The evolving landscape of immune-related adverse events that follow immune checkpoint immunotherapy in cancer patients

Abstract

Immune checkpoint inhibitors (ICIs), including antibodies targeting anti-cytotoxic T lymphocyte antigen 4 (CTLA-4), anti-programmed cell death 1 (PD-1), anti-programmed cell death 1 ligand 1 (PD-L1) have revolutionized cancer treatment.^{1, 2} Both PD-1 and CTLA-4 are upregulated upon T-cell activation and function to inhibit peripheral T-cell responses. These receptors maintain immune homeostasis, limiting tissue damage during infection and helping to prevent autoimmune disease. CTLA-4 suppresses T-cell activation by cell-extrinsic competition with CD28 for binding the same ligands and cell-intrinsic signaling mechanisms.³ The PD-1:PD-:L1/L2 pathway also suppresses T-cell responses by inhibiting TCR/CD28 signaling, target cell lysis and cytokine production.⁴ PD-1 ligand is expressed across a diverse set of hematopoietic and non-hematopoietic cells, suggesting it may have more impact in peripheral tissues, whereas CTLA-4 and its ligands are primarily expressed in secondary lymphoid organs.³ Collectively, ICI immunotherapy leverages this latter biology to regulate T-cell activation and tolerance, allowing the immune system to overcome cancer's evasion of immunity and reestablish its ability to attack tumor tissues.^{1, 2} CTLA-4 was the first target to be approved for ICI therapy (i.e., Ipilimumab) in 2011^5-8 to treat advanced melanoma. Since then, eight agents targeting PD-1/PD-L1 or CTLA-4 pathways have been approved by the Food and Drug Administration (FDA) for over 20 cancer types across 80 indications.⁹ Importantly, the utilization of ICIs in the adjuvant setting is also on the rise, resulting in further growth of the number patients receiving ICI therapy. Cancer patients treated with ICIs are now anticipated to live longer; but prolonged exposure to ICI treatment also puts them at greater risk of long-term complications associated with this line of therapy.^10-14

Unfortunately, the success of ICIs is limited by the emergence of treatment-induced inflammatory toxicities, termed immune-related adverse events (irAEs). irAEs can affect any organ system in the body with a range of severity, resulting in significant morbidity for patients and considerable cost to the healthcare system. These irAEs can develop during or after treatment and often require immunosuppressive therapy or cessation of ICI treatment, thereby limiting the potential lifesaving benefits from ICI therapy.^{1, 2, 15, 16} Depending on the treatment regimen (i.e., drug types, combination, and doses), 66%–96% of patients develop irAEs, and 7%–59% experience severe irAEs.^{17, 18} Fulminant and fatal irAEs—the most common being myocarditis/myositis, pneumonitis, and neurologic events (e.g., Guillain–Barre syndrome and myasthenia gravis)—are experienced by approximately 0.4% of cancer patients treated with anti-PD-1/PD-L1 monotherapy, and 1.2% of patients treated with combination PD-1/CTLA-4 therapy.^{19, 20} Toxicities can manifest at any point during the treatment, but they tend to occur most frequently within the initial 3 months following the initiation of ICI.^{15, 16} Immuno-oncology remains a rapidly evolving field, with many more co-inhibitory pathways beyond CTLA-4 and PD-1 likely to become valuable therapeutic targets in the near future. For example, relatlimab, a compound that blocks lymphocyte antigen 3 (LAG-3) has recently been approved for use in combination with anti-PD1 therapy.²¹ This target is predominantly found in exhausted T cells and could potentially address T cell-related anergy associated with the tumor microenvironment. As ICIs become first- and second-line cancer treatment for more tumor types in the years to come, together with the increased use of ICIs in the adjuvant setting as well as in combination therapy to increase the anti-tumor response efficacy and overcome treatment resistance,^{22, 23} it is expected that the incidence of irAEs will continue to rise and severely limit the life-saving potential of ICIs.

