Introduction

Urothelial carcinoma (UC) represents the predominant histological subtype of malignancies arising from the renal pelvis, ureter, and bladder. UC is anatomically classified into upper urinary tract UC (UTUC) and bladder UC (BUC). While BUC constitutes over 90% of UC cases, UTUC is relatively rare, accounting for 5–10% of diagnoses (Siegel et al. 2024). Up to 65% of UTUC patients present with muscle-invasive disease at diagnosis, compared to 15–25% of BUC patients (Rouprêt et al. 2023). Prognosis remains suboptimal, as approximately 50% of localized UC patients experience disease relapse or recurrence (Bianchi et al. 2023). Despite cisplatin-based chemotherapy remaining the first-line standard for advanced UC, 40% of patients are cisplatin-ineligible due to comorbidities, and most metastatic UC (mUC) cases develop therapeutic resistance (Powles et al. 2020). Immunotherapy has garnered substantial interest in UC, with pivotal trials demonstrating durable responses to immune checkpoint inhibitors (ICIs) as monotherapy or in combination (Necchi et al. 2018; Meeks et al. 2023; Alfred Witjes et al. 2024; Grivas et al. 2023; van der Heijden et al. 2023). Biomarkers of immune evasion and ICI responsiveness are under active investigation to optimize prognostic stratification and therapeutic algorithms (Hui et al. 2023; Liang et al. 2022).

Microsatellite instability (MSI) results from defective DNA mismatch repair (dMMR), manifested by loss of expression in ≥ 1 MMR proteins (MLH1, MSH2, MSH6, PMS2). MSI/dMMR status has been established as a significant biomarker of responses to immunotherapy (Wilbur et al. 2024). The European Association of Urology (EAU) guidelines on UTUC have updated the criteria for selecting UTUC patients for MSI testing (Rouprêt et al. 2023). Reported MSI prevalence varies widely in UTUC (1.6–28.1%) compared to BUC (0.7–1.5%), likely reflecting heterogeneous study populations and testing methodologies (Schneider et al. 2020; Kullmann et al. 2023; Gayhart et al. 2020; Hartmann et al. 2002; Catto et al. 2003; Ericson et al. 2005; Necchi et al. 2021; Audenet et al. 2019; Peak et al. 2023; Gerald et al. 2023).

Nectin-4 has emerged as a highly promising biomarker and therapeutic target in UC. Enfortumab vedotin (EV), an antibody–drug conjugate (ADC) targeting Nectin-4, has demonstrated clinically significant and durable responses (Challita-Eid et al. 2016; Rosenberg et al. 2020). In cisplatin-ineligible patients with locally advanced or metastatic UC, first-line EV combined with pembrolizumab showed enhanced clinical efficacy compared to EV monotherapy (O’Donnell et al. 2023). The combined biomarker profile (dMMR/Nectin-4) warrants further investigation as a predictive tool for therapeutic strategies.

This study aims to define the prevalence of dMMR in UC and correlate MSI status with clinicopathologic features and Nectin-4 expression.

Methods

Patients and materials

A consecutively collected cohort of 339 patients with histologically confirmed UTUC (renal pelvis or ureter) diagnosed between December 2016 and September 2023 at Peking University Third Hospital was retrospectively analyzed. Inclusion criteria: (1) Pathologically confirmed UTUC; (2) Sufficient preserved tumor tissue for immunohistochemistry (IHC); and (3) Complete clinicopathological data. Exclusion criteria: (1) Prior history of BUC.

Immunohistochemistry and evaluation

Formalin-fixed, paraffin-embedded (FFPE) tumor specimens were sectioned into 4-μm slides. After deparaffinization with xylene and rehydration with graded ethanol, endogenous peroxidase activity was blocked with 3% H2O2 for 10 min. Heat-induced epitope retrieval (HIER) was performed using a pressure cooker with EDTA buffer (pH 9.0). Slides were incubated overnight at 4 °C with the following primary antibodies: MSH2 (ZSGB-BIO, ZA-0622, 1:100), MSH6 (ZSGB-BIO, ZA-0541, 1:100), MLH1 (ZSGB-BIO, ZM-0154, 1:25), and PMS2 (ZSGB-BIO, ZA-0542, 1:20), Nectin-4 (Abcam, ab192033, 1:800) at 4 °C overnight. After incubation with secondary antibody (ZSGB-BIO, PV-6000) at 37 °C for 25 min, staining was visualized using 3,3′-diaminobenzidine (DAB; Gene Tech, GK600705) and counterstained with hematoxylin. Slides were scanned at 400 × magnification (Olympus, Japan), and two independent pathologists blinded to clinical data scored staining intensity and distribution.

dMMR: Defined as complete loss of nuclear staining in tumor cells for ≥ 1 MMR protein, with intact staining in internal controls (lymphocytes/stroma).

