Overexpression of Functional SLC6A3 in Clear Cell Renal Cell Carcinoma

Transporter proteins are an integral part of cellular homeostasis by controlling the import and export of essential substances, like amino acid, ions, vitamins, and drugs, across cellular membranes. The regulation of these proteins is thus fundamental for cell survival and organ function. There are two major transporter families in the human genome: the ATP-binding cassettes (ABCs) and the solute carriers (SLCs). The SLC family comprises more than 400 members organized into 52 subfamilies, predominantly involved in molecule uptake (1). The SLC families vary in substrate specificity; some have a very narrow range of substrates, whereas others, like the SLCO family (OATP), handle a broad spectrum of substrates (2). The ABC transporters consist of 48 members divided into seven subfamilies and use ATP to drive the efflux of molecules through cellular membranes (3). Expression of these transporters is associated with stem and progenitor cells (4) and has been linked to drug resistance in human disease, including chemotherapy resistance in cancer (5). There are connections between several transporters and human malignancies. For example, polymorphisms of the bicarbonate transporter SLC4A7 have been associated with breast cancer (6), and the urea transporter SLC14A1 has been linked to bladder cancer susceptibility (7).

The epithelial cells in the renal tubular system are endowed with a large variety of membrane transporter proteins that control the kidneys' reabsorptive processes (8). Renal cell carcinomas (RCCs) account for around 2.5% of malignancies worldwide (9). Clear cell RCC (ccRCC) is the most common form of renal cancer constituting around 75% of the cases (10). ccRCC is characterized by loss of function of the von Hippel Lindau (VHL) gene (11). VHL normally acts as a regulator of the hypoxia-inducible factors (HIF) by targeting these proteins for destruction in the presence of oxygen (12). Under hypoxic conditions, stabilized HIFs control the transcription of genes in the cellular adaptive response to hypoxia, such as VEGF, GLUT-1, CAIX, and TGFβ. Collectively, these changes lead to induction of angiogenesis and a metabolic switch to anaerobic glycolysis (11, 12). Generalized, ccRCC does not respond to conventional therapies like chemotherapy and radiation (13). Even though improvement has been accomplished by the introduction of antiangiogenesis therapies and mTOR inhibitors, all patients with metastatic ccRCC eventually succumb to the disease (14).

Herein, we performed an explorative analysis of transporter gene expression, using the TCGA database. This analysis highlighted the dopamine transporter SLC6A3 that displays a unique transcriptional induction in ccRCC. We provide experimental evidence that the overexpression is associated with a functional uptake of dopamine and that the gene is regulated by HIF-2α in normal renal cells. Collectively, these results provide ample evidences that SLC6A3 is a novel and highly specific biomarker for ccRCC.

Renal tissue samples were obtained from nephrectomies performed owing to malignancy. Ethical permission was granted by the ethical committee at Lund University (Lund, Sweden; LU680-08 and LU289-07) and informed consent was obtained from all patients. For normal samples, the cortical tissue farthest from the tumor was selected, and all samples were dissociated as described in Supplementary Methods. All primary tumor samples were pathologically confirmed as ccRCC by an expert kidney pathologist (M. Johansson). Pathologic assessment of the primary tumor samples can be viewed in Supplementary Table S1.

Renal cancer cell lines SKRC7, SKRC10, SKRC17, SKRC21 (kindly provided by E. Oosterwijk, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands), SNU-349, RCB1963 (Korean Cell Line Bank), breast cancer cell line MCF7 (ATCC), and normal prostate cell line PNT-2 (ATCC) were cultured in RPMI-1640 (IMR-32) medium supplemented with 10% FBS. Primary renal samples and RCC cell lines 786-O (ATCC), Wt7 (kindly provided by W. G. Kaelin, Department of Medicine at the Dana-Farber Cancer Institute, Harvard Medical School, USA), KMRC3 (Japanese Collection of Research Bioresources Cell Bank), and RCC-4 (empty vector or +VHL; Sigma-Aldrich) cells were cultured in DMEM supplemented with 10% FBS. Geniticin (0.5 mg/mL; Gibco) was added to RCC-4 growth medium for selection. Normal breast cell line MCF10A was cultured in DMEM/F12 (Gibco, Life Technologies) supplemented with 5% horse serum, 100 ng/mL cholera toxin, 20 ng/mL EGF, 0.01 mg/mL insulin, and 500 ng/mL hydrocortisone, all from Sigma Aldrich. Normal endothelial cell line human umbilical vein endothelial cell (HUVEC) was cultured on gelatin-coated plates in endothelial cell growth medium 2 (Promo Cell). All medium was supplemented with antibiotics (PEST 1%). Cell line authentication of SNU-349, SKRC10, KMRC3, 786-O, and PNT2 was performed by DDC Medical. The remaining cell lines were thawed from the authenticated cell stock and used within eight passages and authenticated by examination of morphology and growth characteristics. All cell lines were confirmed to be mycoplasma-free.

