CTR2 Identifies a Population of Cancer Cells with Stem Cell-like Features in Patients with Clear Cell Renal Cell Carcinoma Vanessa Galleggiante,* Monica Rutigliano,* Fabio Sallustio, Domenico Ribatti, Pasquale Ditonno, Carlo Bettocchi, Francesco Paolo Selvaggi, Giuseppe Lucarelli* and Michele Battaglia*,† From the Andrology and Kidney Transplantation Unit, Department of Emergency and Organ Transplantation-Urology (VG, MR, PD, CB, FPS, GL, MB), Dialysis and Transplantation Unit, Department of Emergency and Organ Transplantation-Nephrology (FS) and Department of Basic Medical Sciences, Neurosciences and Sensory Organs (DR), University of Bari, Bari, Italy
Purpose: In clear cell renal cell carcinoma tissue samples we identified and characterized a population of renal cell carcinoma derived CD133þ/CD24þ cancer cells. We studied differences between these cells and their nonneoplastic counterpart, tubular adult renal progenitor cells. Materials and Methods: CD133þ/CD24þ renal cell carcinoma derived cells were isolated from 40 patients. The mesenchymal phenotype and stemness proteomic profile of these renal cell carcinoma derived cells were characterized. Colony forming efficiency and self-renewal ability were tested by limiting dilution. Tumorigenic properties were evaluated in vitro by soft agar assay. The angiogenic response was evaluated in vivo by the chorioallantoic membrane angiogenic assay. Microarray analysis was performed on 6 tubular adult renal progenitor cell and 6 renal cell carcinoma derived cell clones. Membrane protein expression was evaluated by flow cytometry and immunofluorescence staining. Results: CD133þ/CD24þ cells were isolated from normal and tumor kidney tissue. Fluorescence activated cell sorting revealed that renal cell carcinoma derived cells did not express mesenchymal stem cell markers. CD133þ/CD24þ tumor cells were more undifferentiated than tubular adult renal progenitor cells. Renal cell carcinoma derived cells were clonigenic and could differentiate into adipocytes, epithelial and osteogenic cells. They could also regenerate tumor cells in vitro and induce angiogenesis in vivo. Gene expression profile identified CTR2 as a membrane marker for this neoplastic population. CTR2 was involved in renal cell carcinoma derived cell cisplatin resistance. Conclusions: Our results indicate the presence of a CD133þ/CD24þ/CTR2þ cancer cell population in clear cell renal cell carcinoma. These cells have some stem cell-like features, including in vitro self-maintenance and differentiating capabilities, and they can induce an angiogenic response in vivo. Key Words: kidney; carcinoma, renal cell; antigens, CD24; CTR2 protein, human; AC133 antigen
RENAL cell carcinoma accounts for approximately 3% of all adult malignancies. It was estimated that in 2013 about 65,000 new cases would be diagnosed and almost 14,000 patients
would die of RCC in the United States.1 Emerging evidence shows that the capacity of a tumor to grow and propagate resides in a small population of low proliferating tumor
0022-5347/14/1926-1831/0 THE JOURNAL OF UROLOGY® © 2014 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2014.06.070 Vol. 192, 1831-1841, December 2014 Printed in U.S.A.
Abbreviations and Acronyms 3-D ¼ 3-dimensional CAM ¼ chorioallantoic membrane ccRCC ¼ clear cell RCC CK-19 ¼ cytokeratin-19 CSC ¼ cancer SC FC ¼ fold change FDR ¼ false discovery rate FGF-2 ¼ fibroblast growth factor-2 HCG ¼ human chorionic gonadotropin MSC ¼ mesenchymal SC RCC ¼ renal cell carcinoma RDC ¼ CD133þ/CD24þ RCC derived cell RPTEC ¼ renal proximal tubular epithelial cell SC ¼ stem cell tARPC ¼ tubular adult renal progenitor cell ZO-1 ¼ zonula occludens 1 Accepted for publication June 17, 2014. Study received institutional review board approval. Supported by Italian Ministry of University and Research Grant MiUR-FIRB: RBAP11B2SX. * Equal study contribution. † Correspondence: Andrology and Kidney Transplantation Unit, Department of Emergency and Organ Transplantation-Urology, Piazza G. Cesare 11, 70124 Bari, Italy (telephone: þ39.080.5478880; FAX: þ39.080.5478880; e-mail: [email protected]).
