Background
The advent of immune checkpoint inhibitor development has offered clinical benefits in a variety of malignancies including melanoma. Nivolumab is a fully humanized monoclonal IgG4 antibody directed against programmed cell death 1 (PD-1), which is expressed on activated T cells and functions as a co-inhibitory receptor. Despite their encouraging efficacies, however, immune checkpoint inhibitors carry risks of treatment-related complications associated with harmful autoimmune responses, which are referred to as immune-related adverse events (irAEs). While the safety profile of nivolumab monotherapy is generally acceptable, with common adverse toxicities including fatigue, rash, pruritus, and diarrhea, there are reports of patients requiring treatment interruption and corticosteroid administration [
1,
2]. In a previous phase II clinical trial, severe irAEs (grade 3/4 according to NCI CTCAE guidelines) occurred in 16.3% of treated patients [
3]. Although colitis is the most common irAE in patients treated with anti-cytotoxic T-lymphocyte antigen 4 (CTLA-4) antibodies, the rate of grade 3/4 diarrhea in those given PD-1/ programmed cell death ligand 1 (PD-L1) agents is very low (1 to 2%) [
4‐
6]. However, autoimmune colitis can be severe with potentially fatal perforations [
7]. Although irAEs associated with nivolumab have gradually been recognized, the mechanisms underlying these irAEs have not as yet been fully clarified. Herein, we report 2 melanoma patients who developed severe colitis during nivolumab treatment and whose pathological findings of colon we could compare between before and after corticosteroid treatment. We analyzed biological samples from the patients and discuss, with a review of the literature, the pathophysiology of this complication.
Discussion and conclusions
In the two cases reported herein, nivolumab caused colitis with marked infiltration of CD8+ cells and T-bet expressing CD4+ T cells, indicating helper T cells (Th) 1 to be responsible for the dominant response. To the best of our knowledge, no previous reports have described Th1-dominant response was associated with irAEs caused by nivolumab. In addition, serum CRP levels and IL-6 levels were proportionate to the severity of colitis apparently caused by nivolumab more clearly than other laboratory data and serum cytokines.
Accumulating evidence suggests that immune checkpoint inhibitors confer a survival benefit in patients with several types of cancer, including melanoma. With the extension of indications for immune checkpoint inhibitors, the recognition of serious irAEs becomes more and more vital for the safe use of these agents. Although the frequency and severity of irAEs following nivolumab monotherapy appear to be lower than those associated with CTLA-4 blockade, only limited information is currently available as to the diagnosis and management of, and the risk factors for, irAEs during nivolumab monotherapy [
9]. Although it is generally accepted that most irAEs arise from immune activation, the mechanistic details are poorly understood [
10]. Several studies have reported immunological analyses of colitis caused by immune checkpoint inhibitors. Immunohistochemical analysis of colon biopsy specimens from anti-CTLA-4 antibody recipients who developed colitis showed no evidence of Foxp3
+ regulatory T cell depletion [
11]. In contrast, however, another study revealed a marked increase in all T-cell subsets (CD3
+, CD4
+, and CD8
+ cells) and of CD4
+CD25
+ regulatory T cells in patients with gastroenteritis due to anti-CTLA-4 antibody [
12]. Ulcerative colitis and Crohn’s disease, which are the two most common inflammatory bowel diseases (IBD), are thought to have different pathogeneses [
13]. Small bowel inflammation in Crohn’s disease is generally associated with increased IFNγ and IL17A expressions (indicative of Th1 and T helper 17 [Th17] cells, respectively), whereas type 2 T helper (Th2) cytokines (e.g., IL-4, IL-5, IL-13) predominate in ulcerative colitis [
14]. Our studies of tissue specimens from these two cases suggested marked infiltrations of CD8
+ cells and T-bet
+ cells in CD4
+ cell as well as CD8
+ cells, indicating the Th1 dominant response to be a causative factor in irAEs associated with nivolumab. Although there are some differences between etiology of colitis induced by nivolumab and that of IBD, both diseases share some clinical findings [
15,
16]. Previous report showed that both diseases have similarities in endoscopic and histopathological findings [
15]. Interestingly, pathophysiology of both diseases is thought to be associated with gut microbiome [
16]. Immune related colitis induced by nivolumab could be managed by treatment similar to that for IBD [
15]. To study the similarities and differences between colitis induced by nivolumab and ulcerative colitis will lead to elucidate the pathophysiology of the both diseases. Therefore, it might be worth doing a comparative study of both diseases by immunoprofiling using multiplex IHC.
