βArrestin2 Mediates Renal Cell Carcinoma Tumor Growth

Renal Cell Carcinoma (RCC) is one of the most lethal urological cancers worldwide. The disease does not present early clinical symptoms and is commonly diagnosed at an advanced stage. Limited molecular drivers have been identified for RCC, resulting in the lack of effective treatment for patients with progressive disease. Ubiquitous βArrestin2 (βArr2) is well established for its function in the desensitization and trafficking of G protein-coupled receptors. More recently, βArr2 has been implicated in the regulation of fundamental cellular functions, including proliferation and invasion. We used bioinformatic and genetic approaches to determine role of βArr2 in RCC tumor growth. Analysis of published human datasets shows that ARRB2 (gene encoding βArr2) expression is increased in RCC tumor compared to normal tissue and that high levels of ARRB2 correlate with worse patient survival. Experimentally, we show that knockout of ARRB2 decreases rate of RCC cell proliferation and migration in vitro and xenograft tumor growth in animals. Mechanistically, βArr2 regulates c-Src activity, Cyclin A expression and cell cycle progression that are involved in tumor growth. These results show that βArr2 is a critical regulator of RCC tumor growth and suggest its utility as a potential marker and drug target to treat advanced disease.

downregulation of βArr2 promotes hepatocellular carcinoma tumor invasion 23 . These seemingly contradictory results imply that βArr2 function may be cell context-and cancer type-dependent.
While βArr1 and βArr2 exhibit high degree of sequence homology and function overlap, their subcellular distribution is distinct. βArr1 is expressed in the cytosol and nucleus and has been shown to exert its mitogenic function, at least in part, through the regulation of gene expression 10,11,16,24 . Less is known about how βArr2, which is strictly detected in the cytosol, regulates mitogenesis. Moreover, to date no studies have been reported on the role of βArr2 in RCC. Here, we show that βArr2 controls c-Src activation and Cyclin A expression and regulates RCC localized and metastatic tumor growth.

ARRB2 is abundantly expressed in human RCC.
To determine potential significance of βArr2 in RCC, we first analyzed available human datasets of ccRCC patients for ARRB1 and ARRB2 genes expression. We found that ARRB2, but not ARRB1, gene is significantly more expressed in ccRCC tumor compared to normal tissue (Fig. 1A,B). Similar results were observed for other kidney cancer subtypes, including papillary (Fig. 1C,D) and chromophobe (Supplemental Fig. S1) RCC that were reported in additional datasets. Notably, ARRB2 levels correlated with ccRCC disease stage (Fig. 1E), and survival data showed that patients (diagnosed with ccRCC) with high ARRB2 expression levels had significantly lower survival outcomes compared to patients with low ARRB2 expression (Fig. 1F). Collectively, these results demonstrate that increased ARRB2, but not ARRB1, expression correlates with advanced ccRCC stage and decreased patient survival. βArr2 expression is upregulated in aggressive human RCC cell lines. To test the functional relevance of increased ARRB2 in advanced RCC, we analyzed ARRB2 (and ARRB1 as a control) expression in commonly used RCC cell lines, including CAKI-1, 786O, SN12C, and ACHN. The human kidney epithelial cell line HK2 (immortalized proximal tubule epithelial cells from normal adult kidney) served as a control 25 . The results show that ARRB2, but not ARRB1, expression was significantly higher in SN12C and ACHN, in comparison to  ) and (E) the P values were calculated using a log-rank test. Box plot line (from top to bottom): maximum; Q3, third quartile; median; Q1, first quartile; and minimum. P values were calculated by using two-tailed paired Student's t test. Data shown are mean ± SEM, **P < 0.001 and ***P < 0.0001.
Next, we characterized these cell lines for their growth properties using in vitro and in vivo assays. Results showed that SN12C and ACHN cells formed faster growing spheres in matrigel (Fig. 2C) and tumors in animals ( Fig. 2D-F), in comparison to HK2, 786O and CAKI-1 cells. These results lead us to conclude that increased βArr2 levels correlate with high proliferative capacity of RCC cells.
Knockout of βArr2 inhibits the cell proliferation. We created stable βArr2 knockout SN12C lines using CRISPR/Cas9 system. Cells were infected with a single lentiviral CRISPR/Cas9 plasmid containing specific sgRNAs that targeted exon 3 or exon 4 of ARRB2 gene. Control cells were also generated using the same CRISPR/Cas9 plasmid that, however, lacked sgRNA. Single cell expansion yielded multiple clones, including ARRB2ex3-14 (βArr2 ko Clone1) and ARRB2ex4-12.19 (βArr2 ko Clone2) that did not show detectable βArr2 protein on Western blot, but exhibited normal expression of endogenous βArr1 protein (Fig. 3A).
