CA-125 as a Biomarker in Renal Medullary Carcinoma: Integrated Molecular Profiling, Functional Characterization, and Prospective Clinical Validation

Abstract

Purpose:

Renal medullary carcinoma (RMC) is a highly aggressive malignancy defined by the loss of the SMARCB1 tumor suppressor. It mainly affects young individuals of African descent with sickle cell trait, and it is resistant to conventional therapies used for other renal cell carcinomas. This study aimed to identify potential biomarkers for early detection and disease monitoring of RMC.

Experimental Design:

Integrated profiling of primary untreated RMC tumor tissues and paired adjacent kidney controls was performed using RNA-sequencing (RNA-seq) and histone Chromatin Immunoprecipitation Sequencing (ChIP-seq). The expression of serum cancer antigen 125 (CA-125), was prospectively evaluated in 47 patients with RMC. Functional studies were conducted in RMC cell lines to assess the effects of SMARCB1 re-expression.

Results:

MUC16, encoding for CA-125, was identified as one of the top upregulated genes in RMC tissues, with concomitant enrichment of active histone marks H3K4me3 and H3K27ac at its promoter. Elevated serum CA-125 levels were found in 31 of 47 (66%) RMC patients and correlated significantly with metastatic tumor burden (p = 0.03). Functional studies in RMC cell lines demonstrated that SMARCB1 re-expression significantly reduced MUC16 expression.

Conclusions:

The correlation between serum CA-125 levels and metastatic burden suggests that CA-125 is a clinically relevant biomarker for RMC. These findings support further exploration of CA-125 for disease monitoring and targeted therapeutics in RMC.

INTRODUCTION

Renal medullary carcinoma (RMC) is a rare but highly aggressive malignancy that mainly affects young individuals of African descent carrying the sickle cell trait and is resistant to the therapies used for the more common renal cell carcinomas (RCC) such as clear cell RCC (1–6). The defining molecular hallmark of RMC is loss of the SMARCB1 tumor suppressor in all cases as determined by immunohistochemistry (2,7). SMARCB1 protects cells from hypoxic stress induced by the sickling of red blood cells in the renal medulla of individuals with sickle hemoglobinopathies such as sickle cell trait (4,7). There are approximately 3 million individuals with sickle cell trait in the United States alone, and more than 300 million individuals with sickle cell trait worldwide (8), all of whom are at risk for developing RMC particularly when exposed to risk factors that can increase red blood cell sickling in the renal medulla (1,4,9). There is clearly a need to develop biology-driven biomarker strategies for the early and accurate detection of RMC that can be cost-effectively deployed at a global scale. Furthermore, biomarkers of tumor burden can guide tailored management strategies such as the combination of chemotherapy with definitive radiation therapy in selected patients with oligoprogressive or oligometastatic RMC (10).

Herein, we performed integrated profiling of primary untreated RMC tumor tissues and adjacent kidney controls to determine whether differentially expressed genes and histone modifications could reveal clinically relevant biomarkers for disease monitoring and inform the development of novel therapeutic strategies. This led us to identify MUC16, which encodes serum cancer antigen 125 (CA-125), as one of the top upregulated genes in RMC tumor tissues. We prospectively validated this finding and functionally determined in RMC cell line models that SMARCB1 loss is necessary but not sufficient for MUC16 expression, which explains why most, but not all, patients with RMC express elevated serum CA-125 levels. These findings provide critical insights towards using serum CA-125 as a clinical biomarker and pave the way towards future exploration of cell surface MUC16 as a potential therapeutic target in patients with RMC.

MATERIALS AND METHODS

Patients

We conducted a single-institution prospective evaluation of serum CA-125 levels in all 47 patients with RMC seen at The University of Texas MD Anderson Cancer Center (MDACC) from July 2019 until January 2024. The study was approved by the MDACC Institutional Review Board (protocols PA11–1045 and PA16–0736) and was conducted according to the Declaration of Helsinki guidelines. Baseline demographic and clinical data for each patient were gathered via individual chart reviews from the institution’s electronic medical records.

Clinical grade serum assays and tissue immunohistochemistry

All serum assays in patient samples were performed at MDACC using clinical grade institutional laboratory tests certified by the Clinical Laboratory Improvement Amendments (CLIA). The normal range thresholds used have been calibrated to ensure analytic validity and clinical reliability. SMARCB1 negativity in all RMC samples was confirmed through immunohistochemistry (IHC) using the CLIA-certified purified mouse anti-BAF47 Clone 25/BAF47 (BD Biosciences, San Jose, CA, USA, Cat# 612111, RRID: AB_2191717), which is utilized for clinical diagnosis of SMARCB1 loss at our institution (3,7) and other clinical groups (11). Sickle cell status was assessed by hemoglobin electrophoresis. IHC for MUC16 was performed in a CLIA certified clinical laboratory at MDACC via a ready to use (RTU) assay using mouse anti-MUC16 clone OV185:1 (Leica Biosystems, Nussloch, Germany, Cat# PA0539-U, RRID: AB_3674374). The intensity of MUC16 protein expression by IHC was recorded as negative (0), weak (1+), moderate (2+), or strong (3+). The proportion of MUC16 protein expression in percentage was also documented. The H-score was then estimated using the following formula: [(% of positive cells) × (intensity of protein expression)].

