Discussion:
BP is an autoimmune disease that causes bullae, erosions, and erythema on the skin and mucosal surfaces. It is characterized by autoantibodies targeting hemidesmosomal proteins BP180 and BP230 involved in adhesion at the epidermal-dermal junction [1]. A wide variety of drugs (diuretics, spironolactone, furosemide, antihypertensives, and antibiotics) have been associated with development of BP in elderly persons [2]. Recently, there have been increasing reports that use of DPP-4 inhibitors is also associated with the development of BP [3-6]. Several retrospective case-control studies have compared the frequency of DPP-4 inhibitor use between BP patients with diabetes and control diabetes patients without BP, demonstrating an association of DPP-4 inhibitors with the development of BP [8-10]. Accordingly, patients who are using DPP-4 inhibitors should be warned to report new-onset diffuse itching, urticarial lesions, or blisters.
Izumi et al. reported that DPP-4 inhibitor-associated BP featured less prominent erythema clinically and showed little evidence of histological inflammation with scanty infiltration of eosinophils into the skin, suggesting it was an atypical form of non-inflammatory BP [11]. It was also suggested that drug-induced BP could have a relatively favorable outcome compared with typical BP. Our Cases 1 and 3 were examples of this non-inflammatory phenotype of DPP-4 inhibitor-associated BP, presenting with less erythema and minor infiltration of eosinophils into the skin on histological examination (Figure 1 and Table 1). On the other hand, there have been reports that most patients with DPP-4 inhibitor-associated BP have typical clinical manifestations and histological features of this disease [9,12]. According to these authors, there are no significant differences of either symptoms or cutaneous eosinophil infiltration between BP arising in diabetic patients with or without prior DPP-4 inhibitor therapy [9,12].
Benzaquen et al. reported that discontinuation of DPP-4 inhibitor therapy had a favorable impact on the outcome of BP in 19 patients with diabetes, because 95% of them achieved clinical remission after stopping DPP-4 inhibitors and receiving first-line treatment for BP [8]. A retrospective case-control study also demonstrated that clinical outcomes were less favorable among 13 patients with diabetes who continued DPP-4 inhibitors compared to 19 patients who discontinued DPP-4 inhibitors, with eight of the 13 patients who continued DPP-4 inhibitors dying between 2 months and 4.9 years after the initial diagnosis of BP [9]. These findings suggested that discontinuation of DPP-4 inhibitor therapy may be associated with better clinical outcomes. Taking the results from these studies together [8, 9], it can be suggested that administration of DPP-4 inhibitors should be discontinued immediately when a diagnosis of BP is made. However, we experienced two cases of DPP-4 inhibitor-associated BP with unfavorable outcomes despite immediate withdrawal of DPP-4 inhibitor therapy and initiation of first-line treatment for BP with an oral steroid and a high-potency topical steroid. Both patients were elderly women over 80 years old.
In Case 1, skin lesions showed multiple relapses despite discontinuation of DPP-4 inhibitor therapy and initiation of treatment for BP. She received intravenous steroid pulse therapy, IVIG, and plasmapheresis in hospital on several occasions. Eventually, she died after a fall, probably from cerebral herniation associated with a subdural hematoma. We speculate that repeated hospitalization and steroid therapy for treatment of BP can cause muscle weakness, resulting in a fall. In Case 2, symptoms of BP were improved by treatment with oral prednisolone and cyclosporine, but the patient died after deterioration of her general condition. Immunosuppressive therapy can lead to compromised immunity in older patients with diabetes, triggering opportunistic infections. Immunosuppressive therapy may also promote the development or progression of cancer. Although we could not identify the pulmonary nodular lesions detected by computed tomography in this patient, she may have died of lung metastasis from primary colon cancer, although the possibility that opportunistic pulmonary infection was associated with her death cannot be excluded. We speculate that persistent immunomodulatory effects of DPP-4 inhibitor despite its withdrawal and use of immunosuppressants may induce carcinogenesis or promote development of colon cancer with pulmonary metastasis. It was recently reported that control of BP and relapse of this condition did not differ between patients who stopped or continued treatment with DPP-4 inhibitors [13]. This report and our experience suggest that it is important for clinicians to pay close attention to the clinical course of DPP-4 inhibitor-associated BP, even after discontinuation of DPP-4 inhibitor therapy.
The mechanisms responsible for BP associated with DPP-4 inhibitor therapy remain to be determined. DPP-4 (CD26) is highly expressed by T cells, especially CD4+ T cells. It is possible that inhibition of DPP-4 may be associated with development of autoimmune skin diseases, because autoreactive T cells are involved in the pathogenesis of BP [14]. A previous study demonstrated that the HLA-DQB1*03:01 allele is a biomarker for genetic susceptibility to BP associated with DPP-4 inhibitors in a Japanese population [15], suggesting an association between HLA class II and this drug-induced autoimmune disease. Our Case 3 was positive for the HLA-DQB1*03:01 allele, in agreement with this report (Table 1). We previously demonstrated that sitagliptin, another DPP-4 inhibitor, reduced circulating CD4+ T cells in patients with type 2 diabetes, especially causing a decline of regulatory T cells [16]. These cells play a very important role in the immune system by suppressing immune responses and maintaining tolerance, and prevent the development of autoimmune diseases by suppressing T cell activation [17]. In fact, a very recent study demonstrated that dysfunction of regulatory T cells is associated with induction of autoantibodies to bullous pemphigoid antigens in mice and humans [18]. Another possibility is that inhibition of DPP-4 augments the activity of eotaxin (CCL11), a DPP-4 substrate, resulting in recruitment of eosinophils to the skin [19].
In conclusion, we experienced 3 patients with type 2 diabetes who developed DPP-4 inhibitor-associated BP. Despite prompt discontinuation of DPP-4 inhibitor therapy and initiation of first-line treatment for BP, the outcome was unfavorable in two patients. Accordingly, clinicians should carefully monitor the course of DPP-4 inhibitor-associated BP even after withdrawal of DPP-4 inhibitor therapy, especially in very elderly patients.