 8 did not outcompete neurotensin, hence did not bind to sortilin at this binding site. Next, we explored if the CSC effect caused by the cytokine treatment was dependent on, or regulated by the binding to sortilin. Results showed that only the IL - 6 - induced mammosphere formation in MCF7 and MDA - MB - 231 cells were dependent on sortilin, using the so rtilin- binding molecule, AF38 46 9. Although, more studies are needed to evaluate if the IL - 6 - and progranulin - induced sphere formation acts directly or indirectly via sortilin, or if there are other pathways and mechanisms involved. Correlation between pr ogranulin and IL- 6 in an in vivo - like model system Primary derived scaffolds ( PDS models ) are emerging as a model to study the influence of the tumor microenvironment on various types of cancer, by mimicking the growth of the cells in a more in vivo- like situation [93, 95, 97] . A previous study form our group which included RNA sequencing where we compared cells grown in the PDS model with cells derived from in vivo- like xenografts, as well as normal 2D growth conditions [93] . 31 Analysis revealed that cells grown in PDS cultures were more similar to in vivo conditions compared to 2D. The PD S model contains intact ECM components and signaling molecules, and has properties associated with the clinical parameters of the original tumor [93] . The cells grown on PDSs have a genomic profile, as well as a proteomic profile of secreted proteins showing involvement in, among others, differentiation, EMT and stemness markers compared to cells grown as conventional 2D cultures [93, 95, 139] . In addition, when studying the data in more detail, w e observed an upregulation of IL - 6, IL - 8 and GRN (progranulin) messenger RNA ( mRNA) expression in the PDS system , as well as the receptors SORT 1 (sortilin) and IL - 6 receptors (IL - 6R and GP1 30) compared to 2D. This confirms the importance of these genes in association with CSC characteristics and involvement in the priming of the pre - metastatic niche [59, 13 3] . Further, when looking more specifically at the secretion induced by cells grown in the PD S model, we observed that cells grown in specific PDSs mimicked clinical features of the original tumor [92] . Focusing on the correlation between secretions of relevant molecules in this study, we observed a positive correlation between IL - 6 and progranulin in both MCF7 and MDA - MB - 231 cells. In MCF7 cells, there was a positive correlation between IL - 6 and IL - 8, w hile this correlation was negative in the MDA - MB - 23 1 cell line. In the MDA - MB - 231 cell line, there was also a negative correlation between IL - 8 and progranulin . Interestingly, MDA - MB - 23 1 cells grown in PDSs had a higher basal sec retion, which is consistent with other studies [140] . 30 line with this, other researchers have observed that treatment with IL - 6 in liver and bile duct cancer increased the expression of progranulin in vitro, as well as that progranulin - induced production of IL - 6 in adipose tissue [134- 13 6] . However, to our knowledge, no feedback loop have been documented in the same system. Importantly , treating bre ast cancer cell lines with increasing concentrations of IL - 6 or IL - 8 elevated the levels of progranulin in the cells in a dose - dependent manner, suggesting a crosstalk between the expression of progranulin and IL - 6 , and IL - 8. IL- 6- induced CSC propagation and its dependence on sortilin IL - 6 and IL - 8 have both been associated with the induction of CSC characteristics in cancer [71, 13 7] . Therefore, we wanted to explore if IL - 6 and IL - 8 affect the amount of CSCs in MCF7 and MDA - MB - 23 1 breast cancer cells. When treating the cells with recombinant IL - 6 or IL - 8, we observed an increase in mammosphere formation, similar to what we could see with progranulin. This indicates that both IL - 6 and IL - 8 mediate breast CSC expansion in breast cancer. Interestingly, sortilin has also been suggested as a high - affinity receptor for IL - 6 in immune cells [64, 138] . Using a FPA , we were able to show that IL - 6 outcompeted the fluorescently labeled neurotensin bound to the soluble sortilin receptor , confirming this interaction. However, IL- 8 did not outcompete neurotensin, hence did not bind to sortilin at this binding site. Next, we explored if the CSC effect caused by the cytokine treatment was dependent on, or regulated by the binding to sortilin. Results showed that only the IL - 6 - induced mammosphere formation in MCF7 and MDA - MB - 231 cells were dependent on sortilin, using the so rtilin- binding molecule, AF38 46 9. Although, more studies are needed to evaluate if the IL - 6 - and progranulin - induced sphere formation acts directly or indirectly via sortilin, or if there are other pathways and mechanisms involved. Correlation between pr ogranulin and IL- 6 in an in vivo - like model system Primary derived scaffolds ( PDS models ) are emerging as a model to study the influence of the tumor microenvironment on various types of cancer, by mimicking the growth of the cells in a more in vivo- like situation [93, 95, 97] . A previous study form our group which included RNA sequencing where we compared cells grown in the PDS model with cells derived from in vivo- like xenografts, as well as normal 2D growth conditions [93] . 31 Analysis revealed that cells grown in PDS cultures were more similar to in vivo conditions compared to 2D. The PD S model contains intact ECM components and signaling molecules, and has properties associated with the clinical parameters of the original tumor [93] . The cells grown on PDSs have a genomic profile, as well as a proteomic profile of secreted proteins showing involvement in, among others, differentiation, EMT and stemness markers compared to cells grown as conventional 2D cultures [93, 95, 139] . In addition, when studying the data in more detail, w e observed an upregulation of IL - 6, IL - 8 and GRN (progranulin) messenger RNA ( mRNA) expression in the PDS system , as well as the receptors SORT 1 (sortilin) and IL - 6 receptors (IL - 6R and GP1 30) compared to 2D. This confirms the importance of these genes in association with CSC characteristics and involvement in the priming of the pre - metastatic niche [59, 13 3] . Further, when looking more specifically at the secretion induced by cells grown in the PD S model, we observed that cells grown in specific PDSs mimicked clinical features of the original tumor [92] . Focusing on the correlation between secretions of relevant molecules in this study, we observed a positive correlation between IL - 6 and progranulin in both MCF7 and MDA - MB - 231 cells. In MCF7 cells, there was a positive correlation between IL - 6 and IL - 8, w hile this correlation was negative in the MDA - MB - 23 1 cell line. In the MDA - MB - 231 cell line, there was also a negative correlation between IL - 8 and progranulin . Interestingly, MDA - MB - 23 1 cells grown in PDSs had a higher basal sec retion, which is consistent with other studies [140] . 32 Paper I V - Granulin peptide domains induce breast cancer stem cell propagation via sortilin The balance between progranulin and its peptide domains are regulated by naturally occurring proteases and protease inhibitors [60] . As of today, an unde rstanding about the role of the eight progranulin cleaved peptide domains, including their function and binding partners, are still modest [60, 14 1] . The enzyme human neutrophil elastase cleaves progranulin into smaller fragments In Paper I, we showed that granulin A was an active progranulin domain affecting CSC activity in several breast cancer cell line s. This suggests that some of the granulin domains themselves are important mediators of the progranulin effect. In this part of the project, we aim to better define how progranulin is cleaved and delineate the CSC functions of the individual peptide domains compared to the full - length progranulin. The growth factor progranulin is thought to have anti - inflammatory properties, while the granulin domains are thought to be pro - inflammatory, suggesting that they might have contrasting functions in the cells [141] . Cleavage of pro granulin occurs naturally by several enzyme s, such as elastase, and can be blocked by various protease inhibitors, thereby controlling the balance between progranulin and the granulin domains [60, 14 1 ] . In this study, we hypothesize that this balance might be lost during the formation of breast cancer. Western blot analysis confirmed that progranulin could be cleaved into smaller peptide fragments by human neutrophil elastase. Further, progranulin cl eavage was blocked by adding the protease inhibitor secretory leukocyte protease inhibitor ( SLPI ) , leaving the full- length progranulin intact. Moreover, sphere formation i n MCF7 cells treated with progranulin or elastase - cleaved progranulin revealed an additive effect of the cleaved peptides compared to full - length progranulin, suggesting that the small peptide fragment s indeed have potent CSC activity. C leaved progranulin peptide domains increase breast cancer stem cell activity To delineate which of the granulin peptides that were responsible for affecting the mammosphere - inducing ability of the cells, we treated breast cancer cell lines with individual peptide domains s ynthesized according to 33 their amino acid sequences. We also included the small C - terminal p art of progranulin, as it has been shown to be required for progranulins binding to sortilin and may have a n effect of its own [111] . Results showed that of the eight peptides testes, only para - granulin, granulin A, granulin C and the C - terminal part of progranuli n induced mammosphere formation, suggesting that these granulin domains indeed have tumor - progressing properties. Progranulin domains binds to sortilin To identify the exact mechanism on how individual granulin domains induce CSC propagation, we investig ated the role of the progranulin receptor sortilin on these domains. From our FPA, we were able to show that para- granulin, granulin A and the C - terminal part of progranulin bind to sortilin at the same site as progranulin , with the C - terminal part of progranulin showing the highest binding potency . This is in line with data published by other researchers [110, 111] . Interestingly , these sortilin - binding domains were also the granulins capable of inducing CSC activity in vitro. Distinctively , gr anulin C induced sphere formation in vitro, but did not seem to bind to sortilin (data not shown). Moreover, treatment with a sortilin degrader effectively reduced the peptide - induced mammosphere increase in MCF7, seen with para- granulin, granulin A and C - terminal part of progranulin. Although, further studies are required to identify and evaluate other binding partners or binding at different sites on sortilin. In summary, the individual granulin peptides para - granulin, granulin A and the C - terminal part of progranulin induce CSC growth in a sortilin - dependent manner in breast cancer. 32 Paper I V - Granulin peptide domains induce breast cancer stem cell propagation via sortilin The balance between progranulin and its peptide domains are regulated by naturally occurring proteases and protease inhibitors [60] . As of today, an unde rstanding about the role of the eight progranulin cleaved peptide domains, including their function and binding partners, are still modest [60, 14 1] . The enzyme human neutrophil elastase cleaves progranulin into smaller fragments In Paper I, we showed that granulin A was an active progranulin domain affecting CSC activity in several breast cancer cell line s. This suggests that some of the granulin domains themselves are important mediators of the progranulin effect. In this part of the project, we aim to better define how progranulin is cleaved and delineate the CSC functions of the individual peptide domains compared to the full - length progranulin. The growth factor progranulin is thought to have anti - inflammatory properties, while the granulin domains are thought to be pro - inflammatory, suggesting that they might have contrasting functions in the cells [141] . Cleavage of pro granulin occurs naturally by several enzyme s, such as elastase, and can be blocked by various protease inhibitors, thereby controlling the balance between progranulin and the granulin domains [60, 14 1 ] . In this study, we hypothesize that this balance might be lost during the formation of breast cancer. Western blot analysis confirmed that progranulin could be cleaved into smaller peptide fragments by human neutrophil elastase. Further, progranulin cl eavage was blocked by adding the protease inhibitor secretory leukocyte protease inhibitor ( SLPI ) , leaving the full- length progranulin intact. Moreover, sphere formation i n MCF7 cells treated with progranulin or elastase - cleaved progranulin revealed an additive effect of the cleaved peptides compared to full - length progranulin, suggesting that the small peptide fragment s indeed have potent CSC activity. C leaved progranulin peptide domains increase breast cancer stem cell activity To delineate which of the granulin peptides that were responsible for affecting the mammosphere - inducing ability of the cells, we treated breast cancer cell lines with individual peptide domains s ynthesized according to 33 their amino acid sequences. We also included the small C - terminal p art of progranulin, as it has been shown to be required for progranulins binding to sortilin and may have a n effect of its own [111] . Results showed that of the eight peptides testes, only para - granulin, granulin A, granulin C and the C - terminal part of progranuli n induced mammosphere formation, suggesting that these granulin domains indeed have tumor - progressing properties. Progranulin domains binds to sortilin To identify the exact mechanism on how individual granulin domains induce CSC propagation, we investig ated the role of the progranulin receptor sortilin on these domains. From our FPA, we were able to show that para- granulin, granulin A and the C - terminal part of progranulin bind to sortilin at the same site as progranulin , with the C - terminal part of progranulin showing the highest binding potency . This is in line with data published by other researchers [110, 111] . Interestingly , these sortilin - binding domains were also the granulins capable of inducing CSC activity in vitro. Distinctively , gr anulin C induced sphere formation in vitro, but did not seem to bind to sortilin (data not shown). Moreover, treatment with a sortilin degrader effectively reduced the peptide - induced mammosphere increase in MCF7, seen with para- granulin, granulin A and C - terminal part of progranulin. Although, further studies are required to identify and evaluate other binding partners or binding at different sites on sortilin. In summary, the individual granulin peptides para - granulin, granulin A and the C - terminal part of progranulin induce CSC growth in a sortilin - dependent manner in breast cancer. 