Lichen Planopilaris Pemphigoides: A Novel Bullous Dermatosis Due to Programmed Cell Death Protein-1 Inhibitor Therapy

Abstract

Lichen planus pemphigoides (LPPemph), apart from bullous pemphigoid, is a rare bullous dermatosis that can be induced by programmed cell death protein-1 (PD-1)/PD ligand 1 (PD-L1) inhibitors. The primary location of PD-1/PD-L1 inhibitor–induced LPPemph has previously only been reported at the nonfollicular dermal–epidermal junction. We present a case of nivolumab-induced LPPemph with an intense perifollicular lichenoid reaction, prominent multifocal perifollicular clefting, which in addition, was also accompanied by linear IgG and C3 immunofluorescence deposits along the dermal–epidermal junction as well as demonstrating a perifollicular pattern. Intriguingly, the serological study of BP180 and BP230 antibodies was negative, suggesting the presence of additional novel antibodies, which primarily favor hair follicles and may contribute to the pathogenesis. Therefore, we consider this entity a novel variant of PD-1/PD-L1 inhibitor–induced bullous dermatosis. To the best of our knowledge, this is the first report that highlights perifollicular bullae accompanied by immunofluorescence findings in a PD-1/PD-L1 inhibitor–induced lesion. We propose a new immunotherapy associated entity, lichen planopilaris pemphigoides, and emphasize the significance of perifollicular changes in the pathogenesis.

INTRODUCTION

PD-1/PD-L1 inhibitors are monoclonal antibodies that amplify the immune response to counteract tumor immune-evasive mechanisms and are an increasingly administrated therapy for primary and metastatic malignancies. Although these immunomodulators have shown significant clinical efficacy in the treatment of several malignancies, upregulated immune response sometimes triggers immune-related adverse events (irAEs). A study reported 49% of patients treated with anti–PD-1 antibodies developed dermatologic toxicities, including vitiligo, maculopapular, lichenoid, and eczematous reactions.¹ Over the past several years, immunobullous reactions and Steven–Johnson syndrome (SJS) have emerged as additional cutaneous irAEs.^1–4

LPPemph is an immunobullous dermatosis histologically characterized by a combination of lichen planus (LP) and subepidermal bullae, which may resemble bullous pemphigoid (BP). However, the pathogenesis of LPPemph is due to circulating antibodies targeting an epitope (MCW-4) at the C-terminal NC16A domain of the BP180 kDa antigen, which is different from the targeted epitope of BP.^5,6 The exact prevalence of LPPemph is unknown, however, using the available data of the ICD 10 code of LPPemph (L43.1), the estimated prevalence is approximately 1 in 1,000,000 patients.⁷ Although extremely rare, LPPemph has been reported in association with PD-1/PD-L1 inhibitors. It is proposed that the blockage of PD-1/PD-L1 signaling could induce the imbalance between T helper cells and T regulatory cells, which further leads to the production of autoantibodies against BP180.^8,9 To the best of our knowledge, the subepidermal bullae in these cases has only been reported at the nonfollicular dermal–epidermal junction.^10–19 Furthermore, positive serological tests for BP180 and/or BP230 antibodies have been reported in all tested cases.^10–16 In this report, we present a case of a follicular immunobullous dermatosis with a negative serological test of BP180 and BP230 antibodies in a nivolumab-treated patient and further discuss the possible underlying pathogenesis.

CASE REPORT

A 60-year-old African American man with a medical history of transitional cell carcinoma of the bladder and hepatocellular carcinoma arising in the background of hepatitis C-induced liver cirrhosis presented to our clinic after being initiated on nivolumab therapy. The patient completed 13 cycles of immunotherapy over the course of 30 weeks with a positive treatment response indicated by decreased alpha fetal protein (AFP). However, for approximately 1 month, the patient had developed recurrent, targetoid plaques and erythematous to dusky cyanotic patches distributed on the right frontal scalp, arms, hands, and legs with progression to oral mucosal erosions and bullous lesions on the back, forearms, and bilateral palms. The oncologist discontinued nivolumab because of the possibility of immunotherapy-induced SJS. After 2 courses of steroids and 1 dose of infliximab, the rash disappeared, and the treatment with nivolumab was resumed.

The rash re-emerged 3 weeks later, nivolumab was again withheld, and the patient was referred to the dermatologist for further evaluation. The scalp showed raised, flaccid blisters and exfoliative skin with erythema, associated with a positive Nikolsky sign (Fig. 1). A punch biopsy of the right parietal scalp was performed. Histology revealed a lichenoid infiltrate with no clefting along the dermal–epidermal junction (Figs. 2A–D). There was, however, prominent multifocal perifollicular clefting associated with an intense inflammatory infiltrate and mild fibrosis (Figs. 3A–D). Direct immunofluorescence (DIF) revealed a linear pattern of IgG and C3 deposition at the dermal–epidermal junction and in a perifollicular distribution (Figs. 4A–D). The ELISA study of patient's serum was negative for autoimmune disease markers, including lupus erythematosus antibodies (ANA, RNP, anti-Sm, anti-DNA, antichromatin, anti-SSA/Ro and anti-SSB/La, and complement C4 and C3), pemphigus antibodies (desmoglein-1 and 3), and pemphigoid markers (BP 180 and BP 230 antibodies). The indirect immunofluorescence study was not performed.

FIGURE 1.: Clinical photography shows flaccid blisters and superficial erosion.