Hypoxia was generated in a Whitley H35 Hypoxystation (Don Whitley Scientific). Cells were placed in the hypoxia chamber at 1% pO₂ and media were immediately changed to hypoxic preconditioned media. As control, cells were incubated in 21% pO₂.

Total RNA was extracted and cDNA synthesis was performed as described in Supplementary Methods. Amplifications were run using a GeneAmp 7300 sequence detector (Applied Biosystems). Real-time detection of the PCR product was performed using the SYBR Green Master Mix (Applied Biosystems). All reactions were performed in triplicate and the comparative C_t method was used to quantify relative mRNA. The geometric mean of three endogenous reference genes (UBC, YWHAZ, and SDHA) was used for normalization. Primer sequences are provided in Supplementary Methods.

Cells were transfected with siRNA targeting HIF-1α (silencer select s6541 #3490821), HIF-2α (EPAS1 #4390824), SLC6A3_1 (id 12976, silencer select, Ambion), (id 12977, silencer select, Ambion), and siNeg ctrl #2 (silencer select #4390847, Ambion) and siNeg ctrl #1 (silencer select #4390843, Ambion) at 2.5, 5, or 10 nmol/L concentration using Lipofectamine 2000 (Invitrogen; 3 μL/well for 12-well plate) according to manufacturer's instruction. Transfections were carried out in serum- and penicillin-free OPTI-MEM medium (Gibco). Effect of knockdown was measured 48 hours after transfection.

Cells were lysed in RIPA supplemented with Complete Protease Inhibitor (Roche Diagnostics). Proteins were separated by SDS-PAGE and transferred to 0.2-μm polyvinylidene difluoride (PVDF) membranes (BioRad) using transblot turbo transfer system. Membranes were blocked in 5% milk in PBS-T and incubated with primary antibody anti–DAT1 (Abcam, ab128848), anti-HIF-1α (NB100-479), or anti-HIF-2α (Abcam ab199) over night and β-actin antibody (MP Medicals, #691001) or SDHA antibody (Abcam, ab14715) as control.

For uptake experiments, cells were seeded at a density corresponding to confluence at 48 hours. At confluence, assay buffer containing 7 nmol/L dihydroxyphenylethylamine 3,4-,[ring-2,5,6,³H]-(Dopamine) ([³H]-dopamine; 56.8 Ci/mmol; Perkin Elmer) and 20 μg/ml l-ascorbic acid (Sigma Aldrich) was added for 5 minutes, as detailed in Supplementary Methods. Competitive inhibition of [³H]-dopamine uptake was performed by incubation with either 2 or 20 μmol/L unlabeled dopamine hydrochloride (Sigma Aldrich). [³H]-Dopamine uptake experiments were also performed in presence of the specific SLC6A3 inhibitor 1-(2-[bis(4-fluorophenyl)methoxy]ethyl)-4-(3-phenylpropyl)piperazine dihydrochloride (GBR12909; Sigma Aldrich). Here, cells were pretreated for 60 minutes with varying concentrations (10 nmol/L to 10 μmol/L) of GBR12909 in the growth media, and thereafter uptake was performed in presence of GBR12909 at the same concentrations. The amount of radioactivity was measured with a Tri-carb 2810TR liquid scintillation analyzer (Perkin Elmer). Uptake was assessed by radioactivity in DPM related to pmol of dopamine per assay well or pmol/μg protein from cell lysate from one assay well. Protein concentration of lysates was determined using BCA Protein Assay Kit (Pierce Biotechnology).

