Kidney injury molecule-1 inhibits metastasis of renal cell carcinoma

Metastasis is present in approximately 30% of patients diagnosed with renal cell carcinoma (RCC) and is associated with a 5-year survival rate of < 15%. Kidney injury molecule 1 (KIM-1), encoded by the HAVCR1 gene, is a proximal tubule cell-surface glycoprotein and a biomarker for early detection of RCC, but its pathophysiological significance in RCC remains unclear. We generated human and murine RCC cell lines either expressing or lacking KIM-1, respectively, and compared their growth and metastatic properties using validated methods. Surprisingly, KIM-1 expression had no effect on cell proliferation or subcutaneous tumour growth in immune deficient (Rag1−/−) Balb/c mice, but inhibited cell invasion and formation of lung metastasis in the same model. Further, we show that the inhibitory effect of KIM-1 on metastases was observed in both immune deficient and immune competent mice. Transcriptomic profiling identified the mRNA for the pro-metastatic GTPase, Rab27b, to be downregulated significantly in KIM-1 expressing human and murine RCC cells. Finally, analysis of The Cancer Genome Atlas (TCGA) data revealed that elevated HAVCR1 mRNA expression in the two most common types of RCC, clear cell and papillary RCC, tumours correlated with significantly improved overall patient survival. Our findings reveal a novel role for KIM-1 in inhibiting metastasis of RCC and suggests that tumour-associated KIM-1 expression may be a favourable prognostic factor.

Invasion assay. Transwell culture plates with 8.0-μm-pore-size polycarbonate membrane filter inserts with 6.5 mm diameter (Corning, NY) were used to perform invasion assays. Plate wells were filled with serum-free media (SFM) or complete media with 10% FBS (Renca in RPMI; 786-O in DMEM). Renca cells were resuspended in SFM at a concentration of 2 × 10 5 cells/ml. 786-O cells were resuspended in SFM at a concentration of 1 × 10 5 cells/mL. Cells were seeded in the Transwell inserts coated with 200 µL of Matrigel (1:100 dilution in SFM; BD Biosciences, NJ) 12 h prior to seeding of cells. The assembled plates were incubated for 24 h in a 37 °C, 5% (v/v) CO 2 incubator. After incubation, insert membranes were dyed with eosin (Sigma-Aldrich, Oakville CA) and toluidine blue (Sigma-Aldrich, Oakville CA). Light microscopy was used to quantify the migrated and invaded cells (Leica Microsytems).

Experimental metastasis model. Renca cells were injected (5 × 10 5 cells/mouse) into the lateral tail
veins of 8-10-week-old female WT BALB/c mice. To study whether KIM-1-mediated inhibition of metastasis is dependent on adaptive immunity, Renca or 786-O cells were instead injected into the tail veins of immunedeficient BALB/c (Rag1 −/− ) mice in the same manner as described above. All mice were sacrificed at day 17 post injection.
Western blot. Cell lysates from 786-O or Renca cell lines were collected with 4% (w/v) sodium dodecyl sulfate (SDS; Bio Basic Inc.) in 1 × phosphate buffered saline. Total protein was quantified using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL). The lysates were boiled for 5 min at 95 °C to denature proteins. Samples were separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA) for 50 min at 90 V. Membranes were blocked with 3% (w/v) BSA (bovine serum albumin; Bio Basic Cat No. AD0023) in TBST (Tris-buffered saline, 0.2% Tween-20) for 30 min and then incubated overnight at 4 °C with anti-murine KIM-1 goat primary antibody (dilution 1:2000; Cat No. AF1817, R&D Systems, Minneapolis, MN) or anti-human monoclonal KIM-1 antibody (AKG, a kind gift from Dr. Bonventre) and anti-GAPDH monoclonal monoclonal or I-19 anti-β-Actin polyclonal antibody (dilution 1:1; Cat No. Sc-32233, and Cat No. Sc-1616, Santa Cruz Biotechnology). After incubating with the appropriate horse radish peroxidase (HRP)-conjugated secondary antibodies, the proteins were visualized using chemiluminescent HRP substrate (Millipore, Billerica, MA). The images for Renca cell lines were captured and analyzed using the Licor C-digit imaging device and Image Studio Lite, respectively. Western blots of human RCC cell lines (769-P and 786-O shControl and shKIM-1) reactive bands were observed by Super Signal West Pico (ThermoFisher Scientific, Waltham, MA). Blots were exposed onto blue X-Ray film and developed using the Kodak M35A-X-OMAT. Statistical analysis. Differences in means between KIM-1 pos and KIM-1 neg groups for all results were analyzed using unpaired two-tailed t tests. Statistical significance was defined as *p < 0.05; **p < 0.01; ***p < 0.001. Statistical analyses were performed using the GraphPad Prism, version 8 software (GraphPad Software Inc.; La Jolla, CA).