www.jurology.com
j
1831
1832
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
cells termed CSCs or tumor initiating cells.2,3 CSCs have self-renewal ability, form a colony in semisolid medium and are tumorigenic at low numbers. Because many properties of CSCs resemble those of normal SCs, similar methodologies may be used to identify normal SCs and their neoplastic counterparts. Specific sets of surface markers are an essential part of most SC enrichment procedures. They express distinct molecular markers that can be used for SC enrichment. Of the latter markers the surface molecule CD133 (prominin) was identified as a SC marker and used to enrich brain tumor initiating cells.4 Recently studies in normal adult human kidneys revealed a population of renal progenitor cells characterized by CD133 and CD24 co-expression.5,6 These CD133þ/CD24þ cells, termed tARPCs, have self-renewal potential as well as the capacity to differentiate into tubular cells. We identified and characterized a population of CD133þ/CD24þ cancer cells in ccRCC tissue samples (RDCs) and studied differences compared to their normal counterpart (tARPCs). In addition, we studied the gene expression profile of these RDCs to identify a marker that could discriminate this neoplastic population from its normal tubular counterpart.
MATERIALS AND METHODS We collected healthy and tumor specimens from 40 patients who underwent radical or partial nephrectomy for ccRCC (see table). Cells were isolated from healthy and neoplastic renal tissue as previously described.7 Neoplastic and normal tubular fraction cells were incubated with CD133Ab conjugated magnetic microbeads. CD133þ cells were then selected using CD24Ab conjugated magnetic nanoparticles (fig. 1). Fluorescence activated cell sorting was done to detect CD133þ and CD24þ in all isolated cell lines (RDCs and tARPCs) and characterize phenotypes. One million cells were labeled with CD133, CD34, CD45, CD14, CD105 and CD24. Mesenchymal phenotyping was performed using Clinical and pathological characteristics of 40 patients No. male No. female Mean SD age No. smokers Mean SD body mass index (kg/m2) No. arterial hypertension No. Ca family history No. pathological stage: pT1 pT2 pT3 pNþ pMþ Fuhrman grade: G1-G2 G3-G4
29 11 64 14 23 25.5 3.91 24 22 23 5 12 0 0 28 12
the MSC Phenotyping Kit (Miltenyi Biotec, San Diego, California). At least 4 independent experiments were performed per cell line. To evaluate protein expression of several SC markers we used the Human Pluripotent Stem Cell Array (R&D SystemsÒ). Certain proteins were identified, including Oct-3/4, NANOG, SOX2, E-cadherin, a-fetoprotein, GATA-4, HNF-3b/FoxA2, PDX-1/IPF1, SOX17, Otx2, TP63/TP73L, Goosecoid, Snail, VEGF R2/KDR/Flk-1 and HCG. Cell cloning was performed using the limiting dilution technique. Epithelial differentiation experiments were done as previously reported.7 The other differentiation experiments were performed using adipogenic induction/maintenance medium and osteogenic differentiation medium according to manufacturer instructions. RDC anchorage independent growth and tumorigenic potential were monitored by soft agar colony formation assay. We used the in vivo CAM technique according to Ribatti et al for angiogenic assays.8 For microarray analysis total RNA was extracted from RDCs, tARPCs, MSCs and RPTECs. Analysis was performed using the HumanHT-12 v4 Expression BeadChip Kit (IlluminaÒ). Quantitative realtime polymerase chain reaction was performed using iQÒ SYBRÒ Green Supermix buffer to validate microarray data. Supplementary table 1 (http://jurology.com/) lists the gene specific primers used. The expression of CD133, CTR2, ZO-1 and CK-19 was evaluated by indirect immunofluorescence and confocal microscopic analysis on tARPCs and RDCs. CD133 and CTR2 protein expression in vivo was evaluated by indirect immunofluorescence and confocal microscopy on 5 mm frozen sections of normal and tumor human renal tissues. Finally, we assayed RDC viability after exposure to cisplatin (2.5 mmol/l) and anti-CTR2 blocking antibody using trypan blue dye exclusion and MTT assay.