RNA sequencing

RNA was extracted from fresh frozen RMC (n=11 cases), adjacent normal kidney (n=6 cases), and cell lines using the RNeasy Kit (Qiagen, Hilden, Germany, Cat#74106) according to the manufacturer’s instructions. Normal kidney tissues were obtained from locations at least 2 cm away from the primary tumors and the absence of metastatic cells was confirmed by a genitourinary pathology expert (Priya Rao) as previously described (7). After ensuring the quality of the initial samples, rRNA depletion was performed on the total RNA from each sample. This was followed by random-primed, stranded cDNA preparation, and quality control. The total RNA was then converted into a library of template molecules for sequencing on the Illumina HiSeq2000, with paired-end read lengths of 100 to 125 nucleotides. RNA-Seq data was trimmed using trimGalore (RRID:SCR_011847), then mapped using STAR (RRID:SCR_004463) (12) to the human genome build hg38. Gene expression was quantified using featureCounts (RRID:SCR_012919) (13) against using the ENSEMBLE gene reference (RRID:SCR_002344). Differential gene expression was derived using upper quartile normalization, RUVSeq (RRID #: SCR_006263) (14), and EdgeR (RRID:SCR_012802) (15); significance was considered achieved at FDR<0.05 and fold change exceeding 1.5x. Enriched pathways were determined using over-representation analysis (ORA), using a hypergeometric distribution and the approach implemented by the Molecular Signature Database (MSigDB) (RRID:SCR_016863 ) (16), with significance achieved at FDR<0.05. Pathway enrichment used the Gene Ontology Biological Processes (GOBP) compendium (RRID:SCR_002811) compiled by MSigDB (v7.5.1). Transcription factor targets enrichment used the CollecTRI compendium (17).

ChIP-Seq data analysis

Details on chromatin extraction and ChIP-seq library preparation are provided in the Supplementary Methods. ChIP-seq data was trimmed for low quality reads using trimGalore (RRID:SCR_011847), then mapped using bowtie2 (RRID:SCR_016368) (18) against the human genome hg38. Differentially enriched peaks (DEP) were derived using diffReps (RRID:SCR_010873) (19) with significance achieved at FDR<0.05 and fold change exceeding 1.5x. ChIP-seq signal maps were generated using bedtools (RRID:SCR_006646) (20) and Deeptools (RRID:SCR_016366) (21). Circos plots were created using the circos package (RRID:SCR_011798) (22). Overlap with the human enhancers was generated using bedtools (RRID:SCR_006646) and the Enhancer Atlas database (EnhAtl) (23). EnhAtl provides enhancer-gene chromatin loops; since we are assessing regulation across a mixture of tissues, we first used bedtools to merge enhancers from EnhAtl, thus generating enhancer anchors; next, we used an in-house Python code to determine enhancers that overlap with DEPs via bedtools and then associated affected enhancer anchors with their target genes. Additional integration of RNA-Seq and ChIP-seq signatures was performed using bedtools (RRID:SCR_006646), determining genes with a DEP within 10kB from their gene bodies. Genome wide chromatin states were derived using the ChromHMM packages (RRID:SCR_018141) (24).

Cell culture

All cell lines were refreshed from frozen early-passage stock after approximately 20 passages. The RMC2C1 cell line (RRID:CVCL_B3PS) was previously characterized (7) and was maintained at 37°C in minimum essential medium (MEM) supplemented with 1X MEM non-essential amino acids, penicillin–streptomycin (100 U/mL). The RMC219 cell line (RRID:CVCL_B3PV) was previously characterized (7) and was grown at 37°C in DMEM:F12 (50:50) medium supplemented with sodium pyruvate and L-glutamine. UOK360 (RRID:CVCL_B3PR) and UOK353 (RRID:CVCL_B3PQ) cell lines were generously provided by Dr. Marston Linehan’s laboratory at National Cancer Institute (Bethesda, Maryland) and were grown as previously described (25) in DMEM high glucose media supplemented 2 mM L-glutamine. OVCAR3 (RRID:CVCL_0465), which is an ovarian cancer cell line, was graciously provided by MD Anderson Traction center and was cultured in RPMI media. 786-O (RRID:CVCL_1051) and A-498 (RRID:CVCL_1056) clear cell RCC cell lines were purchased from American Type Culture Collection (ATCC, Manassas, VA), and the HEK293FT cell line (RRID:CVCL_6911) was acquired from ThermoFisher Scientific (Waltham, MA). 786–0 and HEK293FT were cultured in DMEM while A-498 was cultured in MEM media. All the media were supplemented with 100 U/mL penicillin–streptomycin and 10% heat-inactivated fetal bovine serum. All the cell lines used in this study were routinely tested and consistently confirmed negative for mycoplasma by the Universal Mycoplasma Detection Kit (ATCC Cat#30–1012K) as of 1/1/2025. In addition, short tandem repeat (STR) fingerprinting was performed to confirm cell type.

