Recent advances, patient selection & challenges in managing cancer patients undergoing treatment with immune checkpoint inhibitors

Cancer immunotherapy with humanized monoclonal antibodies (mAbs) that target co-inhibitory immune checkpoint molecules (ICMs) is the most meaningful advance in the management of malignant diseases in recent years1. This has coincided with the acquisition of eloquent, cutting edge insights into the molecular mechanisms, which regulate cell–cell interactions that are fundamental to maintain a balanced, well-synchronized human immune system. These developments have also revitalized the practice of immunotherapy, especially the realization of novel immunomodulatory therapeutic modalities that have the potential to restore weakened anti-cancer immune responses.

Only in recent times, however, has the notion been dispelled that spontaneous human cancers arise and progress due to their apparent lack of immunogenicity, implying redundancy of the immune system in this setting. This misplaced belief has been overtaken by a compelling awareness not only of the key role played by the immune system in anti-tumour host defence but also of the capability of human malignancies to subdue and subvert innate and adaptive cellular immune mechanisms. In this context, the most significant trigger driving the development of the field of clinical oncoimmunology was the discovery that co-inhibitory and co-stimulatory ICMs were the key regulators of cellular immune reactivity. Co-inhibitory ICMs include programmed cell death protein-1 (PD-1), programmed cell death protein ligand-1 (PDL-1), cytotoxic T lymphocyte-associated antigen-4 (CTLA-4), T cell immunoglobulin and mucin-domain containing-3 (TIM-3), lymphocyte activation gene-3 (LAG-3), B and T lymphocyte attenuator (BTLA) and herpesvirus entry mediator. Prominent co-stimulatory ICMs include CD27, CD40, CD134 (OX40), glucocorticoid-induced tumor necrosis factor receptor (TNFR) related protein (GITR), inducible T cell co-simulator and CD1371.

Co-inhibitory checkpoint ICMs are expressed most prominently by T cells, especially regulatory T (Treg) cells and antigen-presenting dendritic cells, as well as by other cell types, including cells of the innate and adaptive immune systems, structural cells in the tumour microenvironment (TME) and, somewhat ominously, by cancer cells per se2. The primary function of co-inhibitory ICMs is to downregulate T cell activation. This immunoregulatory function of co-inhibitory ICMs is essential in maintaining immune homoeostasis, thereby preventing the loss of peripheral self-tolerance that underpins the pathogenesis of autoimmune disease. Tumour cells, on the other hand, utilize the expression of co-inhibitory ICMs, such as CTLA-4 and PDL-1 in particular, as a strategy to outmanoeuvre anti-tumour host defences. This results in interference with the recruitment of cytotoxic T cells to the TME, as well as the inhibition of their anti-tumour activity, enabling immune evasion3.

The co-inhibitory ICMs, PD-1 and CTLA-4, were discovered by Tasuku Honjo4 and James P. Allison5 in 1992 and 1996, respectively, for which these renowned medical scientists were awarded the 2018 Nobel Prize for Physiology and Medicine. These discoveries provided the impetus for the development and pre-clinical/clinical evaluation of the first-generation ICM-targeted therapeutic mAbs, specifically ipilimumab (blockade of CTLA-4) and pembrolizumab (blockade of PD-1) that heralded new standards in the care of various types of malignancies1. This was followed by FDA approval of these and other subsequently engineered, novel co-inhibitory ICM-targeted ICMs, including nivolumab, cemiplimab, durvalumab, atezolizumab and avelumab1. Cancers responsive to these various co-inhibitory ICM-targeted therapeutic ICMs include lung cancer (small and non-small cell types), malignant melanoma, bladder cancer, Hodgkin’s disease, renal cell carcinoma and others1. Of note, however, is the issue that although anti-cancer immunotherapy with co-inhibitory ICM-targeted mAbs is associated with durable responses in some patients, many patients fail to respond to these therapeutic modalities, with some malignancies remaining refractory to these treatments. Although of concern, these limitations of co-inhibitory ICM-targeted immunotherapy may be overcome by the discovery of additional types of ICM with different mechanisms of induction of immunosuppression, raising the potential option of combination of immunotherapy with mAbs, which target different types of ICM. These novel co-inhibitory ICMs currently under pre-clinical/clinical evaluation include research, TIM-3, LAG-3, V-domain Ig-containing suppressor of T cell activation and BTLA1.

In a recently reported study, treatment of patients with previously untreated metastatic or unresectable melanoma with the combination of an anti-LAG-3-targeted mAb (relatlimab) with an anti-PD-1-targeted agent (nivolumab) was associated with a significant prolongation of progression-free survival relative to that observed with nivolumab alone6. Notably, the combination of both agents showed no new safety signals6.

In addition, as a strategy to improve efficacy and outcomes, numerous clinical studies are currently investigating the therapeutic potential of combinations of ICM-targeted therapy with conventional anti-cancer treatment modalities such as chemotherapy and radiotherapy and novel targeted therapies and anti-angiogenic agents (small molecules or mAbs). In this context, several different types of combination treatment are currently approved and in routine use in clinical settings, such as non-small cell lung cancer (NSCLC), renal cell cancer and triple-negative breast cancer (TNBC). Some of these novel combination treatments are currently used in the metastatic malignant melanoma and the neoadjuvant settings, potentially translating into curing some of these patients7.