RESULTS RDC Isolation and Characterization We isolated tubular and cancer cells from healthy and tumor tissue kidney specimens obtained at nephrectomy from 40 patients with ccRCC. Cytofluorimetry and immunofluorescence showed that the cell lines were CD133þ (fig. 2, a and b). To exclude possible contamination of hematopoietic and endothelial progenitor cells we performed flow cytometry to evaluate the presence of the CD24 kidney specific membrane marker in CD133þ cells isolated from cancer and tubular fractions (fig. 2, c). The incidence of CD133þ/CD24þ RDCs was 85% (fig.2, d ). In addition, RDCs expressed the mesenchymal SC marker CD73 but not the markers CD105 and CD90 (fig. 2, e to g). RDCs and tARPCs did not express CD20, CD45, CD14 or CD34 (fig. 2, e to g). These results suggest that RDCs and tARPCs do not have a mesenchymal origin. We also evaluated stemness potential in RDCs and tARPCs, and compared them to CD133e tumor
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
1833
Figure 1. CD133þ/CD24þ cell isolation from healthy and neoplastic renal tissue. Neoplastic and normal tubular fraction cells were selected for CD133 and CD24 markers
fraction cells and RPTECs. Protein array demonstrated that in RDCs and CD133e tumor fraction cells certain SC markers were statistically significantly over expressed, including a-fetoprotein, GATA4, HCG, HNF-3b/FoxA2, NANOG, Oct3/4, PDX-1/IPF1, Snail, SOX2, SOX17 and VEGFR2/ KDR/Flk-1. We also found higher expression of the other proteins, including Goosecoid, OTX2, E-cadherin and TP63/TP73L, but the difference was not statistically significant (fig. 3, a). Comparing RDCs to RPTECs showed statistically significant higher expression in RDCs of certain stemness related proteins, including HCG, HNF-3b/FoxA2, NANOG, PDX-1/IPF1, Snail and SOX2 (fig. 3, a). Moreover, when comparing RDCs and tARPCs, we
found significantly higher expression of a-fetoprotein, HCG, HNF-3b/FoxA2, GATA4, SOX2, Goosecoid, OTX2, SOX17, TP63/TP73L and VEGFR2/KDR/ Flk-1 in RDCs while E-cadherin, Oct3/4, PDX-1/ IPF1 and Snail expression was not significantly higher (fig. 3, b). Cell Cloning and Differentiation Homogenous clonal CD133þ/CD24þ populations were obtained using the limiting dilution technique (supplementary fig. 1, http://jurology.com/). Three individual clones of tumor origin were evaluated for the ability to differentiate into renal and nonrenal cell types. They showed the same differentiation potential. When exposed to epithelial cell
Figure 2. Characterization of CD133þ/CD24þ cells from human ccRCC and representative cytofluorimetry of mesenchymal markers after immunomagnetic sorting for CD133 and CD24. a and b, flow cytometry and immunofluorescence revealed that cells extracted from RCC expressed CD133 marker. c, flow cytometry showed CD24 kidney specific membrane marker in RDCs. d, about 85% of all cells expressed CD133 and CD24 markers. e, MSC positive control. f and g, tARPCs and RDCs expressed CD73 but not CD90 or CD105 and were negative for lymphocyte markers CD45, CD14, CD20 and CD34.