SMARCB1 Re-expression Experiments

We generated tetracycline-inducible SMARCB1 re-expressing RMC219 and UOK360 cell lines as previously described (7), using the tetracycline-inducible pIND20-fSNF5-HA vector (26), kindly provided by Dr. Bernard E. Weissman. The pInducer20 empty backbone (27) was obtained from Stephen Elledge (Addgene plasmid #44012, RRID: Addgene_44012). Lentivirus was produced in HEK-293FT cells to create stable tet-inducible cell lines as previously described (28). All plasmid vectors were propagated in the E. coli strain DH5α (Invitrogen Cat#18265017). For all SMARCB1 re-expression experiments, unless otherwise noted, a doxycycline concentration of 0.5 μg/mL was used for 3 days in cells containing the tetracycline-inducible pIND20-fSNF5-HA vector or the pInducer20 empty backbone control.

Protein extraction and western blot analyses

For cellular protein lysates, cells were scraped on ice using cold 1X RIPA lysis buffer (ThermoFisher Scientific Cat# PI89901) supplemented with Pierce Protease and Phosphatase inhibitor cocktail (ThermoFisher Scientific Cat# A32961), sonicated for five cycles with 30 sec on and 30 sec off per cycle, and centrifuged at maximum speed for 10 min at 4 °C. The pellet was removed, and supernatants were collected and processed further for quantitation of protein concentration using Pierce^™ BCA Protein Assay Kits (ThermoFisher Scientific Cat# 23227) following manufacturer’s instructions.

All the protein samples were resolved by using SDS–PAGE (5–15%) using PROTEAN TGX Precast Gels (Bio-Rad) under reducing and denaturing conditions. Resolved proteins were transferred to polyvinyl difluoride (PVDF) membranes at either 300 mA for 2 to 3 hours or at 15V overnight in transfer buffer (25 mM Tris, 192 mM glycine, pH 8.3, 20% methanol (v/v)). Membranes were then blocked in 5% non-fat dry milk dissolved in tris-buffered saline [50 mM tris-HCl (pH 7.6) and 150 mM NaCl] containing 0.5% Tween 20 (TBST) for 1h at room temperature, followed by incubation with primary antibody with gentle rocking overnight at 4°C. The primary antibodies used included: mouse monoclonal anti-MUC16 (Cell Signaling Technology Cat# 19017, RRID:AB_2924774), rabbit monoclonal anti-SMARCB1 clone D9C2 (Cell Signaling Technology Cat# 8745, RRID:AB_10950321), mouse monoclonal anti-β-Actin (Santa Cruz Biotechnology Cat# sc-47778, RRID:AB_626632), mouse monoclonal anti-α-Tubulin ((Santa Cruz Biotechnology Cat# sc-32293, RRID:AB_628412). Membranes were washed three times with TBST, incubated with the appropriate horseradish peroxidase (HRP)–conjugated secondary antibodies for 1 hour at room temperature, washed three times with TBST, and developed using Pierce ECL substrate (ThermoFisher Scientific Cat# PI35050) or Amersham ECL Prime (Cytiva Cat# RPN2236) for 3 and 5 min, respectively. Quantitation of immunoblots was performed via densitometric analysis using ImageQuant TL software (Cytiva), with data normalized to β-Actin. Relative fold changes in MUC16 levels between control and treated conditions (doxycycline for SMARCB1 induction) were normalized to β-actin and control conditions as previously described (29).

CA-125 enzyme-linked immunosorbent assay (ELISA)

To detect the CA-125 released into the cell culture supernatant, a commercially available CA-125 ELISA kit (RayBiotech Cat# ELH-CA125) was used, and the assay was performed following manufacturer instructions. Briefly, cell culture supernatant samples (untreated or treated either with doxycycline) were collected and centrifuged at 1,250 rpm for 10 minutes to remove any cellular debris. Samples were then stored at −80°C until further analysis. All the samples and the standards were thawed on ice and brought to room temperature prior running the assay. 100 μL of standards, and samples were added to a 96-well microtiter plate pre-coated with anti-CA-125 antibodies. Plates were incubated overnight at 4°C temperature and then washed four times with washing buffer. A secondary enzyme-linked antibody was added, followed by incubation, and washing as per the manufacturer’s instructions. 100 μL of substrate solution was added, incubated in the dark for 20 minutes, and the reaction was stopped with 50 μL of stop solution. Immediately thereafter, OD of the samples in each well was measured at 450 nm using a microplate reader. A standard curve was generated by plotting the OD values against the concentrations of CA-125 standards and CA-125 concentrations were calculated from the standard curve using linear regression. All the samples were assayed in triplicate or more, and data were analyzed using GraphPad prism software.