34 Paper V - Reduction of p rogranulin- induced breast cancer stem cell propagation by sortilin- targeting cyclotriazadisulfonamide (CADA) c ompounds We have demonstrated that by restricting the binding of progranulin to sortilin, we could block progranulin - induced mammosphere formation in vitro, as well as granulin A - induced lung metastasis in an in vivo xenograft model (Paper I) [11 5] . In addition, clinical data proposes progranulin and sortilin as prognostic markers in breast cancer and other cancer types (Paper II) [58, 122 , 14 2] . Subsequently , sortilin could be a potential novel drug target for breast cancer. In order to block the interaction between progranulin and sortilin we used the antiviral agent cyclotriazadisulfonamide (CADA). CADA has been shown to down- regulate cluster of differentiation 4 (CD4) expression in T - cell lines and peripheral blood mononuclear cells, resulting in a marked inhibition of the human immunodeficiency virus (HIV) entry to the host cell [143] . Importantly, CADA also modulates sortilin expression by binding to its signal peptide and inhibits the co - translational translocation of sortilin to the lumen of the endoplasmic reticulum, giving a 50% reduction in sortilin protein expression [144] . Therefore, in collaboration with the inventors of the CADA molecules , we aimed to identify CADA analogs that have a more than 50% downregulation of sortilin expression and could effectively reduce breast cancer stem cell propagation induced by progranulin. CADA molecules down - modulate sortilin in breast cancer cell lines A r ange of synthesized compound similar to CADA were tested for cellular toxicity, looking at cell viability/proliferation and for potency, in both breast cancer cell lines and HEK239 cells. In addition, the compounds ability to downregulate sortilin expression were also tested in two different EUHDVWFDQFHUFHOO OLQHVRQH(5αSRVLWLYH 0&) DQGRQH(5αQHJDWLYH ( MD A - MB - 23 1). Seven of these compounds were included in Paper V. Results showed that out of the seven CADA compounds; compounds 2 , 5 and 6 were the most potent analogs, efficiently downregulating the H[SUHVVLRQRIVRUWLOLQLQERWKWKH(5αSRVLWLYHEUHDVWFDQFHUFHOOOLQH0&) DQGWKH(5αQHJDWLYHFHOOOLQH0'$ - MB - 231. 35 CADA molecules reduce prog ranulin- induced CSC propagation in breast cancer cell lines The most potent sortilin down - modulating compounds were further tested for their ability to reduce progranulin - induced CSC propagation . We chose a FRQFHQWUDWLRQ RI  Ǎ0 of the compounds, due to their good down - modulatory effect on sortilin expression, while at the same time showing no significan t reduction in cell viability. Importantly, several of the CADA compounds showed a significant inhibitio n on the progranulin - ind uced mammosphere formation, independent on hormone status. The most potent sortilin - down - modulators were found to be compound 2 , 5 and 6 , where c ompound 2 was selected as the top candidate, due to the slightly higher toxicity seen with compounds 5 and 6 , especially in the HEK293 cell line. R esults also showed that less sortilin expression on its own did not lead to a reduction in the mammosphere forming capacity of the cells, suggesting that the sphere inhibitory effect is due to the progranulin stimulatio n being blocked. Conclusively, based on flow cytometry data on sortilin expression and toxicity data on HEK293 cells, together with the western blot protein expression data on the breast cancer cell lines, substance 2 was identified as the most selective compound showing least toxicity . Furthermore, compound 2 efficiently blocked progranulin - induced breast CSC expansion in vitro, indepenGHQWRQ(5αVWDWXV and we are therefore selecting this compound for further optimization and upcoming in vivo studies . 34 Paper V - Reduction of p rogranulin- induced breast cancer stem cell propagation by sortilin- targeting cyclotriazadisulfonamide (CADA) c ompounds We have demonstrated that by restricting the binding of progranulin to sortilin, we could block progranulin - induced mammosphere formation in vitro, as well as granulin A - induced lung metastasis in an in vivo xenograft model (Paper I) [11 5] . In addition, clinical data proposes progranulin and sortilin as prognostic markers in breast cancer and other cancer types (Paper II) [58, 122 , 14 2] . Subsequently , sortilin could be a potential novel drug target for breast cancer. In order to block the interaction between progranulin and sortilin we used the antiviral agent cyclotriazadisulfonamide (CADA). CADA has been shown to down- regulate cluster of differentiation 4 (CD4) expression in T - cell lines and peripheral blood mononuclear cells, resulting in a marked inhibition of the human immunodeficiency virus (HIV) entry to the host cell [143] . Importantly, CADA also modulates sortilin expression by binding to its signal peptide and inhibits the co - translational translocation of sortilin to the lumen of the endoplasmic reticulum, giving a 50% reduction in sortilin protein expression [144] . Therefore, in collaboration with the inventors of the CADA molecules , we aimed to identify CADA analogs that have a more than 50% downregulation of sortilin expression and could effectively reduce breast cancer stem cell propagation induced by progranulin. CADA molecules down - modulate sortilin in breast cancer cell lines A r ange of synthesized compound similar to CADA were tested for cellular toxicity, looking at cell viability/proliferation and for potency, in both breast cancer cell lines and HEK239 cells. In addition, the compounds ability to downregulate sortilin expression were also tested in two different EUHDVWFDQFHUFHOO OLQHVRQH(5αSRVLWLYH 0&) DQGRQH(5αQHJDWLYH ( MD A - MB - 23 1). Seven of these compounds were included in Paper V. Results showed that out of the seven CADA compounds; compounds 2 , 5 and 6 were the most potent analogs, efficiently downregulating the H[SUHVVLRQRIVRUWLOLQLQERWKWKH(5αSRVLWLYHEUHDVWFDQFHUFHOOOLQH0&) DQGWKH(5αQHJDWLYHFHOOOLQH0'$ - MB - 231. 35 CADA molecules reduce prog ranulin- induced CSC propagation in breast cancer cell lines The most potent sortilin down - modulating compounds were further tested for their ability to reduce progranulin - induced CSC propagation . We chose a FRQFHQWUDWLRQ RI  Ǎ0 of the compounds, due to their good down - modulatory effect on sortilin expression, while at the same time showing no significan t reduction in cell viability. Importantly, several of the CADA compounds showed a significant inhibitio n on the progranulin - ind uced mammosphere formation, independent on hormone status. The most potent sortilin - down - modulators were found to be compound 2 , 5 and 6 , where c ompound 2 was selected as the top candidate, due to the slightly higher toxicity seen with compounds 5 and 6 , especially in the HEK293 cell line. R esults also showed that less sortilin expression on its own did not lead to a reduction in the mammosphere forming capacity of the cells, suggesting that the sphere inhibitory effect is due to the progranulin stimulatio n being blocked. Conclusively, based on flow cytometry data on sortilin expression and toxicity data on HEK293 cells, together with the western blot protein expression data on the breast cancer cell lines, substance 2 was identified as the most selective compound showing least toxicity . Furthermore, compound 2 efficiently blocked progranulin - induced breast CSC expansion in vitro, indepenGHQWRQ(5αVWDWXV and we are therefore selecting this compound for further optimization and upcoming in vivo studies . 36 37 Conclusions Breast cancer is a very heterogeneous disease, containing different cell populations interacting in complex networks consisting of various secretion systems and components in the surrounding tumor microenvironment. It is necessary to understand the molecular mechanisms and signaling pathways involved in cancer progression and metastasis formation, and more evidence are emerging that the CSC niche and the tumor microenvironment are partly responsible for this effect. In this thesis, we have identified several markers and potential pathways that are important for the maintenance of the CSC population in breast cancer. These factors can pot entially be used as biomarkers for cancer progression and treatment prediction or as possible future drug targets in order to improve existing treatment approaches. More specifically: P ap er I : Here, we identified progranulin as a hypoxia- induced secreted factor influencing breast CSC propagation in vitro, as well as promoting tumor progression and mediating a more metastasizing cellular subtype in vivo. In addition, by targeting the progranulin receptor sortilin, through small sortilin- binding molecules or by degradation, we could demonstrate that sortilin is a functional receptor of progranulin in breast cancer and is responsible for driving the progranulin induced breast CSC propagation in vitro and in vivo. P ap er I I : We found that elevated progranulin and sortilin tumor protein levels are associated with unfavorable clinical prognosis and poor outcomes in breast cancer patients. In summary, these results suggest that high co - expression of progranulin and sortilin defines a highly malignant subgroup of breast cancer patients and can be used as a prognostic biomarker in breast cancer patients. P ap er I I I : Here, we further described how progranulin induced secretion of cytokines involved in CSC propagation. In conclusion, the interplay between IL - 6 and the progranulin- sortilin axis in breast cancer is responsible for driving CSC activity and are linked to aggressive features in breast cancer. P ap er I V : During different biological circumstances, p rogranulin can be cleaved into smaller peptide fragments by human neutrophil elastase. In this paper, we demonstrated that specific granulin peptide domains induce 36 37 Conclusions Breast cancer is a very heterogeneous disease, containing different cell populations interacting in complex networks consisting of various secretion systems and components in the surrounding tumor microenvironment. It is necessary to understand the molecular mechanisms and signaling pathways involved in cancer progression and metastasis formation, and more evidence are emerging that the CSC niche and the tumor microenvironment are partly responsible for this effect. In this thesis, we have identified several markers and potential pathways that are important for the maintenance of the CSC population in breast cancer. These factors can pot entially be used as biomarkers for cancer progression and treatment prediction or as possible future drug targets in order to improve existing treatment approaches. More specifically: P ap er I : Here, we identified progranulin as a hypoxia- induced secreted factor influencing breast CSC propagation in vitro, as well as promoting tumor progression and mediating a more metastasizing cellular subtype in vivo. In addition, by targeting the progranulin receptor sortilin, through small sortilin- binding molecules or by degradation, we could demonstrate that sortilin is a functional receptor of progranulin in breast cancer and is responsible for driving the progranulin induced breast CSC propagation in vitro and in vivo. P ap er I I : We found that elevated progranulin and sortilin tumor protein levels are associated with unfavorable clinical prognosis and poor outcomes in breast cancer patients. In summary, these results suggest that high co - expression of progranulin and sortilin defines a highly malignant subgroup of breast cancer patients and can be used as a prognostic biomarker in breast cancer patients. P ap er I I I : Here, we further described how progranulin induced secretion of cytokines involved in CSC propagation. In conclusion, the interplay between IL - 6 and the progranulin- sortilin axis in breast cancer is responsible for driving CSC activity and are linked to aggressive features in breast cancer. P ap er I V : During different biological circumstances, p rogranulin can be cleaved into smaller peptide fragments by human neutrophil elastase. In this paper, we demonstrated that specific granulin peptide domains induce 38 CSC propagation in breast cancer and that these peptides bind and act through sortilin in a similar manner as progran ulin. P ap er V : In this paper, we identified chemical compounds ( CADA compounds) with high potency towards down- modulating sortilin protein expression and subsequently reducing progranulin- induced CSC activity. These CADA compounds were highly selective wi th low cellular toxicity at optimal sortilin down - modulating conditions and can be further evaluated as therapeutic compounds in relevant in vivo models . Taken together, targeting progranulin through its receptor sortilin could be a potential novel breast cancer therapeutic approach, used as a combination therapy together with chemotherapy or other conventional subtype - based treatments. 