1834
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
Figure 3. Protein array histograms show mean SEM RDC stemness potential. a, high expression of stemness markers a-fetoprotein, GATA4, HCG, HNF-3b/FoxA2, NANOG, Oct3/4, PDX-1/IPF1, Snail, SOX2, SOX17 and VEGFR2/KDR/Flk-1 in 3 RDC lines vs 3 CD133e samples (CD133-TUM) from tumor portion. RDC and RPTEC comparison demonstrated high expression of HCG, HNF-3b/FoxA2, NANOG, PDX-1/IPF1, Snail and SOX2. b, protein levels of stemness marker in 3 RDCs compared to 3 tARPCs. Asterisk indicates p
Purpose: In clear cell renal cell carcinoma tissue samples we identified and characterized a population of renal cell carcinoma derived CD133þ/CD24þ cancer cells. We studied differences between these cells and their nonneoplastic counterpart, tubular adult renal progenitor cells. Materials and Methods: CD133þ/CD24þ renal cell carcinoma derived cells were isolated from 40 patients. The mesenchymal phenotype and stemness proteomic profile of these renal cell carcinoma derived cells were characterized. Colony forming efficiency and self-renewal ability were tested by limiting dilution. Tumorigenic properties were evaluated in vitro by soft agar assay. The angiogenic response was evaluated in vivo by the chorioallantoic membrane angiogenic assay. Microarray analysis was performed on 6 tubular adult renal progenitor cell and 6 renal cell carcinoma derived cell clones. Membrane protein expression was evaluated by flow cytometry and immunofluorescence staining. Results: CD133þ/CD24þ cells were isolated from normal and tumor kidney tissue. Fluorescence activated cell sorting revealed that renal cell carcinoma derived cells did not express mesenchymal stem cell markers. CD133þ/CD24þ tumor cells were more undifferentiated than tubular adult renal progenitor cells. Renal cell carcinoma derived cells were clonigenic and could differentiate into adipocytes, epithelial and osteogenic cells. They could also regenerate tumor cells in vitro and induce angiogenesis in vivo. Gene expression profile identified CTR2 as a membrane marker for this neoplastic population. CTR2 was involved in renal cell carcinoma derived cell cisplatin resistance. Conclusions: Our results indicate the presence of a CD133þ/CD24þ/CTR2þ cancer cell population in clear cell renal cell carcinoma. These cells have some stem cell-like features, including in vitro self-maintenance and differentiating capabilities, and they can induce an angiogenic response in vivo. Key Words: kidney; carcinoma, renal cell; antigens, CD24; CTR2 protein, human; AC133 antigen
RENAL cell carcinoma accounts for approximately 3% of all adult malignancies. It was estimated that in 2013 about 65,000 new cases would be diagnosed and almost 14,000 patients
would die of RCC in the United States.1 Emerging evidence shows that the capacity of a tumor to grow and propagate resides in a small population of low proliferating tumor
0022-5347/14/1926-1831/0 THE JOURNAL OF UROLOGY® © 2014 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2014.06.070 Vol. 192, 1831-1841, December 2014 Printed in U.S.A.
Abbreviations and Acronyms 3-D ¼ 3-dimensional CAM ¼ chorioallantoic membrane ccRCC ¼ clear cell RCC CK-19 ¼ cytokeratin-19 CSC ¼ cancer SC FC ¼ fold change FDR ¼ false discovery rate FGF-2 ¼ fibroblast growth factor-2 HCG ¼ human chorionic gonadotropin MSC ¼ mesenchymal SC RCC ¼ renal cell carcinoma RDC ¼ CD133þ/CD24þ RCC derived cell RPTEC ¼ renal proximal tubular epithelial cell SC ¼ stem cell tARPC ¼ tubular adult renal progenitor cell ZO-1 ¼ zonula occludens 1 Accepted for publication June 17, 2014. Study received institutional review board approval. Supported by Italian Ministry of University and Research Grant MiUR-FIRB: RBAP11B2SX. * Equal study contribution. † Correspondence: Andrology and Kidney Transplantation Unit, Department of Emergency and Organ Transplantation-Urology, Piazza G. Cesare 11, 70124 Bari, Italy (telephone: þ39.080.5478880; FAX: þ39.080.5478880; e-mail: [email protected]).