Intriguingly, irAEs occurring in the setting of ICI therapy closely resemble idiopathic autoimmune diseases, including rheumatoid arthritis (RA), autoimmune hepatitis, and inflammatory bowel disease.¹ Further, polymorphisms of PD-1 and CTLA-4 are associated with various autoimmune disease conditions—such as Type 1 diabetes, thyroiditis, Graves' disease, celiac disease, systemic lupus erythematosus, and rheumatoid arthritis^{15, 16, 24, 25}—with which the observed treatment-related irAEs can share clinical features. irAEs are treated with many of the same therapeutic interventions as autoimmune diseases, including steroids, biologics agents, and sometimes surgery. Insights into the pathogenesis of irAEs, a setting in which the exact disease trigger and timing of disease initiation are known, may thus have implications not only for the care of cancer patients but also for our understanding of idiopathic autoimmune diseases. Studying the cellular and associated molecular patterns associated with irAE may provide a rare window into the exact moment when a patient's own immune system starts mounting an inflammatory response against normal tissue. Such study may also enable generating novel insights into the mechanisms of immune regulation throughout the body and could further our understanding of how CTLA-4 and PD-1 function to maintain immune tolerance.²⁶ As such, deciphering irAE pathogenesis may have implications not only for the care of cancer patients but also provide critical insights into the earliest steps driving and sustaining spontaneous autoimmune diseases, which are challenging to discern and study when such diseases are fully established.

While we can hypothesize that the trigger of the irAE onset is the manipulation of T-cell signaling pathways and the breaking of tissue-specific immune tolerance,^{1, 2, 15, 16} the downstream mechanisms involved in driving and sustaining excessive immune activation against normal tissue remain poorly understood. This is in part due to the fact that preclinical animal models have yet to recapitulate the full spectrum of irAE clinical presentations.^{15, 16} Examples of disease mechanisms proposed thus far have included shared antigens between tumor and affected tissue, preexisting subclinical inflammation, preexisting autoantibodies, antibody-dependent cytotoxicity, and environmental insults.^{15, 16, 26, 27} Given the heterogeneity and breadth of irAE clinical presentations, it is likely that irAEs affecting different organ systems have different pathological drivers. Among the irAEs sharing similar histological characteristics within a specific tissue, it is also possible that multiple mechanistic endotypes result in comparable clinical manifestations. Our limited mechanistic understanding of irAEs to ICI has severely restricted the development of targeted treatment solutions. Hitherto, treatment management guidelines for irAE have been developed almost entirely based on small retrospective clinical studies and expert opinion. Because the mechanisms driving toxicity and severity across different organ systems will likely differ, it will require that the medical community conceive prospective treatment trials for each irAE individually, as opposed to studying multiple irAEs together. Over the last few years, limited progress has been made in understanding the biological drivers of ICI-related irAEs, and in developing effective mitigation strategies. Further research into the cellular and molecular underpinning of irAEs will be critical to define better predictors of toxicity to enable the prevention of their occurrence, as well as the development of new, targeted, and improved treatment solutions that can mitigate these toxicities while preserving the anti-tumor immune response.

This issue comprises a compilation 13 review pieces from top scientists and medical experts in the field who discuss our current understanding of these ICI associated inflammatory toxicities across organ systems. Reviews include several in-depth overviews of tissue-specific irAE presentations. They address our current understanding of cellular and molecular drivers of irAEs and their relationship to anti-tumor immune response, including discussions on available preclinical models to study these mechanisms. They cover potential predictive biomarkers and genetic risk factors identified to-date, and discuss current management strategies being used, as well as clinical trials being developed using therapeutic candidates intended to limit irAE onset and improve patient outcomes. Collectively, these reviews summarize the state-of-the field of this evolving and complex subject, the critical challenges to ICI therapies posed by irAEs, along with the many opened questions and the strategies that should be the focus of future investigations.

The first set of six reviews delves deeply into our understanding of organ-specific irAEs. It is crucial to better understand the immunopathology of irAEs across organ-systems to allow for a more personalized immune suppression therapy and to spare negative effects on anti-tumor immunity. Although irAEs can target any organ system, toxicities occur more frequently at barrier tissues, including the gastrointestinal tract, skin, liver, and the lungs,^{15, 16, 26} and these are the focus of the first three reviews. The fact that barrier organs are so commonly involved suggests that the antigenic targets of the immune response may be the commensal microbiome, though this has not yet been demonstrated.