All values are reported as mean ± SEM from at least three independent experiments unless otherwise stated. The two-sided Student unpaired t test was used for statistical analyses, unless otherwise stated. Three levels of significance were used: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Level 3 RNA-sequencing data processed using the “UNC V2 RNA-Seq” workflow were downloaded from The Cancer Genome Atlas (TCGA) data portal (https://portal.gdc.cancer.gov/). Gene-level RSEM scaled estimates were processed by log2(RSEM × 1M + 1) to create gene expression levels for 20,531 Entrez gene IDs as annotated in the TCGA hg19 June2011 Generic Annotation Files (GAF). Data were obtained for 25 different types of solid tissue tumors. For 16 of these tumor types, three or more corresponding normal tissue samples were available and included in the analyses. A list of all 7,155 samples used in this study is given in Supplementary Table S2.

For validation purposes, we downloaded the Gene Expression Omnibus (GEO) entry GSE3526 comprising gene expression patterns of 353 normal tissue samples from 20 anatomically distinct sites of the human central nervous system (CNS), as well as 45 non-CNS tissues, measured on the Affymetrix HGU133 plus 2.0 platform (15). We also acquired mRNA expression data from 1,036 different cell lines as analyzed on Affymetrix HGU133 plus2.0 chips as provided by the Broad Cancer Cell Line Encyclopedia (CCLE; ref. 16).

SLC genes were selected from the BioParadigms Genomic Transporter Database (www.bioparadigms.com). In total, 391 unique genes from 52 different SLC transporter families were present in the TCGA data. We further included 51 genes encoding ABC transporters. The complete gene list of 442 analyzed transporter genes is provided in Supplementary Table S3.

Statistical analyses were performed using the R software (http://www.r-project.org). Hierarchical Cluster Analyses (HCA), principal component analyses (PCA), and Kaplan–Meier plots were performed using the stats package in R. For HCA, 1-Pearson correlation was used as distance measure and Ward method for agglomeration. For PCA and HCA (Fig. 1A–C), the mean expression value for each transporter gene within the respective TCGA tumor or normal tissue cohorts was used. Differential gene expression between clear cell kidney carcinoma (KIRC) tumor and normal tissues was determined using the Limma Bioconductor package (17). For the analysis of taxonomy groups, samples from all 3 TCGA kidney cancer projects [chromophobe RCC (KICH), KIRC, papillary kidney carcinoma (KIRP)] were merged into one single dataset. Molecular RCC taxonomy groups as well as clinical and histopathologic parameters were used as presented in the article by Chen and colleagues (18).

We set out to explore the expression patterns of cellular transporters of the SLC and ABC families in ccRCC and normal kidney in relation to other normal and tumor tissue types. For this purpose, we performed PCA analyses on mRNA sequencing data from TCGA. As illustrated in Fig. 1A, the transcriptional pattern of the 442 SLC and ABC transporter genes stood out in samples collected from normal kidney and liver tissues relative to samples obtained from the remaining 14 tissues analyzed. Likewise, when analyzing data from 25 different tumor types, two kidney carcinomas, KIRC and papillary RCC (pRCC; KIRP), as well as liver hepatocellular carcinomas, clearly differed from remaining tumors; possibly reflecting their cellular origin from cells endowed with considerable transport capacities (Fig. 1B). Also, neuronally derived tumors, such as glioblastoma multiforme, lower-grade glioma, and pheochromocytoma and paraganglioma, displayed large differences in expression patterns compared with other tumors. We further noted that KICH did not deviate as much from remaining tumor types, most likely explained by the notion that this tumor type is derived from a more distal and less transporter-rich region of the nephron (19, 20).

We next performed unsupervised HCA of transporter gene expression in the TCGA data (Fig. 1C and Supplementary Fig. S1). We observed that tumors and their corresponding normal tissues often clustered together, possibly reflecting that tumors often maintain characteristic transporter expression found in their tissue of origin. The HCA further showed a high degree of tissue specificity among groups of transporter genes. We identified five separate gene clusters of transporters with apparent tissue- and/or tumor-specific expression (colored bars, Fig. 1C and Supplementary Fig. S1).

