Overexpression of KMT9α Is Associated with Aggressive Basal-like Muscle-Invasive Bladder Cancer

Muscle-invasive bladder cancer (MIBC) is associated with limited response rates to systemic therapy leading to a significant risk of recurrence and death. A recently discovered histone methyltransferase KMT9, acts as an epigenetic regulator of carcinogenesis in different tumor entities. In this study, we investigated the presence and association of histological and molecular subtypes and their impact on the survival of KMT9α in MIBC. We performed an immunohistochemical (IHC) analysis of KMT9α in 135 MIBC patients undergoing radical cystectomy. KMT9α was significantly overexpressed in the nucleus in MIBC compared to normal urothelium and low-grade urothelial cancer. Using the HTG transcriptome panel, we assessed mRNA expression profiles to determine molecular subtypes and identify differentially expressed genes. Patients with higher nuclear and nucleolar KMT9α expression showed basal/squamous urothelial cancer characteristics confirmed by IHC and differentially upregulated KRT14 expression. We identified a subset of patients with nucleolar expression of KMT9α, which was associated with an increased risk of death in uni- and multivariate analyses (HR 2.28, 95%CI 1.28–4.03, p = 0.005). In conclusion, basal-like MIBC and the squamous histological subtype are associated with high nuclear KMT9α expression. The association with poor survival makes it a potential target for the treatment of bladder cancer.


Introduction
Bladder cancer (BCa) causes 213,000 deaths worldwide every year, and > 90% of cases are urothelial carcinoma [1]. Twenty-five percent of patients present with muscle-invasive bladder cancer (MIBC) at the time of diagnosis, which is associated with 5-year overall survival rates of approximately 50%. Despite radical surgical treatments and cisplatin-based chemotherapies, the recurrence rates are high [2]. The introduction of immune-checkpoint inhibitors in the treatment of MIBC and metastatic urothelial carcinoma has improved the survival and management of patients [3][4][5][6][7][8]. However, only a subset of patients responds to immunotherapies and/or are resistant to chemotherapies. This requires the development of novel biomarkers to stratify patients, the combination of therapies, and the exploration of further effective therapeutic approaches for advanced BCa. Lab) and KRT14 (Clone: Poly9060; 1:10000; BioLegend Inc., San Diego, CA, USA) were performed manually.
Quantitative and semiquantitative analysis of IHC was annotated by genitourinary pathologists. For KMT9α, we determined the percentage of nuclear-positive tumor cells. Nucleolar KMT9α positivity was observed when more than 10% of tumor cells showed clear nucleolar KMT9α expression. TMA cores with either an absence of representative tumor tissue or the presence of staining artifacts were excluded from the analysis.

RNA-Isolation and Molecular Subtype Calling
We took a 1mm punch from FFPE blocks of a representative tumor area with ≥ 50% tumor content, isolated the RNA with the "truXTRAC FFPE total NA Kit-Column" (Covaris, Woburn, MA, USA), and measured the RNA concentration using the QuantiFluor RNA System (Promega, Madison, WI, USA) according to the manufacturer's protocol. The mRNA expression of 19,398 mRNA targets was determined using the HTG Transcriptome Panel (HTG Molecular Diagnostics, Tuscon, AZ, USA) on an Illumina NextSeq 550 system (Illumina, San Diego, CA, USA), as described before [20]. Gene counts were normalized using median normalization and log2-transformed for further analysis. Molecular subtypes of MIBC were determined using the R-based consensus MIBC classification tool and the Bioconductor package for R [24]. Differentially expressed genes were identified using DESeq2 (version 1.30.1) [25]. p value < 0.001 and expression fold-change > 2 or < (− 2) were considered statistically significant.

Statistical Analysis
For the survival analysis, we included patients with radical cystectomy. Patients with distant metastases or neoadjuvant chemotherapy were excluded from the analyses. We primarily assessed the overall survival (OS) as the endpoint, which was defined as the time interval between surgery and death.
The Kaplan-Meier method was used to estimate and illustrate survival probabilities. We used uni-and multivariable Cox's proportional hazards models to estimate the hazard ratio (HR) and the corresponding 95% confidence interval for the survival. All tests were two-tailed; we used a significance level of α = 5%. Statistical analyses were performed using JMP (SAS Institute Inc., Cary, NC, USA) version 16.2.0 and R Studio (version 2022.02.3).

Patient Characteristics
Overall, we included specimens from 135 patients with MIBC [20]. The median age was 70 years (IQR 59-76). In total, 105 (78%) patients were male, and 30 (22%) were female. The clinicopathological details of the cohort are summarized in Tables 1 and 2, stratified for patients with and without nucleolar expression of KMT9α and nuclear KMT9α < 5%, 5-15%, and ≥ 15%, respectively. The squamous histological subtype was significantly associated with nucleolar and high nuclear KMT9α expression, whereas the micropapillary histological subtype showed lower KMT9α expression values. Regarding the molecular subtypes, we noticed a significant association with the basal squamous subtype for nucleolar KMT9α expression. For higher nuclear KMT9α expression, such a trend towards the basal squamous subtype was also observed (

Patient Characteristics
Overall, we included specimens from 135 patients with MIBC [20]. The median age was 70 years (IQR 59-76). In total, 105 (78%) patients were male, and 30 (22%) were female. The clinicopathological details of the cohort are summarized in Tables 1 and 2, stratified for patients with and without nucleolar expression of KMT9α and nuclear KMT9α < 5%, 5-15%, and ≥ 15%, respectively. The squamous histological subtype was significantly associated with nucleolar and high nuclear KMT9α expression, whereas the micropapillary histological subtype showed lower KMT9α expression values. Regarding the molecular subtypes, we noticed a significant association with the basal squamous subtype for nucleolar KMT9α expression. For higher nuclear KMT9α expression, such a trend towards the basal squamous subtype was also observed ( Table 2).