One of the first irAEs reported was vitiligo, which is a condition characterized by immune-mediated destruction of melanocytes that can occur in cancer patients with melanoma undergoing going treatment with various immunotherapies.^{15, 16, 26} Vitiligo is driven by T cells and antibodies targeting melanocyte antigens, which are also produced by melanoma. This immune response leads to the destruction of melanocytes, resulting in the characteristic depigmentation observed in affected individuals.²⁸ Given these early observations, it was originally hypothesized that other irAEs might also result from targeting antigens common to healthy and malignant tissue, similarly to the mechanisms observed in vitiligo. Nonetheless, the fact that the majority of irAEs are not specific to a particular tumor type cast doubt on this hypothesis. Alternative hypotheses have now emerged that are discussed in the first three reviews. These posit a role for tissue-resident memory cells (TRM cells), which are long-lived memory effectors that reside in peripheral tissues and make up the majority of T cells in barrier tissue. TRM cells express PD-1 and other immune checkpoint molecules that control their activation. These cells play important roles in tissue surveillance and recall responses to pathogens.^29-31 Recent studies have investigated whether ICI therapy triggers irAE by promoting the expansion and activation of local resident TRM cells. This work is discussed in the first three reviews covering irAE affecting the gastrointestinal track, the skin, and the lung.

The review by Michael Dougan provides a comprehensive review on the gastrointestinal (GI) toxicities caused by ICIs.³² Although most of the GI irAEs are mild in nature, their high occurrence rate makes them the leading cause of severe toxicities associated with ICIs. In his review, he provides an in-depth description of the clinical manifestations along with a thorough discussion of the immune mechanisms known to-date (e.g., gut CD8⁺ TRMs) associated with ICI-colitis, which is the most common GI toxicity associated with the use of ICIs.³³ He also discusses risk factors associated with ICI-colitis, current standard of care, upcoming clinical trials, and the challenges related to the current lack of clinical trial data to guide management of patient care. Importantly, he highlights multiple unanswered questions in the field that require further investigation.

The second review from Blair Allais and colleagues is focused on dermatologic toxicities, which are the most frequently reported, the most visible, and among the earliest seen in cancer patients on ICI therapy.³⁴ While most of these cutaneous immune-related adverse effects (cirAEs) are mild, severe presentations occur.³⁵ Allais et al. provide a clear and thorough discussion of what is known about cirAEs through the different lenses being used to study them. These include harnessing the power of large datasets to understand clinical presentations and their association with outcomes. They establish specific definitions for each cirAEs, and their need for more targeted therapies. They describe leveraging cellular and molecular strategies to infer novel insights into disease mechanisms. They also describe the current understanding of immunosuppression treatment for managing cirAEs, as well as future opportunities in the field to help improve patient care.

Fiamma Berner and Lukas Flatz focused their review on two emerging classes of self-antigens involved in skin toxicity and pneumonitis in non-small cell lung cancer patients undergoing ICI therapy.³⁶ They discuss how these self-antigen-specific T cells are believed to play a dual role by mediating irAE to skin and lung, and by promoting anti-tumor immune responses. Such dual association creates a challenge in distinguishing between the immune signature associated with irAE from those associated with treatment response. The review also highlights methods for the identification of self-antigen targets and discusses the role of antigens shared by organs affected by irAEs and tumors. Finally they discuss current challenges and solutions that may lead to successful cancer patient treatment with ICIs in the absence of problematic irAEs.

Gary Reynolds delves into rheumatological irAEs, which are disabling, can limit treatment efficacy and evolve to become chronic.³⁷ These presentations can superficially share features of common immune-mediated rheumatic diseases (IMRDs),³⁸ such as rheumatoid arthritis, psoriatic arthritis, Sjogren's syndrome, and polymyalgia rheumatica. This review discusses the parallels and differences between rheumatic irAEs and IMRDs, and highlight that, based on current knowledge, they are clinically and immunologically distinct entities.

