Immune checkpoint inhibitors (ICIs) have represented an innovation in the treatment of several malignancies since the approval of ipilimumab in 2011 (1). These are monoclonal antibodies targeting the cytotoxic T-lymphocyte antigen-4 (CTLA-4) or programmed cell death-1/ligand 1 (PD-1/PD-L1) which are involved in the negative regulation of T-cell immune function. This binding causes the failure of the tumoral evasion mechanisms, and, consequently, an increased triggering of the immune system against cancer. However, this immune activation is non-specific and it can affect many different organ systems leading to the so-called immune-related adverse events (irAEs) in up to 70% of treated patients, such as pneumonitis, colitis, endocrinopathies, and myocarditis (2).

The skin is the most involved organ being affected in approximately 30% of patients treated with anti PD-(L)1 and 50% with anti CTLA-4 drug, respectively (3, 4). Furthermore, patients who develop a cutaneous irAE (cirAE) need close monitoring for signs or symptoms of extracutaneous ones as it may be a predictive factor (5). It has been proposed to divide the cirAEs into four histopathological categories which are inflammatory, immunobullous, keratinocyte changes and melanocyte changes. The inflammatory ones are the most frequent and appear as maculopapular, lichenoid, psoriatic and eczematous eruptions (6). Even though the development of cirAEs has been associated with an increased survival and tumor response (7–9) their prognostic significance remains unclear (10).

Immunobullous eruptions have been increasingly reported in the literature mostly linked to anti PD-1/PD-L1 drugs. The estimated overall incidence varies from 1 to 5% (11, 12), with bullous pemphigoid (BP) being the most commonly observed phenotype. Lichen planus pemphigoides (LPP), pemphigus vulgaris (PV), and mucous membrane pemphigoid (MMP) have also been described in association with ICIs, albeit uncommonly. Bullous irAEs represent a therapeutic challenge for clinicians because they might result in significant morbidity and mortality if untreated. Moreover, immunosuppressive systemic therapies and/or ICI interruption are often required to reach adequate control of the cutaneous involvement resulting in a worsening of the cancer prognosis (11).

Herein, we reported our Italian case series about ICI-induced autoimmune blistering disorders, especially ICI-BP. In addition, we conducted a comprehensive review of bullous cirAEs to compare our data with the cases already described and to better understand what is known so far on this issue.

This is a case series based on a national multicenter, retrospective, observational cohort including all patients treated with ICIs and developed an immunobullous cirAE during treatment or up to 12 months after discontinuation. Data were collected from 14 Italian hospitals between September 2021 and February 2023, after the institutional review board approval obtained from the ethic committee of the Turin University hospital. Information reported included patient demographics (age, sex), oncology history, ICI therapy, cirAE presentation and severity, histopathological findings, direct and indirect immunofluorescence (DIF/IIF), antibodies detected by enzyme-linked immunosorbent assay (ELISA), cirAE treatment, and tumor outcome. No patient identifiable data were used. The immunobullous cirAE diagnosis had to be supported by at least one positive test including histopathological examination, DIF, IIF or ELISA. Common Terminology Criteria for Adverse Events (CTCAE) ver. 5.0 was used to identify skin toxicity severity. Quantitative values were expressed as the median value and range.

Moreover, we performed a comprehensive review of the English-language medical literature about immunobullous cirAE. We used the databases PubMed, Embase, Scopus and Web of Science. Search strategy identified articles with the terms “bullous pemphigoid,” “lichen planus pemphigoides,” “pemphigus,” and “mucous membrane pemphigoid” combined with “cancer immunotherapy,” “immune checkpoint inhibitors,” “nivolumab,” “pembrolizumab,” “cemiplimab,” “ipilimumab,” “avelumab,” “atezolizumab,” “durvalumab.” The search involved all fields including title, abstract, keywords, and full text. Articles without available full text or with limited and inconsistent data from individual patients were excluded. We considered papers published by January 2023.

