Identification of therapeutic targets for renal medullary carcinoma via integrated genomic and transcriptomic profiling

Abstract

Renal medullary carcinoma (RMC) is a rare but highly aggressive kidney cancer that resists conventional therapies. To identify therapeutic targets, this study employs histopathologic, genomic, and transcriptomic profiling of 25 RMC samples. TROP2, EPCAM, CLDN6, and CDH6 are significantly overexpressed compared with other renal and solid tumors. Pathway analyses indicate Hippo pathway upregulation and a tumor microenvironment rich in fibroblasts and neutrophils. We subsequently explore treatment of four heavily pretreated patients, all with high TROP2 expression, using sacituzumab govitecan, a TROP2-targeted antibody-drug conjugate. Of these four patients, one patient achieves a partial response with symptom improvement, two patients maintain stable disease, and the median progression-free survival reaches 2.9 months. This study represents the most extensive molecular characterization of RMC to date, identifying TROP2 and other potential therapeutic targets. Sacituzumab govitecan demonstrates potential clinical benefit, warranting further evaluation in prospective trials to confirm its efficacy and explore additional targets identified herein.

Keywords: CDH6; CLDN6; EPCAM; Hippo pathway; SMARCB1; TROP2; renal medullary carcinoma; sacituzumab govitecan.

Graphical abstract

Figure 1

Genomic and transcriptomic profiling of RMC (A) Location of profiled tumor tissues for the RMC discovery cohort (n = 15). The main source was untreated primary kidney tumor tissues (n = 11), whereas the remaining tissues were obtained from metastatic sites. (B) Principal-component analysis (PCA) plots of the RMC samples (n = 15) and gene expressions of KIRC (n = 491), KIRP (n = 281), and KICH (n = 63) from TCGA (top) or ccRCC (n = 23) and PRCC (n = 7) from internal reference PanSolid cohort (bottom). (C) Pathway comparisons of RMC and KIRC transcriptomic signatures using ssGSEA and PROGENy scaled with KIRC cohort. (D) Tumor microenvironment (TME) cellular deconvolution of RNA-seq profiles by the Kassandra algorithm. (E) Mechanisms of SMARCB1 loss in RMC tissues determined by WES. Low tumor purity samples are less reliable for copy number and somatic mutation analysis from WES. (F) Median values of ssGSEA gene signature scores and cell-type proportions for each of the four TME profiles on KIRC cohort: immune-enriched fibrotic (n = 39), immune-enriched non-fibrotic (n = 140), fibrotic (n = 106), and immune desert (n = 206).

Figure 2

Transcriptomic pathways deregulated in the RMC discovery cohort compared with other malignancies (A) ssGSEA enrichment scores for the Hippo and cell-cycle pathways in RMC compared with TCGA transcriptomes from clear cell RCC (KIRC), papillary RCC (KIRP), colorectal adenocarcinoma (COREAD), lung adenocarcinoma (LUAD), sarcoma (SARC), and mesothelioma (MESO). (B) ssGSEA enrichment scores for the Hippo and cell-cycle pathways in RMC compared with internal reference transcriptomes from clear cell RCC (ccRCC), papillary RCC (PRCC), colorectal adenocarcinoma (CRC), lung adenocarcinoma (LuAd), sarcoma, and mesothelioma. (C) PROGENy scores of WNT, NF-κB, EGFR, pseudohypoxia, TNF-α, PI3K, MAPK, JAK-STAT, and TGF-β signatures between RMC and internal reference transcriptomes from ccRCC, PRCC, CRC, LuAd, sarcoma, and mesothelioma. All p values comparing RMC with each malignancy were calculated by the Mann-Whitney U test. For boxplots, the upper whisker indicates the maximum value or 75th percentile +1.5 IQR; the lower whisker indicates the minimum value or 25th percentile −1.5 IQR. The number of samples used for the internal reference transcriptomes is listed in Table S2.

Figure 3

Comparison of TME cell composition between the RMC discovery cohort and other malignancies TME cell proportions were reconstructed using the Kassandra algorithm in transcriptomic data from RMC and clear cell RCC (ccRCC), papillary RCC (PRCC), colorectal adenocarcinoma (CRC), lung adenocarcinoma (LuAd), and mesothelioma samples from our internal reference PanSolid cohort. All p values comparing RMC with each malignancy were calculated by the Mann-Whitney U test. For boxplots, the upper whisker indicates the maximum value or 75th percentile +1.5 IQR; the lower whisker indicates the minimum value or 25th percentile −1.5 IQR. The number of samples used for the internal reference transcriptomes is listed in Table S2.