Immunostaining, Image Acquisition, and quantification

RMC2C1, RMC219, UOK360, UOK353 and OVCAR3 cells were placed onto six-well culture dishes with coverslips and incubated overnight at 37°C incubator with 5% CO₂. Subsequently, the cells were fixed with 4% paraformaldehyde (PFA) in PBS for 15 min at room temperature and permeabilized with 0.5% Triton X-100 in PBS for 20 min. The cells were then blocked using 3% BSA in PBS for 1 h and blocked with anti-MUC16 antibody (Cell Signaling Technology Cat# 19017, RRID:AB_2924774) prepared in a 3% BSA solution at ratios of 1:500 and incubated overnight at 4°C. After three washes with PBS, cells were stained with mouse Alexa Fluor^® 594 labelled secondary antibodies (Invitrogen, Eugene, OR) at a 1:2000 dilution at room temperature for 1 h. After three washes, cells were postfixed with 4% PFA for 10 min to stabilize the signal followed by two washes with PBS and counterstaining with DAPI (ThermoFisher Scientific Cat# 62248; diluted 1:10000) and Alexa Fluor^® 488 Phalloidin (ThermoFisher Scientific Cat# A12379; diluted at 1:2000) for 10–20 min at room temperature. Coverslips were mounted using Prolong^™ Gold or Glass Antifade Mountant (ThermoFisher Scientific Cat# P36934 or P36984, respectively). For acquisition of the high-resolution images, an HC PL APO 100×/1.40 CS2 laser scanning confocal microscope (Leica Sp8) with spectral emission detection on an inverted stand with a fully motorized stage and fast z movement (Leica Super Z Galvo stage) and allowed up to ~20 nm resolution for observation of subcellular structures. For quantification of MUC16 protein expression, exposure time was kept unchanged across all the conditions. All the acquired images were then processed using Leica’s LASX software. The acquired fluorescent images were then analyzed using FIJI software (Version 2.0.0-rc-49/1.51a) by selecting one cell at a time in an image and measuring the area, integrated density and mean gray value. Using the calculation for corrected total cell fluorescence (CTCF) = integrated density (area of selected cell × mean fluorescence of background readings), as previously described (30), the fluorescence intensity of each cell was calculated using Excel (Microsoft Office 2011 for Mac). For each image, three background areas were used to normalize against autofluorescence/noise.

Statistical analysis

Unless otherwise specified, all graphs and statistical significance was determined using GraphPad Prism 10.0.3 software (Boston, MA, USA). Linear regression with 95% confidence intervals was used to correlate the number of metastatic sites with CA-125 levels. Adjusted P values < 0.05 were considered significant.

Data availability

ChiP-seq and RNA-seq data from the RMC2C1 and RMC219 cells as well as ChiP-seq from the RMC tumor tissues and adjacent kidney can be found at the Gene Expression Omnibus repository (GEO accession number GSE284396). RNA-seq data from patient tumor samples were generated by Msaouel et al. (7) and can be found at the NCBI Sequence Read Archive (SRA) hosted by the NIH (SRA accession: PRJNA605003). Additional Supplementary Data files and computational analysis scripts are available at Zenodo (https://doi.org/10.5281/zenodo.14374700). All raw data used to generate figures and tables in this study are available from the corresponding author at PMsaouel@mdanderson.org and will be promptly reviewed to verify whether the request is subject to any intellectual property or confidentiality obligations.

RESULTS

Integrated chromatin and transcriptomic profiling of RMC

We integrated ChIP-seq for histone modifications with RNA-seq data from untreated primary RMC tumors and adjacent kidney controls (Figure 1A and Supplementary Table 1). Of the four histone markers assessed by ChIP-seq (Figure 1B), the histone modification H3K27ac is typically associated with active gene transcription and often found at active enhancers and promoters, H3K4me1 is generally found at enhancers that are either active or at an intermediate state marking regions that are accessible but not necessarily actively transcribing, H3K4me3 is predominantly located at the promoters of actively transcribed genes, whereas H3K27me3 is associated with gene repression (32–35). RMC tumor tissues were most prominently enriched for the H3K4me3 and H3K27ac histone modifications compared with adjacent kidney control, as indicated by the number of increased differentially enriched peaks (DEPs) (Figure 1B, C). Further, DEPs of both histone modifications within 10kb of gene body (Figure 1D) and at enhancers (Figure 1E) were frequently associated with differentially expressed genes by RNA-seq in RMC tissues compared with adjacent kidney control (Supplementary Figure 1). Figure 1F shows illustrative examples of the concordance between transcriptional upregulation and increased DEPs for H3K4me3 and H3K27ac in RMC tissues compared with adjacent normal control for the MYC and CD70 genes, previously established to be highly expressed in RMC (7,36).

Figure 1. Integrated chromatin and transcriptomic profiling of RMC.

A. Tumor and adjacent normal tissue samples were collected from patients with RMC prior to treatment and used for RNA-Seq and ChIP-seq analysis. B. Bar plot with number of DEPs for each histone mark. C. Bar plot with number of differentially expressed genes (DEGs) that have a DEP within +/–10kb from the gene body. D. Bar plot with number of DEGs that have a DEP in an enhancer, as defined by Enhancer Atlas. E. Integrative Genome Viewer (IGV) tracks for three genes showing signal intensities for RNA-seq or ChIP-seq (H3K27ac or H3K4me3) reads comparing normal and tumor samples. Black lines under peaks indicate DEPs called by DiffReps. F. Schematic of the MUC16 protein structure, with the putative cleavage site indicated in red.

Using the histone modification marks in RMC and adjacent normal tissues, we derived genome-wide chromatin maps using ChromHMM (Supplementary Figure 2A). Compared to normal, we determined an enrichment of poised/bivalent TSS over CpGs (Supplementary Figure 2B), but overall similar sizes for the various epigenomic states (Supplementary Figure 2C). However, a focused examination of the chromatin at MYC and CD70 revealed that chromatin upstream and downstream of these genes changed from Quiescent and Polycomb repressed states to enhancer states, EnhWk, EnhA1, and EnhA2 (Supplementary Figure 2D), suggesting that epigenetic reprogramming of the cis-regulatory elements of MYC and CD70 occurs in RMC tumorigenesis.

