Research ArticleImmunologyOncology Open Access | 10.1172/jci.insight.157839

Kazi J. Nahar,1,2,3 Felix Marsh-Wakefield,2,3,4 Robert V. Rawson,1,5 Tuba N. Gide,1,2,3 Angela L. Ferguson,2,3,4 Ruth Allen,2,3 Camelia Quek,1,2,3 Ines Pires da Silva,1,2,3 Stephen Tattersal,6 Christopher J. Kiely,6 Neomal Sandanayake,6 Matteo S. Carlino,1,7 Geoff McCaughan,2,4 James S. Wilmott,1,2,3 Richard A. Scolyer,1,2,3,5 Georgina V. Long,1,2,6,8 Alexander M. Menzies,1,2,6,8 and Umaimainthan Palendira1,2,3

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Nahar, K. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Marsh-Wakefield, F. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Rawson, R. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Gide, T. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Ferguson, A. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Allen, R. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Quek, C. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by da Silva, I. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Tattersal, S. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Kiely, C. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Sandanayake, N. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Carlino, M. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by McCaughan, G. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Wilmott, J. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Scolyer, R. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Long, G. in: JCI | PubMed | Google Scholar |

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Menzies, A. in: JCI | PubMed | Google Scholar

1Melanoma Institute Australia,

2Faculty of Medicine and Health,

3Charles Perkins Centre, and

4Centenary Institute, The University of Sydney, Sydney, New South Wales, Australia.

5Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.

6Royal North Shore Hospital, Sydney, New South Wales Australia.

7Crown Princess Mary Cancer Centre and Westmead Hospitals, New South Wales, Australia.

8Mater Hospital, North Sydney, New South Wales, Australia.

Address correspondence to: Umaimainthan Palendira or Alexander Menzies, Melanoma Institute Australia, The Poche Centre, 40 Rocklands Road, Wollstonecraft NSW 2065, Australia. Phone: 612 86276115; Email: umaimainthan.palendira@sydney.edu.au (UP). Email: alexander.menzies@sydney.edu.au (AM).

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Find articles by Palendira, U. in: JCI | PubMed | Google Scholar |

Authorship note: KJN and FMW are co–first authors. AMM and UP are co–senior authors.

Published September 29, 2022 - More info

Published in Volume 7, Issue 21 on November 8, 2022
JCI Insight. 2022;7(21):e157839. https://doi.org/10.1172/jci.insight.157839.
© 2022 Nahar et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published September 29, 2022 - Version history
Received: December 21, 2021; Accepted: September 21, 2022 View PDF Abstract

Immune-related adverse events represent a major hurdle to the success of immunotherapy. The immunological mechanisms underlying their development and relation to antitumor responses are poorly understood. By examining both systemic and tissue-specific immune changes induced by combination anti–CTLA-4 and anti–PD-1 immunotherapy, we found distinct repertoire changes in patients who developed moderate-severe colitis, irrespective of their antitumor response to therapy. The proportion of circulating monocytes were significantly increased at baseline in patients who subsequently developed colitis compared with patients who did not develop colitis, and biopsies from patients with colitis showed monocytic infiltration of both endoscopically and histopathologically normal and inflamed regions of colon. The magnitude of systemic expansion of T cells following commencement of immunotherapy was also greater in patients who developed colitis. Importantly, we show expansion of specific T cell subsets within inflamed regions of the colon, including tissue-resident memory CD8+ T cells and Th1 CD4+ T cells in patients who developed colitis. Our data also suggest that CD8+ T cell expansion was locally induced, while Th1 cell expansion was systemic. Together, our data show that exaggerated innate and T cell responses to combination immunotherapy synergize to propel colitis in susceptible patients.

