Expression of DNA mismatch repair proteins in melanoma patients treated with immune checkpoint inhibitors

Purpose To investigate the protein expression of DNA mismatch repair (MMR) proteins in patients with cutaneous melanoma (CM) under immune checkpoint inhibitor (ICI) therapy. Methods Immunohistochemistry was performed on tumor tissue for MMR proteins MLH1, MSH2, MSH6, and PMS2 in 50 metastatic CM patients treated with ICI (ipilimumab, nivolumab, pembrolizumab). Results Best overall response (BOR) rate was 48% (24/50). Reduced MMR protein expression (nuclear expression in < 80% of tumor cells) was observed in 8 patients (16%). Compared to other clinical parameters, baseline neutrophil/lymphocyte ratio and reduced intratumoral MMR protein expression (P = 0.0033) were determined as the only parameters significantly associated with favorable BOR. However, in this small study population, reduced MMR protein expression did not reach statistical significance in multivariate analysis. Conclusion Reduced MMR protein expression is observed in CM and might predict favorable BOR in patients treated with ICI, as was observed for other entities. However, these findings need to be substantiated in larger patient cohorts.


Introduction
The incidence of cutaneous melanoma (CM) is increasing worldwide, with more than 55,000 deaths per year attributed to this malignancy. Immune checkpoint inhibitors (ICI), specifically antibodies against programmed death protein 1 (PD-1, pembrolizumab, nivolumab) and cytotoxic T lymphocyte associated protein 4 (CTLA-4, ipilimumab) have become first-line treatment regimens for patients with advanced CM. However, approximately 50% of patients do not respond to ICI and it is still difficult to predict who will or will not respond. Hence, there is an urgent need for biomarkers predicting treatment outcome to ICI, particularly considering ICI-mediated adverse events and high costs (Whiteman et al. 2016;Bai et al. 2021;Schadendorf et al. 2018). For many other malignancies, such as colorectal, endometrial, prostate, and bladder cancers, deficiencies in the DNA mismatch repair (MMR) machinery, evident in microsatellite instability (MSI), carry prognostic significance. For example, the prognosis of most cancers with high-level MSI resulting from MMR deficiency is comparatively favorable 1 3 and indicates responsiveness to treatment with ICI (Breakstone 2021; Germano et al. 2021;Mandal et al. 2019;Cho et al. 2021). At least in part, this is thought to be due to an increased number of frameshift mutations in these tumors resulting in the presentation of neoantigens. In the present study, we aimed to determine protein expression of MMR components in patients with unresectable or metastatic CM and to evaluate whether protein expression is associated with response to ICI treatment.

Patients
We included patients with inoperable stage III and IV CM who received treatment with ICI. Therapy and staging procedures were performed in accordance with national guidelines for the management of CM and interdisciplinary tumor board decisions (Schadendorf et al. 2018). ICI, including mono-nivolumab, mono-pembrolizumab, mono-ipilimumab, and ipilimumab plus nivolumab were administered according to label (Marconcini et al. 2018). Complete staging was performed including lymph node ultrasound, thoracic and/ or abdominal computed tomography or positron emission tomography in combination with computer tomography, and cranial magnetic resonance tomography (Schadendorf et al. 2018). The criteria for best overall response (BOR) were used in accordance with RECIST 1.1 (Eisenhauer et al. 2009). Laboratory parameters assessed at the beginning of ICI treatment included full blood count (neutrophil/lymphocyte ratio), BRAF status, lactate dehydrogenase (LDH) levels, and serum S100B.

Microscopic evaluation
Microscopic evaluation was carried out using slides scanned at 20-fold magnification with the Nanozoomer Whole Slide Scanner (Hamamatsu, Herrsching am Ammersee, Germany). All scans were assessed by means of the NDP. view2 software (Hamamatsu Photonics, Hamamatsu City, Japan). Microscopic evaluation was performed as previously described (Gambichler et al. 2021a, b). Briefly, nuclear staining of each MMR protein was evaluated in all tumor cells in 5-10 field of views. Protein expression was expressed as the percentage of nuclear-stained tumor cells relative to all tumor cells assessed. Per definition, MMR deficiency was declared when there was complete absence of nuclear staining for at least one protein (Umar et al. 2004;Hashmi et al. 2017). Furthermore, cases with an MMR protein expression of less than 80% were classified as diminished MMR protein expression status.