We noted that MUC16 was the 24^th most highly expressed gene in RMC tissues compared with adjacent kidney control (Supplementary Table 2) with concomitant enrichment for H3K4me3 and H3K27ac in its promoter region (Figure 1F). MUC16 is a transmembrane protein with a heavily glycosylated extracellular domain that undergoes proteolytic cleavage to release the CA-125 glycoprotein fragment into circulation while the transmembrane domain and cytoplasmic tail remain on the cell surface (Figure 1G). We focused on MUC16 due to the clinical potential of serum CA-125 as a tumor biomarker and the potential to target cell surface MUC16 therapeutically (37). Analysis of ChromHMM epigenome states revealed a transition from Quiescent and Polycomb repressed states to enhancer states, suggesting that MUC16 is also epigenetically upregulated in part (Supplementary Figure 2D). Our previously published whole exome sequencing data also revealed recurrent focal chromosomal amplifications at chromosome 19p13.2, leading to copy number gains of MUC16 in RMC tissues (7). Interrogation of our previously published transcriptomic comparisons (7), confirmed high upregulation of MUC16 gene expression in RMC tissues compared with adjacent normal kidney (log2fold change = 6.97, false discovery rate [FDR] < 0.0001), SMARCB1-proficient collecting duct carcinoma (CDC; log2fold change = 2.79, FDR < 0.0001) and SMARCB1-proficient upper tract urothelial carcinoma (UTUC; log2fold change = 4.47, FDR < 0.0001). Of note, RMC histomorphologically resembles and is anatomically close to CDC and UTUC and is often treated with similar therapeutic strategies despite being a biologically distinct entity characterized by SMARCB1 loss (2,7). We then interrogated data from the Cancer Genome Atlas (TCGA) (38–40) and noted that MUC16 upregulation is also uncommon in the three most common RCC subtypes: clear cell, papillary, and chromophobe RCC (Supplementary Figure 3). This prompted us to prospectively investigate the role of CA-125 in patients with RMC.

Prospective Evaluation of Clinically Relevant Serum Tumor Markers in RMC

We prospectively evaluated the expression of serum CA-125 in an independent cohort of 47 patients with RMC presenting in our clinic from July 2019 until January 2024 (Figure 2 and Supplementary Table 3). As shown in Supplementary Tables 3 and 4, this cohort is representative of the established baseline demographic and clinical characteristics in RMC, with the exception that there was no predilection towards the right kidney (2,3,41,42). Patients had a median age of 33 years, ranging from 14 to 75 years, primarily identified as black (78.7%), male (66%), had sickle cell trait (89.4%), and presented with stage IV disease at diagnosis (85.1%).

Figure 2. Prospective Evaluation of CA-125 in patients with RMC.

A. Measurements of serum markers from patients with RMC and healthy controls. Dotted line indicates upper threshold of normal range. B. CA-125 serum levels from RMC patient samples categorized by number of metastatic sites (left) with the correlation plot and 95% confidence interval (right). C. CA-125 levels and tumor burden measurements over time for patient RMC61. Systemic therapies (ST) are indicated below. D. CT scans of the tumor at the beginning of treatment (Day 0) and at Day 140 for patient RMC61. E. Immunohistochemistry (IHC) of representative RMC tissue sections using a CLIA-certified antibody against MUC16 (see Methods). Scale bar equals 100μm. F. Plots of peak CA-125 serum levels along with a histology H-score for MUC16 staining. G. Expression of MUC16 by RNA-seq for 11 patients used to stratify RMC tumors with high or low MUC16 gene expression. Arrows indicate samples used for ChIP-seq. H. Volcano plot showing number of DEGs comparing RMC tumors with high MUC16 over low MUC16 expression. I. Over-representation analysis (ORA) to identify enriched pathways from the Gene Ontology Biological Process compendium.

We also determined the levels of other clinically relevant serum tumor markers and noted that serum lactate dehydrogenase (LDH) was frequently elevated in our RMC cohort (Figure 2A) with the hypoxia-associated LDH5 being the predominantly elevated serum isoenzyme (Supplementary Figure 4A). Each of the five LDH isoenzymes (LDH1–5) is a tetramer composed of different combinations of LDHA and/or LDHB subunits (43). LDH1 consists of four LDHB subunits and was either below or within normal limits in the serum of all but 1/16 patients with RMC, whereas LDH5 is composed of four LDHA subunits and was elevated above normal limits in the serum of 14/16 patients with RMC (Supplementary Figure 4A). This is consistent with our RNA-seq findings showing significant upregulation of LDHA (log2foldchange 1.82, FDR < 0.0001) and downregulation of LDHB (log2foldchange −2.16, FDR < 0.0001) in RMC versus adjacent kidney (Supplementary Figure 4B).

Tumor-associated serum markers such as alpha fetoprotein (AFP), carbohydrate antigen 19–9 (CA-19.9), and carcinoembryonic antigen (CEA) were rarely elevated in RMC and further evaluation for these markers was therefore not pursued in patients seen from January 2022 onwards (Figure 2A). While cancer antigen 15–3 (CA 15–3) was frequently above normal limits in patients with RMC (Figure 2A), it did not exceed 60 U/ml in the 16 patients profiled and did not show a consistent association with metastasis burden to warrant further investigation (Supplementary Figure 4C).