Introduction

Antibodies against the immune checkpoints CTLA-4, PD-1, and its ligand PD-L1 are now important treatments across oncology, and new drugs targeting other checkpoints are in development. The combination of ipilimumab (IPI; anti–CTLA-4) and nivolumab (anti–PD-1) demonstrates superior response and survival in melanoma compared with single-agent nivolumab or IPI and, therefore, has become the preferred first-line regimen for many patients with advanced disease (1). Apart from melanoma, combination immune checkpoint inhibitors are now approved by the U.S. Food and Drug Administration (FDA) for treatment of several cancers, including advanced non–small cell lung cancer, advanced renal cell cancer, hepatocellular cancer, malignant mesothelioma, and metastatic colorectal cancer with mismatch repair–deficient/microsatellite instability-high aberrations (2–6).

While immunotherapy intends to restore antitumor immunity, it also often leads to immune activation in normal host tissue, resulting in immune-related adverse events (irAEs) (7). Such irAEs are frequent, can be severe, and result in morbidity and rarely death. While they may mimic idiopathic autoimmune diseases, their clinical course, management, and outcomes appear distinct. The incidence of irAEs is high, ranging from 54% to 76% depending on the treatment, and is highest in patients treated with combination anti–CTLA-4 and anti–PD-1 therapy, where at least one-third of the patients develop moderate-severe (grade 3 or 4, according to the Common Terminology Criteria for Adverse Events, CTCAE v.5.0) toxicity (8, 9). Although irAEs tend to develop early, between 4 and 6 weeks from the start of immunotherapy, some cases have developed 3 years later (10). IrAEs can affect any organ of the body but most commonly affect the gastrointestinal tract, endocrine glands, skin, and liver. Notably, irAEs involving the gastrointestinal tract were the most common cause of toxicity, leading to treatment discontinuation in clinical trials, occurring up to 13.6% with combination therapy (11).

Mechanisms underlying the development of irAEs are not well understood, and whether they are similar to classical autoimmune diseases is unclear (12). Whether antitumor responses play a role in the development of irAEs and whether irAEs could be treated without compromising tumor responses are poorly understood. Moreover, early recognition of patients prone to develop severe irAEs could allow early intervention and minimize discontinuation rate. Early diversification of T cell repertoire and increased T cell receptor clonotypes were associated with irAEs in patients with metastatic prostate cancer treated with anti–CTLA-4 therapy (13). In a similar study, clonal expansion of CD8+ T cells was associated with moderate-severe toxicity (14). Gene expression profiling in IPI-treated patients suggested high expression of CD177, a neutrophil marker, to be predictive of colitis development (15). High levels of IL-17 at baseline have been associated with colitis in patients with metastatic melanoma treated with IPI (16). A recent study performed single-cell RNA-Seq of colon biopsies in combination immunotherapy–treated patients and demonstrated a potential role for CD8+ tissue-resident memory (Trm) T cells in colitis (17). Most studies thus far have focused on either the systemic immune responses or local responses independently and, therefore, have not determined the full impact of immune perturbations associated with the development of irAEs.

In this study, we have utilized high-dimensional mass cytometry to interrogate circulating immune cells and the systemic immune responses in patients with melanoma who did and did not develop colitis on-treatment with combination therapy of anti–CTLA-4 and anti–PD-1 (18). Paired colon biopsies from patients who developed colitis (including areas with and without clinically apparent colitis) were investigated by multiplex IHC (mIHC) to map the immune landscape of the end-organ. We show a complex interaction between multiple innate and adaptive immune cells that likely drive the immunopathology of moderate-severe colitis. In particular, we show that the innate immune repertoire of patients prior to immunotherapy could make patients susceptible to colitis and that immunotherapy-induced changes associated with colitis are independent of antitumor responses.

Results

Study design. Thirty-seven patients with metastatic melanoma treated with combination immunotherapy (IPI and either nivolumab or pembrolizumab) were selected for this study; 19 who had developed moderate-severe (≥grade 2–4) colitis on-treatment and 18 who did not develop any signs of colitis or other significant toxicity (no-colitis; ≤ grade 2 rash and thyroiditis were permitted). Mass cytometry analysis (by CyTOF) of peripheral blood mononuclear cells (PBMC) taken from these patients at baseline (T0, pretreatment) and at the time of colitis (T1, or a matched time point for the no-colitis group) was performed to define systemic immune composition and therapy-induced changes.