FIGURE 2.: H&E sections show a lichenoid interface dermatitis with prominent and multifocal perifollicular clefting (A). Medium power (100X) shows orthokeratotic scale, wedge-shaped hypergranulosis, saw tooth–like change of the dermal–epidermal junction (B) which shows tendency to involve the follicle (C). High-power (400X) examination reveals a florid interface process with numerous dyskeratotic keratinocytes and colloid bodies (D).

FIGURE 3.: The interface process involves the follicles circumferentially and leads to separation of the follicular epithelium from the dermis (A). In addition, there is an intense mainly perifollicular and perivascular infiltrate containing occasional eosinophils. Figures B–D shows higher-power (B and C: 100X, D: 400X) views of the perifollicular reaction, fibroplasia, and clefting. Scattered apoptotic bodies and focal melanin incontinence can be identified.

FIGURE 4.: Direct immunofluorescence studies show linear deposition of IgG (A and C) and complement C3 (B and D) along both the dermal-epidermal junction (A and B) and perifollicular (C and D) areas.

At the 1-week follow-up, the patient reported significant improvement of the rash with 40 mg dose of prednisone in the morning and 20 mg in the evening. At the 1-month follow-up, there was complete resolution of the previously seen bullous reaction. However, the patient was no longer a candidate for additional immunotherapy due to the side effects of the medication.

DISCUSSION

In this report, we describe a histologically distinct variant of LPPemph, which is characterized by perifollicular bullae and unremarkable ELISA study of serum, and designate it PD-1/PD-L1 inhibitor–induced lichen planopilaris pemphigoides.

Serological BP180 antibody is consistently positive in all 14 reported cases of PD-1/PD-L1 inhibitor–induced LPPemph.^10–16 There are 3 plausible explanations to account for the absence of detectable BP180 antibodies in this case. First, negative ELISA studies may be due to the prior and ongoing steroid therapy at the time of serology testing, resulting in fluctuating titers with subsiding disease severity leading to systematically undetectable levels of circulating antibodies. An additional explanation is that the patient's relatively limited cutaneous reaction may correlate with a diluted systemic effect (equilibrate the antibody within the patient's extracellular space) and thus have resulted in negative ELISA studies. A third possible explanation raising the possibility of a “novel antibody” in the current case. We hypothesize that the “novel antibody” primarily favors the perifollicular basement membrane and triggers a lichenoid reaction. This “novel antibody” is responsible for the distinct pathogenesis of lichen planopilaris pemphigoides and explains hair follicles as the primary location of clefting in this entity; in contrast to the clefting at the nonfollicular dermal–epidermal junction in LPPemph. Given the rarity of this entity, the ability to investigate different phases of the lesions, which could better elucidate the pathogenesis of lichen planopilaris pemphigoides, remains challenging.

A case–control study found 12 of 22 patients (55%) with LP had anti-HCV antibodies, considerably higher than the 25% of 40 psoriatic patients or the 0.17% of blood donors who tested positive.²⁰ Chronic hepatitis C in our patient may have served as a contributing factor toward rash development. Because LPPemph may occur in longstanding LP, it is conceivable that if hepatitis C increases the risk of developing LP, it could also increase the risk of LPPemph. Nevertheless, our patient was diagnosed with hepatitis C more than 11 years ago and had no history of a similar rash before starting nivolumab therapy. In addition, the classic shaggy band pattern of fibrinogen in LP does not present in the current case.

Paraneoplastic autoimmune multiorgan syndrome (PAMS) or paraneoplastic pemphigus (PNP) is clinically defined as a polymorphous cutaneous eruption.²¹ DIF of PAMS usually shows intraepidermal and linear basement membrane zone patterns with IgG antibody. These IgG autoantibodies target the members of plakin family and desmogleins. However, the ELISA study of this patient's serum for desmoglein 1 and 3 was negative. Furthermore, treatment of the tumor in PAMS typically leads to a clinical improvement of the rash whereas in the current case development of LPPemph shortly followed by nivolumab therapy, with rash resolution after drug discontinuation and rash reemergence after drug rechallenge. This proposes causal relationship to the medication and a meaningful difference between PAMS and LPPemph and consequently favors our diagnosis of a drug-induced LPPemph.⁷

Lichen planopilaris (LPP), also known as follicular LP, is primarily a cicatricial alopecia and histologically presents with perifollicular fibroplasia, lichenoid perifollicular infiltrate, and vacuolar interface changes involving the follicular infundibulum and isthmus of hair follicles.²² Although this case shows similar histological finding as LPP on H&E imaging, the DIF findings is not consistent with LPP. DIF of LPP demonstrates nonspecific grouped globular deposits, usually IgM, adjacent to the follicular epithelium, indicating colloid bodies.^23,24 PD-1/PD-L1 inhibitor has been reported to induce LPP.^25–29 However, limited immunogenic evidence was provided in these reports. Therefore, the histological and DIF findings in this case is unlikely to be explained solely by LPP.

Reports of PD-1/PD-L1 inhibitor–associated LPPemph are extremely rare, with 15 reported case studies, including ours. Of these, only 5 were induced by nivolumab. We believe this case has striking features and represents a previously undescribed unique presentation. We termed it PD-1 inhibitor–induced lichen planopilaris pemphigoides. Furthermore, the distinct histopathology suggests a novel follicular-centric pathogenesis for this entity. Here, we highlight the importance of a multidisciplinary approach in managing the dermatologic toxicities of immune checkpoint inhibitors and emphasize the significance of perifollicular changes in the differential diagnosis of bullous dermatoses.