Conversely, serum CA-125 was found to be elevated at presentation in 31/47 (66%) patients with RMC with levels often reaching above 400 U/mL (Figure 2A) and a significant positive correlation with metastatic burden at presentation (Figure 2B), as well as longitudinal association with radiologic response and progression as shown in Figures 2C–D using patient RMC61 as illustrative example of longitudinal changes in serum CA-125 levels. To rule out the possibility that negative CA-125 results were due to low disease burden at presentation, negative CA-125 levels were confirmed longitudinally in the 15/16 patients at subsequent peaks of disease burden (one patient, RMC87, remained disease-free throughout all subsequent follow-up). Cell surface expression levels of MUC16 in RMC tissues correlated well with corresponding peak serum CA-125 levels (Figures 2E–F). These results prompted further investigation into the potential biologic roles of MUC16 expression in RMC.

Differential transcriptomic and chromatin profiles of MUC16 expressing RMC tissues

We looked back again in our retrospective RMC cohort assayed with RNA-seq and ChIP-seq and noted that 5/11 primary untreated RMC samples expressed high MUC16 mRNA (group as “MUC16 high”) whereas 6/11 RMC samples (grouped as “MUC16 low”) and all adjacent kidney control samples expressed low MUC16 mRNA levels (Figure 2G). RNA-seq comparison of differentially expressed genes (DEGs) (Figure 2H) showed that MUC16 high RMC samples had significant deregulation of pathways related to cell adhesion, migration, and development (Figure 2I). DEPs for MUC16 high over MUC16 low were integrated with corresponding DEGs. Using the DEPs overlapping with enhancers and the Enhancer Atlas database, we determined that upregulated genes were more commonly enriched for active histone marks (H3K27ac and H3K4me3) whereas downregulated genes were enriched for the repressive histone mark H3K27me3 (Supplementary Figure 5A–C). The differentially expressed genes were targets of transcription factors, such as SP1 and JUN (Supplementary Figure 5D), known to be mediators of cell cycle regulation, proliferation, and cellular reprogramming (44,45). This prompted us to further investigate the role of MUC16 in RMC cell line models.

MUC16 expression in RMC cell lines

We next evaluated the protein expression of MUC16 in four previously established and characterized RMC cell lines: RMC2C1, RMC219, UOK360, and UOK353 (7,25). Using the ovarian cancer cell line OVCAR3 as positive control (46), we established the presence of cleaved serum CA-125 in the supernatant by ELISA (Figure 3A), total MUC16 by Western blot (Figure 3B), and cell surface MUC16 by immunofluorescent microscopy (Figures 3C–D) in the RMC219, UOK360, and UOK353 but not the RMC2C1 cell line. RMC219 and UOK360 consistently showed similar or slightly lower MUC16 and CA-125 levels to OVCAR3 and higher than UOK353 (Figure 3). We then generated RMC219 and UOK360 cells capable of conditionally re-expressing SMARCB1 upon doxycycline treatment and found that SMARCB1 re-expression led to decreased levels of cleaved serum CA-125 in the supernatant by ELISA (Figure 4A), total MUC16 by Western blot (Figure 4B), and cell surface MUC16 by immunofluorescent microscopy (Figures 4C–D) in both cell line models.

Figure 3. Expression of MUC16 and CA-125 in RMC cell lines.

A. Quantitation of CA-125 levels obtained by ELISA assay performed using the supernatants of four different RMC cell lines. Ovarian cell line (OVCAR3) supernatant was used as a positive control. 786-O, A-498 and HEK293FT supernatants were used as negative controls. Mean +/− SEM shown for N=8 samples per cell line (or 9 samples for each of the three negative control cell lines). B. Western blot (WB) showing the protein expression of MUC16 in various RMC cell lysates, with OVCAR3 as a positive control. Actin was used as a protein loading control. C. Quantification of WB via densitometric analysis normalized to β-Actin. D. Representative confocal immunofluorescence (IF) images of RMC cells lines and of OVCAR3 cell line showing cellular localization of MUC16 (red), actin (green), DAPI-stained nuclei (blue), and merged images. Scale bar= 25μm. E. Quantitation of fluorescent intensities of MUC16 from immunofluorescence images. Units are corrected total cell fluorescence (CTCF) intensity of MUC16 using ImageJ. Mean +/− SEM shown for N=40–133 fields per cell line.

Figure 4. MUC16 and CA-125 levels upon re-expression of SMARCB1 in RMC cell lines.

A. Quantitation of CA-125 levels obtained by ELISA assay performed using the supernatants collected after re-expression of SMARCB1 in two RMC cell lines. Mean +/− SEM shown for N=14–16 samples per group. B. Western blot showing MUC16 and SMARCB1 protein levels in RMC219 and UOK360 cells with or without re-expression of SMARCB1. Tubulin was used as a protein loading control. C. Representative confocal immunofluorescence images of RMC219 and UOK360 cells with or without re-expression of SMARCB1 showing cellular localization of MUC16 (red), actin (green), DAPI-stained nuclei (blue), and merged images. Scale bar = 25μm. D. Quantitation of fluorescent intensities of MUC16 from immunofluorescence images. Units are corrected total cell fluorescence (CTCF) intensity of MUC16 using ImageJ. Mean +/− SEM shown for N=85–137 fields per cell line. Asterisks indicate significance determined by p-value of <0.05 (*), <0.01 (**), <0.001 (***), or <0.0001 (****) using a Mann-Whitney t-test.