39 F uture perspectives This thesis has provided substantial evidence for the involvement of progranulin and sortilin in breast CSC propagation and patient outcome. These findings are supported by studies from other researchers, highlighting the importance of targeting the progranulin - sortilin axis as a potential treatment strategy for cancer patients expressing high levels of progranulin or sortilin. However, more experiments are needed to furt her elucidate the mechanisms and pathways involved. Sortilin targeted therapy in breast cancer Neutralizing antibodies for progranulin are currently under preclinical development for treating cancers, such as lung and breast cancer [141] . A different therapeutic approach could be by targeting the progranulin receptor sortilin, as sortilin overexpression is seen in various types of cancers compared to normal cells, and has been shown to play a role in CSC activity, differentiation, EMT and invasion [62] . In cancer, anti - sortilin antibody treatment has been shown to induce apoptosis in leukemic cells without affecting normal cells [14 5] , and i n pancreatic cancer, sortilin knockdown by siRNA or inhibition by AF3 846 9, obstructs pancreatic cancer cell adhesion and invasion [146] . Moreover, our studies have shown that, in vivo, AF38 46 9 blocks the progranulin - induced lung metastasis in mice [115] . This implies that targeting sortilin may be a therapeutic option to treat various types of cancer, and might be worth exploring further [62] . Moreover, sortilin is also a receptor for the neuronal growth factor neurotensin. In several types of cancer, sortilin has been shown to induce neurotensin- mediated cancer cell growth and proliferation , emphasizing the need to study the role of neurotensin in breast cancer [62] . Currently, in collaboration with SciLifeLab drug discovery and development platform, external chemists, Sortina Pharma AB , GU Venture and the Grants and Innovation Office at the University of Gothenburg, we are identifying and developing new sortilin targeted compounds, such as small molecule sortilin inhibitors and sortilin - targeting antibodies. W e aim to identify sortilin - targeted drugs for treatment of patients with imbalance in the progranulin- sortilin signaling pathway in order to reduce the progranulin - induced CSC spreading and metastasis form ation. Novel synthesized compounds are tested using our sortilin- binding F P A setup. Potential sortilin - binding compounds are further v alidated using the 38 CSC propagation in breast cancer and that these peptides bind and act through sortilin in a similar manner as progran ulin. P ap er V : In this paper, we identified chemical compounds ( CADA compounds) with high potency towards down- modulating sortilin protein expression and subsequently reducing progranulin- induced CSC activity. These CADA compounds were highly selective wi th low cellular toxicity at optimal sortilin down - modulating conditions and can be further evaluated as therapeutic compounds in relevant in vivo models . Taken together, targeting progranulin through its receptor sortilin could be a potential novel breast cancer therapeutic approach, used as a combination therapy together with chemotherapy or other conventional subtype - based treatments. 39 F uture perspectives This thesis has provided substantial evidence for the involvement of progranulin and sortilin in breast CSC propagation and patient outcome. These findings are supported by studies from other researchers, highlighting the importance of targeting the progranulin - sortilin axis as a potential treatment strategy for cancer patients expressing high levels of progranulin or sortilin. However, more experiments are needed to furt her elucidate the mechanisms and pathways involved. Sortilin targeted therapy in breast cancer Neutralizing antibodies for progranulin are currently under preclinical development for treating cancers, such as lung and breast cancer [141] . A different therapeutic approach could be by targeting the progranulin receptor sortilin, as sortilin overexpression is seen in various types of cancers compared to normal cells, and has been shown to play a role in CSC activity, differentiation, EMT and invasion [62] . In cancer, anti - sortilin antibody treatment has been shown to induce apoptosis in leukemic cells without affecting normal cells [14 5] , and i n pancreatic cancer, sortilin knockdown by siRNA or inhibition by AF3 846 9, obstructs pancreatic cancer cell adhesion and invasion [146] . Moreover, our studies have shown that, in vivo, AF38 46 9 blocks the progranulin - induced lung metastasis in mice [115] . This implies that targeting sortilin may be a therapeutic option to treat various types of cancer, and might be worth exploring further [62] . Moreover, sortilin is also a receptor for the neuronal growth factor neurotensin. In several types of cancer, sortilin has been shown to induce neurotensin- mediated cancer cell growth and proliferation , emphasizing the need to study the role of neurotensin in breast cancer [62] . Currently, in collaboration with SciLifeLab drug discovery and development platform, external chemists, Sortina Pharma AB , GU Venture and the Grants and Innovation Office at the University of Gothenburg, we are identifying and developing new sortilin targeted compounds, such as small molecule sortilin inhibitors and sortilin - targeting antibodies. W e aim to identify sortilin - targeted drugs for treatment of patients with imbalance in the progranulin- sortilin signaling pathway in order to reduce the progranulin - induced CSC spreading and metastasis form ation. Novel synthesized compounds are tested using our sortilin- binding F P A setup. Potential sortilin - binding compounds are further v alidated using the 40 mammosphere assay, as well as additional functional assays in vitro, before validation of top candidates in vivo. Our long - term goal is to develop a sortilin- based drug for clinical trials. We anticipate that the developed drug will probably not eradicate tumor cells itself, but will limit the fraction of CSCs that in the long term will suppress the tumor and metastasis formation by limiting the repopulation capacity of the tumor (Figure 9). These drugs can be used as an adjuvant therapy i n combination with other cancer- specific targeting drugs such as tamoxifen. Figure 9: Cancer stem cell -targeted therapy. Progranulin or sortilin could be an attractive target for breast cancer therapy, targeting the resistant cancer stem cells. Combined with conventional treatment to prevent resistance and cancer recurrence. Image adapte d from [147 ] and created with BioRender.com. Further , sortilin also exists naturally in a soluble form, termed soluble sortilin (sSortilin), derived from transmembrane sortilin cleaved by v arious proteases [62] . In neuroendocrine tumors and colon cancer, sSortilin have been shown to be involved in tumorigenesis and cancer progression through the activation of focal adhesion kinase (FAK)- Scr signaling , through interaction with an unidentified rece ptor on the target cells, leading to FAK - Scr phosphorylation [62] . This phosphorylation avtivates pathways, including Akt and PI3K that are involved in cell survival, migration and tumor invasion [62] . In addition, sSortilin leads to loss of cellular matrix contact due to a ctin microfilament modifications in the plasma membrane 41 in some cancers [62] . More investigations are needed to determine the importance of sSortilin in breast cancer and its role as a potential plasma biomarker for various diseases [62] . Besides sortilin , additional receptors are known to bind to progranulin , such as the inflammatory receptors TN F receptor 1 and 2, enlightening progranulins role in inflammation, as well as EphA 2 that are expressed in several types of cancer and are associated with cancer formation and progression [148] . As these receptors also bind to the growth factor progranulin, we aim to further investigate the TN F receptors and EphA2 to delineate their role in breast cancer and involvement in triggering cellular plasticity. The PDS model as a tool to improve personalized medicine As mentioned in the methods section, the PDS model can be used as a drug- screening tool for more patient - specific studies to investigate the therapeutic effect by novel compounds and drug combinations, as well as performing efficacy and toxicity studies on cancer cells growing in a patient - specific tumor microenvironment. In this project, we aim to use the PDS model to combine sortilin - targeting drugs with chemotherapy, as well as in combination with endocrine treatment, providing us with the possibility to evaluate treatment efficiency and responses in tumor microenvironments derived from individual patients. Here, we can also correlate the PDS response with liquid biopsies from the patients, where you use the serum of patients to study secreted proteins or even circulating tumor DNA (ctDNA) looking for altered genes or identification of biomarkers , such as progranulin, sortilin and IL - 6 . The role of the progranulin peptides As progranulin is known to be cleaved into smaller peptide domains, we aimed to delineate the role of the individual progranulin peptides on CSC activity in breast cancer (Paper IV). Little is known about the functions and mechanisms regulated by t hese individual peptide domains in cancer, and seems to be dependen t on the cell type tested [60, 149] . In epithelial cells, g ranulin A and B lead to growth inhibition and , in contrast, granulin D has shown proliferative ef fects in glioma cells [60] . Some studies even suggest that epithelial cells induce production of IL - 8 in response to granulin B stimulation [60] . This could indicate that the increase in IL - 8 secretion described in Paper III is due to progranulin being cleaved by 40 mammosphere assay, as well as additional functional assays in vitro, before validation of top candidates in vivo. Our long - term goal is to develop a sortilin- based drug for clinical trials. We anticipate that the developed drug will probably not eradicate tumor cells itself, but will limit the fraction of CSCs that in the long term will suppress the tumor and metastasis formation by limiting the repopulation capacity of the tumor (Figure 9). These drugs can be used as an adjuvant therapy i n combination with other cancer- specific targeting drugs such as tamoxifen. Figure 9: Cancer stem cell -targeted therapy. Progranulin or sortilin could be an attractive target for breast cancer therapy, targeting the resistant cancer stem cells. Combined with conventional treatment to prevent resistance and cancer recurrence. Image adapte d from [147 ] and created with BioRender.com. Further , sortilin also exists naturally in a soluble form, termed soluble sortilin (sSortilin), derived from transmembrane sortilin cleaved by v arious proteases [62] . In neuroendocrine tumors and colon cancer, sSortilin have been shown to be involved in tumorigenesis and cancer progression through the activation of focal adhesion kinase (FAK)- Scr signaling , through interaction with an unidentified rece ptor on the target cells, leading to FAK - Scr phosphorylation [62] . This phosphorylation avtivates pathways, including Akt and PI3K that are involved in cell survival, migration and tumor invasion [62] . In addition, sSortilin leads to loss of cellular matrix contact due to a ctin microfilament modifications in the plasma membrane 41 in some cancers [62] . More investigations are needed to determine the importance of sSortilin in breast cancer and its role as a potential plasma biomarker for various diseases [62] . Besides sortilin , additional receptors are known to bind to progranulin , such as the inflammatory receptors TN F receptor 1 and 2, enlightening progranulins role in inflammation, as well as EphA 2 that are expressed in several types of cancer and are associated with cancer formation and progression [148] . As these receptors also bind to the growth factor progranulin, we aim to further investigate the TN F receptors and EphA2 to delineate their role in breast cancer and involvement in triggering cellular plasticity. The PDS model as a tool to improve personalized medicine As mentioned in the methods section, the PDS model can be used as a drug- screening tool for more patient - specific studies to investigate the therapeutic effect by novel compounds and drug combinations, as well as performing efficacy and toxicity studies on cancer cells growing in a patient - specific tumor microenvironment. In this project, we aim to use the PDS model to combine sortilin - targeting drugs with chemotherapy, as well as in combination with endocrine treatment, providing us with the possibility to evaluate treatment efficiency and responses in tumor microenvironments derived from individual patients. Here, we can also correlate the PDS response with liquid biopsies from the patients, where you use the serum of patients to study secreted proteins or even circulating tumor DNA (ctDNA) looking for altered genes or identification of biomarkers , such as progranulin, sortilin and IL - 6 . The role of the progranulin peptides As progranulin is known to be cleaved into smaller peptide domains, we aimed to delineate the role of the individual progranulin peptides on CSC activity in breast cancer (Paper IV). Little is known about the functions and mechanisms regulated by t hese individual peptide domains in cancer, and seems to be dependen t on the cell type tested [60, 149] . In epithelial cells, g ranulin A and B lead to growth inhibition and , in contrast, granulin D has shown proliferative ef fects in glioma cells [60] . Some studies even suggest that epithelial cells induce production of IL - 8 in response to granulin B stimulation [60] . This could indicate that the increase in IL - 8 secretion described in Paper III is due to progranulin being cleaved by 42 proteinases in the cells , alt hough further investigations are needed in order to determine this . In Paper IV, we demonstrated that i n breast cancer, some of these cleaved progranulin peptides also bind to sortilin, and granulin A induce d lung metastasis in in vivo breast xenograft models (Paper I) [115] . However, more mechanistic data and knowledge about pathways regulated by the individual granulin peptides are needed . Interestingly, Z hang and colleagues are developing antibodies specifically targeting the individual granulin peptides, making it possible to study the individual peptides in more detail [150] . Notably , as progranulin is cleaved by various proteases in the cells , it is not always a complete cleavage into the individual granulins, as various lengths and combination s of the peptides are possible. These combinations of peptide lengths can also have an effect on the cells and need to be further investigated. Progranulin and sortilin relevance in other cancer types Progranulin and sortilin expression are thought to be elevated not only in breast cancer, but also in other cancer types, such as prostate and pancreatic cancers, melanoma, as well as kidney and lung carcinomas, and are linked to more aggressive clinical phenotypes [58, 62] . To further support our findings and to expand our research field, we will incorporate other cancer types to our research methods, in order to explore the progranulin - sortilin axis and their relevance in other types of cancer. 