We noted that transcript levels of several transporters were reduced or abolished in some tumor samples, as compared with corresponding normal tissues. This may reflect a dedifferentiation process that takes place during oncogenic transformation or may be caused by differences in cellular composition between tumor and normal tissue such as infiltration of tumor-associated lymphocytes and fibroblasts. Alternatively, the absence of gene expression in a tumor could be explained by cellular heterogeneity within the corresponding tissue of origin, that is, the transporter may be expressed in cell types different from which the cancer stems from. This was particularly obvious in case of the kidney-specific gene cluster. Here, expression of numerous transporters was absent or low in pRCC and ccRCC than in normal renal samples (light green, Fig. 1C). Both these tumor entities are believed to arise from proximal tubule cells. Normal kidney tissue, however, is composed of a large number of cell types, each with its own specialized setup of transporter proteins. Genes with high expression in the normal kidney samples, but not in pRCC or ccRCC samples, included RHCG, an important ammonia transporter specifically expressed in distal tubuli (21), the HCO3 transporter SLC4A1 expressed by intercalating cells of the distal nephron (22), and the Na–K–Cl co-transporter SLC12A1 (NKCC2) that is expressed in thick ascending limb of the loop of Henle (23). Hence, genes solely expressed in the distal regions of the nephron were not expressed in pRCC and ccRCC, possibly reflecting their cellular origin from cells of the proximal tubule.

We next sought to identify transporter genes that were differentially expressed between ccRCC and normal kidney tissue (Fig. 1D and Supplementary Table S4). Surprisingly, the most upregulated gene was identified as the dopamine reuptake transporter SLC6A3, which increased almost 25-fold in ccRCCs relative to normal kidney tissue. The human kidney is an important target organ for dopamine action and the entire nephron expresses various receptors for dopamine. Both circulating and renally produced dopamine affect the kidney in several ways, including vasodilation, control of diuresis, and sodium homeostasis. However, there are no reports indicating a functional role of SLC6A3 in the kidney. We therefore scrutinized SLC6A3 expression in the panel of normal and tumor tissues obtained from TCGA. The high expression observed in ccRCC was not reflected in any other tissue or tumor group; in fact, the vast majority of samples, including the two other kidney tumor types, display near-absent levels of SLC6A3 expression (Fig. 1E). Interestingly, lung adenocarcinoma and mesothelioma, within which a small fraction of samples displayed elevated SLC6A3 levels, are known to harbor focal genomic gains of the TERT gene located close to SLC6A3 at chromosome band 5p15.33 (24–26), raising the possibility that increased expression of SLC6A3 in these samples is a passenger event to increased TERT expression.

We next analyzed SLC6A3 expression in relation to the molecular RCC taxonomy groups described by Chen and colleagues (18). Increased expression was almost exclusively found in samples of the three ccRCC taxonomy subtypes (CC-e1, -e2, and -e3), thus substantiating the possible connection between pseudohypoxia and SLC6A3 overexpression (Supplementary Fig. S2A). The highest expression was noted in the CC-e2 subtype, which is the subtype enriched for ccRCCs of lower stage. SLC6A3 expression was found slightly higher in VHL- and PBRM1-mutated ccRCC samples, which is in line with the notion that the CC-e.2 subtype is enriched for these two mutations (Supplementary Fig. S2B and S2C). Decreased expression of SLC6A3 was noted in stage IV tumors (Supplementary Fig. S2F) and was also significantly associated with adverse outcome (Supplementary Fig. S2G). Importantly, among the majority of patients who succumbed to the disease, SLC6A3 expression in the primary tumor was readily detectable and higher than what is seen in other tumor types and normal tissues (Supplementary Fig. S2A and S2H).

The SLC6A3 dopamine transporter, also known as DAT1, is normally expressed in dopaminergic neurons of the CNS. We therefore downloaded microarray expression data of 20 different anatomically distinct sites of the human CNS and 45 non-CNS tissues (15). As expected, SLC6A3 was only found expressed at appreciable levels in substantia nigra and the ventral segmentum, where dopaminergic cells reside (Fig. 1F). Thus, we conclude that SLC6A3 expression in normal tissues is restricted to dopaminergic cell types of the brain and that ccRCC represents the only tumor entity, within the 25 different types of malignancies investigated here, that consistently display high levels of this transporter.