To elucidate role of βArr2 in the RCC cell growth, we used the control and βArr2 ko clones in in vitro cell proliferation assays. Results show that the knockout of βArr2 expression decreased the proliferation rate of SN12C cells cultured in serum-containing growth medium (Fig. 3B). To add support to this observation, we engineered CAKI-1 cells (that have low levels of endogenous βArr2; Fig. 2B) to express empty vector (EV) and Flag-tagged βArr2 (Flag-βArr2) (Supplemental Fig. S2A). We observed that the forced expression of βArr2 significantly increased proliferation rate of the Flag-βArr2 CAKI-1 cells, in comparison to the control EV cells (Fig. 3C).
We used sphere formation in matrigel assay as another way to further implicate βArr2 in RCC growth. Results show that whereas control SN12C cells formed large numbers of fast-growing spheres, cells with βArr2 ko formed fewer and slow-growing spheres (Fig. 3D,E). In addition to differences in the sphere growth rate (Fig. 3D,E), we also observed morphological differences between the control and βArr2 ko colonies growing in matrigel (Fig. 3F). We observed that outer cells in control colonies displayed spike-like extensions whereas colonies originating from cells with βArr2 ko had smooth edges, like those of the non-tumorigenic HK2 cells (Fig. 3F). suggests a role in cell invasion and migration. Cell attachment to surrounding matrix is a prerequisite for the cell migration and invasion, and we used in vitro assay to test whether βArr2 plays a role in the cell attachment. Results show that as early as one hr after seeding, the number of attached βArr2 ko cells was significantly less than control cells (Fig. 4A). To overcome the possibility that the knockout of βArr2 yields compensatory mechanisms responsible for the reduction in cell attachment, we used shRNA in SN12C (Fig. 4B) and siRNA in ACHN (Supplemental Fig. S2B) cells to knockdown βArr2. We measured rate of cell migration in transwell chambers and observed a 50% reduction in the migration of both cell types when βArr2 was knocked down, in comparison to respective controls (Fig. 4C,D and Supplemental Fig. S2C,D). We also tested impact of βArr2 on the cell invasion of matrigel and could show that the knockdown of βArr2 inhibited ACHN cell invasion by 60% compared to control cells (Supplemental Fig. S2E). Together, these results establish a role for βArr2 in RCC cell migration and invasion.  26 . Tumors were allowed to grow for 5 weeks before animals were sacrificed and organs (tumor-implanted and contralateral normal kidney, lymph nodes, lung, spleen, liver, and intestine) harvested. For duration of the experiment, tumor size was monitored weekly by palpation 26 and ultrasound imaging (Supplemental Fig. S3A). In agreement with the effect of βArr2 on cell (Fig. 3B,C) and sphere ( Fig. 3D-F) growth, the results showed that tumors from βArr2 ko clones grew significantly less, in comparison to tumors originating from control cells (Fig. 5A,B).
Macroscopic analysis of the sacrificed animals evidenced presence of metastatic growth throughout the viscera of mice harboring control SN12C tumors, but not in animals implanted with βArr2 ko SN12C cells (Supplemental Fig. S3B). Concordantly, staining of kidney tissue sections with a human LDHA antibody revealed that tumors from βArr2 ko SN12C cells failed to invade through the cortex of the kidney, in contrast to control tumors where majority of the kidney cortex was infiltrated by the SN12C cells (Fig. 5C). Cancer cell local invasion and migration are a prerequisite for the tumor metastasis. Lymph node infiltration by cancer cells is an early mark for dissemination and metastatic growth. We observed an increase in renal lymph node size (Supplemental Fig. S3C,D) in animals implanted with control, but not βArr2 ko SN12C cells. The harvested organs were also inspected for the presence of visible metastatic nodules that were successfully seen in the spleen, intestine and liver obtained from animals implanted with control SN12C cells, but not with βArr2 ko SN12C cells (Fig. 5D). Moreover, the staining of lymph node (Supplemental Fig. S3E) and lung (Fig. 5E) tissues with human LDHA antibodies revealed presence of metastatic growth in control but not βArr2 ko tissues. For lungs, 6 out of 7 animals implanted with control SN12C cells showed metastatic growth, but no animals harboring βArr2 ko SN12C tumors evidenced the tumor growth at metastatic sites (Fig. 5F). We repeated the subrenal capsule implantation experiment once more but this time extended duration of the experiment from the original 5 week to a 9-week period. The rationale was to test whether the prolonged implantation time would allow for the detection of localized and metastatic tumor growth. Unexpectedly, the almost doubling of the original experiment duration still failed to produce an increase in βArr2 ko tumor growth (Supplemental Fig. S4A,B). Furthermore, the animals did not show any visible metastatic tumor growth to the viscera or soft tissues, implying βArr2 is involved primarily in RCC tumor growth.  Supplemental Fig. S4). We determined expression levels of Ki67, a marker for cell proliferation, in tumor tissues originating from control and βArr2 ko SN12C cells. The sections from βArr2 ko tumors evidenced fewer Ki67-positive cells compared to control cells (Supplemental Fig. S5A), suggesting involvement of βArr2 in the cell cycle progression. LDHA, which is detected only in the human cells, was expressed similarly among control and βArr2 ko cells. To support the idea that βArr2 impacts the cell cycle, control and βArr2 ko cells were stained with DAPI, that reflects DNA content, and analyzed 48 hr after seeding. The majority of βArr2 ko cells were detected in sub-G1, outside the range of what was considered live for the control cells ( Fig. 6A and Supplemental Fig. S5B). Of the live cells, we calculated the cell fraction in G1, S, and G2 phases and could observe the clear reduction in the number of βArr2 ko cells in G2 phase, in comparison to control cells (Fig. 6B).
