Nectin-4 is an immunoglobulin-like transmembrane protein important for cell-cell adhesion [8] and is almost ubiquitously expressed in UC, making it a desirable target. The intensity of nectin-4 expression can be characterized with the immunohistochemical (IHC) scoring (H-score) system, which is the sum of the staining intensity characterized by pathology (0–3) multiplied by the percentage of cells (0–100) stained at the given intensity, resulting in a range of 0–300. In an IHC analysis of a wide range of solid tumor specimens, there were 524 bladder cancer specimens analyzed for nectin-4 expression, with 31% having strong intensity (H-score 200–300), 29% having moderate intensity (H-score 100–199), 23% with weak intensity (H-score 15–99), and only 17% of samples staining negative (H-score 0–14) [9]. Similarly, nectin-4 expression for 99 samples of upper tract UC revealed 10.1% with strong intensity, 24.2% with moderate intensity, and 31.3% with weak intensity, with 34.3% of samples with negative staining [10]. In the EV-101 phase 1 trial, which for the first time evaluated the safety, tolerability, and antitumor activity of EV for patients with aUC, a documented nectin-4 H-score was initially required prior to enrollment [11]. The median H-score for the 152 tested tumor samples was 290, with only five samples having an H-score < 150 including one sample with H-score 0. Because of the overwhelming positivity, nectin-4 scoring was not required for the EV trials afterwards.

The surface expression of nectin-4 may differ between primary UC and metastatic sites, potentially driving resistance. In a recent analysis, 137 patient samples of primary tumor with matched synchronous or metachronous metastases were stained [12]. For primary tumor specimens, median H-score was 110 (interquartile range 25–200), including 19.7% negative samples. For corresponding metastatic specimens, median H-score was 40 (interquartile range 0–140) with 39.4% negative samples. Decrease of nectin-4 expression with metastatic spread was statistically significant (P < 0.001). In this analysis, 47 patients received EV, and absent or weak nectin-4 staining was associated with significantly worse progression-free survival (PFS) compared to moderate/strong staining. High nectin-4 expression also significantly correlated with radiographic partial or complete response.

Therefore, nectin-4 expression is a potential predictive marker of response to EV. Tumors with absent or weak expression of nectin-4 in primary or metastatic tissue may be associated with primary refractory disease, and downregulation of nectin-4 expression with treatment could mediate secondary resistance. Because EV binds selectively to nectin-4, EV intuitively may bind tumor cell surfaces that overexpress nectin-4 more effectively compared to cell surfaces that underexpress or do not express nectin-4.

This principle was evaluated in an exploratory analysis of nectin-4 expression using immunohistochemistry assay of primary or metastatic tissue from patients enrolled in EV-302 [13]. For 800 available samples, the median H-score was 275, with a median of 280 (interquartile range 230–298) specifically for the EVP arm. In the EVP arm, only 9.6% of samples exhibited H-score < 150. Patients on the EVP arm with H-score < 275 had median PFS (mPFS) of 10.4 months (95% CI, 8.2–14.5) and mOS of 26.1 months (95% CI, 22.3–NR), compared to mPFS of 15.3 months and mOS of NR (95% CI, 25.6–NR) for patients with H-score ≥ 275. Objective response rates were 52.6% for H-score < 150, 58.0% for H-score ≥ 150 to < 225, and 71.8% for H-score ≥ 225. An exploratory analysis evaluating patients with available nectin-4 expression levels (H-score ≥ 275 or < 275) and programmed cell death protein ligand 1 (PD-L1) expression [based on the combined positive score (CPS), with < 10 considered low and ≥ 10 considered high] is summarized in Table 1. All subsets were significantly better for EVP compared to the PBC control arm (mPFS of 6.2–6.4 months with little variation), regardless of nectin-4 expression and PD-L1 status. This analysis did suggest there were better outcomes for patients on EVP with higher levels of nectin-4 expression.