Statistics
Data analysis was performed using the statistical package MedCalc Software version 20.014 (MedCalc Software, Ostend, Belgium). Where appropriate, univariate analysis was performed using the ROC curve analysis of continuous variables (e.g., NRL), Kaplan-Meier statistics, Spearman correlation procedure, and χ 2 test. Data analysed with a P value < 0.1 was included in the multivariate analysis. The latter was carried-out by means of logistic regression. P values < 0.05 were considered significant.
Taken together, nuclear staining of all four MMR proteins assessed (PMS2, MSH6, MLH1 and MSH2) was very high with median nuclear expression levels of over 95%. Consequently, expression of each MMR protein was strongly correlated with expression of the other ones (P < 0.001). Thus, none of the patients showed MMR-deficiency as defined by complete loss of nuclear expression. However, reduced MMR protein expression as defined by nuclear expression in less than 80% of tumor cells was observed in 8 patients (16%), of which 5 showed diminished MSH2 expression, 3 MLH1, and 1 reduced PMS2 expression (Fig. 1). In contrast to gender, age, S100B, LDH, BRAF-status, high-tumor thickness, and stage of disease, low neutrophil/lymphocyte ratio at baseline [median 3.1 (1-9.7), associated criterion ≤ 3.1, AUC 0.70)] and reduced MMR protein expression were determined as the only parameters significantly associated with BOR (P = 0.0092 and P = 0.0033, respectively). However, the latter two parameters did not significantly correlate with progression-free survival and CMSS (Fig. 2). Using multivariate logistic regression analysis, we found that low neutrophil/lymphocyte ratio remained in the model as significant predictor for BOR, whereas reduced MMR protein expression did not reach statistical significance (P = 0.0023 vs. P = 0.061, respectively).

Discussion
The American Society for Clinical Pathology and similar authorities strongly recommend the assessment of MSI/ MMR biomarkers in colorectal cancer for better prognostic stratification of patients. This recommendation is emphasized by recent evidence of MSI as a positive predictive factor for response to ICI (Vatrano et al. 2020). Hence, the Food and Drug Administration has recently approved pembrolizumab as first-line therapy for MSI-high/deficient MMR metastatic colorectal cancer (Zhu et al. 2021). MSIhigh/deficient MMR status is observed most frequently in colorectal (up to 15%) and gastric cancer (about 10%) (Zhu et al. 2021), and less frequently in hepatocellular carcinoma and esophageal and pancreatic adenocarcinoma (< 5%) (Zhu et al. 2021). Because of the dramatic response of MSIhigh/deficient MMR tumors to ICI, MSI/MMR testing has increased significantly in many solid malignancies.
By contrast, data on MSI/MMR in cutaneous malignancies are relatively scarce. In non-melanoma skin cancer, complete loss of MMR protein expression has been rarely observed, whereas decreased expression similar to the data presented here has been observed in cutaneous squamous cell carcinoma and Merkel cell carcinoma (Gambichler et al. 2021a, b). Moreover, MSI-high/deficient MMR status might also be relevant in CM and its responsiveness to ICI (Buder-Bakhaya and Hassel 2018; Kubeček and Kopecký 2016). Korabiowska et al. (2004) studied MMR proteins both by immunohistochemistry and in situ hybridization in 59 CMs.

3
In line with our data, they proposed that in CM, a decreased expression of MMR proteins, rather than a total loss, might be prognostic significance (Korabiowska et al. 2004). Alvino et al. (2014) also reported a reduction of MLH1, MSH2, and PMS2 protein expression in CM compared to melanocytic nevi. Notably, high MSH6 expression in CM was significantly associated with an increased risk of CM mortality, a finding not corroborated by our study. Roncati (2018) Wang et al. (2022) recently described a novel computational tool (CODEFACS) and a supporting immune interactions framework (LIRICS), enabling an averaged "virtual single-cell" characterization of the tumor microenvironment from bulk tumor expression data. They detected a shared repertoire of cell-type-specific ligand-receptor interactions unique to the tumor microenvironment of MMR-deficient malignancies. Wang et al. (2022) focused on CM, which is known for one of the best response rates to ICI, and where a wealth of publicly available bulk expression datasets exists of patients receiving ICI. Using machine learning techniques, they detected a subset of intercellular tumor Fig. 2 The figure shows the Kaplan-Meier curves of the progression-free survival (PFS, a) and the cutaneous melanomaspecific survival (CMSS, b)  1 3 microenvironment interactions that stratified survival outcomes of CM patients receiving ICI therapy better than some recently published transcriptomics-based methods. Thus, hypermutated CM with deficient MMR showed the best response to ICI treatment (Wang et al. 2022).
The main limitations of the present study include the small sample size and absence of testing for microsatellite instability using multiplex PCR. The data presented here support other findings that MMR deficiency, defined by complete loss of nuclear expression in tumor cells, appears to be a rare event, since it was not observed in this cohort of 50 CM patients. Similar to Merkel cell carcinoma (Gambichler et al. 2021b), however, we detected a decrease of single MMR proteins in 16% of cases, whereas diminished MMR expression was most frequently observed for MSH2. Importantly, all patients with reduced MMR protein expression responded to ICI. Specifically, decreased MMR protein expression was significantly associated with BOR rates in univariate analysis. In multivariate analysis, however, this parameter did not reach statistical significance, which was very likely due to the limited number of patients investigated. Thus, in contrast to Wang et al. (2022), we did not observe a correlation between MMR expression status and PFS or CMSS. In conclusion, reduced MMR protein expression is not uncommon in CM and might be a predictor for improved response to treatment with ICI. However, the present data have to be substantiated in larger trials.