www.jurology.com
j
1831
1832
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
cells termed CSCs or tumor initiating cells.2,3 CSCs have self-renewal ability, form a colony in semisolid medium and are tumorigenic at low numbers. Because many properties of CSCs resemble those of normal SCs, similar methodologies may be used to identify normal SCs and their neoplastic counterparts. Specific sets of surface markers are an essential part of most SC enrichment procedures. They express distinct molecular markers that can be used for SC enrichment. Of the latter markers the surface molecule CD133 (prominin) was identified as a SC marker and used to enrich brain tumor initiating cells.4 Recently studies in normal adult human kidneys revealed a population of renal progenitor cells characterized by CD133 and CD24 co-expression.5,6 These CD133þ/CD24þ cells, termed tARPCs, have self-renewal potential as well as the capacity to differentiate into tubular cells. We identified and characterized a population of CD133þ/CD24þ cancer cells in ccRCC tissue samples (RDCs) and studied differences compared to their normal counterpart (tARPCs). In addition, we studied the gene expression profile of these RDCs to identify a marker that could discriminate this neoplastic population from its normal tubular counterpart.
MATERIALS AND METHODS We collected healthy and tumor specimens from 40 patients who underwent radical or partial nephrectomy for ccRCC (see table). Cells were isolated from healthy and neoplastic renal tissue as previously described.7 Neoplastic and normal tubular fraction cells were incubated with CD133Ab conjugated magnetic microbeads. CD133þ cells were then selected using CD24Ab conjugated magnetic nanoparticles (fig. 1). Fluorescence activated cell sorting was done to detect CD133þ and CD24þ in all isolated cell lines (RDCs and tARPCs) and characterize phenotypes. One million cells were labeled with CD133, CD34, CD45, CD14, CD105 and CD24. Mesenchymal phenotyping was performed using Clinical and pathological characteristics of 40 patients No. male No. female Mean SD age No. smokers Mean SD body mass index (kg/m2) No. arterial hypertension No. Ca family history No. pathological stage: pT1 pT2 pT3 pNþ pMþ Fuhrman grade: G1-G2 G3-G4
29 11 64 14 23 25.5 3.91 24 22 23 5 12 0 0 28 12
the MSC Phenotyping Kit (Miltenyi Biotec, San Diego, California). At least 4 independent experiments were performed per cell line. To evaluate protein expression of several SC markers we used the Human Pluripotent Stem Cell Array (R&D SystemsÒ). Certain proteins were identified, including Oct-3/4, NANOG, SOX2, E-cadherin, a-fetoprotein, GATA-4, HNF-3b/FoxA2, PDX-1/IPF1, SOX17, Otx2, TP63/TP73L, Goosecoid, Snail, VEGF R2/KDR/Flk-1 and HCG. Cell cloning was performed using the limiting dilution technique. Epithelial differentiation experiments were done as previously reported.7 The other differentiation experiments were performed using adipogenic induction/maintenance medium and osteogenic differentiation medium according to manufacturer instructions. RDC anchorage independent growth and tumorigenic potential were monitored by soft agar colony formation assay. We used the in vivo CAM technique according to Ribatti et al for angiogenic assays.8 For microarray analysis total RNA was extracted from RDCs, tARPCs, MSCs and RPTECs. Analysis was performed using the HumanHT-12 v4 Expression BeadChip Kit (IlluminaÒ). Quantitative realtime polymerase chain reaction was performed using iQÒ SYBRÒ Green Supermix buffer to validate microarray data. Supplementary table 1 (http://jurology.com/) lists the gene specific primers used. The expression of CD133, CTR2, ZO-1 and CK-19 was evaluated by indirect immunofluorescence and confocal microscopic analysis on tARPCs and RDCs. CD133 and CTR2 protein expression in vivo was evaluated by indirect immunofluorescence and confocal microscopy on 5 mm frozen sections of normal and tumor human renal tissues. Finally, we assayed RDC viability after exposure to cisplatin (2.5 mmol/l) and anti-CTR2 blocking antibody using trypan blue dye exclusion and MTT assay.