We then performed RNA-seq and ChIP-seq in the MUC16 negative RMC2C1 cell line and the MUC16 positive RMC219 cell line (Figure 5A). Integration of differentially expressed genes by RNA-seq (Figure 5B) with reprogrammed enhancers based on histone ChIP-seq (Figures 5C–D) revealed deregulation of pathways related to cell adhesion, migration, and development (Figures 5E) with 450 of these pathways overlapping with the corresponding pathways enriched in RMC tissues with high MUC16. (Supplementary Table 5). Notably, the RMC219 cell line expressing MUC16 was significantly enriched for the active H3K27ac and H3K4me3 histone markers in the promoter of MUC16 compared with the MUC16-negative RMC2C1 cell line, whereas the repressive histone marker H3K27me3 was significantly enriched in the MUC16 promoter of RMC2C1 cells (Figure 5G).

Figure 5. Integrated chromatin and transcriptomic profiling of MUC16 positive negative RMC cell lines.

A. RMC-derived cell lines used for RNA-seq and ChIP-seq. B. Volcano plot showing number of DEGs comparing cell line samples MUC16+ over MUC16. C. Bar plot with number of DEPs for each histone mark. D. Bar plot with number of DEGs that have a DEP in an enhancer, as defined by Enhancer Atlas. E. Over-representation analysis (ORA) to identify enriched pathways using a total of 4970 DEGs that had reprogrammed enhancers. Pathways are from the Gene Ontology Biological Process (GOBP) collection. F. Overlap of enriched GOBP pathways from MUC16 high over MUC16 low samples using either cell lines (turquoise) or RMC tumors (green). G. IGV tracks for MUC16 showing signal intensities for ChIP-seq (H3K27ac, H3K4me3, H3K4me1, H3K27me3) reads comparing RMC2C1 MUC16− and RMC219 MUC16+ cell lines. Lines under peaks indicate DEPs called up (black) or down (grey) by DiffReps.

DISCUSSION

This is the first study to evaluate the use of serum biomarkers in RMC. We identified MUC16, encoding serum cancer antigen 125 (CA-125), as a highly upregulated gene in RMC through integrated profiling of primary untreated RMC tumor tissues and paired adjacent kidney controls using RNA-seq and histone modification ChIP-seq. This upregulation was associated with the enrichment of active histone marks H3K4me3 and H3K27ac at the MUC16 promoter, indicating a transcriptionally active chromatin state; this observation was further strengthened by inspecting genome wide chromatin states derived using ChromHMM. Subsequent prospective evaluation of serum CA-125 levels in 47 patients with RMC revealed that elevated serum CA-125 levels were noted in 66% of patients and correlated with metastatic tumor burden. This finding is significant given the lack of effective biomarkers for RMC and the poor prognosis associated with this disease (2,3). The detection of elevated serum CA-125 in the majority of patients with RMC suggests its potential utility as a biomarker for disease monitoring and possibly early detection, especially in populations at risk, such as those with sickle cell trait (1,2).

The observation that not all patients with RMC exhibited elevated serum CA-125 levels, despite SMARCB1 loss, suggests that additional regulatory mechanisms may influence MUC16 expression. Our data indicate that SMARCB1 loss is necessary but not sufficient for MUC16 expression, pointing to the involvement of other factors in the regulation of this gene. This highlights the complexity of RMC biology and the need for further research to fully understand the pathways driving MUC16 expression and its role in tumor progression. Notably, serum CA-125 levels are also elevated in 60–80% SMARCB1-deficient epithelioid sarcoma cases (47,48), suggesting a broader relevance of this biomarker across different SMARCB1-deficient malignancies. Future studies should assess the potential utility of CA-125 as a biomarker for other malignancies characterized by SMARCB1 loss such as malignant rhabdoid tumors, atypical teratoid rhabdoid tumors, and SMARCB1-deficient sinonasal carcinomas (49).

Our results informed our activation of a phase 2 clinical trial (NCT06444880 at clinicaltrials.gov) evaluating the use of the MUC16 x CD3 bispecific antibody ubamatamab in patients with SMARCB1-deficient RMC or epithelioid sarcoma expressing MUC16. The trial employs the MUC16 IHC assay and scoring system established in the present study, along with the same serum CA-125 assay used here, to enroll patients with H score ≥ 25 and/or serum CA-125 ≥ 70 units/ml. Based on our preliminary findings, such as the correlation between longitudinal monitoring of serum CA-125 levels and tumor response demonstrated in Figure 2C, the clinical trial will monitor serum CA-125 levels every 3 weeks. This frequency was selected to capture dynamic changes in tumor burden while aligning with the typical cycle length of systemic therapy. The prospective data generated from this trial will allow us to refine and define the optimal time intervals for measuring changes in serum CA-125 levels in the context of treatment response.