Several biomarkers have been identified in recent years, which improve the selection of patients for ICM-based treatments. For example, PDL-1 testing using immunohistochemical assays is currently the most widely utilized, validated and accepted biomarker to select patients for treatment with anti-PD-1 or anti-PDL-1-targeted mAbs. A challenge for the clinical use of these assays is the necessity of using different companion diagnostic assays for specific ICM agents8.

In this context, it is essential to point out that the response to ICM-targeted mAbs differs depending on the numbers and composition of cells in the TME. Among responders to ICM-targeted ICMs, tumours have a high neoantigen load, high numbers of tumour-infiltrating lymphocytes (TILs), particularly effector cells, and an increased T effector cell: Treg ratio favouring secretion of interferon-γ, as well as low levels of Tregs and myeloid-derived suppressor cells (MDSCs)9. In non-responding patients, the TME is associated with high infiltration of immunosuppressive cells, such as Tregs, tumour-associated macrophages and MDSCs, in the setting of low numbers of natural killer (NK) cells and activated lymphocytes10.

Co-inhibitory ICMs, including CTLA-4 and PD-1, are present on activated T cells and PD-1 on antigen-presenting dendritic cells. These interactions lead to T cell exhaustion, defined as a state of T cell dysfunction, most commonly seen in chronic infections and cancer. T cell exhaustion is characterized by poor effector function, sustained expression of inhibitory receptors and a transcriptional state distinct from functional effector or memory T cells. Exhaustion is associated with suboptimal control of infection and tumours. These abnormalities can be eliminated by targeting these inhibitory checkpoints with mAbs, leading to the re-activation of T cells targeting cancer cells through the secretion of effector cytokines and cytotoxic granules10.

Three major categories of TME have been characterized across various tumour types: immune-desert (cold tumours largely devoid of lymphocytes), immune-excluded (lymphocytes are present in the peri-tumoural stroma only) and immune-infiltrated/inflamed (hot tumours)11. Immune-inflamed tumours have a higher response to ICM-targeted immunotherapy than cold tumours or immune-excluded tumours. Reports suggest that inflamed immune tumours have higher sensitivity because therapeutic ICM-targeted mAbs inhibit immune evasion12. Studies have demonstrated that favourable responses to ICI-based treatments are associated with the presence of TILs and other anti-tumour immune cells in the TME12.

In breast cancer (BC), the molecular subtype of the tumour has a substantial impact on its interaction with the immune system. In this context, higher numbers of TILs more frequently infiltrate TNBC and HER2-positive BC than hormone receptor-positive tumours13.

Tumour mutational burden (TMB) is defined as the number of mutations per DNA megabase and was first evaluated as a biomarker for co-inhibitory ICM-based immunotherapy following observations of favourable responses in tumours with high TMBs, including NSCLC, melanoma and bladder cancer. Pre-clinical data indicated that the association between TMB and the efficacy of ICM-based immunotherapy could be explained by the expression of neoantigens that increase tumour immunogenicity and responses to ICM-based treatments14. Neoantigens may develop during DNA replication via DNA mismatch repair (MMR). MMR is a physiological mechanism that identifies and repairs erroneous insertion, deletion and misincorporation of bases that appear during DNA recombination and DNA damage15. In mammalian cells, five MMR proteins have been identified. These proteins include MutS-homologs 2, 3 and 6, MutL-homolog 1 and post-meiotic segregation 216. Some cases of colorectal cancer are associated with microsatellite instability (MSI), characterized by elevated rates of small indels (insertion/deletions) and point mutations of one to six or more base pairs described as microsatellites17.

In metastatic colorectal cancer, mismatch repair-deficient/microsatellite instability (dMMR/MSI) represents a small subset of approximately five per cent and is associated with a prognosis18. In this patient population, the combination of the co-inhibitory ICM-targeted mAbs, nivolumab and ipilimumab, was investigated in 119 dMMR/MSI-H metastatic colorectal cancer patients in the CheckMate 142 trial. The treatment was associated with a consistent clinical benefit manifested as an objective response rate of 55 per cent19. Responses were durable, with a progression-free survival rate of 71 per cent and overall survival of 85 per cent after one year. Another group of metastatic colorectal patients responsive to ICM-based treatment are those who have tumours known as hypermutated tumours. This condition has been described as tumours harbouring polymerase proofreading domain mutations20. In addition, MSI can be found in other types of malignancy, including gastric cancer, endometrial cancer and many other types of cancer21. Despite these advances, novel biomarkers are urgently needed to improve patient selection, assess prognosis and monitor treatment. In this context, biomarkers should ideally be accessed via minimally invasive procedures, using matrices such as plasma or liquid biopsies22.