KIM-1 expression does not alter RCC cell proliferation or tumour growth in vivo.
We explored the role of tumour-associated KIM-1 in RCC using human and murine models. We silenced endogenous KIM-1 in human 786-O cells 23,24 using shRNA (Fig. 1A,B), and expressed exogenous murine KIM-1 (KIM-1 pos ) in murine Renca (RCC) cells using lentiviruses [Renca cells do not express endogenous KIM-1] (Fig. 1C,D). The effectiveness of knockdown or transfection was assessed at the mRNA and protein level as shown in Fig. 1. Targeted silencing of KIM-1 or KIM-1 overexpression had no effect on RCC cell proliferation in both human (786-O) and murine (Renca) cells in vitro ( Fig. 2A,C). Moreover, we did not observe any significant differences in KIM-1-dependent tumour growth upon subcutaneous implantation of either 786-O or Renca cells into the flanks of immune-deficient (Rag-1 −/− ) Balb/c mice (Fig. 2B,D). Taken together, these data indicate that KIM-1 does not promote RCC tumour growth characteristics.

KIM-1 expression inhibits RCC cell invasion.
Given the absence of KIM-1-dependent effects on RCC cell proliferation and tumour growth in vivo, we investigated the potential role of KIM-1 in altering the metastatic potential of both human and murine RCC cells. First, we used the Transwell invasion assay to compare invasion differences between shKIM-1 and shControl 786-O cells. shKIM-1 786-O cells showed significantly reduced invasive capability compared with shControl 786-O cells (Fig. 3A). Similar differences were observed with invasion differences between KIM-1 pos and KIM-1 neg Renca cells. KIM-1 pos Renca cells showed significantly reduced invasive capability compared with KIM-1 neg Renca cells (Fig. 3B). Given that the Matrigel invasion assay provides a physiologically relevant model to study the invasive potential of tumorigenic cells 25 27 ) resulted in increased cell proliferation but inconclusive results in terms of migration-migration in KIM-1 overexpressing cells was delayed at early time points but proceeded to migrate faster than control cells at later time points 13 .