Harish Seethapathy, Kavita Mistry, and Meghan Sise elegantly describe irAE manifestations in the kidney, which present as acute interstitial nephritis (ICI-AIN).³⁹ They thoroughly explore the clinical manifestation of ICI-AIN and highlight its distinctive presentation compared to conventional drug-induced AIN. The authors provide a detailed description of current mechanistic understanding of ICI-AIN, and underscore the diagnostic and therapeutic implications that arise from these findings.

To conclude this first section of the issue, Jingyi Gong, Tomas Neilan, and Daniel A. Zlotoff present a thorough review on immune-related myocarditis (irMyocarditis), which, among all the cardiovascular toxicities, has been the most extensively studied due to its substantial risk of mortality.⁴⁰ While irMyocarditis occurs in 1% of ICI recipients, the incidence is expected to continue rising with the increasing use of combination regimens to treat cancer patients.^{19-21, 41-43} It is fatal in 20%–40% of cases, which is approximately 10-fold higher than in non-ICI lymphocytic myocarditis.^41-44 Nearly half of the cases result in major adverse cardiac events, such as severe arrhythmias, cardiogenic shock, and sudden cardiac death.^41-44 Furthermore, the development of irMyocarditis is a near-absolute contraindication for future use of ICI therapy, which in turn may harm the oncologic prognosis.⁴⁵ The authors frame their review around two primary objectives. First, they provide a comprehensive summary of the current understanding of irMyocarditis pathogenesis, and they compare it to other types of myocardial inflammation, which they examine across three categories: the cellular mediators, participating molecular signals and soluble factors, and T-cell receptor (TCR) clonality and specificity. Second, they discuss different hypotheses explaining its heightened morbidity compared to other forms of cardiac inflammation and why it only affects a limited number of cancer patients receiving ICI therapy. They conclude their elegant review by highlighting the many knowledge research and clinical gaps in the field requiring further investigation to enable developing better solutions for patient care.

While the above six reviews each focus on the clinical manifestations associated with specific organ system irAEs, the next section includes four reviews that delve more deeply into our biological understanding of irAE presentations across tissues, the known cellular drivers, their relationship to autoimmunity, and the available preclinical models that can be leveraged to study their mechanistic underpinnings. Importantly, defining the underlying disease mechanisms presents an opportunity not only to alleviate these detrimental side effects but to also improve the efficacy of ICI therapy itself.

Namrata Singh, Anne Hocking, and Jane Buckner provide an interesting review on the intersection between irAEs to ICI and autoimmune diseases.⁴⁶ While several irAEs can phenocopy the clinical presentation of well-defined autoimmune diseases, it remains unclear whether these diseases are truly the same syndrome, especially as in many cases the ICI-associated presentation is more fulminate.⁴⁷ This review explores whether the mechanisms responsible for inducing ICI-related irAEs are similar or different from those underlying spontaneous autoimmunity, focusing on ICI-diabetes, ICI-thyroiditis, and ICI-arthritis. The authors present a summary of our current understanding of the role of genetics and autoantibodies in the development of irAEs, and discusses the use of single-cell genomic strategies that have enabled the identification of specific T-cell signatures associated with irAEs. Ultimately, deciphering the relationship between these irAEs and spontaneous autoimmune diseases holds significant importance in terms of understanding basic mechanisms and clinical implications. Indeed, from a basic science standpoint, elucidating the connection between these clinical presentations will generate valuable insights into the fundamental mechanisms of autoimmunity and the biological principles of immune tolerance in human. From a clinical perspective, unraveling the relationship between irAEs and autoimmunity could help nominate therapeutic strategies for irAE based on therapies already clinically approved to mitigate autoimmunity.

Kavita Dhodapkar, Alyssa Duffy, and Madhav Dhodapkar discuss how the use of immune checkpoint inhibitor therapy can impact B-cell tolerance and drive alterations in the B-cell compartments.⁴⁸ Their review explores the potential involvement of humoral immunity, and the role of specific B-cell subsets and autoantibodies in driving the irAE pathogenesis associated with the use of ICI. They also highlight the need to invest more resources in investigating the cross talk between T and B cells that drives the activation of pathogenic B cells, contributing to the development of irAEs.