Cyclin A is involved in the cell cycle progression and together with Cdk2 allows dividing cells to synthesize DNA and enter mitosis 27 . Indeed, Cyclin A expression is detectable in late G1 phase, continues to increase through S phase, reaches maximal levels in early G2 phase at which time the protein begins to degrade in the proteasome and becomes undetectable during early mitosis. Western blot analysis revealed that Cyclin A expression was reduced in βArr2 ko cells compared to the control cells (Fig. 6C). Distinctly, the levels of Cyclin D1, which is highly expressed in mid-G1 to early mitosis phases, were similar in the control and βArr2 ko cells. These observations are consistent with our results showing that the cell cycle arrest observed in βArr2 ko cells is at the S/G2 check point (Fig. 6B). Progression through the cell cycle involves multiple intermediates, notable among them is the βArr effector tyrosine kinase c-Src 20,28,29 . We reasoned, therefore, that βArr2 regulates Cyclin A expression and cell cycle progression through c-Src activation. Indeed, we found that the knockout of βArr2 blunts the activation, but not expression, of c-Src as evidenced by the decreased c-Src phosphorylation on Y416 (Fig. 6D). Moreover, we observed that the treatment with PP2, a selective inhibitor of Src family kinases, decreased the expression of Cyclin A (Fig. 6E), implying βArr2 regulates the cell cycle progression through c-Src activation and Cyclin A expression.

Discussion
Identification of molecular drivers of RCC is critical for the development of therapeutic strategies to more effectively treat the advanced disease. Here, we showed that βArr2 regulates the RCC cell proliferation in vitro and localized and metastatic tumor growth. Several studies have implicated βArr2 in mitogenic signaling 30 and established its increased expression in human tumor versus normal tissue 31,32 . However, just how βArr2 expression and signal deregulation contributes to RCC remains unknown. Herein, we show that inhibition of βArr2 attenuates c-Src activation, Cyclin A expression, and the ensuing cell cycle progression that controls the cell growth (Fig. 6F).
The knockout of βArr2 inhibited the cell cycle progression and cell proliferation in vitro, and xenograft tumor formation in animals. In cell model systems, βArr2-mediated mitogenic signaling downstream of stimulated G protein-coupled receptors involves the activation of ERK MAP kinases and tyrosine kinases 10,33 . Consistent with these findings, our results here show that decreased expression of βArr2 inhibited c-Src activation. In the case of βArr1, it was reported that it binds directly to inactive c-Src leading to c-Src activation 20,28 . Our results show that downregulation of βArr2 inhibits c-Src activation but not expression in RCC cells, presumably due the lack of opportunity to form a βArr2-Src complex. In addition to the regulation of c-Src activation that is critical for cell proliferation, the knockout of βArr2 inhibited Cyclin A expression and cell cycle progression. The βArr2-dependent regulation of Cyclin A expression has not been reported before and the responsible mechanisms remain unclear. However, it has been reported that activated c-Src regulates cell cycle progression through Myc and Cyclin A/Cdk2, thereby permitting the S-G2 progression of the cell cycle 34 . Together, our results put forth an expanded role for βArr2 function in RCC cell growth; through the regulation of Cyclin A expression and cell cycle progression 35 . RCC is a heterogeneous disease and the most associated genetic malfunction in ccRCC subtype is VHL inactivation 36 . However, VHL inactivation alone may not be sufficient to drive tumor formation as evidenced in genetically engineered Vhl −/− animal models 37 . In addition to VHL inactivation, recent studies in mice have identified mutations in other genes, including cell cycle regulating genes, in the development of ccRCC 38 . Our study showed that knockout of βArr2 inhibits the cell cycle programs and RCC xenograft tumor growth in mice, suggesting deregulated expression of βArr2 may collaborate with disease-driving genetic mutations in the development of RCC. In a recent study, it was reported that ccRCC can develop in mice with deletion of Vhl, Trp53 and Rb1 genes 39 , but inactivation of Trp53 and Rb1 is rare in human ccRCC 3 . It is worth mentioning that tumor tissues from the Vhl/Trp53/Rb1 mouse expressed recurrent mutations in other genes, including Kif3a 39 , and βArr2 appears to regulate Kif3a function and cilium formation 40,41 . As such, our results are consistent with the idea that increased expression of βArr2 provides a supportive role in the development of ccRCC.