43 Acknowledgements This thesis would not have been possible without all the help and support from my colleagues and friends. I will miss you all. First of all, I wou ld like to thank my main supervisor , Göran Landberg , for giving me the opportunity to work with such an interesting project and inspiring research team. By letting me work independently and try out new ideas , you have helped me develop and grow as a researcher. My co - supervisor, Sara, for your enco uragement and support. For being by my side and guiding me through this PhD at times when I felt lost. Your knowledge and kindness is inspiring. I am grateful for all the help and support, I have learned so much from you . My second co - supervisor, Anders, for sharing your exceptional knowledge and passion for science and research. I would like to thank all current and former members of the Landberg lab. For all your help and lab expertise, but most of all, for all the good discussions and fun times togethe r (both in and outside the lab). Susann, my mentor during my master thesis project. F or sharing your knowledge and believing in me, but most of all you have been a great inspiration and role model. You were the one introducing me to research, and for that , I am forever grateful. Pernilla, for your continuous support, enc ouragement and friendship along the way , from my time as a master student and associate researcher, and especially now during the last year of my PhD. Thank you for all the help in the lab, interesting discussions and all the enjoyable talks and fikas . Emma P, for your friendship, and being someone I could always talk to. Thank you for all the statistics discussi ons and nice coffee breaks. Anna, for friendship, for always being honest , and all the good times when we did our master thesis. Elena, Mamen, Simona, Andreas, Jennifer, André , Paul (we do miss your singing), Èamon, Gabrielle, Svanheiður, Ylva , and others . It has been fun working with all of you! T he Ståhlberg group, for all the collaborations and group meetings. Soheila, Daniel, Emma J, Gustav, Lisa, Parmida, and others. Espec ially, Stefan, for being an awesome bioinformatician and friend. Sorry for all the stupid questions and for you being “forced” to act as my “Bioinformatics for Dummies” guide throughout these years. 42 proteinases in the cells , alt hough further investigations are needed in order to determine this . In Paper IV, we demonstrated that i n breast cancer, some of these cleaved progranulin peptides also bind to sortilin, and granulin A induce d lung metastasis in in vivo breast xenograft models (Paper I) [115] . However, more mechanistic data and knowledge about pathways regulated by the individual granulin peptides are needed . Interestingly, Z hang and colleagues are developing antibodies specifically targeting the individual granulin peptides, making it possible to study the individual peptides in more detail [150] . Notably , as progranulin is cleaved by various proteases in the cells , it is not always a complete cleavage into the individual granulins, as various lengths and combination s of the peptides are possible. These combinations of peptide lengths can also have an effect on the cells and need to be further investigated. Progranulin and sortilin relevance in other cancer types Progranulin and sortilin expression are thought to be elevated not only in breast cancer, but also in other cancer types, such as prostate and pancreatic cancers, melanoma, as well as kidney and lung carcinomas, and are linked to more aggressive clinical phenotypes [58, 62] . To further support our findings and to expand our research field, we will incorporate other cancer types to our research methods, in order to explore the progranulin - sortilin axis and their relevance in other types of cancer. 43 Acknowledgements This thesis would not have been possible without all the help and support from my colleagues and friends. I will miss you all. First of all, I wou ld like to thank my main supervisor , Göran Landberg , for giving me the opportunity to work with such an interesting project and inspiring research team. By letting me work independently and try out new ideas , you have helped me develop and grow as a researcher. My co - supervisor, Sara, for your enco uragement and support. For being by my side and guiding me through this PhD at times when I felt lost. Your knowledge and kindness is inspiring. I am grateful for all the help and support, I have learned so much from you . My second co - supervisor, Anders, for sharing your exceptional knowledge and passion for science and research. I would like to thank all current and former members of the Landberg lab. For all your help and lab expertise, but most of all, for all the good discussions and fun times togethe r (both in and outside the lab). Susann, my mentor during my master thesis project. F or sharing your knowledge and believing in me, but most of all you have been a great inspiration and role model. You were the one introducing me to research, and for that , I am forever grateful. Pernilla, for your continuous support, enc ouragement and friendship along the way , from my time as a master student and associate researcher, and especially now during the last year of my PhD. Thank you for all the help in the lab, interesting discussions and all the enjoyable talks and fikas . Emma P, for your friendship, and being someone I could always talk to. Thank you for all the statistics discussi ons and nice coffee breaks. Anna, for friendship, for always being honest , and all the good times when we did our master thesis. Elena, Mamen, Simona, Andreas, Jennifer, André , Paul (we do miss your singing), Èamon, Gabrielle, Svanheiður, Ylva , and others . It has been fun working with all of you! T he Ståhlberg group, for all the collaborations and group meetings. Soheila, Daniel, Emma J, Gustav, Lisa, Parmida, and others. Espec ially, Stefan, for being an awesome bioinformatician and friend. Sorry for all the stupid questions and for you being “forced” to act as my “Bioinformatics for Dummies” guide throughout these years. 44 To everyone at the Sahlgrenska Center for Cancer Research , at floor 4, 5 and 6 , for being helpful and crea ting a nice work environment. Notably, to Jana, Agnieszka and Dorota, for your invaluable friendship. Members of the Åman group, Malin, Christoffer, Mandy and Pernilla. Also , Gülay, Ágota, Kristell, Emil, Elin, Mohamed, Maryam, and everyone else at SCCR. To my f riends from the master´s program, Davide (and Rasmus) , Cristiania, Rebecca, Angelica, Dimitra, Mercé, Martin, George, Katrin, for all the fun times we have had together. Especially Stefanie, for your kindness and friendship. I would also like to thank our animal technicians, Jessica and Hannah , for helping out in the animal research facility, as well as all the research foundations for their funding support. Thanks to all the collaborators and co- authors not already mentioned. To my family and friends, without you, none of this would have been possible! My mother, growing up, you always let me be curious and n urtured my passion for science . You have been encouraging me all the way, even though you didn´t like the idea of me moving so far away. Stig, for being the best stepfather I had never dreamed of getting. To my sister, Marthe, I know it has been tough for you when I have been in another country for so long. To my brother, Kristian , for all the fun times and fishing trips . Without all the love and support from all of you, I would not be the person I am today. My parents in law, Helge and Violetta, for the love and support, and a ll the nice talks and dinners . Lilian, for becoming my S wedish family , and l etting me stay in your home. To my friends back in Norway , Linn, Jørgen and Maja (and their little sunshine Alma), Silje , Ida and Øyvind. I am so happy and grateful to have you all. For bringing me back to reality when I needed a break , hanging out and having so much fun together. To all my friends that I studied with in Oslo, Thanuja, Anja, Vilde, Betty, Freja, Bergitte, Ida, Anne Marte , Hilde, Stine, and others. Finally , I would like to thank Adrian, for always being there and supporting me all these years . Pushing me to follow my dreams, even though it led to us being further apart. Sorry for always being so far away. I love you. 45 References 1. Bray F, F erlay J , Soerjomataram I, Siegel RL, Torre LA, Jemal A: Gl obal canc e r stat i st ic s 2018: GLOB O CAN esti m at e s of inci de nc e and mort ali t y worl dw i de f or 36 canc er s in 185 count r ie s . CA Cancer J Clin 2018, 68(6):394- 424. 2. Social styrelsen: Canc er i sif f r or 2018 . ISBN 978- 91-88161- 18-5 2018. 3. Social styrelsen: St at is ti k om nyupptäkt a canc erf al l 2019 . 2020. 4. McPherson K, Steel CM, Dixon JM: AB C of breas t dise as es . Breas t cance r - e pi de mi ology, risk f act or s , and geneti c s . BMJ 2000, 321( 7 61):624- 628. 5. Harbeck N, Penault - Llorca F, Cortes J, Gnant M, Houssami N, Poortmans P, Ruddy K, Tsang J, Cardoso F: Br e ast canc e r . Nature Reviews Disease Primers 2019, 5(1):66. 6. Howard BA, Gusterson BA: H um an breast deve l opme nt . J Mammary Gland Biol Neoplasia 2000, 5( 2):119- 137. 7. Bierie B, Moses HL: TGFβ: the molecular Jekyll and Hyde of canc e r . Nature Reviews Cancer 2006, 6 ( 7): 506- 520. 8. Hanahan D, Weinberg RA: The Hallmarks of Cancer . Cell 2000, 100 (1):57- 70. 9. Coussens LM, Werb Z: I nf l amm at i on and cance r . Nature 2002, 420 ( 6917):860- 867. 10. Hanahan D, Weinberg Robert A: Hallmarks of Cancer: The Ne xt Gene r ati on . Cell 2011, 144( 5) :646- 674. 11. Kothari C, Ouellette G, Labrie Y, Jacob S, Diorio C, Durocher F: I de nt if ic ati on of a gene signat ur e for dif f er e nt stage s of breas t canc er development that could be used for early diagnosis and specific therapy. Oncotarget 2018, 9(100):37407- 37420. 12. Lanigan F, O’Connor D, Martin F, Gallagher WM: Com m on Molecular Mechanisms of Mammary Gland Development and Breast Canc e r . Cell Mol Life Sci 2007, 64(24):3159- 3184. 13. Bombonati A, Sgroi DC: The molecular pathology of breast c anc e r progre s si on . J Pathol 2011, 223 (2):307- 317. 14. Li CI, Uribe DJ , Daling JR: Clinical characteristics of different histologic types of breast cancer . B r J Cancer 2005, 93 ( 9):1046- 1052. 15. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack J R, Ross DT , Johnsen H, Akslen LA et al : M ole c ul ar portraits of human breast tumours . Nature 2000, 406 ( 6797) : 747- 752. 44 To everyone at the Sahlgrenska Center for Cancer Research , at floor 4, 5 and 6 , for being helpful and crea ting a nice work environment. Notably, to Jana, Agnieszka and Dorota, for your invaluable friendship. Members of the Åman group, Malin, Christoffer, Mandy and Pernilla. Also , Gülay, Ágota, Kristell, Emil, Elin, Mohamed, Maryam, and everyone else at SCCR. To my f riends from the master´s program, Davide (and Rasmus) , Cristiania, Rebecca, Angelica, Dimitra, Mercé, Martin, George, Katrin, for all the fun times we have had together. Especially Stefanie, for your kindness and friendship. I would also like to thank our animal technicians, Jessica and Hannah , for helping out in the animal research facility, as well as all the research foundations for their funding support. Thanks to all the collaborators and co- authors not already mentioned. To my family and friends, without you, none of this would have been possible! My mother, growing up, you always let me be curious and n urtured my passion for science . You have been encouraging me all the way, even though you didn´t like the idea of me moving so far away. Stig, for being the best stepfather I had never dreamed of getting. To my sister, Marthe, I know it has been tough for you when I have been in another country for so long. To my brother, Kristian , for all the fun times and fishing trips . Without all the love and support from all of you, I would not be the person I am today. My parents in law, Helge and Violetta, for the love and support, and a ll the nice talks and dinners . Lilian, for becoming my S wedish family , and l etting me stay in your home. To my friends back in Norway , Linn, Jørgen and Maja (and their little sunshine Alma), Silje , Ida and Øyvind. I am so happy and grateful to have you all. For bringing me back to reality when I needed a break , hanging out and having so much fun together. To all my friends that I studied with in Oslo, Thanuja, Anja, Vilde, Betty, Freja, Bergitte, Ida, Anne Marte , Hilde, Stine, and others. Finally , I would like to thank Adrian, for always being there and supporting me all these years . Pushing me to follow my dreams, even though it led to us being further apart. Sorry for always being so far away. I love you. 45 References 1. Bray F, F erlay J , Soerjomataram I, Siegel RL, Torre LA, Jemal A: Gl obal canc e r stat i st ic s 2018: GLOB O CAN esti m at e s of inci de nc e and mort ali t y worl dw i de f or 36 canc er s in 185 count r ie s . CA Cancer J Clin 2018, 68(6):394- 424. 