Noah Earland and colleagues lay out a thorough perspective⁴⁹ on their recently published work that analyzed a cohort of advanced melanoma using a multimodal strategy involving the use of single-cell RNA sequencing, as well as mass cytometry of peripheral blood cells to identify features predictive of irAE development.⁵⁰ Their work showed that increased TCR diversity together with increased circulating levels of activated CD4 effector memory T cells are both associated with the development of severe irAEs, independent of the tissue affected.⁵⁰ Through their review, they contextualize their results in comparison to other published work on irAE blood-based biomarkers. They examine possible mechanistic links between CD4 memory T cells and effector memory T cells, along with other factors that could contribute to irAE development, such as autoreactive tissue-resident memory T cells located at the tissue site of toxicity. Finally, they discuss factors that should be considered to empower future clinical translation of findings to develop and establish blood-based irAE-related clinical tests.

Presently, the pathophysiological mechanisms of irAE are poorly understood in part because preclinical mouse models underestimate the occurrence of irAEs compared to patient experience,⁵¹ or do not accurately replicate critical aspects of the human irAE biology.^{15, 16, 26, 52-54} Nevertheless, there is a multitude of available animal systems that could be tailored and leveraged to further model biological pathways identified in patients. For the last manuscript of this section, Morgan Cina, Jessica Venegas, and Arabella Young put together a very comprehensive review on the preclinical models available that capture some aspects of the irAE etiology across different organ systems. These can be used to study the biological underpinnings of irAEs, along with their relationship to anti-tumor immune responses.⁵⁵ They examine the contexts for which these models are most useful, how they could be improved, as well as the advantages and disadvantages of different model systems. Approaches include in silico modeling of published datasets, in vitro organoid models, in vivo mouse and non-human primate models, as well as canine patients. They review why the development of model systems to study irAEs has not progressed rapidly. Particularly in comparison to efforts invested in establishing models to study therapeutic combination strategies. The authors end by discussing what the future holds for developing new preclinical models that capture with higher fidelity the breadth irAE presentations. Better models are critical to ultimately improve our mechanistic understanding of irAEs and their associated treatments.

The third and last section of this issue includes three reviews that examine what factors have been identified to-date that may help determine who will develop irAEs to ICI, whether we can predict which patient is predisposed to irAEs, and what is the relationship between irAE presentations and anti-tumor immune response. The ability to predict the occurrence of specific irAEs to ICI will be a critical step towards the development of prophylactic therapies and could prove valuable in making informed decisions regarding tailoring cancer treatment options for each patient.

One of the strategies used to identify disease predictors is to focus on uncovering germline susceptibility variants to ICI-related irAEs, which is elegantly reviewed in this issue by Alexander Gusev.⁵⁶ Because these irAEs can phenocopy the clinical presentation of some sporadic autoimmune diseases, it has been hypothesized that they could be linked to autoimmunity through germline genetics and would be heritable.^57-59 In this review, he summarizes current work mapping susceptibility germline variants to irAEs, outcome, and responses to ICIs. He also provides perspectives on recently published work from his group that leverages genome-wide association analysis to identify IL7 susceptibility variants associated with irAEs,⁶⁰ findings that were subsequently independently replicated.⁶¹ He also discusses extant questions in the field, as well as opportunities to leverage germline genetic association studies to further our understanding of irAE etiology, in addition to improving our ability to predicting them.

There are currently no FDA approved biomarkers that predict the occurrence of irAEs, making it challenging to identify cancer patients for whom the risk of suffering from toxicity to ICI therapy is outweighed by the potential lifesaving benefits of receiving them. Moreover, the timing of these toxicities is not as predictable as side effects associated with other cancer treatments, such as chemotherapy and targeted therapy.⁶² Nevertheless, several efforts to map such biomarkers for irAE toxicities that could guide treatment selection, prevention, and monitoring have been initiated. These are thoroughly review by Rachel Goodman, Seungyeon Jung, Justin Balko, and Douglas Johnson.⁶³ The authors describe different molecular and clinical biomarkers across three categories, including those that can help: (i) risk stratify patient pre-treatment to identify subjects likely to develop toxicities; (ii) diagnose irAEs, which remain challenging to do in many cases; and (iii) predict occurrence of irAEs during treatment course. These could enable tailoring clinical decision-making, including whether to alter treatment course or to increase the frequency of patient monitoring.