Ubiquitously expressed βArrestin proteins comprise βArr1 and βArr2 that are highly homologous and, for the most part, functionally redundant 33,42 . Our bioinformatics analyses revealed an association between increased expression of βArr2, and not βArr1, in RCC, implying functional specificity among the two related proteins. Indeed, proteomics analysis revealed specificity in the interaction between βArr1 and βArr2 and binding partners 43 . Using public databases we found that heightened expression of βArr2 positively correlates with RCC disease stage and poor prognosis thereby validating the clinical implication of βArr2 as a marker for advanced disease. Also, the increased expression of βArr2 in RCC patients is not restricted to race or gender and is detected in young patients 44 , reinforcing the idea that βArr2 may serve as an effective marker in the diagnosis of RCC.
In summary, we provide here clear evidence that βArr2 is highly expressed in human RCC tissue and exerts a key role in cancer formation. Inhibition of βArr2 expression reduced localized and metastatic RCC tumor growth. These results give rationale for the use of βArr2 as a prognostic biomarker and potential therapeutic target to combat the insofar lethal advanced RCC.

Materials and Methods
Reagents. Antibodies  RNA extraction and expression. Total RNA was extracted with High Pure RNA isolation kit (Roche), and 1 μg in a final volume of 20 μl was reverse-transcribed with iScript ™ Reverse Transcription Supermix per the manufacturer's instructions. Quantitative PCR reactions containing 400 ng of cDNA and 5 μl of iQ SYBR Green Supermix 5× in a total volume of 10 μl were performed in triplicate using Bio-Rad CFX detection system and target gene expression was normalized to 18S RNA. The primers used for gene amplification of human 18S (QT00199367), ARRB1 (QT00071197) and ARRB2 (QT00058051) were obtained from Qiagen.
Migration and invasion assays. Cell migration assays were done using 8 μm pore transwell chambers (Fischer; 07-200-150). Briefly, cells were serum starved for 16 hr, detached, re-suspended in starvation medium, and added to the transwell chambers (2.5 × 10 4 cells/well for ACHN, and 3 × 10 4 cells/well for SN12C). Starvation medium containing 1% FBS was added to the lower chambers and incubated at 37 °C for 8 hr for ACHN and 24 hr for SN12C cells. For invasion assay, cells (1 × 10 5 cells/well for ACHN, and 5 × 10 4 cells/well for SN12C) were seeded on pre-coated matrigel inserts (BD Biosciences; 08-774-122) and incubated for 24 hr. At termination point, cells were fixed and stained with 0.1% crystal violet in 20% ethanol. Cells that remained at the top of filter were removed and migrated cells were counted through a 10× objective lens with an Axioskop microscope (Zeiss).
Scientific REPORts | (2018) 8:4879 | DOI:10.1038/s41598-018-23212-w Tumor implantation. All experiments involving mice were done according to a protocol reviewed and approved by the Institutional Animal Care and Use Committee at the University of Florida. Subrenal capsule implantation protocol is detailed in Zhang et al. 26 . Briefly, male hsd: athymic nude-Fox 1nu (Envigo), 6 weeks old mice were grouped according to body weight. Soft collagen pellets containing 1 × 10 6 cells were placed under the capsules of the left kidney. Tumor growth was monitored by palpation and ultrasound imaging (GE Medical System InSiteExC) and, at the end of the experiment, both kidneys, draining lymph nodes, spleen, liver, intestine, and lungs were harvested, weighed, and fixed in 10% buffered formalin phosphate for IHC analysis.
Immunohistochemistry. Harvested organs were embedded in paraffin and sectioned (5 μm), deparaffinized in xylene, rehydrated in graded alcohol, subjected to heat-induced antigen retrieval, and blocked. Sections were probed with the indicated antibody at 4 °C overnight, and then incubated with Signal Stain Boost. Samples were developed with AEC substrate, counterstained with hematoxylin, and mounted with Faramount. Microscopic images were taken at 10× or 40× using Nikon Eclipse 50i microscope equipped with a DS-Fi1 camera and NIS-Elements BR3.1 software.
Statistics. Data are expressed as means ± SEM or SD. Statistical analysis was performed with either one-way ANOVA with Tukey's post-test or two-tailed paired Student's t test, and a P < 0.05 was considered statistically significant. Graphs were generated using Graph Pad Prism 6 software and axis labels were generated using Adobe Illustrator CS5.1.