Tumor-infiltrating CD8
+ lymphocytes appear to be a favorable prognostic factor in the vast majority of cancers [
17]. On the other hand, previous study showed that the subepithelial layer was enriched with CD8
+ T cells in colitis induced by anti-PD-1 antibodies, whereas CD4
+ T cells were predominant in colitis associated with anti-CTLA-4 antibodies [
18]. The cytotoxic T lymphocyte expressing CD8 may play a major role in both the efficacy of and the adverse events caused by nivolumab. However, the precise roles of Th cells in irAEs remain unknown. The results of our study suggest that Th1 cells have a considerable impact on irAEs. In addition, infiltration of Foxp3
+ regulatory T cells was prominent in case 1 whose colitis persisted even after corticosteroid administration. Activated human CD4
+ and CD8
+ T cells transiently express Foxp3 [
19]. In our 2 cases, however, Foxp3 expression was almost independent of T-bet and CD8 expression according to immunohistochemical findings, indicating most Foxp3
+ cells reflected regulatory T cells. Those findings suggest that regulatory T cells might play a role in the recovery from colitis due to nivolumab. However, our findings are based on analysis of only two patients. Thus, further study is needed to confirm our findings.
Biomarkers possibly predicting the development of toxicities have been explored in patients receiving immune checkpoint inhibitor treatment. An increase from baseline in IL-17 after treatment was shown to be associated with irAEs [
20]. IL-17 is one of the central inflammatory cytokines upregulated in the inflammatory bowel diseases [
21]. In our study, although cytokines were not measured at baseline, specific biomarkers including IL-17 that reflect the severity of symptoms or pharmacological responses to nivolumab were not identified. The pathological findings also revealed mild infiltration of Th17 cells (RORγt
+ cells) into the colon, indicating that Th17 cells may only play relatively minor roles in irAEs. On the other hand, in our 2 cases, CRP and IL-6 elevations were blunted by corticosteroid administration, in parallel with the resolution of colitis. CRP is an acute phase protein synthesized by the liver that serves as an early marker of inflammation or infection. The synthesis of CRP is stimulated by IL-6 and levels of CRP are strongly correlated with serum levels of IL-6. Although it is difficult to examine IL-6 routinely in daily clinical practice, assays for measuring CRP are available in routine clinical practice and are inexpensive. Several reports have suggested CRP to be a potential biomarker for autoimmune disorders including inflammatory bowel diseases [
22,
23]. Previous study showed that IL-6 level in melanoma patients was elevated compared to healthy donor and tended to increase in patients with advanced stage [
24]. The clinical courses and findings of our cases suggest that CRP as well as IL-6 reflects the treatment responses of irAEs caused by immune checkpoint inhibitors. Although many reports suggested CRP reflect the disease state in various cancers including melanoma [
25,
26], examination of CRP kinetics could help for treatment of irAE in cancer patients. In addition, CRP is easier to measure than IL-6 in clinical setting. Thus, routine measurement of CRP may facilitate the prediction of the clinical course of irAEs. However, it should be noted that CRP could be affected by several clinical factors such as malignancies, infection, and administration of corticosteroid. Baseline CRP level was diverse especially in patients with malignancies. CRP levels were elevated (≥ 10 mg/dL) in 10.7% of melanoma patients at baseline [
27]. Therefore, when we confirm recovery from irAEs, it is thought to be important to check whether decrease of CRP is in parallel with improvement of the other factors.
In conclusion, the cases presented herein suggested robust infiltration of CD8+ cells and T-bet+ cells in CD4+ as well as CD8+ T cells, indicating a Th1 dominant response, to be associated with the mechanism underlying the development of irAEs due to nivolumab. Additionally, CRP as well as IL-6 was found to be a potential biomarker reflecting treatment responses in patients with irAEs. Further pathological studies are needed to enrich our understanding of irAEs.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.