Figure 4

Expression of cell surface targets in the RMC discovery cohort Comparison of gene expression levels for MUC16 (CA-125), ERBB2 (HER2), PVLR4 (Nectin-4), CLDN6 (claudin-6), CDH6 (cadherin 6), FOLR1 (folate receptor alpha), CLDN18 (claudin-18), CD276 (B7-H3), DLL3 (delta-like ligand 3), MSLN (mesothelin), EPCAM (epithelial cell adhesion molecule), and TACSTD2 (TROP2) between RMC (n = 15 samples) and 1,020 reference samples from the internal PanSolid cohort. All p values were calculated by the Mann-Whitney U test. For boxplots, the upper whisker indicates the maximum value or 75th percentile +1.5 IQR; the lower whisker indicates the minimum value or 25th percentile −1.5 IQR. The number of samples used for the internal reference transcriptomes is listed in Table S2.

Figure 5

Expression of cell surface targets in the RMC discovery cohort compared with relevant malignancies (A) MUC16 expression in RMC versus ovarian cancer (OV) tissues. (B) ERBB2 (HER2) expression in RMC versus HER2+ breast cancer tissues. (C) PVRL4 (Nectin-4) expression in RMC versus bladder cancer tissues. (D) CLDN6 (claudin-6) expression in RMC versus HER2+ breast cancer tissues. (E) CDH6 (cadherin 6) in RMC versus ovarian cancer (OV), clear cell RCC (RCC), and cholangiocarcinoma (CHOL) tissues. (F) FOLR1 (folate receptor alpha) in RMC versus ovarian cancer (OV) tissues. (G) DLL3 (delta-like ligand 3) in RMC versus neuroendocrine carcinoma plus small cell lung carcinoma (BG_NEC) tissues. (H) MSLN (mesothelin) in RMC versus mesothelioma tissues. (I) EPCAM (epithelial cell adhesion molecule) in RMC versus colorectal adenocarcinoma (CRC) and pancreatic adenocarcinoma (PAAD) tissues. (J) TACSTD2 (TROP2) in RMC versus triple-negative breast cancer (TN breast cancer) and bladder cancer tissues. (K) Distribution of TACSTD2 (TROP2) expression the 15 RMC samples (each sample corresponds to a red dashed line) compared with the reference PanSolid bimodal distribution (purple bars). All p values comparing RMC with each cohort were calculated by the Mann-Whitney U test. For boxplots, the upper whisker indicates the maximum value or 75th percentile +1.5 IQR; the lower whisker indicates the minimum value or 25th percentile −1.5 IQR. The number of samples used for the internal reference transcriptomes is listed in Table S2.

Figure 6

Efficacy of SN-38 in RMC (A) Multivariable analysis by multiple linear regression of gene expression differences between RMC and clear cell RCC (ccRCC) (left panel) and between RMC and triple-negative breast cancer (TNBC) (right panel) from our internal reference PanSolid cohorts. The models were adjusted for tumor stage, biological sex, tumor purity, and age. Genes are grouped by functional category: therapeutic targets (TOP1), ABC drug efflux pumps (ABCB1, ABCC1, ABCC2, and ABCG2), metabolic enzymes (GUSB, UGT1A1, CYP3A4, and CYP3A5), and DNA damage response (DDR) genes (ATM, ATR, and PARP1). Adjusted p values (Benjamini-Hochberg correction) are provided for each gene. (B) Viability curves and half-maximal inhibitory concentrations (IC50) of RMC cell lines (RMC2C1, RMC219, UOK353, and UOK360) and the TNBC cell line MDA-MB-231 after 120-h exposure to SN-38. IC₅₀ concentrations are listed for each cell line and are presented as the mean IC₅₀ determined from three independent experiments (biological replicates), with error bars indicating standard error of the mean (SEM) for the triplicate measurements.

Figure 7

Efficacy of sacituzumab govitecan in patients with RMC (A) Waterfall plot of best overall response, sorted by sacituzumab govitecan dose in four patients with RMC. The number of prior systemic therapies and TROP2 expression levels by RNA-seq and IHC are also shown. IHC, immunohistochemistry; N/A, not available; PR, partial response; SD, stable disease; PD, progressive disease; RNA-seq, RNA sequencing. (B) Swimmer plot showing the durability of responses to therapy. The number of prior systemic therapies and TROP2 expression levels by RNA-seq and IHC are also shown. (C) CT scans of the primary RMC tumor at the beginning of treatment (day 0) and after three infusions of sacituzumab govitecan for patient RMC96.