To confirm the data obtained from TCGA, we assessed mRNA expression of SLC6A3 using quantitative real-time PCR (qPCR) in ccRCC cell lines. As control, we used the breast cancer cell line MCF7. Quite surprisingly, and in stark contrast to tumor sample expression, conventional ccRCC cell lines, such as 786-O, RCC4, and the SKRCs, showed very low or nondetectable mRNA expression of SLC6A3 (Fig. 2A). We therefore examined SLC6A3 expression in the CCLE database comprising more than 1,000 established cancer cell lines (16). We observed that two of 36 kidney cancer cell lines, KMRC3 and SNU-349, expressed exceptionally high levels of SLC6A3 (Supplementary Fig. S3A). Importantly, both these cell lines are derived from pathologically confirmed ccRCC tumors and display DNA copy number profiles that are highly characteristic for primary ccRCC tumors, including chromosome 3p deletion and gain of 5q (Supplementary Fig. S3B). A subsequent qPCR confirmed overexpression of SLC6A3 in these two cell lines (Fig. 2A).

Western blot analyses confirmed that cell lines with high (KMRC3 and SNU-349) or moderate (SKRC10) SLC6A3 mRNA expression displayed appreciable levels of SLC6A3, whereas the commonly used 786-O cell line expressed very low levels of the protein. No protein expression could be detected in extracts from MCF-7 cells (Fig. 2B). KMRC3 and SNU-349 are not widely used and grow very slow than the more conventional ccRCC cell lines. For example, KMRC3 display a doubling time of approximately 7 days compared with 24 hours for 786-O (27). The growth pattern of SNU-349 and KMRC3 is more in line with that of primary ccRCC cultures.

Short-term cell cultures of primary ccRCC displayed high expression of SLC6A3 in contrast to cultures from matched normal renal cortical cells that showed very low levels (Fig. 2C). Primary tumor samples further displayed a higher protein level of SLC6A3 than cells from normal kidney (Fig. 2D). On the basis of these results, we conclude that SLC6A3 is expressed at very high levels in primary ccRCC and in cell lines with growth properties similar to that of primary ccRCC cell cultures. In contrast, the expression is lost in most established ccRCC cell lines.

ccRCC is characterized by a state of pseudohypoxia due to its functional loss of VHL. We therefore wanted to clarify whether SLC6A3 display properties of an HIF target gene. For this purpose, we first compared the expression of SLC6A3 with that of the well-established bona fide HIF target gene, CAIX. As shown in Fig. 3A, CAIX is induced in all solid tumors comprised in the TCGA data in comparison to their corresponding normal tissue. This most likely reflects the well-documented varying state of hypoxia that characterizes solid tumors. This stands in stark contrast to the ccRCC-specific induction of SLC6A3 (Fig. 1E). Thus, SLC6A3 is not a conventional HIF target gene.

We reasoned that the high levels of SLC6A3 in ccRCC could be related to cell-of-origin specific hypoxic induction. We therefore exposed primary normal human renal cells to hypoxia and noticed a consistent upregulation of the SLC6A3 transcript (Fig. 3B). In an attempt to compare this with other nonmalignant cells, we cultured endothelial cells (HUVECs), mammary epithelial cells (MCF10A), and prostate epithelial cells (PNT-2) under hypoxic conditions. Neither HUVEC nor MCF10A displayed hypoxic induction of SLC6A3, whereas PNT-2 displayed a modest increase (Fig. 3C). Primary ccRCC cells that display high levels of SLC6A3 (Fig. 2C) showed no further induction when exposed to hypoxia (Supplementary Fig. S4A). Likewise, commonly known hypoxia targets were not further induced in ccRCC tumor samples treated with hypoxia (Supplementary Fig. S4B). Interestingly, both RCC4 and 786-O cells reconstituted with VHL displayed significantly lower SLC6A3 levels than the corresponding VHL-negative cells (Supplementary Fig. S4C), indicating that long-term normalization of VHL status in ccRCC cells is associated with downregulation of SLC6A3. To further dissect the role of HIF-1α and HIF-2α, the two main effectors of hypoxic signaling, we performed knockdown experiments using siRNA in KMRC3, SNU-349, and SKRC10 (Fig. 3D). Western blot analyses showed that all three cell lines expressed both HIF-1α and HIF-2α (Supplementary Fig. S5A). Interestingly, SLC6A3 transcripts were decreased by siHIF-2α and dual siHIF-1α and siHIF-2α but not with siHIF-1α alone in all three cell lines (Fig. 3D). In KMRC3 and SNU-349, the HIF-1α knockdown displays a corresponding decrease of the HIF-1α–associated target gene CAIX, whereas GLUT1 displays a less HIF subtype–specific regulatory pattern in response to the respective siRNA (Supplementary Fig. S5B). Altogether, these results points toward a link between VHL loss and dysregulation of SLC6A3 in renal cells and ccRCC and that HIF-2α might be an important mediator of this response.