Twenty-six colon biopsy specimens were taken from patients who developed moderate-severe colitis, and they were examined by mIHC. A subset of patients from whom colon biopsies were taken (7 of 26) had additional biopsies taken from regions deemed endoscopically and histopathologically normal for mIHC staining (Table 1 and Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.157839DS1).

Table 1

Cohort characteristics of healthy controls and colitis patients

Circulating innate immune profile at baseline is associated with immunotherapy-induced colitis. Only a subset of patients receiving combination therapy develop severe colitis. In order to determine whether the immune profile of patients prior to the start of immunotherapy had any impact on the development of colitis, we performed multiparameter analysis by CyTOF on PBMC taken before and during treatment. Samples were divided into 4 groups: baseline no-colitis (no-colT0), treatment no-colitis (no-colT1), baseline colitis (colT0), and treatment colitis (colT1). Unsupervised FlowSOM clustering of the data generated 40 metaclusters covering both innate and adaptive immune populations (Figure 1, A and B). To determine whether the overall immune cell profiles of individual patients were different between the 4 groups, principal component analysis (PCA) and permutational multivariate ANOVA (PERMANOVA) were used. These revealed that patients who developed colitis had the most significantly different profile, and this was more apparent during treatment (Figure 1C). In order to determine whether any differences at baseline were associated with colitis, we performed deeper immune phenotyping of the immune repertoire by separating T cells from other immune populations — namely, innate immune populations. Unsupervised FlowSOM clustering was used to generate a further 40 metaclusters for both T cells and innate immune cells. The PCA and PERMANOVA of these clusters revealed that the baseline innate immune cell makeup was significantly different between those who developed colitis and those who did not develop any symptoms of colitis (Figure 1, D and E). Interestingly, a similar analysis of T cell populations showed no significant differences at baseline (Figure 1, F and G). In order to validate these findings, we used manual gating to identify 56 immune populations from the entire immune profile and examined the magnitude of the differences that were statistically significant. This showed that multiple innate populations were indeed significantly different at baseline between patients who developed colitis when compared with those who did not (Figure 2 and Supplemental Figure 1E). Importantly, total CD14+ monocytes (Figure 2B) were among those that were different at baseline. There was also a difference in a subset of NK cells, with CD56hiCD16dim NK cells significantly higher in patients who did not develop colitis (Figure 2B and Supplemental Figure 1E). There were no significant differences in total CD3, CD4, and CD8 T cells, nor were there significant differences in any of their major subsets (Figure 2, A and B, and Supplemental Figure 1F). Together, these data demonstrate that the innate immune cell composition, particularly monocyte subsets, are potentially critical for the development of immunotherapy-induced colitis.

Figure 1

Patients who develop colitis have a distinct innate immune repertoire prior to treatment. CyTOF analysis followed by FlowSOM clustering was performed on PBMCs to determine the immune repertoire of patients. (A) UMAP plot visualizing major cell populations as annotated from all samples (n = 37) on both time points and the colors represent the 40 metaclusters generated. (B) Heatmap showing relative median signal intensity of markers (columns) for each metacluster (rows). Myeloid (blue) and T cell (red) subsets are annotated. (C) A PCA plot was generated on the data of patients (n = 37) with metacluster levels (as proportion of PBMC) as variables. A PERMANOVA was performed on the scaled Euclidean distance of all patients. Patients were grouped based on colitis before and after treatment. Myeloid cells and T cells were then gated on manually, followed by FlowSOM clustering. (D and F) UMAP of myeloid cells (D) and T cells (F) from all samples (n = 37) on both time points with annotated subsets. (E and G) PCA plot and PERMANOVA of myeloid cells (E) and T cells (G). Brown-Forsythe and Welch ANOVA with Dunnett T3 multiple-comparison test (unpaired groups) or mixed-effects analysis with Šídák’s multiple-comparison test (paired time points) were used. Median shown. *P ≤ 0.05. no‑colT0, no-colitis baseline; no‑colT1, no-colitis treatment; colT0, colitis baseline; colT1, colitis treatment.