Serum tumor biomarkers, such as prostate-specific antigen (PSA) in prostate cancer, CA-125 in ovarian cancer, and AFP or beta-HCG in germ cell neoplasms, are widely used for disease detection, monitoring treatment response, and assessing recurrence (50). Similarly, serum CA-125 in RMC offers potential as a non-invasive biomarker for longitudinal monitoring of disease burden. Unlike some of these established markers such as PSA in prostate cancer, CA-125 in RMC appears to be less universally elevated, reflecting the biological heterogeneity of this disease. This highlights the importance of combining serum biomarker data with tissue-based analyses to improve diagnostic and therapeutic precision in RMC. Notably, there are currently no published data on plasma circulating tumor DNA (ctDNA) in RMC. Given the aggressive nature of RMC and its limited biomarker landscape, prospective studies investigating the utility of ctDNA alone and in combination with CA-125 for disease detection, monitoring, and therapeutic response are urgently needed. Furthermore, while we observed lower MUC16 and CA-125 levels in the RMC2C1 cell line compared to other RMC cell lines, the reasons for this remain unclear. Future studies should investigate potential genetic or epigenetic differences in this cell line that may explain these observations.

Limitations of our analysis include the moderate sample size as well as potential selection and confounding biases which cannot be ruled out, given the retrospective nature of certain analyses and the focus on patients from a single institution. Furthermore, while the observed associations and findings provide insights into the biology of RMC, they are hypothesis-generating and require external validation in independent cohorts and diverse settings. There is also a potential concern that the “adjacent normal” tissue used in our study may harbor pre-malignant biological changes, although histologic normality was confirmed by an expert pathologist. In addition, the comparison of MUC16+ and MUC16− RMC cell lines in Figure 5 provides preliminary insights but is limited by the small number of cell lines analyzed, making it challenging to attribute observed differences solely to MUC16 status. These limitations underscore the need for larger, multicenter studies and additional experimental models to confirm and extend our conclusions. Despite these challenges, our findings represent an important step forward in understanding and addressing the unmet needs in RMC research and management.

In conclusion, our study has several clinical implications. The identification of serum CA-125 as a potential biomarker for RMC could lead to the development of non-invasive strategies for early detection and monitoring of disease progression, which is crucial for improving patient outcomes in this deadly disease. The significant correlation between serum CA-125 levels and metastatic burden in patients with RMC highlights the potential of CA-125 as a reliable biomarker for disease monitoring and therapeutic response. Furthermore, the therapeutic targeting of MUC16 could provide a novel approach for treating RMC, particularly for patients who do not respond to conventional therapies. Future studies should explore the efficacy of MUC16-targeted therapies in preclinical models and clinical trials.

TRANSLATIONAL RELEVANCE.

Renal medullary carcinoma (RMC) is a highly aggressive and therapy-resistant malignancy characterized by loss of the SMARCB1 tumor suppressor and mainly affecting young individuals of African descent with sickle cell trait. We comprehensively profiled gene expression and histone chromatin modifications in RMC tissues and identified MUC16 and its cleaved serum product, CA-125, as significantly upregulated providing a potential biomarker for disease monitoring. This was prospectively validated in an independent cohort of patients with RMC, where elevated serum CA-125 levels were found to correlate with metastatic burden, suggesting the utility of this biomarker for non-invasive monitoring of disease progression. Functional studies indicated that SMARCB1 expression was inversely associated with MUC16 expression. These findings support the development of CA-125 as a biomarker and the exploration of MUC16-targeted therapies for RMC and other SMARCB1-deficient malignancies.

Abbreviations list:

AFP

Alpha fetoprotein

ATCC

American Type Culture Collection

ATGC

Advanced Technology Genomics Core

CA-125

Cancer antigen 125

CA 15–3

Cancer antigen 15–3

CA-19.9

Carbohydrate antigen 19–9

CDC

Collecting duct carcinoma

CEA

Carcinoembryonic antigen

ChIP-seq

Chromatin Immunoprecipitation Sequencing

CLIA

Clinical laboratory improvement amendments

CTCF

Corrected total cell fluorescence

ctDNA

Circulating tumor DNA

DAPI

4′,6-diamidino-2-phenylindole

DEG

Differentially expressed gene

DEP

Differentially enriched peak

ELISA

Enzyme-linked immunosorbent assay

EnhAtl

Enhancer Atlas database

FDR

False discovery rate

GOBP

Gene Ontology Biological Processes

HRP

Horseradish peroxidase

IF

Immunofluorescence

IGV

Integrative genome viewer

IHC

Immunohistochemistry

LDH

lactate dehydrogenase

MDACC

MD Anderson Cancer Center

MSigDB

Molecular Signature Database

ORA

Over-representation analysis

PFA

Paraformaldehyde

PSA

Prostate-specific antigen

PVDF

Polyvinyl difluoride

RCC

Renal cell carcinoma

RMC

Renal medullary carcinoma

RNA-seq

RNA-sequencing

RTU

Ready to use

SRA

Sequence read archive

STR

Short tandem repeat

TBST

Tris-buffered saline [50 mM tris-HCl (pH 7.6) and 150 mM NaCl] containing 0.5% Tween 20

TCGA

The cancer genome atlas

UTUC

Upper tract urothelial carcinoma

WB

Western blot