Nectin-4 expression was quantitatively assessed using the validated histochemical score (H-score) system. This scoring metric was calculated by multiplying the staining intensity (graded 0–3: 0 = negative, 1 = weak, 2 = moderate, 3 = strong) by the corresponding percentage of positively stained tumor cells (0–100%), followed by summation of these products to yield a final H-score range of 0–300. Rigorous validation included human placental tissue as a positive control and parallel sections with primary antibody omission as negative controls. Samples were then classified as negative (H-score 0–14), low (H-score 15–99), medium (H-score 100–199), and high (H-score 200–300), as previously described (Calandrella et al. 2022; Tekin et al. 2023).

Tumors were also categorized into three immune phenotypes as follows: inflamed phenotype, excluded phenotype and desert phenotype (Takahara et al. 2021).

Statistical analysis

Continuous variables are presented as mean ± standard deviation (SD). Categorical variables are summarized as frequencies and percentages. Group comparisons were performed using: Student’s t-test for normally distributed continuous variables, Mann–Whitney U test for non-normally distributed continuous variables, and Chi-square or Fisher’s exact test for categorical variables, as appropriate. Statistical analysis was performed using SPSS version 24.0 (SPSS, Inc.). Significance was defined as P < 0.05.

Results

A total of 339 consecutive, treatment-naïve UTUC tumor samples were included in this study. Clinicopathologic characteristics of the cohort are summarized in Table 1. 25 patients (7.4%) exhibited dMMR. The patterns of MMR proteins IHC expression are shown in Fig. 1.

Table 1 Baseline clinicopathological characteristics of UTUC patients
Full size table
Fig. 1
figure 1figure 1

Immunohistochemical expression of MMR proteins in UTUC specimens. A The positive expression of MSH2. B The negative expression of MSH2. C The positive expression of MSH6. D The negative expression of MSH6. E The positive expression of MLH1. F The negative expression of MLH1. G The positive expression of PMS2. H The negative expression of PMS2. Note: MMR, mismatch repair; UTUC, upper tract urothelial carcinoma

Full size image

We then analyzed correlation between MMR status and clinicopathologic features of the UTUC patients. Analysis of associations between MMR status and clinicopathologic features is detailed in Table 2. The patterns of Nectin-4 IHC expression are shown in Fig. 2. Notably, dMMR status was significantly associated with medium/high Nectin-4 expression (P = 0.044). However, no significant differences were observed between dMMR and proficient mismatch repair (pMMR) groups in gender distribution, age, BMI (body mass index), tumor grade, or immunophenotypic profiles. MMR protein deficiency patterns are detailed in Table 3. The predominant deficiency pattern was concurrent MSH2/MSH6 loss (56%), followed by combined MLH1/PMS2 deficiency (24%). Furthermore, analysis of MMR protein deficiency correlations with clinicopathological characteristics (Table 4) revealed that MSH6-negative expression was significantly associated with moderate/high Nectin-4 expression (P = 0.013). Comparisons with previous studies are presented in Supplementary Table 1.

Table 2 Correlation between MMR status and clinicopathological features of the UTUC patients
Full size table
Fig. 2
figure 2

Immunohistochemical expression of Nectin4 in UTUC specimens. A Tissue with negative expression. B Tissue with weak expression. C The Tissue with moderate expression. D Tissue with strong expression. Note: UTUC, upper tract urothelial carcinoma

Full size image
Table 3 Loss patterns of MMR protein (s)
Full size table
Table 4 Correlation between MMR protein loss and clinicopathological features of the UTUC patients
Full size table

Discussion

Cisplatin-based chemotherapy continues to serve as the cornerstone of adjuvant therapy for UTUC. However, cisplatin eligibility mandates adequate baseline health status, yet a significant proportion of real-world patients are ineligible. Approximately 50% of patients are ineligible due to post-nephroureterectomy renal impairment, advanced age, or preexisting chronic kidney disease (Huang et al. 2022). Furthermore, among patients receiving cisplatin-based regimens, treatment-related adverse events (TRAEs) are frequently reported. In the POUT trial, Grade ≥ 3 acute treatment-emergent adverse events occurred in 50% of participants, with 33% reporting serious adverse events (Birtle et al. 2020).

The therapeutic landscape for UTUC has shifted toward systemic therapies. Immunotherapy has gained prominence both as an adjuvant treatment for high-risk UTUC and in the second-line management of metastatic UTUC (mUTUC) (Rouprêt et al. 2023). Although PD-L1 expression is a widely used biomarker for predicting response to immune checkpoint inhibitors (ICIs), durable clinical benefit is not guaranteed in PD-L1-high patients, and responses may occur even in PD-L1-negative cases.