RESULTS RDC Isolation and Characterization We isolated tubular and cancer cells from healthy and tumor tissue kidney specimens obtained at nephrectomy from 40 patients with ccRCC. Cytofluorimetry and immunofluorescence showed that the cell lines were CD133þ (fig. 2, a and b). To exclude possible contamination of hematopoietic and endothelial progenitor cells we performed flow cytometry to evaluate the presence of the CD24 kidney specific membrane marker in CD133þ cells isolated from cancer and tubular fractions (fig. 2, c). The incidence of CD133þ/CD24þ RDCs was 85% (fig.2, d ). In addition, RDCs expressed the mesenchymal SC marker CD73 but not the markers CD105 and CD90 (fig. 2, e to g). RDCs and tARPCs did not express CD20, CD45, CD14 or CD34 (fig. 2, e to g). These results suggest that RDCs and tARPCs do not have a mesenchymal origin. We also evaluated stemness potential in RDCs and tARPCs, and compared them to CD133e tumor
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
1833
Figure 1. CD133þ/CD24þ cell isolation from healthy and neoplastic renal tissue. Neoplastic and normal tubular fraction cells were selected for CD133 and CD24 markers
fraction cells and RPTECs. Protein array demonstrated that in RDCs and CD133e tumor fraction cells certain SC markers were statistically significantly over expressed, including a-fetoprotein, GATA4, HCG, HNF-3b/FoxA2, NANOG, Oct3/4, PDX-1/IPF1, Snail, SOX2, SOX17 and VEGFR2/ KDR/Flk-1. We also found higher expression of the other proteins, including Goosecoid, OTX2, E-cadherin and TP63/TP73L, but the difference was not statistically significant (fig. 3, a). Comparing RDCs to RPTECs showed statistically significant higher expression in RDCs of certain stemness related proteins, including HCG, HNF-3b/FoxA2, NANOG, PDX-1/IPF1, Snail and SOX2 (fig. 3, a). Moreover, when comparing RDCs and tARPCs, we
found significantly higher expression of a-fetoprotein, HCG, HNF-3b/FoxA2, GATA4, SOX2, Goosecoid, OTX2, SOX17, TP63/TP73L and VEGFR2/KDR/ Flk-1 in RDCs while E-cadherin, Oct3/4, PDX-1/ IPF1 and Snail expression was not significantly higher (fig. 3, b). Cell Cloning and Differentiation Homogenous clonal CD133þ/CD24þ populations were obtained using the limiting dilution technique (supplementary fig. 1, http://jurology.com/). Three individual clones of tumor origin were evaluated for the ability to differentiate into renal and nonrenal cell types. They showed the same differentiation potential. When exposed to epithelial cell
Figure 2. Characterization of CD133þ/CD24þ cells from human ccRCC and representative cytofluorimetry of mesenchymal markers after immunomagnetic sorting for CD133 and CD24. a and b, flow cytometry and immunofluorescence revealed that cells extracted from RCC expressed CD133 marker. c, flow cytometry showed CD24 kidney specific membrane marker in RDCs. d, about 85% of all cells expressed CD133 and CD24 markers. e, MSC positive control. f and g, tARPCs and RDCs expressed CD73 but not CD90 or CD105 and were negative for lymphocyte markers CD45, CD14, CD20 and CD34.
1834
CTR2 IDENTIFIES CANCER CELLS WITH STEM CELL-LIKE FEATURES
Figure 3. Protein array histograms show mean SEM RDC stemness potential. a, high expression of stemness markers a-fetoprotein, GATA4, HCG, HNF-3b/FoxA2, NANOG, Oct3/4, PDX-1/IPF1, Snail, SOX2, SOX17 and VEGFR2/KDR/Flk-1 in 3 RDC lines vs 3 CD133e samples (CD133-TUM) from tumor portion. RDC and RPTEC comparison demonstrated high expression of HCG, HNF-3b/FoxA2, NANOG, PDX-1/IPF1, Snail and SOX2. b, protein levels of stemness marker in 3 RDCs compared to 3 tARPCs. Asterisk indicates p