Table 2 summarizes the characteristics of the patients reported in our case series. These are 45 cases, 41 males (91%) and 4 females (9%), with a median age of 74 years (range 46–90). In all cases, they developed an ICI-induced BP while they were receiving the cancer treatment. The ICI identified were mainly nivolumab [28 of 45 (62%)] and pembrolizumab [11 of 45 (25%)], in the remaining cases combination ipilimumab with nivolumab, cemiplimab, spartalizumab, and atezolizumab were reported [6 of 45 (13%)]. They were used for the treatment of non-small-cell lung cancer (NSCLC) [18 of 45 (40%)], melanoma [12 of 45 (27%)], colorectal adenocarcinoma [5 of 45 (11%)], renal clear cell carcinoma [5 of 45 (11%)], head and neck squamous cell carcinoma (HNSCC) [4 of 45 (9%)], urothelial carcinoma [1 of 45 (2%)], and in 73% of cases they were stage IV (33 of 45). The median time to onset of cutaneous symptoms after ICI initiation was 35 weeks (range 4–260), while the median time to BP diagnosis was 48 weeks (range 5–286). In 19 patients (42%) pruritus without any cutaneous eruption was the first clinical manifestation, while bullous lesions appeared since the beginning in 9 patients (20%). The mucosal involvement was reported in 8 patients (18%). According to CTCAE grading for bullous dermatitis, BP affected more than 30% of the body surface in 40% of patients (G3 and G4 in 18 of 45 cases). Skin biopsy for histopathological examination was performed in 28 patients (62%) and it confirmed the diagnosis in all cases, while DIF and IIF were carried out in 31 (69%) and 26 (58%) patients, respectively, showing positive results in all cases. The ELISA for BP180 autoantibody was performed in 42 cases and was positive in 30 patients (66%), while the ELISA for BP230 autoantibody was carried out in 41 cases and was positive in 15 patients (33%). The first-line therapy of the cutaneous toxicity was topical steroids in 4 patients (9%), topical and systemic steroids in 41 (91%). In 18% of cases (n = 8) a second-line therapy was needed such as doxycycline (n = 5) and dapsone (n = 3). One patient required a third-line therapy with dupilumab. Immunotherapy was permanently discontinued for 17 patients (38%), while it was temporarily held for 16 patients (36%) of which about 50% (n = 7) experienced a relapse after rechallenging with the same ICI. The median time between the BP diagnosis and the control of symptoms was 9 weeks (range 1–68) in the 38 patients (84%) with a partial or complete response. In the remaining 16% (n = 7) the ICI-BP was refractory to the treatment. Tumor response of the 36 cases with available data revealed that 9 patients (20%) had a complete or partial response, 16 (36%) had stable disease, and 11 (24%) had progression disease.

Bullous autoimmune dermatoses are an uncommon cirAE whose prevalence is challenging to determine because it varies among several studies. Nevertheless, it is well known to be more infrequent than inflammatory eruptions or vitiligo (6, 92) since its reported incidence rates are less than 5% (11, 93). The exact underlying pathogenesis of bullous cirAEs has not fully understood but it potentially involves both the innate and adaptive immune systems.

In the tumor microenvironment, anti PD-1/PD-L1 antibodies enhance exhausted T effector cells function leading to immune activation against cancer. The lysis of tumoral cells releases numerous antigens whose presentation by antigen presenting cells causes the abnormal priming of both cytotoxic T lymphocytes and T helper 1 (Th1) cells. These mechanisms interfere with the immune tolerance resulting in the attack even against self-tissues (6). It is not surprising that lichenoid skin reaction is one of the most common cirAEs because this Th1-polarized reaction can cause intense interface dermatitis. It has been proposed that the lichenoid inflammation might expose antigens in the basal layer, making them targets for antibody development. In fact, research has demonstrated a role for T-cell trigger in enhancing the humoral response as well as a T-cell independent PD-1+ B-cell activation resulting in an aberrant antibodies production (94, 95). This theory could also explain why BP is the most common ICI-induced bullous dermatosis. In fact, the hemidesmosomes (BP180 and BP230) at the DEJ are more exposed than the desmogleins or other intercellular adhesion molecules to antibodies formation following the interface damage. In addition, the substantial male predominance among patients affected by ICI-induced BP, unlike classic BP, it could be associated with some sex-associated molecular differences. Indeed, it has been reported a higher tumor mutational burden and the presence of more immunogenic neoantigens in male patients with melanoma, so this may contribute to an increased incidence of irAE such as ICI-BP (96).