2. Social styrelsen: Canc er i sif f r or 2018 . ISBN 978- 91-88161- 18-5 2018. 3. Social styrelsen: St at is ti k om nyupptäkt a canc erf al l 2019 . 2020. 4. McPherson K, Steel CM, Dixon JM: AB C of breas t dise as es . Breas t cance r - e pi de mi ology, risk f act or s , and geneti c s . BMJ 2000, 321( 7 61):624- 628. 5. Harbeck N, Penault - Llorca F, Cortes J, Gnant M, Houssami N, Poortmans P, Ruddy K, Tsang J, Cardoso F: Br e ast canc e r . Nature Reviews Disease Primers 2019, 5(1):66. 6. Howard BA, Gusterson BA: H um an breast deve l opme nt . J Mammary Gland Biol Neoplasia 2000, 5( 2):119- 137. 7. Bierie B, Moses HL: TGFβ: the molecular Jekyll and Hyde of canc e r . Nature Reviews Cancer 2006, 6 ( 7): 506- 520. 8. Hanahan D, Weinberg RA: The Hallmarks of Cancer . Cell 2000, 100 (1):57- 70. 9. Coussens LM, Werb Z: I nf l amm at i on and cance r . Nature 2002, 420 ( 6917):860- 867. 10. Hanahan D, Weinberg Robert A: Hallmarks of Cancer: The Ne xt Gene r ati on . Cell 2011, 144( 5) :646- 674. 11. Kothari C, Ouellette G, Labrie Y, Jacob S, Diorio C, Durocher F: I de nt if ic ati on of a gene signat ur e for dif f er e nt stage s of breas t canc er development that could be used for early diagnosis and specific therapy. Oncotarget 2018, 9(100):37407- 37420. 12. Lanigan F, O’Connor D, Martin F, Gallagher WM: Com m on Molecular Mechanisms of Mammary Gland Development and Breast Canc e r . Cell Mol Life Sci 2007, 64(24):3159- 3184. 13. Bombonati A, Sgroi DC: The molecular pathology of breast c anc e r progre s si on . J Pathol 2011, 223 (2):307- 317. 14. Li CI, Uribe DJ , Daling JR: Clinical characteristics of different histologic types of breast cancer . B r J Cancer 2005, 93 ( 9):1046- 1052. 15. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack J R, Ross DT , Johnsen H, Akslen LA et al : M ole c ul ar portraits of human breast tumours . Nature 2000, 406 ( 6797) : 747- 752. 46 16. Sorlie T, Perou CM, Ti bshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS et al: Ge ne expr e ss i on patterns of breast carcinomas distinguish tumor subclasses Zith clinical i m pl ic ati ons . Proc Natl Acad Sci U S A 2001, 98(19):10869- 10874. 17. Bert os NR, Park M: B r e as t canc er - one term, many ent it ie s ? J Clin Invest 2011, 121( 0):3789- 3796. 18. Edge SB, Compton CC: The $merican Joint Committee on Cancer: the th edition of the $JCC cancer staging manual and the f ut ure of TNM . Ann Surg Oncol 2010,  ( 6):1471- 1474. 19. Carels N, Spinasse LB, Tilli TM, Tuszynski J A: Tow ar d prec is i on medi ci ne of breas t canc e r . Theor Biol Med Model 2016, 13 : 7. 20. Guiu S, Michiels S, André F, Cortes J, Denkert C, Di Leo A, Hennessy BT, Sorlie T, Sotiriou C, Turner N et al: M ol e c ul ar subc las se s of breast cancer: hoZ do Ze define them" The ,M3$.T  :orking Gr oup State me nt . Ann Oncol 2012, 23 (12):2997- 3006. 21. Polyak K: H e te r oge ne it y in breast canc e r . The Journal of Clinical Investigation 2011, 121( 0):3786- 3788. 22. Hu Z , Fan C, Oh DS, Marron JS, He X, Qaqi sh BF, Liva s y C, Carey LA, Reynolds E, Dressler L et al : The molecular portraits of breast t um or s are cons e r ve d acros s micr oarr ay platf orm s . BMC Genomics 2006, : 96. 23. Perou CM: M ole c ular strat if i c at ion of tri ple - negati ve breast cance r s . Oncologist 2011, 16 Suppl 1 : 61- 70. 24. Carey LA, Perou CM, Livas y CA, Dressler LG, Cowan D, Conway K, Karaca G, Troester MA, Tse CK, Edmi ston S et al : Rac e , breast cancer subtypes and survival in the Carolina Breast Cancer St udy . JAMA 2006, 295 (21):2492- 2502. 25. Yerushalmi R, Woods R, Ravdin PM, Hayes MM, Gelmon KA: .i in breast cancer: prognostic and predictive potential . Lancet Oncol 2010, 11 (2):174- 183. 26. Zelnak AB, O'Regan RM: 2ptimi]ing (ndocrine Therapy f or Breas t Canc e r . J Natl Compr Canc Netw 2015, 13 ( 8):e56- 64. 27. Reis - Fi lho JS, Pusztai L: Ge ne expr e s si on prof i li ng in breast canc e r: class if i cati on, prognos ti c at i on, and predi ct ion . Lancet 2011, ( 9805):1812- 1823. 28. Lumachi F, Brunello A, Maruzzo M, Basso U, Basso SM: Tr e a tm e nt of estr oge n rece pt or - posi ti ve breast cance r . Curr Med Chem 2013, 20 ( 5): 596- 604. 29. Paterni I, Granchi C, Katzenellenbogen JA, Minutolo F: (strogen receptors alpha (5Į and beta (5β : subtype - s el e ct i ve li gands and cl i ni c al pot e nt i al . Steroids 2014, 90 :13- 29. 47 30. Davoli A, Hocevar BA, Brown TL: 3rogression and t r e at me nt of HER2 - posi tive breas t canc er . Cancer Chemother Pharmacol 2010, 65 ( 4): 611- 623. 31. Wong CC, Gilkes DM, Zhang H, Chen J, Wei H, Chaturvedi P, Fraley SI, Wong CM, Khoo US , Ng IO et al : H ypoxi a - i nduci ble f act or 1 is a master regulator of breast cancer metastatic niche formation. Proc Natl Acad Sci U S A 2011, 108( 39) : 6369- 16374. 32. Dawson SJ , Provenzano E, Cal da s C: Tri ple negat i ve breast canc e rs : cli nic al and prognos ti c impl i cati ons . Eur J Cancer 2009, 45 Suppl 1 :27 - 40. 33. Badve S, Dabbs DJ , Schnitt SJ, Baehner FL, Decker T, Eusebi V, Fox SB, Ichihara S, Jacquemier J, Lakhani SR et al: B asal - li ke and tri pl e - negative breast cancers: a critical revieZ Zith an emphasis on the implications for pathologists and oncologists . Mod Pathol 2011, 24( ) :157- 167. 34. Tkac zuk KR, Yue B, Zhan M, Tait N, Yarlagadda L, Dai H, Serrero G: ,ncreased Circulating /evel of the 6urvival Factor G3 3rogranulin in the 6erum of Breast Cancer 3atients :hen Compared to Healthy 6ubMects . Breast Cancer : Basic and Clinical Research 2011, 5:155- 162. 35. Meacham CE, Morrison SJ: Tumour heterogeneity and c anc e r cel l plast ic it y . Nature 2013, 501 : 328. 36. Badve S, Nakshatri H: B reas t - c a nce r stem cel ls —be yond sem ant ic s . The Lancet Oncology 2012, 13 (1):e43 - e48. 37. Visva der JE, Lindeman GJ: Canc e r Stem Cell s: Curre nt Stat us and Evol vi ng Compl e xi ti e s . Cell Stem Cell 2012, 10( 6) :7 7- 728. 38. Korkaya H, Liu S, Wicha MS: B re ast cance r stem cel ls , cyt oki ne netw o rks and the tumor microenvironment . The Journal of Clinical Investigation 2011, 121( 0):3804- 3809. 39. Petrova V, Annicchiarico - Petruzzelli M, Melino G, Amelio I: The hypo[ic tumour microenvironment . Oncogenesis 2018,  (1):10. 40. Muz B, de la Puente P, A zab F, Azab AK: The role of hypo[ia in cancer progression angiogenesis metastasis and resistance to therapy. Hypoxia (Auckland, NZ) 2015, 3 : 83- 92. 41. Harrison H, Rogerson L, Gregson HJ , Brennan KR, Clarke RB, Landberg G: Contrasting hypo[ic effects on br eas t canc e r stem cel l hierarchy is dependent on (5- alpha status . Cancer Res 2013, ( 4) :1420- 1433. 42. Dean M, Fojo T, Bates S: Tum o ur stem cell s and drug resis t anc e . Nat Rev Cancer 2005, 5 ( 4):275- 284. 43. Shan NL, Shin Y, Yang G, Furmanski P, Suh N: B r e ast canc e r stem cell s: A revieZ of their characteristics and the agents that affect them . Mol Carcinog 2021, 60 (2):73- 100. 46 16. Sorlie T, Perou CM, Ti bshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS et al: Ge ne expr e ss i on patterns of breast carcinomas distinguish tumor subclasses Zith clinical i m pl ic ati ons . Proc Natl Acad Sci U S A 2001, 98(19):10869- 10874. 17. Bert os NR, Park M: B r e as t canc er - one term, many ent it ie s ? J Clin Invest 2011, 121( 0):3789- 3796. 18. Edge SB, Compton CC: The $merican Joint Committee on Cancer: the th edition of the $JCC cancer staging manual and the f ut ure of TNM . Ann Surg Oncol 2010,  ( 6):1471- 1474. 19. Carels N, Spinasse LB, Tilli TM, Tuszynski J A: Tow ar d prec is i on medi ci ne of breas t canc e r . Theor Biol Med Model 2016, 13 : 7. 20. Guiu S, Michiels S, André F, Cortes J, Denkert C, Di Leo A, Hennessy BT, Sorlie T, Sotiriou C, Turner N et al: M ol e c ul ar subc las se s of breast cancer: hoZ do Ze define them" The ,M3$.T  :orking Gr oup State me nt . Ann Oncol 2012, 23 (12):2997- 3006. 21. Polyak K: H e te r oge ne it y in breast canc e r . The Journal of Clinical Investigation 2011, 121( 0):3786- 3788. 22. Hu Z , Fan C, Oh DS, Marron JS, He X, Qaqi sh BF, Liva s y C, Carey LA, Reynolds E, Dressler L et al : The molecular portraits of breast t um or s are cons e r ve d acros s micr oarr ay platf orm s . BMC Genomics 2006, : 96. 23. Perou CM: M ole c ular strat if i c at ion of tri ple - negati ve breast cance r s . Oncologist 2011, 16 Suppl 1 : 61- 70. 24. Carey LA, Perou CM, Livas y CA, Dressler LG, Cowan D, Conway K, Karaca G, Troester MA, Tse CK, Edmi ston S et al : Rac e , breast cancer subtypes and survival in the Carolina Breast Cancer St udy . JAMA 2006, 295 (21):2492- 2502. 25. Yerushalmi R, Woods R, Ravdin PM, Hayes MM, Gelmon KA: .i in breast cancer: prognostic and predictive potential . Lancet Oncol 2010, 11 (2):174- 183. 26. Zelnak AB, O'Regan RM: 2ptimi]ing (ndocrine Therapy f or Breas t Canc e r . J Natl Compr Canc Netw 2015, 13 ( 8):e56- 64. 27. Reis - Fi lho JS, Pusztai L: Ge ne expr e s si on prof i li ng in breast canc e r: class if i cati on, prognos ti c at i on, and predi ct ion . Lancet 2011, ( 9805):1812- 1823. 28. Lumachi F, Brunello A, Maruzzo M, Basso U, Basso SM: Tr e a tm e nt of estr oge n rece pt or - posi ti ve breast cance r . Curr Med Chem 2013, 20 ( 5): 596- 604. 29. Paterni I, Granchi C, Katzenellenbogen JA, Minutolo F: (strogen receptors alpha (5Į and beta (5β : subtype - s el e ct i ve li gands and cl i ni c al pot e nt i al . Steroids 2014, 90 :13- 29. 47 30. Davoli A, Hocevar BA, Brown TL: 3rogression and t r e at me nt of HER2 - posi tive breas t canc er . Cancer Chemother Pharmacol 2010, 65 ( 4): 611- 623. 31. Wong CC, Gilkes DM, Zhang H, Chen J, Wei H, Chaturvedi P, Fraley SI, Wong CM, Khoo US , Ng IO et al : H ypoxi a - i nduci ble f act or 1 is a master regulator of breast cancer metastatic niche formation. Proc Natl Acad Sci U S A 2011, 108( 39) : 6369- 16374. 32. Dawson SJ , Provenzano E, Cal da s C: Tri ple negat i ve breast canc e rs : cli nic al and prognos ti c impl i cati ons . Eur J Cancer 2009, 45 Suppl 1 :27 - 40. 33. Badve S, Dabbs DJ , Schnitt SJ, Baehner FL, Decker T, Eusebi V, Fox SB, Ichihara S, Jacquemier J, Lakhani SR et al: B asal - li ke and tri pl e - negative breast cancers: a critical revieZ Zith an emphasis on the implications for pathologists and oncologists . Mod Pathol 2011, 24( ) :157- 167. 34. Tkac zuk KR, Yue B, Zhan M, Tait N, Yarlagadda L, Dai H, Serrero G: ,ncreased Circulating /evel of the 6urvival Factor G3 3rogranulin in the 6erum of Breast Cancer 3atients :hen Compared to Healthy 6ubMects . Breast Cancer : Basic and Clinical Research 2011, 5:155- 162. 35. Meacham CE, Morrison SJ: Tumour heterogeneity and c anc e r cel l plast ic it y . Nature 2013, 501 : 328. 36. Badve S, Nakshatri H: B reas t - c a nce r stem cel ls —be yond sem ant ic s . The Lancet Oncology 2012, 13 (1):e43 - e48. 37. Visva der JE, Lindeman GJ: Canc e r Stem Cell s: Curre nt Stat us and Evol vi ng Compl e xi ti e s . Cell Stem Cell 2012, 10( 6) :7 7- 728. 38. Korkaya H, Liu S, Wicha MS: B re ast cance r stem cel ls , cyt oki ne netw o rks and the tumor microenvironment . The Journal of Clinical Investigation 2011, 121( 0):3804- 3809. 39. Petrova V, Annicchiarico - Petruzzelli M, Melino G, Amelio I: The hypo[ic tumour microenvironment . Oncogenesis 2018,  (1):10. 40. Muz B, de la Puente P, A zab F, Azab AK: The role of hypo[ia in cancer progression angiogenesis metastasis and resistance to therapy. Hypoxia (Auckland, NZ) 2015, 3 : 83- 92. 41. Harrison H, Rogerson L, Gregson HJ , Brennan KR, Clarke RB, Landberg G: Contrasting hypo[ic effects on br eas t canc e r stem cel l hierarchy is dependent on (5- alpha status . Cancer Res 2013, ( 4) :1420- 1433. 42. Dean M, Fojo T, Bates S: Tum o ur stem cell s and drug resis t anc e . Nat Rev Cancer 2005, 5 ( 4):275- 284. 43. Shan NL, Shin Y, Yang G, Furmanski P, Suh N: B r e ast canc e r stem cell s: A revieZ of their characteristics and the agents that affect them . Mol Carcinog 2021, 60 (2):73- 100. 48 44. Sheridan C, Kishimoto H, Fuchs RK, Mehrotra S, Bhat - Nakshatri P, Turner CH, Goulet R, Jr., Badve S, Nakshatri H: CDCD - brea st cancer cells e[hibit enhanced invasive properties: an early step ne c e ss ar y for met ast asi s . Breast Cancer Res 2006, 8(5) : R59. 45. Liu S, Cong Y, Wang D, Sun Y, Deng L, Liu Y, Martin- Trevino R, Shang L, McDermott SP, Landis MD et al: B r ea s t cance r stem cells transition betZeen epithelial and mesenchymal states reflective of their normal counterparts. Stem Cell Reports 2014, 2 (1):78- 91. 46. Marjanovic ND, Weinberg RA, Chaffer CL: Ce ll plas ti c it y and heterogeneity in cancer . Clin Chem 2013, 59 (1):168- 179. 47. Shaw FL, Harrison H, Spence K, Ablett MP, Simões BM, Farnie G, Clarke RB: $ Detailed Mammosphere $ssay 3rotocol for the Q uantif i c at ion of Breast Stem Cell Acti vi t y . J Mammary Gland Biol Neoplasia 2012,  (2):111 - 117. 48. Dontu G, Abdallah WM, Foley JM, Jackson KW, Clarke MF, Kawa mura MJ, Wicha MS: I n vit r o propagat i on and transc r i pt ional profiling of human mammary stemprogenitor cells . Genes Dev 2003, (10):1253- 1270. 49. Bradshaw A, Wickremsekera A, Tan ST, Peng L, Davi s PF, Itinteang T: Cancer 6tem Cell Hierarchy in Glioblastoma Multiforme . Frontiers in Surgery 2016, 3(21). 50. Campbell LL, Polyak K: Breast tumor heterogeneity: c anc e r stem cell s or clonal evol ut i on? Cell Cycle 2007, 6(19):2332- 2338. 51. Greaves M, Maley CC: Cl onal evol uti on in canc e r . Nature 2012, 481 ( 7381):306- 313. 