The MSI phenotype represents a well-validated biomarker across diverse malignancies (Dudley et al. 2016; Naboush et al. 2017). MSI is strongly associated with response to immunotherapy. Across multiple tumor types, MSI-H tumors demonstrate superior therapeutic efficacy and responsiveness to adjuvant chemotherapy and immunotherapy compared to MSS counterparts (Kullmann et al. 2023; Arabi et al. 2009; Sepulveda et al. 2017; Marcus et al. 2019). MSI-H tumors are characterized by elevated expression of immune checkpoint proteins such as PD-1 and PD-L1, rendering them particularly susceptible to PD-1/PD-L1 blockade (Marcus et al. 2019; Sepulveda et al. 2017). However, research on MSI status in UTUC remains limited. Reported incidence rates of MSI in UTUC vary widely (1.6–28.1%), attributable to heterogeneous study populations and racial disparities. Although pembrolizumab has received tissue-agnostic approval for advanced MSI-H or dMMR malignancies, UTUC remains underrepresented in clinical trials, with urothelial carcinomas constituting only 2.1% of cases. In the phase III KEYNOTE-045 trial comparing pembrolizumab to second-line chemotherapy in recurrent advanced urothelial carcinoma, UTUC cases comprised only 14% (n = 75) of the cohort, and MSI status was not assessed. Comparative analysis with prior investigations is presented in Supplemental Table 1. To our knowledge, this represents the largest cohort to date evaluating MMR status in UTUC patients, including substantial representation from China. This investigation revealed a dMMR prevalence of 7.4% in the UTUC cohort. Reported dMMR prevalence ranges from 2.4 to 26.3% across studies (García-Tello et al. 2014; Ito et al. 2020). Previous investigations (García-Tello et al. 2014; Shang et al. 2022) with relatively large sample size (> 100 cases) primarily employed tissue microarray (TMA) platforms. However, MMR deficiency may be underestimated due to limited tissue sampling of TMA. While interpretation of MMR expression in non-tumorous cores may produce false-positive results with limited samples, our methodology utilized full-tissue sections for IHC analysis, minimizing interpretive errors inherent to restricted sampling. Our results indicated that concurrent MSH2/MSH6 loss was the predominant form of MMR protein loss, consistent with previous studies (Schneider et al. 2020; Gayhart et al. 2020; Calandrella et al. 2022; Shang et al. 2022).

Our analysis revealed that tumors with dMMR status exhibited significantly elevated Nectin-4 expression compared to pMMR cases. Whereas prior research (Calandrella et al. 2022) was limited to 27 patients, our cohort of 339 cases provides robust validation of the association between dMMR status and enhanced Nectin-4 expression. Notably, we contribute the discovery that MSH6 deficiency exhibited a statistically significant association with elevated Nectin-4 expression levels (P = 0.013), whereas MSH2 deficiency similarly showed elevated expression trends that did not reach statistical significance. Enfortumab vedotin (EV), a novel antibody–drug conjugate (ADC), targets Nectin-4-expressing cells to deliver a microtubule-disrupting agent, thereby inducing apoptosis (Challita-Eid et al. 2016). Given the high prevalence of Nectin-4 expression in UC, EV monotherapy elicited durable responses in metastatic UC (mUC) cohorts in the EV-101 trial (Rosenberg et al. 2020). The EV-301 trial analysis revealed that EV therapy conferred a clinically meaningful overall survival advantage compared to chemotherapy in patients with advanced, previously treated UC. Beyond monotherapy, combination regimens of EV with ICIs (e.g., pembrolizumab) are under investigation in UC. EV plus pembrolizumab yielded superior outcomes to platinum-based chemotherapy in treatment-naïve advanced UC (Powles et al. 2024). Cisplatin-ineligible patients with locally advanced or metastatic UC receiving first-line EV plus pembrolizumab exhibited enhanced clinical benefit compared to EV monotherapy (O’Donnell et al. 2023). Our findings identify a synergistic biomarker profile in dMMR UTUC, characterized by Nectin-4 overexpression and PD-L1 upregulation, which may potentiate EV/ICI combination efficacy. This approach could circumvent cisplatin resistance mechanisms while leveraging the ‘immune-rich’ dMMR microenvironment. To our knowledge, this represents the largest cohort establishing an association between dMMR status and Nectin-4 overexpression in UTUC. We propose a biomarker-driven trial design: dMMR status for ICI cohort selection and Nectin-4 expression to stratify potential EV responders. In dMMR/MSI-H UTUC, Nectin-4 expression screening could optimize EV therapy selection, particularly for cisplatin-ineligible patients. We also advocate for integrating dMMR screening into ongoing ADC trials to validate its predictive role. This dual biomarker strategy may strategically refine therapy selection for EV-based regimens, addressing unmet needs in cisplatin-ineligible UTUC.

Our study has several limitations. First, this was a retrospective, single-center study. Despite representing the largest East Asian cohort to date, prospective multicenter studies are required to validate our findings. Second, a substantial proportion of patients were lost to follow-up, precluding analysis of associations between dMMR status/Nectin-4 expression and clinical outcomes in UTUC. Future studies should integrate longitudinal clinical data to evaluate the prognostic and predictive utility of these biomarkers. Furthermore, our study assessed dMMR/Nectin-4 status in primary tumors only; metastatic lesions may exhibit discordant biomarker profiles. We did not compare dMMR /Nectin-4 expression status in primary lesions and metastatic lesions. The lack of genomic or transcriptomic data precludes exploration of causal links between dMMR and Nectin-4 regulation.

Conclusion

7.4% of UTUC patients exhibited dMMR status and patients showed different levels of Nectin-4 expression. We also indicated a significant correlation between dMMR status and Nectin-4 high expression in UTUC. The observed pattern of combination of these data may be tested as a potential predictive tool for the use of a combination of immune-drug conjugates and immune checkpoint inhibitors. The need for further studies in this area is warranted.