With exposure to EV, secondary resistance may occur with downregulation of nectin-4 expression. In a case report of one patient who developed resistance to EV, IHC analysis from metastasectomy of two enlarging oligometastatic disease sites in the lung and adrenal gland were analyzed. There was decreased expression of nectin-4 in these metastatic sites compared to their expression on the primary tumor [14]. Nectin-4 scoring over time in the development of resistance to EV from large cohorts is lacking, but there is data from other approved ADCs regarding downregulation of the surface antigen. In the DAISY trial of patients who received trastuzumab deruxtecan (T-DXd), an ADC that binds to HER2, an analysis of 20 patients who developed secondary resistance showed 13 patients had a decrease in HER2 immunohistochemistry expression [15].

UC tumors of subtype histologies have generally been excluded from clinical trials. These tumors tend to be more aggressive and have a worse prognosis [16]. There has been recent evidence that nectin-4 expression is lower in these histologies, which likely has important implications for the effectiveness of EV.

One study characterized nectin-4 expression using IHC from radical cystectomy samples among the three most frequent variant subtypes: squamous cell carcinoma, adenocarcinoma, and sarcomatoid UC [17]. The median H-score of nectin-4 was the lowest for sarcomatoid UC (10, range 0–185), compared to squamous cell carcinoma (150, range 0–250), and adenocarcinoma (140.5, range 30–275).

Another study evaluated nectin-4 expression of 169 patients on several different histologic subtypes [18]. Subtypes with low rates of nectin-4 positivity included sarcomatoid carcinoma (10% of available samples), small cell carcinoma (0%), and micropapillary tumors (28%), all aggressive subtypes. Therefore, low nectin-4 expression may reduce EV efficacy for tumors with a significant proportion of these rare subtypes.

In the Urothelial Cancer Network to Investigate Therapeutic Experiences (UNITE) database, the efficacy of EV monotherapy was assessed in a real-world setting for 566 patients with documented histology (366 with pure UC, 200 with either pure variant or component of variant histologies) [19]. Patients with predominant variant histology (> 50%) had a numerically lower observed response rate (ORR) compared to patients with pure UC, but there were no differences in mPFS or mOS. Notably, patients with pure variant histology had worse outcomes compared to patients with pure UC in terms of ORR (0% versus 52%, P < 0.001), mPFS (HR 1.9, 95% CI 1.06–3.39, P = 0.03), and mOS (HR 2.96, 95% CI 1.160–5.46, P < 0.001). These hypothesis-generating results suggest that majority or pure subtype histologies may be associated with primary refractory disease to EV. Moreover, there may be implications in terms of primary or secondary resistance for tumors with predominant urothelial histology and minority subtype histology.

MMAE, a synthetic derivative of dolastatin 10, is an antimitotic agent that inhibits the polymerization of tubulin, leading to microtubule dysregulation [20]. Although the mechanisms of resistance to MMAE have not been fully elucidated, they may be extrapolated from vinca alkaloids and taxanes, which are also disruptors of microtubules. Resistance to these agents is diverse and can include the upregulation of efflux pumps by the cancer cell, the upregulation of vaults which are cytoplasmic organelles that trap cytotoxic drug to prevent their ability to bind to its target, the upregulation of detoxification enzymes to break down the toxin, and the upregulation of tubulin synthesis or the development of mutations to minimize interaction with the toxin [21, 22]. If there is resistance to MMAE from EV, a subsequent vedotin-based treatment, such as disitamab vedotin (DV), would be less efficacious.

Cancer cells not only can directly become resistant to chemotherapy intracellularly, but studies have also shown they can actively pump out the intracellular toxins. ATP-binding cassette transporters such as multi-drug resistance 1 (MDR1), also known as P-glycoprotein, have been well characterized to pump toxins like MMAE out of cells. These transporters can be upregulated in chemoresistant cancer cells [23]. In a preclinical UC model using the RT112 cell line, secondary resistance to EV was noted to be associated with an increase in MDR1 gene expression and TGF-β gene expression, without much downregulation of nectin-4 [24]. In the case report of one patient who developed resistance to EV as described above, IHC analysis from metastasectomy of two enlarging oligometastatic disease sites in the lung and adrenal gland were analyzed. There was decreased expression of nectin-4 and increased expression of MDR1 in these metastatic sites compared to their expression on the primary tumor, respectively [14]. Like EV, brentuximab vedotin (BV), which is an ADC that binds to CD30, also uses MMAE as its toxin. The mechanisms of resistance for BV against MMAE include increased expression of MDR1, of which MMAE is a substrate that leads to its efflux out of cancer cells [25, 26].