52. Bradshaw A, Wickremsekera A, Tan ST, Peng L, Davi s PF, Itinteang T: Cancer 6tem Cell Hierarchy in Glioblastoma Multiforme . Frontiers in surgery 2016, 3:21. 53. Beck B, Bla npain C: Unrave ll i ng canc er stem cel l pot e nti al . Nat Rev Cancer 2013, 13( 0):727 - 738. 54. Lobo NA, Shimono Y, Qian D, Clarke MF: The biology of c anc e r stem cell s . Annu Rev Cell Dev Biol 2007, 23 : 675 - 699. 55. Mani SA, Guo W, Liao M - J , Eaton EN, Ayyanan A, Z hou AY, Brooks M, Reinhard F, Zhang CC, Shipitsin M et al : The (pithelial - Mesenchymal Transition Generates Cells Zith 3roperties of 6tem Cells . Cell 2008, 133 ( 4) : 704 - 715. 56. Kal luri R, Weinberg RA: The basics of epithelial - mesenchymal transition. J Clin I nvest 2009, 119 ( 6) :1420 - 1428. 57. De Muynck L, Van Damme P: Cel l ul ar Eff ec ts of 3rogranulin in Health and Disease . J Mol Neurosci 2011, 45 ( 3): 549. 58. Abella V, Pino J, Scotece M, Conde J, Lago F, Gonzalez - Gay MA, Mera A, Gómez R, Mobasheri A, Gualillo O: 3rogranulin as a 49 biomarker and potential therapeutic agent . Drug Discovery Today 2017, 22(10):1557- 1564. 59. Arechavaleta- Velasco F, Perez - Juarez CE, Gerton GL, Diaz - Cueto L: 3rogranulin and its biological effects in cancer . Med Oncol 2017, 34(12):194. 60. Zhu J, Nathan C, Jin W, Sim D, Ashcroft GS, Wahl SM, Lacomis L, Erdjument - Bromage H, Tempst P, Wright CD et al: Conver si on of 3roepithelin to (pithelins: 5oles of 6/3, and (lastase in Host Defense and :ound 5epair . Cell 2002, 111 ( 6): 867- 878. 61. Suh H - S, Choi N, Tarassi shin L, Lee SC: Re gul at i on of 3rogranulin ([pression in Human Microglia and 3roteolysis of 3rogranulin by Matri[ Metalloproteinase -  MM3 -  . PLoS One 2012, ( 4):e35115. 62. Ghaemi manesh F, Mehravar M, Mil ani S, Poursani EM, Sal i minejad K: The multifaceted role of sortilinneurotensin receptor  in human cancer development . J Cell Physiol 2021, na (n/a). 63. Nykj aer A, Willnow TE: Sor t ili n: a rece pt or to regulat e neur onal viabil it y and func t i on . Trends Neurosci 2012, 35( 4) :261 - 270. 64. Mortensen MB, Kjolby M, Gunnersen S, Larsen JV, Pal mfeldt J, Fal k E, Nykj aer A, Bentzon JF: Tar get i ng sort il i n in imm une cell s reduces proinflammatory cytokines and atherosclerosis . J Clin Invest 2014, 124 (12):5317- 5322. 65. Goettsch C, Kjolby M, Aika wa E: Sor ti li n and Its Mul ti pl e 5oles in Cardiovascular and Metabolic Diseases . Arterioscler Thromb Vasc Biol 2018, 38 (1):19- 25. 66. Tang W, Lu Y, Tian QY, Zhang Y, Guo FJ, Liu GY, Syed NM, Lai Y, Lin EA, Kong L et al : The groZth factor progranulin binds to TNF receptors and is therapeutic against inflammatory arthritis in mice . Science 2011, 332 ( 6028):478- 484. 67. Neill T, Buraschi S, Goyal A, Sharpe C, Natkanski E, Schaefer L, Morrione A, Iozzo RV: (ph$ is a functional receptor for the groZth f ac tor progr anul i n . The Journal of Cell Biology 2016, 215( 5) :687- 703. 68. Lippitz BE: Cytokine patterns in patients Zith cancer: a s ys te m at ic revi ew . The Lancet Oncology 2013, 14 ( 6):e218- e228. 69. Elinav E, Nowarski R, Thaiss CA, Hu B, J in C, Flavell RA: I nf lamm ati on - i nduce d canc e r : cros s tal k betw e e n tum our s, imm une cell s and micr oor gani sm s . Nature Reviews Cancer 2013, 13: 759. 70. Nguyen DP, Li J , Tewari AK: I nf l am m at ion and prost at e cancer: the role of interleukin  ,/-  . BJU Int 2014, 113( 6) :986- 992. 71. Taniguchi K, K arin M: I L -  and related cytokines as the critical lynchpins betZeen inflammation and cancer. Semin Immunol 2014, 26 (1):54- 74. 48 44. Sheridan C, Kishimoto H, Fuchs RK, Mehrotra S, Bhat - Nakshatri P, Turner CH, Goulet R, Jr., Badve S, Nakshatri H: CDCD - brea st cancer cells e[hibit enhanced invasive properties: an early step ne c e ss ar y for met ast asi s . Breast Cancer Res 2006, 8(5) : R59. 45. Liu S, Cong Y, Wang D, Sun Y, Deng L, Liu Y, Martin- Trevino R, Shang L, McDermott SP, Landis MD et al: B r ea s t cance r stem cells transition betZeen epithelial and mesenchymal states reflective of their normal counterparts. Stem Cell Reports 2014, 2 (1):78- 91. 46. Marjanovic ND, Weinberg RA, Chaffer CL: Ce ll plas ti c it y and heterogeneity in cancer . Clin Chem 2013, 59 (1):168- 179. 47. Shaw FL, Harrison H, Spence K, Ablett MP, Simões BM, Farnie G, Clarke RB: $ Detailed Mammosphere $ssay 3rotocol for the Q uantif i c at ion of Breast Stem Cell Acti vi t y . J Mammary Gland Biol Neoplasia 2012,  (2):111 - 117. 48. Dontu G, Abdallah WM, Foley JM, Jackson KW, Clarke MF, Kawa mura MJ, Wicha MS: I n vit r o propagat i on and transc r i pt ional profiling of human mammary stemprogenitor cells . Genes Dev 2003, (10):1253- 1270. 49. Bradshaw A, Wickremsekera A, Tan ST, Peng L, Davi s PF, Itinteang T: Cancer 6tem Cell Hierarchy in Glioblastoma Multiforme . Frontiers in Surgery 2016, 3(21). 50. Campbell LL, Polyak K: Breast tumor heterogeneity: c anc e r stem cell s or clonal evol ut i on? Cell Cycle 2007, 6(19):2332- 2338. 51. Greaves M, Maley CC: Cl onal evol uti on in canc e r . Nature 2012, 481 ( 7381):306- 313. 52. Bradshaw A, Wickremsekera A, Tan ST, Peng L, Davi s PF, Itinteang T: Cancer 6tem Cell Hierarchy in Glioblastoma Multiforme . Frontiers in surgery 2016, 3:21. 53. Beck B, Bla npain C: Unrave ll i ng canc er stem cel l pot e nti al . Nat Rev Cancer 2013, 13( 0):727 - 738. 54. Lobo NA, Shimono Y, Qian D, Clarke MF: The biology of c anc e r stem cell s . Annu Rev Cell Dev Biol 2007, 23 : 675 - 699. 55. Mani SA, Guo W, Liao M - J , Eaton EN, Ayyanan A, Z hou AY, Brooks M, Reinhard F, Zhang CC, Shipitsin M et al : The (pithelial - Mesenchymal Transition Generates Cells Zith 3roperties of 6tem Cells . Cell 2008, 133 ( 4) : 704 - 715. 56. Kal luri R, Weinberg RA: The basics of epithelial - mesenchymal transition. J Clin I nvest 2009, 119 ( 6) :1420 - 1428. 57. De Muynck L, Van Damme P: Cel l ul ar Eff ec ts of 3rogranulin in Health and Disease . J Mol Neurosci 2011, 45 ( 3): 549. 58. Abella V, Pino J, Scotece M, Conde J, Lago F, Gonzalez - Gay MA, Mera A, Gómez R, Mobasheri A, Gualillo O: 3rogranulin as a 49 biomarker and potential therapeutic agent . Drug Discovery Today 2017, 22(10):1557- 1564. 59. Arechavaleta- Velasco F, Perez - Juarez CE, Gerton GL, Diaz - Cueto L: 3rogranulin and its biological effects in cancer . Med Oncol 2017, 34(12):194. 60. Zhu J, Nathan C, Jin W, Sim D, Ashcroft GS, Wahl SM, Lacomis L, Erdjument - Bromage H, Tempst P, Wright CD et al: Conver si on of 3roepithelin to (pithelins: 5oles of 6/3, and (lastase in Host Defense and :ound 5epair . Cell 2002, 111 ( 6): 867- 878. 61. Suh H - S, Choi N, Tarassi shin L, Lee SC: Re gul at i on of 3rogranulin ([pression in Human Microglia and 3roteolysis of 3rogranulin by Matri[ Metalloproteinase -  MM3 -  . PLoS One 2012, ( 4):e35115. 62. Ghaemi manesh F, Mehravar M, Mil ani S, Poursani EM, Sal i minejad K: The multifaceted role of sortilinneurotensin receptor  in human cancer development . J Cell Physiol 2021, na (n/a). 63. Nykj aer A, Willnow TE: Sor t ili n: a rece pt or to regulat e neur onal viabil it y and func t i on . Trends Neurosci 2012, 35( 4) :261 - 270. 64. Mortensen MB, Kjolby M, Gunnersen S, Larsen JV, Pal mfeldt J, Fal k E, Nykj aer A, Bentzon JF: Tar get i ng sort il i n in imm une cell s reduces proinflammatory cytokines and atherosclerosis . J Clin Invest 2014, 124 (12):5317- 5322. 65. Goettsch C, Kjolby M, Aika wa E: Sor ti li n and Its Mul ti pl e 5oles in Cardiovascular and Metabolic Diseases . Arterioscler Thromb Vasc Biol 2018, 38 (1):19- 25. 66. Tang W, Lu Y, Tian QY, Zhang Y, Guo FJ, Liu GY, Syed NM, Lai Y, Lin EA, Kong L et al : The groZth factor progranulin binds to TNF receptors and is therapeutic against inflammatory arthritis in mice . Science 2011, 332 ( 6028):478- 484. 67. Neill T, Buraschi S, Goyal A, Sharpe C, Natkanski E, Schaefer L, Morrione A, Iozzo RV: (ph$ is a functional receptor for the groZth f ac tor progr anul i n . The Journal of Cell Biology 2016, 215( 5) :687- 703. 68. Lippitz BE: Cytokine patterns in patients Zith cancer: a s ys te m at ic revi ew . The Lancet Oncology 2013, 14 ( 6):e218- e228. 69. Elinav E, Nowarski R, Thaiss CA, Hu B, J in C, Flavell RA: I nf lamm ati on - i nduce d canc e r : cros s tal k betw e e n tum our s, imm une cell s and micr oor gani sm s . Nature Reviews Cancer 2013, 13: 759. 70. Nguyen DP, Li J , Tewari AK: I nf l am m at ion and prost at e cancer: the role of interleukin  ,/-  . BJU Int 2014, 113( 6) :986- 992. 71. Taniguchi K, K arin M: I L -  and related cytokines as the critical lynchpins betZeen inflammation and cancer. Semin Immunol 2014, 26 (1):54- 74. 50 72. Heikki l ä K, Ebrahim S, Lawlor DA: 6ystematic revieZ of the as s oci ati on betw e e n circ ulat i ng inte rl e uki n -  ,/-  and cancer . E ur J Cancer 2008, 44 ( 7): 937- 945. 73. Masjedi A, Hashemi V, Hojjat - Farsangi M, Ghalamf arsa G, Azi zi G, Yousefi M, Jadidi - Ni aragh F: The significant role of interleukin - 6 and its signaling pathZay in the immunopathogenesis and treatment of br e ast cance r . Bio med Pharmacother 2018, 108 :1415- 1424. 74. Kumari N, Dwarakanath BS, Das A, Bhatt AN: Rol e of int er l e uki n -  in cancer progression and therapeutic resistance . Tumor Biology 2016,  ( 9):11553- 11572. 75. Rose - John S, Scheller J, Elson G, Jones SA: I nte r le uki n - 6 biol ogy is coor di nat e d by mem br ane - bound and solubl e rece pt or s: rol e in inf l amm ati on and cance r . J Leukoc Biol 2006, 80(2):227- 236. 76. Heinrich PC, Behrmann I, Haan S, Hermanns HM, Mül ler - Newen G, Schaper F: 3rinciples of interleukin ,/ - 6- t ype cyt oki n e signal li ng and its regul at ion. The Biochemical journal 2003,  ( Pt 1):1 - 20. 77. Zarogoulidis P, Katsikogianni F, Tsiouda T, Sakka s A, Katsikogiannis N, Zarogoulidis K: I nt e rl e uki n - 8 and Int e rl e uki n-  for Canc e r . Cancer Invest 2014, 32( 5) :197 - 205. 78. Semenza GL: The hypo[ic tumor microenvironment: $ dr i vi ng f orc e for breas t canc e r progr es s ion . Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 2016, 1863( 3) :382- 391. 79. Favaro E, Lord S, Harris AL, Buffa FM: Ge ne expr e ss i on and hypo[ia in br e ast canc e r . Genome Med 2011, 3(8) : 55. 80. Chaturvedi P, Gilkes DM, Takano N, Semenza GL: Hypoxi a- i nduc i bl e fact or - de pe ndent signali ng betw e e n tri ple - ne gat ive breast cancer cells and mesenchymal stem cells promotes macrophage r e c r ui tm e nt . Proc Natl Acad Sci U S A 2014, 111(20):E2120- E2129. 81. Gilkes DM, Semenza GL: 5ole of hypo[ia- i nduci ble f actor s in breas t canc er met as t as is . Future Oncol 2013, 9 (11):1623- 1636. 82. Mirabelli P, Coppola L, Salvatore M: Cance r Cel l Line s Are Usef ul Model Syst em s for Medical 5esearch . Cancers (Basel) 2019, 11( 8): 098. 83. Wil ding JL, Bodmer WF: Canc e r cell line s f or drug disc ove r y and devel opm ent . Cancer Res 2014, ( 9):2377- 2384. 84. Chinwall a AT , Cook LL, Delehaunty KD, Fewell GA, Fulton LA, Fulton RS, Graves TA, Hil lier LW, Mardis ER, McPherson JD et al : ,nitial seTuencing and comparative analysis of the mouse genome . Nature 2002, 420 ( 6915):520- 562. 85. Okada S, Vaeteewoottacharn K, Kariya R: Appl ic ati on of Highly ,mmunocompromised Mice for the (stablishment of 3atient- Derived ;enograft 3D; Models . Cells 2019, 8( 8) : 889. 51 86. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB: ,dentification of human br ai n tum our ini ti at i ng cel l s . Nature 2004, 432 ( 7015):396- 401. 87. Au - Han D, Au - Rodriguez - Bravo V, Au - Izad mehr S, Au - Domingo - Domenech J, Au - Cordon- Cardo C: I s ol ati on and Characteri]ation of Tumor - initiating Cells from 6arcoma 3atient- de r i ve d Xenograf t s . JoVE 2019(148):e57011. 88. Rosfjord E, Lucas J , Li G, Gerber HP: Advan ce s in pat i e nt - de r i ve d tum or xenogr af t s: f rom targe t ident if i cati on to predi ct i ng cli ni c al response rat e s in oncol ogy . Biochem Pharmacol 2014, 91 (2):135- 143. 89. Hida lgo M, Amant F, Biankin AV, Budinská E, Byrne AT, Cal da s C, Clarke RB, de Jong S, Jonkers J, Mæl andsmo GM et al: 3atient - Derived ;enograft Models: $n (merging 3latform for Translational Cancer 5esearch. Cancer Discov 2014, 4 ( 9) : 998. 90. Jespersen H, Lindberg MF, Donia M, Söderberg EMV, Andersen R, Keller U, Ny L, Svane IM, Nil s son LM, Nil sson J A: Cli ni cal response s to adopt i ve T - ce l l transf e r can be model ed in an aut ol ogous imm une - humani]ed mouse model . Nature Communications 2017, 8 (1):707. 91. Kersten K, de Visser KE, van Miltenburg MH, Jonkers J: Ge ne t ic all y engi nee r e d mous e model s in oncol ogy research and cancer m e dic i ne . EMBO Mol Med 2017, 9 (2):137- 153. 92. Persson E, Gregersson P, Gustafsson A, Fitzpatrick P, Rhost S, Ståhlberg A, Landberg G: 3atient - der i ve d scaf f ol ds inf l ue nce secr et i on prof il es in canc er cell s mirr ori ng cli ni c al feat ur e s a nd breast canc e r subt ypes . Cell Communication and Signaling 2021, 19 (1):66. 93. Landberg G, Fitzpatrick P, Isa kson P, Jonasson E, Karlsson J, Larsson E, Svanström A, Rafnsdottir S, Persson E, Gustafsson A et al: 3atient - de ri ve d scaf f ol ds uncove r breast canc e r prom oti ng prope rt ie s of the microenvironment . Biomaterials 2020, 235 :119705. 94. Langhans SA: Three - Dimensional in 9itro Cell Culture Models in Drug Discovery and Drug 5epositioning. Front Pharmacol 2018, 9 : 6. 95. Ferreira LP, Gaspar VM, Mano JF: De c el l ul ari ze d ([tracellular Matri[ for Bioengineering 3hysiomimetic D in 9itro Tum o r Model s . Trends Biotechnol 2020. 96. Cruz - Acuña R, Vunjak - Novakovic G, Burdick J A, Rustgi AK: (merging technologies provide insights on cancer e[tracellular matri[ bi ol ogy and therapeutics . iScience 2021, 24( 5) :102475. 97. D'Angelo E, Natarajan D, Sensi F, Aj ayi O, Fass an M, Mammano E, Pil ati P, Pavan P, Bresolin S, Preziosi M et al : 3atient- Derived 6caffolds of Colorectal Cancer Metastases as an 2rganotypic D 50 72. Heikki l ä K, Ebrahim S, Lawlor DA: 6ystematic revieZ of the as s oci ati on betw e e n circ ulat i ng inte rl e uki n -  ,/-  and cancer . E ur J Cancer 2008, 44 ( 7): 937- 945. 