The normal function of SLC6A3 is to transport dopamine from the synaptic cleft back into the presynaptic neuron. To assess the functionality of SLC6A3 in ccRCC, the [³H]-dopamine uptake was measured in cultured ccRCC cells. The levels of [³H]-dopamine uptake mirrored SLC6A3 mRNA levels in the ccRCC cell lines, with SNU-349 and KMRC3 displaying the highest uptake (Fig. 4A). To determine whether the uptake of [³H]-dopamine was mediated by a dopamine transporter, we first performed the experiments in the presence of unlabeled dopamine. The only cell lines that showed a significant effect of the competition were SNU-349 and KMRC3 (Fig. 4B). In addition, we used varying concentrations of a highly specific inhibitor of SLC6A3, the cocaine analogue GBR12909. Again, the only cell lines responding to GBR12909 treatment with a decreased [³H]-dopamine uptake were SNU-349 and KMRC3 (Fig. 4C). These two cell lines displayed a significant decrease of [³H]-dopamine uptake also at levels down to 10 nmol/L of GBR12909 (Fig. 4D).

To further link SLC6A3 levels and uptake of dopamine, we performed knockdown experiments using siRNA in SNU-349 cells. The SLC6A3 knockdown was efficient using two different siRNAs and resulted in a decreased uptake of [³H]-dopamine (Fig. 5A and B).

Consistent with SLC6A3 mRNA expression, primary ccRCC tumor cells displayed strongly enhanced [³H]-dopamine uptake in comparison to normal renal cells from the same patient (Fig. 6A). Short-term cultures of primary ccRCCs also showed a decrease in [³H]-dopamine uptake in the presence of 2 and 20 μmol/L unlabeled dopamine (Fig. 6B), an effect not seen in normal renal cells (Fig. 6A). When exposing primary normal kidney cells to hypoxia, the cells displayed a trend toward increased [³H]-dopamine uptake (Fig. 6C), in line with the results of the hypoxia-induced SLC6A3 mRNA (Fig. 3B). Furthermore, when targeting the HIF subtypes in KMRC3 and SNU-349, only siHIF-2α led to a significant decrease in [³H]-dopamine uptake (Fig. 6D). These results further substantiate the notion that HIF-2α is the main driver of functional SLC6A3 expression in ccRCC.

This study is based on an explorative assessment of transporter gene expression in multiple tumor types and corresponding normal tissues comprised in the TCGA project. As expected, distinct tissue-specific expression patterns were observed in the normal tissues. Many tumors retained expression of transporters that are highly specific for their tissue of origin and may thus represent an important source for understanding tumor biology and development of functional biomarkers that can be applied in, for example, differential diagnosis. Importantly, we could identify tumor-specific expression of transporters that are suppressed in corresponding normal tissues which, based on their functional attributes, could be exploited for diagnostic or therapeutic purposes. One such example is SLC6A3, which is specifically overexpressed in ccRCC compared with normal kidney and other tumor types. We demonstrate that this overexpression is accompanied by active dopamine uptake in ccRCC cells, making SLC6A3 an excellent functional biomarker for ccRCC. In previous gene expression profiling studies of primary ccRCC, a high expression of SLC6A3 has been noted (28, 29). However, to the best of our knowledge, the present report represents the first functional analysis of SLC6A3 in ccRCC. One explanation for this might be that most established cell lines used in experimental studies of ccRCC, such as 786-O, are virtually devoid of SLC6A3 expression. Although established cell lines have provided an essential source for our current understanding of cancer cell behavior, there is a pronounced lack of fidelity to the original tumor in certain aspects of cancer cell biology. This is particularly true with respect to transporter proteins, many of which are lost or downregulated in conjunction with the in vitro establishment procedure (30). We could nevertheless identify two established ccRCC cell lines that retain high expression of the SLC6A3 transporter.

ccRCC is characterized by a functional loss of VHL, causing a cellular state of pseudohypoxia. We thus hypothesized that SLC6A3 could be induced by low oxygen tension. However, in our pan-TCGA comparison between the prototypical hypoxia target CAIX and SLC6A3, it was apparent that the links between SLC6A3 and hypoxia are more complex than the one between CAIX and hypoxia (Fig. 1E and 3A). Nevertheless, we show that SLC6A3 can be induced by hypoxia in normal cortical renal cells, albeit not to the same levels as noted in the primary tumors. Furthermore, the observation that VHL-reconstitution in the ccRCC cell line RCC4 led to a significant decrease in SLC6A3 expression lends support to a link between hypoxia and SLC6A3. Future studies of the chromatin landscape surrounding SLC6A3 will possibly provide more direct clues into why such a prominent and cell type–specific induction of this gene is seen.