We conclude this issue with a comprehensive review by Steven M. Blum, Sherin Rouhani, and Ryan Sullivan on the molecular and clinical relationship between anti-tumor immune response and irAEs, along with a discussion of the role that irAE treatments may play in modulating the anti-tumor immune response.⁶⁴ They examine the relationship between irAEs and anti-tumor immunity, which can be complex⁶⁵ and influenced by factors such as the specific irAE-affected organ, tumor histology, and individual patient's characteristics. The authors highlight the importance of developing therapeutic strategies that can maximize the anti-tumor immunity efficacy of ICIs while minimizing the occurrence of irAEs by modulating both the host and the tumor immune environments, which will ultimately be essential to improve patient outcomes.

The landscape of cancer patient populations being treated with ICI therapies has significantly changed since their FDA approval for the treatment of metastatic melanoma in 2011; ICI is now used in indications with curative intent. For malignancies like metastatic melanoma, which were largely fatal before the introduction of ICI therapy, the benefits of ICIs far surpass the potential risks of developing irAEs in most patients, making such risk acceptable. In the context of adjuvant and neoadjuvant therapy, the risk-to-benefit ratio is less straightforward; while there has been improved disease-free survival, it is important to note that a considerable proportion of patients would be expected to achieve cancer remission without the use of ICIs, highlighting the need to reevaluate the irAEs risk versus the reward of being cancer-free. The illustration included in this introduction (Figure 1) captures the complicated balance between the benefits and risks associated with the use of ICI therapy to treat cancer patient. Despite the emergence of irAEs, it is important to remember that the introduction of ICI drugs has resulted in many good news stories, providing new hopes to cancer patients.

With the increasing utilization of ICI therapy among the cancer patient population, it is becoming increasingly critical to support research efforts aimed at understanding the root cause of irAE development, and how they can be mitigated while promoting anti-tumor immunity. Reviews included in this issue highlight how our understanding of irAEs has evolved over the years, along with the many challenges to be overcome and the remaining opened questions to be answered in the field, such as, which factors play a role in determining patient risk of experiencing irAEs? Can we predict which patient will develop irAEs? What factors contribute to the higher incidence of irAEs in certain organ systems? How are spontaneous autoimmune diseases and irAEs that phenocopy them related? Which factors influence the severity of an irAE in a particular patient, which can range from transient to chronic or fatal? How can we choose the most effective treatment strategies to treat irAEs while sparing the anti-tumor immune response? Once an irAE has improved upon treatment, which patient is at risk for redeveloping irAEs and which patient can be safely rechallenged with ICI therapy? Answering these outstanding questions will likely be the focus of upcoming studies in the immuno-oncology field.

Lastly, a broad theme that emerges from several reviews in this issue is the need to provide more financial support to this research across disciplines and to cross-institutional multidisciplinary teams spanning healthcare systems, academic researchers, federal research agencies, commercial partners, and patient advocacy groups to solve the underpinnings of irAEs and integrate critical discoveries into sustainable clinical solutions and applications. Even once irAE therapeutic targets are identified, the field will still necessitate cross-institutional collaborative endeavors to establish the clinical trial infrastructure required for evaluating these treatments. We owe it to cancer patients to work together as a community to find solutions to mitigate irAEs in order to improve their quality of life and enable them to optimally benefit from the life-saving virtues of ICI-therapy to fight their cancers.

ACV has been a paid consultant to Bristol Myers Squibb.

ACV is funded by the National Institute of Health Director's New Innovator Award (DP2CA247831), the Damon Runyon-Rachleff Innovation Award, The Melanoma Research Alliance Young Investigator Award (https://doi.org/10.48050/pc.gr.143739), the Massachusetts General Hospital (MGH) Transformative Scholar in Medicine Award, and the MGH Howard M. Goodman Fellowship.