73. Masjedi A, Hashemi V, Hojjat - Farsangi M, Ghalamf arsa G, Azi zi G, Yousefi M, Jadidi - Ni aragh F: The significant role of interleukin - 6 and its signaling pathZay in the immunopathogenesis and treatment of br e ast cance r . Bio med Pharmacother 2018, 108 :1415- 1424. 74. Kumari N, Dwarakanath BS, Das A, Bhatt AN: Rol e of int er l e uki n -  in cancer progression and therapeutic resistance . Tumor Biology 2016,  ( 9):11553- 11572. 75. Rose - John S, Scheller J, Elson G, Jones SA: I nte r le uki n - 6 biol ogy is coor di nat e d by mem br ane - bound and solubl e rece pt or s: rol e in inf l amm ati on and cance r . J Leukoc Biol 2006, 80(2):227- 236. 76. Heinrich PC, Behrmann I, Haan S, Hermanns HM, Mül ler - Newen G, Schaper F: 3rinciples of interleukin ,/ - 6- t ype cyt oki n e signal li ng and its regul at ion. The Biochemical journal 2003,  ( Pt 1):1 - 20. 77. Zarogoulidis P, Katsikogianni F, Tsiouda T, Sakka s A, Katsikogiannis N, Zarogoulidis K: I nt e rl e uki n - 8 and Int e rl e uki n-  for Canc e r . Cancer Invest 2014, 32( 5) :197 - 205. 78. Semenza GL: The hypo[ic tumor microenvironment: $ dr i vi ng f orc e for breas t canc e r progr es s ion . Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 2016, 1863( 3) :382- 391. 79. Favaro E, Lord S, Harris AL, Buffa FM: Ge ne expr e ss i on and hypo[ia in br e ast canc e r . Genome Med 2011, 3(8) : 55. 80. Chaturvedi P, Gilkes DM, Takano N, Semenza GL: Hypoxi a- i nduc i bl e fact or - de pe ndent signali ng betw e e n tri ple - ne gat ive breast cancer cells and mesenchymal stem cells promotes macrophage r e c r ui tm e nt . Proc Natl Acad Sci U S A 2014, 111(20):E2120- E2129. 81. Gilkes DM, Semenza GL: 5ole of hypo[ia- i nduci ble f actor s in breas t canc er met as t as is . Future Oncol 2013, 9 (11):1623- 1636. 82. Mirabelli P, Coppola L, Salvatore M: Cance r Cel l Line s Are Usef ul Model Syst em s for Medical 5esearch . Cancers (Basel) 2019, 11( 8): 098. 83. Wil ding JL, Bodmer WF: Canc e r cell line s f or drug disc ove r y and devel opm ent . Cancer Res 2014, ( 9):2377- 2384. 84. Chinwall a AT , Cook LL, Delehaunty KD, Fewell GA, Fulton LA, Fulton RS, Graves TA, Hil lier LW, Mardis ER, McPherson JD et al : ,nitial seTuencing and comparative analysis of the mouse genome . Nature 2002, 420 ( 6915):520- 562. 85. Okada S, Vaeteewoottacharn K, Kariya R: Appl ic ati on of Highly ,mmunocompromised Mice for the (stablishment of 3atient- Derived ;enograft 3D; Models . Cells 2019, 8( 8) : 889. 51 86. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB: ,dentification of human br ai n tum our ini ti at i ng cel l s . Nature 2004, 432 ( 7015):396- 401. 87. Au - Han D, Au - Rodriguez - Bravo V, Au - Izad mehr S, Au - Domingo - Domenech J, Au - Cordon- Cardo C: I s ol ati on and Characteri]ation of Tumor - initiating Cells from 6arcoma 3atient- de r i ve d Xenograf t s . JoVE 2019(148):e57011. 88. Rosfjord E, Lucas J , Li G, Gerber HP: Advan ce s in pat i e nt - de r i ve d tum or xenogr af t s: f rom targe t ident if i cati on to predi ct i ng cli ni c al response rat e s in oncol ogy . Biochem Pharmacol 2014, 91 (2):135- 143. 89. Hida lgo M, Amant F, Biankin AV, Budinská E, Byrne AT, Cal da s C, Clarke RB, de Jong S, Jonkers J, Mæl andsmo GM et al: 3atient - Derived ;enograft Models: $n (merging 3latform for Translational Cancer 5esearch. Cancer Discov 2014, 4 ( 9) : 998. 90. Jespersen H, Lindberg MF, Donia M, Söderberg EMV, Andersen R, Keller U, Ny L, Svane IM, Nil s son LM, Nil sson J A: Cli ni cal response s to adopt i ve T - ce l l transf e r can be model ed in an aut ol ogous imm une - humani]ed mouse model . Nature Communications 2017, 8 (1):707. 91. Kersten K, de Visser KE, van Miltenburg MH, Jonkers J: Ge ne t ic all y engi nee r e d mous e model s in oncol ogy research and cancer m e dic i ne . EMBO Mol Med 2017, 9 (2):137- 153. 92. Persson E, Gregersson P, Gustafsson A, Fitzpatrick P, Rhost S, Ståhlberg A, Landberg G: 3atient - der i ve d scaf f ol ds inf l ue nce secr et i on prof il es in canc er cell s mirr ori ng cli ni c al feat ur e s a nd breast canc e r subt ypes . Cell Communication and Signaling 2021, 19 (1):66. 93. Landberg G, Fitzpatrick P, Isa kson P, Jonasson E, Karlsson J, Larsson E, Svanström A, Rafnsdottir S, Persson E, Gustafsson A et al: 3atient - de ri ve d scaf f ol ds uncove r breast canc e r prom oti ng prope rt ie s of the microenvironment . Biomaterials 2020, 235 :119705. 94. Langhans SA: Three - Dimensional in 9itro Cell Culture Models in Drug Discovery and Drug 5epositioning. Front Pharmacol 2018, 9 : 6. 95. Ferreira LP, Gaspar VM, Mano JF: De c el l ul ari ze d ([tracellular Matri[ for Bioengineering 3hysiomimetic D in 9itro Tum o r Model s . Trends Biotechnol 2020. 96. Cruz - Acuña R, Vunjak - Novakovic G, Burdick J A, Rustgi AK: (merging technologies provide insights on cancer e[tracellular matri[ bi ol ogy and therapeutics . iScience 2021, 24( 5) :102475. 97. D'Angelo E, Natarajan D, Sensi F, Aj ayi O, Fass an M, Mammano E, Pil ati P, Pavan P, Bresolin S, Preziosi M et al : 3atient- Derived 6caffolds of Colorectal Cancer Metastases as an 2rganotypic D 52 Model of the Li ve r Met astat i c Micr oe nvir onme nt . Cancers (Basel) 2020, 12( ). 98. Dunne LW, Huang Z, Meng W, Fan X, Zhang N, Zhang Q, An Z: H um a n decel l ul ari ze d adi pos e tiss ue scaf f ol d as a mode l f or breas t cancer cell groZth and drug treatments . Biomaterials 2014, 35 (1 8):4940- 4949. 99. Parkinson GT, Salerno S, Ranji P, Håkansson J, Bogestål Y, Wettergren Y, Ståhlberg A, Bexe Lindskog E, Landberg G: 3atient - der i ve d scaf f ol ds as a mode l of color e ct al canc er . Cancer Medicine 2021, 10( 3): 867- 882. 100. Gustafsson A, Garre E, Leiva MC, Salerno S, Ståhlberg A, Landberg G: 3atient - de ri ve d scaf f ol ds as a drug - t e sti ng pl atf orm f or endocrine therapies in breast cancer . Sci Rep 2021, 11( ):13334. 101. Reynolds BA, Weiss S: Gene r ati on of neurons and ast r oc yte s from isol at e d cel ls of the adult mammalian central nervous s ys te m . Science 1992, 255(5052):1707- 1710. 102. Harrison H, Simões BM, Rogerson L, Howell SJ, Landberg G, Clarke RB: 2estrogen increases the activity of oestrogen receptor negative breast cancer stem cells through paracrine (GF5 and 1otch s i gnal li ng . Breast Cancer Res 2013, 15(2):R21. 103. VanGuilder HD, Vrana KE, Freeman WM: Tw e nt y - f ive year s of Tuantitative 3C5 for gene e[pression analysis. Biotechniques 2008, 44( 5): 619- 626. 104. Ståhlberg A, Kubista M: The ZorkfloZ of s i ngl e - c el l expr e s s i on prof il i ng usi ng quant it ati ve real - time 3C5. Expert Rev Mol Diagn 2014, 14 ( 3): 323- 331. 105. Bass JJ, Wil kinson DJ , Rankin D, Philli ps BE, Szewczyk NJ , Smith K, Atherton PJ: $n overvieZ of technical considerations for :estern blotting applications to physiological research. Scand J Med Sci Sports 2017, (1):4 - 25. 106. Clark TG, Bradburn MJ , Love SB, Altman DG: Sur vival $nalysis 3art ,: Basic concepts and first analyses . Br J Cancer 2003, 89(2):232- 238. 107. Bradburn MJ, Clark TG, Love SB, Altman DG: Sur vival $nalysis 3art ,,: Multivariate data analysis – an int r oduct i on to conc e pt s and methods . Br J Cancer 2003, 89 ( 3): 431- 436. 108. Prieto - Vila M, Takahashi R - U, Usuba W, Kohama I, Ochiya T: Drug 5esistance Driven by Cancer 6tem Cells and Their 1iche . Int J Mol Sci 2017, 18 (12):2574. 109. Jacobsson H, Harrison H, Hughes E, Persson E, Rhost S, Fitzpatrick P, Gustafsson A, Andersson D, Gregersson P, Magnusson Y et al: H ypoxi a- i nduce d secr et i on stim ul ate s breast cance r stem cell regul at or y signalling pathZays . Mol Oncol 2019. 53 110. Lee WC, Almeida S, Prudencio M, Caulfield TR, Zhang Y - J , Tay WM, Bauer PO, Chew J, Sasaguri H, J ansen - West KR et al: Tar ge te d manipulation of the sortilin–pr ogr anul i n axi s resc ues progr anuli n haploinsufficiency. Hum Mol Genet 2014, 23( 6) :1467 - 1478. 111. Zheng Y, Brady OA, Meng PS, Mao Y, Hu F: C - Term i nus of 3rogranulin ,nteracts Zith the Beta- 3ropeller 5egion of 6ortilin to 5egulate 3rogranulin Trafficking. PLoS One 2011, 6 ( 6):e21023. 112. Tanimoto R, Morcavallo A, Terracci ano M, Xu S - Q, Stefanello M, Buraschi S, Lu KG, Bagley DH, Gomella LG, Scotlandi K et al : Sor ti li n regul at es progr anuli n act i on in castr ati on- r e si st ant prost ate canc e r cel ls . Endocrinology 2015, 156 (1):58 - 70. 113. Ghaem i manesh F, Ahmadian G, Talebi S, Zarnani A - H, Behmanesh M, Hemmati S, Hada vi R, Jeddi - Tehrani M, Farzi M, Akhondi MM et al : The effect of sortilin silencing on ovarian carcinoma cells . Avicenna J Med Biotechnol 2014, 6 ( 3):169- 177. 114. Schrøder TJ, Christe nsen S, Lindberg S, Langgård M, Davi d L, Maltas PJ , Eski l dsen J, Jacobsen J, Tagmose L, Simonsen KB et al : The identification of $F: $n orally bioavailable inhibitor of the 9363 family sorting receptor 6ortilin. Bioorg Med Chem Lett 2014, 24(1):177- 180. 115. Rhost S, Hughes E, Harrison H, Rafnsdottir S, Jacobsson H, Gregersson P, Magnusson Y, Fitzpatrick P, Andersson D, Berger K et al: 6ortilin inhibition limits secretion- i nduc e d progr anul i n - de pe nde nt breast cance r progr es si on and canc er stem cel l ex pans i on. Breast Cancer Res 2018, 20(1):137. 116. Peinado H, Zhang H, Matei IR, Costa - Si l va B, Hoshino A, Rodrigues G, Psail a B, Kapl an RN, Bromberg JF, Kang Y et al : 3re - metastatic niches: organ- specific homes for metastases . Nature Reviews Cancer 2017,  ( 5): 302- 317. 117. Conley SJ , Gheordunescu E, Kakarala P, Newman B, Korkaya H, Heath AN, Clouthier SG, Wicha MS: Anti a ngi oge ni c agent s incr e ase breast cancer stem cells via the generation of tumor hypo[ia. Proc Natl Acad Sci U S A 2012, 109 (8) :2784- 2789. 118. Lu R, Serrero G: ,nhibition of 3C cell - derived groZth factor 3CDGF epithelingranulin precursor e[pression by antisense 3CDGF cD1$ transfection inhibits tumorigenicity of the human breast carcinoma cell line MD$- M B - 468. Proc Natl Acad Sci U S A 2000, ( 8): 3993- 3998. 119. Pizarro GO, Zhou XC, Koch A, Gharib M, Rava l S, Bible K, Jones MB: 3rosurvival function of the granulin- epithelin precursor is important in tumor progression and chemoresponse . Int J Cancer 2007, 120( 1):2339- 2343. 52 Model of the Li ve r Met astat i c Micr oe nvir onme nt . Cancers (Basel) 2020, 12( ). 98. Dunne LW, Huang Z, Meng W, Fan X, Zhang N, Zhang Q, An Z: H um a n decel l ul ari ze d adi pos e tiss ue scaf f ol d as a mode l f or breas t cancer cell groZth and drug treatments . Biomaterials 2014, 35 (1 8):4940- 4949. 99. Parkinson GT, Salerno S, Ranji P, Håkansson J, Bogestål Y, Wettergren Y, Ståhlberg A, Bexe Lindskog E, Landberg G: 3atient - der i ve d scaf f ol ds as a mode l of color e ct al canc er . Cancer Medicine 2021, 10( 3): 867- 882. 100. Gustafsson A, Garre E, Leiva MC, Salerno S, Ståhlberg A, Landberg G: 3atient - de ri ve d scaf f ol ds as a drug - t e sti ng pl atf orm f or endocrine therapies in breast cancer . Sci Rep 2021, 11( ):13334. 101. Reynolds BA, Weiss S: Gene r ati on of neurons and ast r oc yte s from isol at e d cel ls of the adult mammalian central nervous s ys te m . Science 1992, 255(5052):1707- 1710. 102. Harrison H, Simões BM, Rogerson L, Howell SJ, Landberg G, Clarke RB: 2estrogen increases the activity of oestrogen receptor negative breast cancer stem cells through paracrine (GF5 and 1otch s i gnal li ng . Breast Cancer Res 2013, 15(2):R21. 103. VanGuilder HD, Vrana KE, Freeman WM: Tw e nt y - f ive year s of Tuantitative 3C5 for gene e[pression analysis. Biotechniques 2008, 44( 5): 619- 626. 104. Ståhlberg A, Kubista M: The ZorkfloZ of s i ngl e - c el l expr e s s i on prof il i ng usi ng quant it ati ve real - time 3C5. Expert Rev Mol Diagn 2014, 14 ( 3): 323- 331. 105. Bass JJ, Wil kinson DJ , Rankin D, Philli ps BE, Szewczyk NJ , Smith K, Atherton PJ: $n overvieZ of technical considerations for :estern blotting applications to physiological research. Scand J Med Sci Sports 2017, (1):4 - 25. 106. Clark TG, Bradburn MJ , Love SB, Altman DG: Sur vival $nalysis 3art ,: Basic concepts and first analyses . Br J Cancer 2003, 89(2):232- 238. 107. Bradburn MJ, Clark TG, Love SB, Altman DG: Sur vival $nalysis 3art ,,: Multivariate data analysis – an int r oduct i on to conc e pt s and methods . Br J Cancer 2003, 89 ( 3): 431- 436. 108. Prieto - Vila M, Takahashi R - U, Usuba W, Kohama I, Ochiya T: Drug 5esistance Driven by Cancer 6tem Cells and Their 1iche . Int J Mol Sci 2017, 18 (12):2574. 109. Jacobsson H, Harrison H, Hughes E, Persson E, Rhost S, Fitzpatrick P, Gustafsson A, Andersson D, Gregersson P, Magnusson Y et al: H ypoxi a- i nduce d secr et i on stim ul ate s breast cance r stem cell regul at or y signalling pathZays . Mol Oncol 2019. 53 110. Lee WC, Almeida S, Prudencio M, Caulfield TR, Zhang Y - J , Tay WM, Bauer PO, Chew J, Sasaguri H, J ansen - West KR et al: Tar ge te d manipulation of the sortilin–pr ogr anul i n axi s resc ues progr anuli n haploinsufficiency. Hum Mol Genet 2014, 23( 6) :1467 - 1478. 111. Zheng Y, Brady OA, Meng PS, Mao Y, Hu F: C - Term i nus of 3rogranulin ,nteracts Zith the Beta- 3ropeller 5egion of 6ortilin to 5egulate 3rogranulin Trafficking. PLoS One 2011, 6 ( 6):e21023. 112. Tanimoto R, Morcavallo A, Terracci ano M, Xu S - Q, Stefanello M, Buraschi S, Lu KG, Bagley DH, Gomella LG, Scotlandi K et al : Sor ti li n regul at es progr anuli n act i on in castr ati on- r e si st ant prost ate canc e r cel ls . Endocrinology 2015, 156 (1):58 - 70. 113. Ghaem i manesh F, Ahmadian G, Talebi S, Zarnani A - H, Behmanesh M, Hemmati S, Hada vi R, Jeddi - Tehrani M, Farzi M, Akhondi MM et al : The effect of sortilin silencing on ovarian carcinoma cells . Avicenna J Med Biotechnol 2014, 6 ( 3):169- 177. 114. Schrøder TJ, Christe nsen S, Lindberg S, Langgård M, Davi d L, Maltas PJ , Eski l dsen J, Jacobsen J, Tagmose L, Simonsen KB et al : The identification of $F: $n orally bioavailable inhibitor of the 9363 family sorting receptor 6ortilin. Bioorg Med Chem Lett 2014, 24(1):177- 180. 