KIM-1 inhibits the metastatic potential of RCC cells independent of adaptive immunity.
To  www.nature.com/scientificreports/ lung nodules compared to mice injected with shControl 786-O cells (Fig. 4A). Moreover, we found that the immune deficient Rag-1 −/− mice injected intravenously with KIM-1 pos Renca cells developed significantly fewer metastatic lung nodules compared with mice injected with KIM-1 neg Renca cells (Fig. 4B). Taken together, the above data suggest that KIM-1 expression inhibits steps involved in the metastatic cascade of RCC cells including invasion and extravasation. The response to immune checkpoint inhibitors in patients with metastatic RCC underscores the importance of cells of the adaptive immune system in the disease process 29 . To determine if KIM-1 inhibited lung metastasis in the presence of the adaptive immune system in our murine cell line, we repeated injections of KIM-1 pos and KIM-1 neg Renca cells intravenously into immune-competent (wild type) Balb/c mice. Importantly, we once again found that KIM-1 pos Renca cells developed significantly fewer metastatic lung nodules compared with KIM-1 neg Renca cells in Balb/c mice (Fig. 4C). To confirm our findings using an independent human RCC cell line, we silenced endogenous KIM-1 in human 769-P renal adenocarcinoma cells which also express endogenous KIM-1 using the lentiviral system used in the 786-O cells (Fig. S2A). Since 769-P cells did not form tumours in immune deficient mice (as stated above), we used our validated chorioallantoic membrane (CAM) of chicken embryo model to compare extravasation efficiency between 769-P shKIM-1 and 769-P shControl cells. In keeping with the from the intravenous model of metastasis suing the 786-O cells, we found that silencing of KIM-1 in 769-P cells significantly impeded extravasation of RCC cells from the veins of chick embryos (Fig. S2B and S2C).
Taken together, our data indicates that the KIM-1-mediated inhibition of metastasis is inherent to the cancer cells themselves, and not a by-product of the adaptive immune system targeting KIM-1 expressing cells in a more efficient way. These results could be due to loss of KIM-1 expression, which may contribute to the metastatic phenotype observed in RCC cells. However, this remains to be formally studied using human RCC metastatic tissue samples. www.nature.com/scientificreports/ To explore mechanisms underlying KIM-1-dependent inhibition of metastasis, we compared the transcriptomic profiles of KIM-1 pos and KIM-1 neg Renca cells (Fig. S1). The most significant enriched gene was for the pro-metastatic GTPase, Rab27b 35 , which was downregulated 29-fold in KIM-1 pos vs. KIM-1 neg Renca cells. We were able to corroborate these data in human RCC cells-there was approximately 50% reduction in Rab27b mRNA between 786-O shKIM-1 vs. 786-O shControl cells (Fig. S6). Interestingly, high expression of Rab27b mRNA was shown to correlate with worse overall survival in patients with clear cell RCC (ccRCC) and papillary RCC (pRCC) 36 .

KIM-1 expression in RCC patients predicts greater overall survival. Our current findings regard-
ing KIM-1 in RCC contrast that of several groups who have claimed that KIM-1 promotes tumour growth and exacerbates cancer progression. Microvascular invasion (MVI) is defined by the invasion of cancer cells into the endothelial walls of small blood vessels and is associated with higher risks of metastases and death in patients with ccRCC 14,37 . Mijuskovic et al. (2018) showed that high degrees of MVI were associated with significantly increased expression of tumour tissue KIM-1, and that urinary KIM-1 was associated with worse disease prognoses and higher TNM staging 38 . Their study therefore suggests that KIM-1 is associated with greater risks of invasion and metastases in patients with ccRCC. We therefore sought to access the effect of KIM-1 expression on patient survival, so we analyzed The Cancer Genome Atlas (TCGA) RNAseq database for mRNA expression of HAVCR1. We found that RCC patients using KIRC and KIRP patient databases (both combined and individual database analysis) expressed higher HAVCR1 mRNA within their tumour tissues compared with their normal adjacent tissues (Fig. 5A-C, Fig. S3A-C, Fig. S4A-C, Table 1), which supports our understanding that KIM-1 is highly expressed in RCC tumours 6 . When we examined the relationship between KIM-1 expression and RCC patient survival, however, we found that higher HAVCR1 mRNA levels in patients were correlated with greater overall survival (Fig. 5D). The survival curves corresponding to the female and male patients from the same cohort are shown in figure S5. Taken together, these data suggest that although KIM-1 is highly expressed in RCC tumours, increased KIM-1 expression is predictive of greater overall patient survival. www.nature.com/scientificreports/ Overall, our findings propose a novel role for tumour-associated KIM-1 in the pathogenesis of RCC that may be advantageous to patients. Moreover, elevated KIM-1 expression in tumour samples may serve as a positive prognostic factor for patients. Our study is limited in that the KIM-1 mRNA from the TCGA was obtained from primary tumour samples and not from metastatic tissues. Although the heterotopic tumour model used here convincingly suggested that KIM-1 does not promote tumour growth, an orthotopic (intrarenal) strategy may have offered additional insights into mechanisms of metastasis. The above conclusions may be congruent with previous findings from our group demonstrating that KIM-1 binds to and suppresses its activity of Gα 12 10,18 . Gα 12 has been demonstrated to promote metastasis in RCC via its activation by lysophosphatidic acid G-protein-coupled receptor 30,31 . Gα 12 has also been shown to promote the expression of TGF-β1 through a Rho/Rac-dependent