It is not uncommon that tumor cells harness different transporter proteins for maintaining their growth and survival. For example, the HIF target gene SLC2A1 (GLUT1) is frequently upregulated in tumor cells to ensure sufficient energy supply. Similarly, the high-affinity glucose transporter GLUT3 (SLC2A3) is overexpressed in breast, colorectal, and bladder cancers while normally being expressed primarily in neuronal cells (31–33). LAT1 or SLC7A5 transports neutral amino acids into the cell, which can be used as building blocks and energy. SLC7A5 is upregulated in a variety of cancers (34), including ccRCC, where it has been found to activate mTORC1 via HIF-2a induction (35). In addition, tumors are often dependent on glutamine for the maintenance of cellular NADPH pools as well as for synthesizing the antioxidant glutathione (36). Consequently, many SLCs with glutamine-transporting functions are upregulated in cancer cells (37). In above examples, there is an obvious link between the specific function of the transporter and the metabolic need of the tumor cells. In an effort to assess whether SLC6A3 play any such role in ccRCC, we compared the effect on viability upon SLC6A3 inhibition in two cell lines expressing high levels of SLC6A3 (KMRC3 and SNU-349) versus 786-O which do not express appreciable levels of the transporter. We detected a similar effect of GBR12909 treatment in all cell lines regardless of the levels of SLC6A3 (Supplementary Fig. S6). Thus, any potential functional benefit of the specific upregulation of the SLC6A3 transporter in ccRCC is presently not clear. The observation that the SLC6A3 levels are generally lower in high-stage tumors indicates that SLC6A3 may be important for initial stages of tumorigenesis rather than having a role during later stages of tumor progression.

Irrespective of the reasons for the high expression levels of functional SLC6A3 in ccRCC, we postulate that this finding could be exploited for development of novel therapeutic or diagnostic modalities. For example, Ioflupane I123, a cocaine analogue that binds to SLC6A3 and hence accumulates on dopaminergic neurons, is widely used for diagnosing Parkinson disease. The results presented in this study, showing ccRCC-specific expression of functional SLC6A3, call for a clinical trial of this diagnostic tool for ccRCC. In addition, a plethora of molecules that bind to and/or inhibit SLC6A3 have been developed. Given the biologic properties of SLC6A3, there are also several possibilities of exploiting its transporter function for ccRCC therapy based on radionuclide labeling of such compounds. However, to fully appreciate the potential use of SLC6A3 expression in ccRCC, several key issues have to be clarified, including the underlying molecular mechanisms causing the overexpression, and the potential advantage of SLC6A3 overexpression for the tumor cells.

J. Hansson, L. Gustavsson, H. Axelson, D. Lindgren, and E. Johansson are listed as co-inventors on a patent on utilization of SLC6A3 overexpression in ccRCC for diagnosis and treatment of ccRCC that is owned by Akuru Pharma. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J. Hansson, D. Lindgren, L. Gustavsson, H. Axelson

Development of methodology: J. Hansson, D. Lindgren, M. Johansson, L. Gustavsson, H. Axelson

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Hansson, D. Lindgren, H. Nilsson, E. Johansson, H. Axelson

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Hansson, D. Lindgren, H. Axelson

Writing, review, and/or revision of the manuscript: J. Hansson, D. Lindgren, H. Nilsson, E. Johansson, M. Johansson, L. Gustavsson, H. Axelson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Johansson, H. Axelson

Study supervision: H. Axelson

Joakim Larsson, Red Glead Discovery, is gratefully acknowledged for providing experimental resources.

This work was supported by the Swedish Cancer Society, CREATE Health, BioCARE, a Strategic Research Program at Lund University, Gunnar Nilsson's Cancer Foundation, and Regional ALF funds.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.