The inhibition of these drug efflux pumps appears to be effective in resensitizing cancer cells to the original chemotherapy agent. The MDR1 protein inhibitor tariquidar was demonstrated in a murine model with ADC-resistant tumors to safely reverse the ADC-induced resistance [27]. Cyclosporine, which competitively inhibits MDR1 [28], has been evaluated in a phase 1 trial (NCT03013933) in combination with BV for patients with relapsed/refractory Hodgkin lymphoma. In this cohort already refractory to BV, the reported overall response rate was 75% [29]. Therefore, MDR1 inhibitors may be useful for patients whose tumors develop resistance to EV over time, which must be validated prospectively.

Another way of quantifying nectin-4 expression is characterization of the NECTIN4 gene using either DNA or mRNA expression, which may be a more objective method of measurement compared to the H-score. The NECTIN4 mRNA expression was measured within muscle-invasive UC specimens in one study [30]. This study found heterogeneous NECTIN4 expression between molecular subtypes, with significant expression in luminal subtypes and lower expression in basal, neuroendocrine-like, and stroma-rich subtypes. Moreover, knockdown of NECTIN4 in luminal cells led to EV resistance, while upregulation of NECTIN4 in basal cells led to increased EV sensitivity. This suggests the importance of NECTIN4 gene expression predicting treatment response to EV.

In clinical practice, from a cohort of 108 patients with aUC with available tumor tissue samples and treated with EV, NECTIN4 copy number variants were assessed via fluorescence in situ hybridization (FISH) [31]. NECTIN4 amplifications were detected in 26% (28/108) of samples. The NECTIN4-amplified tumors correlated with higher nectin-4 expression than NECTIN4 non-amplified tumors (median H-score 295 versus 90, interquartile range 235–300 versus 20–205). Among 27 available matched primary and metastatic samples, the NECTIN4 copy number variation was stable in 93% of samples, with 7 of 8 NECTIN4-amplified primary tumor tissue samples retaining the amplification in the matched metastatic sample. For the 28 patients with NECTIN4 amplification, the objective response rate on EV was 96%, compared to just 32% for the 80 patients without NECTIN4 amplification. In contrast to changes in nectin-4 H-score between primary and metastatic disease, low NECTIN4 amplification in the primary specimen can be potentially used to predict primary resistance in aUC. The importance of NECTIN4 amplification of the tumor tissue and its association with EV response should be further investigated. Additionally, there may be downregulation of NECTIN4 with EV exposure over time, so this may be a method to monitor secondary resistance.

Next-generation sequencing (NGS) to evaluate specific gene alterations associated with EV resistance has been assessed retrospectively. In the UNITE database, 124 of 488 patients (25.4%) had primary refractory disease, defined as progressive disease as best response to EV monotherapy [32]. For patients with available NGS data, it was noted that patients with primary refractory disease had a higher rate of CDKN2B alterations (19% versus 12%, P = 0.05). Another analysis of the UNITE database identified alterations in ERBB2 and KDM6A, as well as high tumor mutational burden, as biomarkers predictive of better overall survival (OS) in patients who received EV [33]. A small single-institution study of 29 patients with aUC receiving EV found TP53 and MDM2 alterations were predictive of higher ORR, longer PFS, and longer OS [34]. The significance of these genomic alterations for EV resistance would have to be assessed in prospective studies.

Genomic alterations have also been implicated in resistance to EVP. In analysis of 118 patients treated with EVP in the UNITE database, TP53 and KMT2D alterations were associated with worse outcomes on EVP [35]. In multivariate analysis including age, race, histology, clinical performance status, blood counts, and albumin as covariates, patients had worse PFS with KMT2D alterations (HR 2.2; 95% CI, 1.0–4.5; P = 0.04) and TP53 alterations (HR 2.3; 95% CI, 1.2–4.2; P < 0.01], and worse OS with TP53 alterations (HR 2.3; 95% CI, 1.0–5.4, P = 0.05]. How these gene alterations contribute to treatment resistance of EVP needs further investigation.