Survival Analysis
We assessed the survival rates of 135 patients with adequate follow-up that received radical cystectomy in curative intent (cM0) without neoadjuvant chemotherapy and had available data for KMT9α expression. The median follow-up was 51 months (IQR 19-97 months). In this group, 40 patients received at least two cycles of adjuvant chemotherapy. The 12-month OS and disease-free survival (DFS) rates were 60% (95% CI 53-69) and 50% (95% CI 42-60), respectively.
Known predictive factors, such as tumor and lymph node stage as well as the application of adjuvant chemotherapy, were significantly associated with OS (Table 3 and Figure 2). In the total cohort, the nuclear expression of KMT9α was not associated with overall survival. After stratification for patients receiving only the cystectomy vs. patients receiving adjuvant chemotherapy, patients with a nuclear expression of KMT9α showed a significantly lower risk of death with adjuvant chemotherapy (HR 0.34, 95% CI 0.13-0.88, p = 0.03).
Patients showing the nucleolar presence of KMT9α (n = 23) had a significantly increased risk of death in the multivariate Cox regression model adjusting for tumor stage, lymph node status, and the application of adjuvant chemotherapy (Figure 2

Discussion
In this study, we investigated the expression and significance of KMT9, a novel H4K12me1 histone methyltransferase, in a cohort of patients with muscle-invasive bladder cancer. High nuclear expression of KMT9α was found in a significant number of cases and was associated with basal squamous and neuroendocrine histological and molecular subtypes. Nucleolar KMT9α expression was an independent predictor of poor prognosis.
These findings are in line with previously published data regarding KMT9 in other malignancies. For lung cancer, the histone methyltransferase KMT9 was shown to regulate the proliferation and survival of small-cell lung cancer and NSCLC cells [17]. Similarly, Metzger et al. described KMT9α to regulate prostate tumor proliferation [16]. Furthermore, Berlin et al. proposed KMT9 as an important regulator of colorectal carcinogenesis using mouse models and human tumor organoids [18]. Nucleolar KMT9α expression was associated with decreased survival, especially in patients not receiving adjuvant chemotherapy ( Figure S2). The nucleolus is the location of ribosomal biogenesis, which is often increased in cancer cells [33,34]. However, our data did not allow conclusions regarding the functions of KMT9 in the nucleolus.
Higher nuclear KMT9α expression and more KMT9α nucleolar positive cases were observed in patients with abnormal p53 expression, which is a surrogate marker for mutations in the TP53 gene [23]. TP53 mutations are considered a hallmark of highgrade urothelial cancer and have been described to occur in basal squamous and neuronal molecular subtypes in 61% and 94%, respectively, which is in line with our data ( Figure S6) [12,24,35,36]. P53 is a tumor-suppressor and key mediator of stress signaling; thus, it is closely related to ribosomal biogenesis and nucleolar functions [33,34]. So far, the correlation between mutated TP53 and nucleolar KMT9α expression was observed on a case level, but higher resolutions and functional analyses are required to uncover interactions between KMT9 and p53 on a single cell level.
Our results show that the expression of KMT9α is associated with basal gene and protein expression, leading to a squamous phenotype in many cases. This also translates into a survival detriment in patients with nucleolar KMT9α expression ( Figure 2C). Interestingly, the neuroendocrine histological and molecular subtype also shows higher nuclear KMT9α expression but no nucleolar KMT9α. Patients with molecular basal and neuronal MIBC have also been shown to have the worst prognosis in other studies [12,24,37].
On the other hand, we found that patients with high nuclear KMT9α expression had a survival benefit when treated with adjuvant chemotherapy. The data to predict response to (mainly neoadjuvant) chemotherapy based on histological or molecular bladder cancer subtypes is controversial, and so it cannot be concluded whether a basal-molecular subtype of MIBC responds well to chemotherapy [37][38][39][40]. Biomarkers to stratify MIBC patients are required to select patients for the application of perioperative chemotherapy. However, KMT9α will need prospective validation as a predictive marker in a larger cohort.
In addition to their importance as predictive and prognostic markers, novel therapeutic strategies are required as sequenced or combined approaches to improve the survival rates of BCa patients. The high presence of alterations in genes involved in histone modifications and preliminary data make histone methyltransferases a promising target [11,15,41]. As an example, a phase 1/2 trial is testing Tazemetostat, an EZH2 inhibitor, in combination with the immune-checkpoint inhibitor pembrolizumab in patients with advanced urothelial carcinoma (https://clinicaltrials.gov: NCT03854474, accessed on 8 January 2023). Results regarding KMT9α generated in vitro, in xenograft tumors, and in organoids [16,18], together with the overexpression in aggressive MIBC, must be translated into an applicable drug to inhibit KMT9α that can be explored as a potential therapeutic approach.
The limitation of our study is the retrospective design, leading to a selection bias in patients receiving chemotherapy. Furthermore, the number of samples included, especially in the molecular analysis, is rather low. Thus, findings need to be reproduced in larger datasets. Our analysis relies on histology, IHC, gene expression, and clinical data and delivers important insights into the expression of KMT9α in a well-characterized cohort of MIBC patients. However, functional analyses explaining the observed effects in bladder cancer are pending and are important to further explore KMT9 as a therapeutic target for BCa.

Conclusions
In conclusion, KMT9α is highly expressed in bladder cancer cells with aggressive basal and neuroendocrine phenotypes. This needs to be translated into further exploration of KMT9α expression as a predictive marker for chemotherapy response in MIBC. Furthermore, as a histone, methyltransferase KMT9α is a promising target, and the effect of its inhibition should be explored in vitro and in vivo to treat bladder cancer.