Figure 2

Difference in circulating monocytes at baseline is associated with the development of severe colitis. Manual gating was performed on major myeloid cell and T cell subsets to identify immune populations that were associated with the development of immunotherapy-induced colitis. (A) Volcano plot summarizes the changes between baselines of colitis (col) and no-colitis (no-col) patients. Subsets with P ≤ 0.05 are annotated. Blue circles are myeloid cell subsets; red circles are T cells. (B) Proportions of cells expressing the markers were determined by manual gating between patients who developed colitis and those who did not. Brown-Forsythe and Welch ANOVA with Dunnett T3 multiple-comparison test (unpaired groups) or mixed-effects analysis with Šídák’s multiple-comparison test (paired time points) were used. Median shown. *P ≤ 0.05. no‑colT0, no-colitis baseline; no‑colT1, no-colitis treatment; colT0, colitis baseline; colT1, colitis treatment.

A multitude of systemic changes to the immune profile are associated with combination therapy. We next sought to determine what changes were associated with combination immunotherapy and whether any of those changes were specific to patients who developed colitis. To this end, we compared the immune profile at baseline against the immune profile on-treatment (at the development of colitis or a corresponding time point for those without colitis). When we compared the global systemic changes between baseline and on-treatment using both unsupervised FlowSOM clustering (Figure 3A) and manual gating strategies (Figure 3B), immunotherapy-induced changes were apparent in both colitis and no-colitis cohorts. These included changes to innate populations, as well as T cells, although changes to T cells were more dominant (Figure 3, A and B). Among the populations changed during treatment were the CD14+ monocytes. We saw increases in the proportions of HLA-DR+CD14+CD16– monocytes and HLA-DR+CD14+CD16+ monocytes, although the values did not reach significance. However, none of these changes were specific for patients who developed colitis (Figure 3C), suggesting that these are common for all patients undergoing combination treatment. Interestingly, CD56hiCD16dim NK cell proportions were significantly reduced between baseline and on-treatment. This reduction was significant only in those who developed colitis — not those who did not (Figure 3C). Therapy-associated changes to T cells included an expansion of ICOS+Ki67+CD4+ T cells (Supplemental Figure 1F) and Ki67+ proliferative CD4+ and CD8+ T cells (Figure 3D and Supplemental Figure 1F). A significant expansion of proliferative CD8+ T cells was only present in those who developed colitis, suggesting that it could be involved in the development of colitis (Figure 3D). Similarly, a significant expansion of proliferative CD4+ T cells was apparent in those who developed colitis but not in those who did not (Figure 3D). On the other hand, a subset of ICOS+ proliferative CD4+ T cells expanded significantly in both those who developed colitis and those who did not (Supplemental Figure 1F). Interestingly, we also found differences in Th1 cells. While the proportion of Th1 cells declined between baseline and on-treatment time points in the no-colitis group, there were no significant changes in those who developed colitis (Figure 3D).

Figure 3

T cell subsets are altered after treatment, irrespective of colitis. (A) Volcano plots representing metaclusters that are changed between groups. Metaclusters with P ≤ 0.05 are annotated. Blue circles are metaclusters phenotypically resembling innate cells, red circles are T cells, and gray circles are other cell subsets. (B) Volcano plots representing differences between colitis (col) and no-colitis (no-col) groups. Blue circles are innate cell subsets, red circles are T cell subsets, and gray circles are other cell subsets. Subsets with P ≤ 0.05 are annotated. (C and D) Selected innate (C) and T cell (D) subsets are shown. Brown-Forsythe and Welch ANOVA with Dunnett T3 multiple-comparison test (unpaired groups) or mixed-effects analysis with Šídák’s multiple-comparison test (paired time points) were used. Median shown. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. no‑colT0, no-colitis baseline; no‑colT1, no-colitis treatment; colT0, colitis baseline; colT1, colitis treatment; mc, metacluster.

We next wanted to compare the on-treatment immune profile of both colitis and no-colitis patients to determine whether an immunotherapy-induced imbalance could also be associated with colitis (Figure 3, A and B). This also showed differences in b