115. Rhost S, Hughes E, Harrison H, Rafnsdottir S, Jacobsson H, Gregersson P, Magnusson Y, Fitzpatrick P, Andersson D, Berger K et al: 6ortilin inhibition limits secretion- i nduc e d progr anul i n - de pe nde nt breast cance r progr es si on and canc er stem cel l ex pans i on. Breast Cancer Res 2018, 20(1):137. 116. Peinado H, Zhang H, Matei IR, Costa - Si l va B, Hoshino A, Rodrigues G, Psail a B, Kapl an RN, Bromberg JF, Kang Y et al : 3re - metastatic niches: organ- specific homes for metastases . Nature Reviews Cancer 2017,  ( 5): 302- 317. 117. Conley SJ , Gheordunescu E, Kakarala P, Newman B, Korkaya H, Heath AN, Clouthier SG, Wicha MS: Anti a ngi oge ni c agent s incr e ase breast cancer stem cells via the generation of tumor hypo[ia. Proc Natl Acad Sci U S A 2012, 109 (8) :2784- 2789. 118. Lu R, Serrero G: ,nhibition of 3C cell - derived groZth factor 3CDGF epithelingranulin precursor e[pression by antisense 3CDGF cD1$ transfection inhibits tumorigenicity of the human breast carcinoma cell line MD$- M B - 468. Proc Natl Acad Sci U S A 2000, ( 8): 3993- 3998. 119. Pizarro GO, Zhou XC, Koch A, Gharib M, Rava l S, Bible K, Jones MB: 3rosurvival function of the granulin- epithelin precursor is important in tumor progression and chemoresponse . Int J Cancer 2007, 120( 1):2339- 2343. 54 120. Serrero G, Hi cks D: ,mmunohistochemical Detection of 3rogranulin 3G51G3G(3 in Tumor Tissues as a Cancer 3rognostic Biomarker . In: Progranulin: Methods and Protocols. edn. Edited by Bateman A, Bennett HPJ , Cheung ST. New York, NY: Springer New York; 2018: 107- 120. 121. Serrero G, Hawkins DM, Yue B, Ioffe O, Bejarano P, Philli ps JT , Head J F, Ell iott RL, Tkac zuk KR, Godwin AK et al : 3rogranulin G3 tumor tissue e[pression is associated Zith increased risk of recurrence in breast cancer patients diagnosed Zith e s tr oge n recept or posit i ve invas i ve duct al carc i nom a . Breast Cancer Research : BCR 2012, 14( ): R26- R26. 122. Roselli S, Pundavela J , Demont Y, Faulkner S, Keene S, Attia J , Jiang CC, Zhang XD, Wal ker MM, Hondermarck H: So rt il i n is associ at e d Zith breast cancer aggressiveness and contributes to tumor cell adhesion and invasi on . Oncotarget 2015, 6(12):10473- 10486. 123. Tangkeangsirisin W, Haya shi J, Serrero G: 3C Cell - Derived GroZth Factor Mediates Tamo[ifen 5esistance and 3romotes Tumor GroZth of Human Breas t Canc e r Cel ls . Cancer Res 2004, 64 ( 5):1737- 1743. 124. Davies C, Pan H, Godwin J, Gray R, Arriagada R, Raina V, Abraham M, Medeiros Alencar VH, Badran A, Bonfill X et al : Long - te rm effects of continuing adMuvant tamo[ifen to  years versus stopping at  ye ar s af te r diagnos i s of oest r oge n rece pt or - posi ti ve breast cance r: ATLAS, a randomi s e d trial . Lancet 2013, 381 ( 9869): 805- 816. 125. Gray R, Clarke M, Collins R, Peto R: The (BCTCG overvieZ of adMuvant therapy of breast cancer :hat are the i m pl ic ati ons for fut ur e studi e s ? Earl y Breas t Cance r Triali s ts ' Col l abor ati ve Group . Ann N Y Acad Sci 1993, 698 : 339 - 348. 126. Wang Z, Chen J - Q, Liu J - l , Ti an L: Exos om es in tumor micr oe nvi r onm e nt : novel trans por te r s and biom ar ke r s . J Transl Med 2016, 14 (1):297. 127. Wang N , Liu W, Zheng Y, Wang S, Yang B, Li M, Song J, Zhang F, Zhang X, Wang Q et al: CXCL1 deri ve d from tum or - ass oci ate d macrophages promotes breast cancer metastasis via activating 1F- țB62; signaling . Cell Death Dis 2018, 9( 9) :880. 128. Wang D, Sun H, Wei J, Cen B, DuBois RN: CXCL1 Is Critical for 3remetastatic 1iche Formation and Metastasis in Colorectal Canc e r . Cancer Res 2017, (13):3655- 3665. 129. Méndez - García LA, Nava - Ca stro KE, Ochoa - Mercado TL, Pal a cios- Arreola MI, Ruiz- Manzano RA, Segovia - Mendoza M, Solleiro - Vil la vi cencio H, Cázarez - Martínez C, Morales- Montor J: B r e as t Cance r Metastasis: $re Cytokines ,mportant 3layers During ,ts Development and 3rogression" J Interferon Cytokine Res 2019, 39(1):39- 55. 55 130. Jing B, Wang T, Sun B, Xu J, Xu D, Liao Y, Song H, Guo W, Li K, Hu M et al : ,/6T$T 6ignaling 2rchestrates 3remetastatic 1iche F orm at i on and Imm unosuppr e s si ve Trait s in Lung . Cancer Res 2020, 80( 4): 784- 797. 131. Jin F, Miao Y, Xu P, Qiu X: I L -  regulates the stemness pr ope rt i es of cance r stem cel ls in the small - c e ll lung canc er cel l line H446. Onco Targets Ther 2018, 11 : 5723- 5731. 132. Corrò C, Healy ME, Engler S, Bodenmiller B, Li Z, Schraml P, Weber A, Frew IJ , Rechsteiner M, Moch H: I L - 8 and CXCR1 expr e s s i on is associated Zith cancer stem cell- li ke prope rt ie s of clear cel l renal canc e r . The Journal of Pathology 2019, 248(3): 377- 389. 133. Paltridge JL, Belle L, Khew - Goodall Y: The secretome in c anc e r progre s si on . Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics 2013, 1834( 1):2233- 2241. 134. Matsubara T, Mita A, Minami K, Hosooka T, Kitazawa S, Takahashi K, Tamori Y, Yokoi N, Watanabe M, Matsuo E et al: 3G51 is a key adipokine mediating high fat diet- i nduc e d insul i n resi s tanc e and obesity through ,/- 6 in adi pos e tiss ue . Cell Metab 2012, 15( ): 38 - 50. 135. Liu F, Zhang W, Yang F, Feng T, Zhou M, Yu Y, Yu X, Zhao W, Yi F, Tang W et al: I nte r l e uki n - 6- st im ul ate d progr anuli n expr e ss i on contributes to the malignancy of hepatocellular carcinoma cells by ac t ivati ng mTO R signal i ng. Sci Rep 2016, 6:21260. 136. Frampton G, Invernizzi P, Bernuzzi F, Pae HY, Quinn M, Horvat D, Galindo C, Huang L, McMil lin M, Cooper B et al : Int er l e uki n - 6- driven progranulin e[pression increases cholangiocarcinoma groZth by an Akt - dependent mechanism . Gut 2012, 61 (2):268- 277. 137. Dethlefsen C, Høj feldt G, Hojman P: The role of i nt r at um or al and syste mic IL - 6 in breast cance r . Breast Cancer Res Treat 2013, 138 ( 3): 657- 664. 138. Yabe - Wada T, Matsuba S, Takeda K, Sato T, Suyama M, Ohkawa Y, Taka i T, Shi H, Philpott CC, Naka mura A: TLR signals posttranscriptionally regulate the cytokine trafficking mediator sortilin. Sci Rep 2016, 6(1):26566. 139. Bhattacharya S, Calar K, de la Puente P: M im ic ki ng tum or hypo[ia and tumor- immune interactions employing three - di m e ns i onal in vit ro mode ls . J Exp Clin Cancer Res 2020, 39 (1):75. 140. Chavey C, Bibeau F, Gourgou - Bourgade S, Burlinchon S, Boissi ère F, Laune D, Roques S, Lazennec G: O es tr oge n rece pt or negati ve breast cancers e[hibit high cytokine content . Breast Cancer Res 2007, 9( 1):R15. 141. Nguyen AD, Nguyen TA, Martens LH, Mitic LL, Farese RV, Jr.: 3rogranulin: at the interface of neurodegenerative and metabolic di s e as e s . Trends Endocrinol Metab 2013, 24 (12):597- 606. 54 120. Serrero G, Hi cks D: ,mmunohistochemical Detection of 3rogranulin 3G51G3G(3 in Tumor Tissues as a Cancer 3rognostic Biomarker . In: Progranulin: Methods and Protocols. edn. Edited by Bateman A, Bennett HPJ , Cheung ST. New York, NY: Springer New York; 2018: 107- 120. 121. Serrero G, Hawkins DM, Yue B, Ioffe O, Bejarano P, Philli ps JT , Head J F, Ell iott RL, Tkac zuk KR, Godwin AK et al : 3rogranulin G3 tumor tissue e[pression is associated Zith increased risk of recurrence in breast cancer patients diagnosed Zith e s tr oge n recept or posit i ve invas i ve duct al carc i nom a . Breast Cancer Research : BCR 2012, 14( ): R26- R26. 122. Roselli S, Pundavela J , Demont Y, Faulkner S, Keene S, Attia J , Jiang CC, Zhang XD, Wal ker MM, Hondermarck H: So rt il i n is associ at e d Zith breast cancer aggressiveness and contributes to tumor cell adhesion and invasi on . Oncotarget 2015, 6(12):10473- 10486. 123. Tangkeangsirisin W, Haya shi J, Serrero G: 3C Cell - Derived GroZth Factor Mediates Tamo[ifen 5esistance and 3romotes Tumor GroZth of Human Breas t Canc e r Cel ls . Cancer Res 2004, 64 ( 5):1737- 1743. 124. Davies C, Pan H, Godwin J, Gray R, Arriagada R, Raina V, Abraham M, Medeiros Alencar VH, Badran A, Bonfill X et al : Long - te rm effects of continuing adMuvant tamo[ifen to  years versus stopping at  ye ar s af te r diagnos i s of oest r oge n rece pt or - posi ti ve breast cance r: ATLAS, a randomi s e d trial . Lancet 2013, 381 ( 9869): 805- 816. 125. Gray R, Clarke M, Collins R, Peto R: The (BCTCG overvieZ of adMuvant therapy of breast cancer :hat are the i m pl ic ati ons for fut ur e studi e s ? Earl y Breas t Cance r Triali s ts ' Col l abor ati ve Group . Ann N Y Acad Sci 1993, 698 : 339 - 348. 126. Wang Z, Chen J - Q, Liu J - l , Ti an L: Exos om es in tumor micr oe nvi r onm e nt : novel trans por te r s and biom ar ke r s . J Transl Med 2016, 14 (1):297. 127. Wang N , Liu W, Zheng Y, Wang S, Yang B, Li M, Song J, Zhang F, Zhang X, Wang Q et al: CXCL1 deri ve d from tum or - ass oci ate d macrophages promotes breast cancer metastasis via activating 1F- țB62; signaling . Cell Death Dis 2018, 9( 9) :880. 128. Wang D, Sun H, Wei J, Cen B, DuBois RN: CXCL1 Is Critical for 3remetastatic 1iche Formation and Metastasis in Colorectal Canc e r . Cancer Res 2017, (13):3655- 3665. 129. Méndez - García LA, Nava - Ca stro KE, Ochoa - Mercado TL, Pal a cios- Arreola MI, Ruiz- Manzano RA, Segovia - Mendoza M, Solleiro - Vil la vi cencio H, Cázarez - Martínez C, Morales- Montor J: B r e as t Cance r Metastasis: $re Cytokines ,mportant 3layers During ,ts Development and 3rogression" J Interferon Cytokine Res 2019, 39(1):39- 55. 55 130. Jing B, Wang T, Sun B, Xu J, Xu D, Liao Y, Song H, Guo W, Li K, Hu M et al : ,/6T$T 6ignaling 2rchestrates 3remetastatic 1iche F orm at i on and Imm unosuppr e s si ve Trait s in Lung . Cancer Res 2020, 80( 4): 784- 797. 131. Jin F, Miao Y, Xu P, Qiu X: I L -  regulates the stemness pr ope rt i es of cance r stem cel ls in the small - c e ll lung canc er cel l line H446. Onco Targets Ther 2018, 11 : 5723- 5731. 132. Corrò C, Healy ME, Engler S, Bodenmiller B, Li Z, Schraml P, Weber A, Frew IJ , Rechsteiner M, Moch H: I L - 8 and CXCR1 expr e s s i on is associated Zith cancer stem cell- li ke prope rt ie s of clear cel l renal canc e r . The Journal of Pathology 2019, 248(3): 377- 389. 133. Paltridge JL, Belle L, Khew - Goodall Y: The secretome in c anc e r progre s si on . Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics 2013, 1834( 1):2233- 2241. 134. Matsubara T, Mita A, Minami K, Hosooka T, Kitazawa S, Takahashi K, Tamori Y, Yokoi N, Watanabe M, Matsuo E et al: 3G51 is a key adipokine mediating high fat diet- i nduc e d insul i n resi s tanc e and obesity through ,/- 6 in adi pos e tiss ue . Cell Metab 2012, 15( ): 38 - 50. 135. Liu F, Zhang W, Yang F, Feng T, Zhou M, Yu Y, Yu X, Zhao W, Yi F, Tang W et al: I nte r l e uki n - 6- st im ul ate d progr anuli n expr e ss i on contributes to the malignancy of hepatocellular carcinoma cells by ac t ivati ng mTO R signal i ng. Sci Rep 2016, 6:21260. 136. Frampton G, Invernizzi P, Bernuzzi F, Pae HY, Quinn M, Horvat D, Galindo C, Huang L, McMil lin M, Cooper B et al : Int er l e uki n - 6- driven progranulin e[pression increases cholangiocarcinoma groZth by an Akt - dependent mechanism . Gut 2012, 61 (2):268- 277. 137. Dethlefsen C, Høj feldt G, Hojman P: The role of i nt r at um or al and syste mic IL - 6 in breast cance r . Breast Cancer Res Treat 2013, 138 ( 3): 657- 664. 138. Yabe - Wada T, Matsuba S, Takeda K, Sato T, Suyama M, Ohkawa Y, Taka i T, Shi H, Philpott CC, Naka mura A: TLR signals posttranscriptionally regulate the cytokine trafficking mediator sortilin. Sci Rep 2016, 6(1):26566. 139. Bhattacharya S, Calar K, de la Puente P: M im ic ki ng tum or hypo[ia and tumor- immune interactions employing three - di m e ns i onal in vit ro mode ls . J Exp Clin Cancer Res 2020, 39 (1):75. 140. Chavey C, Bibeau F, Gourgou - Bourgade S, Burlinchon S, Boissi ère F, Laune D, Roques S, Lazennec G: O es tr oge n rece pt or negati ve breast cancers e[hibit high cytokine content . Breast Cancer Res 2007, 9( 1):R15. 141. Nguyen AD, Nguyen TA, Martens LH, Mitic LL, Farese RV, Jr.: 3rogranulin: at the interface of neurodegenerative and metabolic di s e as e s . Trends Endocrinol Metab 2013, 24 (12):597- 606. 56 142. Berger K, Rhost S, Rafnsdóttir S, Hughes É, Magnusson Y, Ekholm M, Stål O, Rydén L, Landberg G: Tum or co - e xpr e s s i on of progr anul i n and sort il i n as a prognos ti c biom ar ker in breas t canc er . BMC Cancer 2021, 21( ):185. 143. Vermeire K, Zhang Y, Princen K, Hatse S, Sama l a MF, Dey K, Choi HJ , Ahn Y, Sodoma A, Snoeck R et al: C$D$ inhibits human immunodeficiency virus and human herpesvirus  replication by doZn- modulation of the cellular CD receptor . Virology 2002, 302 (2):342- 353. 144. Van Puyenbroeck V, Claeys E, Schols D, Bell TW, Vermeire K: $ 3roteomic 6urvey ,ndi cates 6ortilin as a 6econdary 6ubstrate of the (5 Translocation ,nhibitor Cyclotria]adisulfonamide C$D$ . Mol Cell Proteomics 2017, 16(2):157 - 167. 145. Farahi L, Ghaemimanesh F, Mil ani S, Raza vi SM, Akhondi MM, Rabbani H: 6ortilin as a 1ovel Diagnostic an d Therapeutic Biomarker in Chronic /ymphocytic /eukemia. Avicenna J Med Biotechnol 2019, 11 ( 4):270 - 276. 146. Gao F, Griffin N, Faulkner S, Li X, King SJ , Jobling P, Denham JW, Jiang CC, Hondermarck H: The Membrane 3rotein 6ortilin Can Be Targeted to ,nhibit 3ancreatic Cancer Cell ,nvasion. The American Journal of Pathology 2020, 190 ( 9) :1931 - 1942. 147. Das S, Srikanth M, Kessler JA: Canc er stem cel ls and gl i om a . Nature Clinical Practice Neurology 2008, 4 : 427. 148. Tandon M, Vemula SV , Mittal SK : Em er gi ng str ate gie s f or (ph$ receptor targeting for cancer therapeutics . Expert Opin Ther Targets 2011, 15 (1):31- 51. 149. Tolkatchev D, Mali k S, Vinogradova A, Wang P, Chen Z, Xu P, Bennett HPJ , Bateman A, Ni F: 6tructure dissection of human pr ogr anul i n ident if ie s we ll - folded granulinepithelin modules Zith uni que f unc ti onal act i vi tie s . Protein Science : A Publication of the Protein Society 2008,  ( 4) : 711 - 724. 150. Zhang T, Du H, Santos MN, Wu X, Reinheckel T, Hu F: Differential regulation of progranulin derived gr anul i n pepti de s . bioRxiv 2021:2021.2001.2008.425959.