BP180 is normally expressed by undifferentiated keratinocytes of the basal layer, and it is ceased as they migrate upwards and differentiate in the epidermis. For this reason, it is expressed in SCC as a hallmark of de-differentiation (97). Therefore, it can also be detected in cells of neural crest origin as in proliferating melanocytes with an undifferentiated phenotype. Sequencing of COL17A1 from melanoma cDNA has revealed a series of aberrations that cause the post-translational degradation of the ectodomain and so its deficiency (98). Consequently, the endodomain accumulates in tumor cells, which has been associated with an invasive phenotype. This aberrant expression of BP180 is an accessible target for in vitro immunotherapy (99). These findings suggest that targeting of BP180 on tumor cells by the activated immune system could lead to a cross-reactive immunogenicity against the DEJ and the development of BP (100). The “same-antigen theory” could also explain ICI-induced PNP in patients affected by SCC as it is a cancer arising from keratinocytes and expressing autoantigens routinely identified in PNP (91). Moreover, anti-BP180 auto-antibodies in the sera of melanoma patients have been reported to be significantly higher than in the sera of healthy, at both early and advanced stages of disease, and this correlates with a higher probability for these patients to develop BP during anti-PD1 therapy (101).

The immunotherapy could also unmask a genetic susceptibility to develop bullous autoimmune disorders activating the immune system and leading to the clinical expression. The HLA-DQB1*03:01 allele which has been associated to BP has been found in higher frequencies in melanoma patients (102).

Data collected in our case series have confirmed what has already been described in the literature on ICI-induced BP. Unlike classic-BP, it has a male predominance (103) and this has been assumed to be associated with gender effects on immunotherapy activity (96). Drug-induced pemphigoid, as well as ICI-BP, is characterized by a younger age of onset (104) compared to the classic type whose incidence increases significantly over 80 years (105). BP-like eruption has been reported most frequently in patients receiving anti PD-1/PD-L1 antibodies for melanoma and NSCLC. It is not a class effect of these drugs but their mechanism of action close to peripheral tissues may be related to an increased reactivity against cutaneous self-antigens (106). ICI-BP appears later than other cirAEs (17, 107) and it is often preceded by a longer prodromal phase than classic BP characterized by persisting pruritus and/or non-specific dermatitis. In fact, it has a significantly longer delay from symptom onset to diagnosis than the classic one despite having similar delays from symptom onset to dermatology referral (47). Even though ICI-BP seems to have some peculiar clinical features, this does not reflect significant differences regarding histopathologic and DIF findings (35). Given the moderate-to-severe clinical presentation (17) and delayed diagnosis, management of ICI-BP often necessitates discontinuation of immunotherapy and treatment with oral/intravenous corticosteroids to control the cutaneous toxicity. Although several studies have suggested an association between the development of ICI-BP and improved cancer outcomes (7, 46, 48), the heterogeneity of information collected in this paper does not allow for confirmation of this theory. Future studies should evaluate the best tumor response in patients with ICI-BP based on cancer types and treatment modalities.

Bullous autoimmune dermatoses have gained increasing interest among cirAEs induced by anti PD-1/PD-L1 autoantibodies. Among the published literature, ICI-induced BP is the most frequently described, while LPP, MMP and pemphigus are reported anecdotally. There are several theories which try to clarify their underlying pathogenesis without a complete success. It potentially involves both the innate and adaptive immune systems, the cross-reactivity immunogenicity, the genetic susceptibility, and other unknown factors. The clinical presentation of these bullous cirAEs can vary, posing a challenge for prompt recognition and appropriate treatment. Immunossuppressive therapy and/or discontinuation of immunotherapy are often necessary for management. Dermatological referral is necessary to establish as soon as possible an appropriate therapeutic algorithm to control the cutaneous toxicity avoiding a negative impact on cancer outcome.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

MM, SR, and PQ contributed to conception and design of the study and organized the database. MM and MAc contributed to literature search. MM, MR, AMa, GZ, FM, EA, RM, EC, AP, GG, PSo, CS, MC, LP, DF, RB, PSe, PV, MT, MAr, CV, AMi, RS, ED, and BM contributed to collection of the case series based on the national multicentre cohort. MM wrote the manuscript. MR, AMa, GZ, EA, EC, SR, and PQ contributed to manuscript revision. All authors read and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.