Co-inhibitory ICM-based treatments are generally well tolerated and are significantly less toxic than standard chemotherapy regimens. Nevertheless, co-inhibitory ICM-targeted ICMs have side effects related to their lack of specificity, resulting in over-activation of the immune system. These side effects are called immune-related adverse events (IrAEs) and include fatigue, dermatological, gastrointestinal, hepatic, pulmonary, endocrine, ophthalmic, neurological and unusual toxicities such as type 1 diabetes mellitus and cardiac haematological disorders23. Dermatological side effects are frequently present and manifest as skin rashes that are often maculopapular and mild23. Severe skin toxicities such as Stevens–Johnson syndrome or toxic epidermal necrolysis are seen in only a minority of patients23. Vitiligo also occurs in only a few patients receiving ICM-targeted mAbs; interestingly, this IrAE is commonly associated with clinical benefit and long-term survival23. Diarrhoea and enterocolitis are the most commonly encountered gastrointestinal toxicities. Rarely, in severe cases, these complications can be associated with toxic megacolon and perforation23. Endocrine IrAE symptoms are non-specific and include fatigue, mental status changes, headaches and dizziness, with hypothyroidism being the most frequent endocrine abnormality. Other endocrinopathies such as hypophysitis or adrenal insufficiency occur more often with ipilimumab. Type 1 diabetes, although uncommon, may present with severe acute symptoms23. Other, less frequently occurring IrAEs include ophthalmological conditions such as episcleritis, uveitis or conjunctivitis23. Neurologic IrAEs include myasthenia gravis, aseptic meningitis, encephalitis, motor and sensory neuropathies, including Guillain–Barre syndrome, and other rare conditions such as enteric or autonomic neuropathies and transverse myelitis23. Musculoskeletal IrAEs include inflammatory arthritis, myositis, polymyalgia rheumatic and osteitis23.

Although the pathogenesis of co-inhibitory ICM-targeted immunotherapy-mediated IrAEs remains unknown, several immune mechanisms have been implicated. These include the emergence/activation of auto-reactive T cells and B cells/autoantibodies, resulting in tissue damage due to complement activation and production of pro-inflammatory cytokines/chemokines, possibly related to changes in the intestinal microbiome. In this context, auto-reactive T cells (CD4+ and CD8+) play a role in the development of co-inhibitory ICM-targeted mAb-induced fulminant myocarditis, myositis and diabetes mellitus24. Moreover, the presence of T cell infiltrates has been demonstrated in cases of colitis, some dermatological complications, nephritis, sicca syndrome, liver injury and pneumonitis24, while dysregulation of Th1 and Th17 cell numbers and reactivity has been suggested as a potential mechanism for co-inhibitory ICM-targeted mAb-induced colitis24.

Das et al25 showed a reduction in the numbers of total circulating B cells in the setting of an increase in a CD21^lo PD-1+ B cell sub-population, as well as the presence of plasmablasts that preceded and was correlated with the frequency and timing of adverse events in melanoma patients undergoing combination ICM-targeted treatment.

There is no clear relationship regarding the role of auto-antibodies in the pathogenesis of IrAE-related thyroid dysfunction. In this setting, auto-antibodies (thyroglobulin and/or thyroid peroxidase antibodies) are detected in only 18-70 per cent of patients, whereas in Hashimoto’s thyroiditis, these are present in 90-95 per cent of patients26. Limited evidence exists that thyroid auto-antibodies at baseline may increase the risk of thyroid dysfunction following the administration of ICM-targeted immunotherapy. With respect to the possible involvement of auto-antibodies in the pathogenesis of co-inhibitory ICM-targeted immunotherapy-induced inflammatory arthritis, it is noteworthy that these patients are usually seronegative27.

In the case of IrAE-related hypophysitis, the proposed mechanism leading to pituitary cell destruction involves the binding of anti-CTLA-4 mAbs to CTLA-4 expressed on pituitary cells, with resultant complement activation, macrophage/phagocyte infiltration, enhanced antigen presentation and infiltration of B and T cells and destruction of pituitary cells28.

The gut microbiome has been proposed to play a role in both the anti-tumour therapeutic efficacy of ICM-targeted mAbs and the development of IrAEs in cancer patients undergoing this type of immunotherapy. In the case of melanoma patients undergoing treatment with anti-CTLA-4-targeted mAbs, those patients whose baseline gut microbiomes were enriched for Faecalibacterium and other Firmicutes had a better response to treatment and more prolonged progression-free survival in the setting of more frequent colitis. In contrast, those enriched with Bacteroidetes had less frequent colitis29.

An additional largely unexplored challenge is represented by the potential threat of the development of severe IrAEs posed to cancer patients by the presence of underlying autoimmune disorders. For this reason, cancer patients with pre-existing autoimmune diseases were not included in the early co-inhibitory ICM-based clinical trials. Although the risk of development of IrAEs or exacerbation of their underlying autoimmune disease is not well understood in these patients, there is evidence that the onset of IrAEs may occur faster in these patients30. Accordingly, ongoing close monitoring of these patients early in the treatment course is necessary30.

Professional organizations such as the American Society of Clinical Oncology31, the National Comprehensive Cancer Network31, the European Society for Medical Oncology32, the Society for Immunotherapy of Cancer33 and the Multinational Association of Supportive Care in Cancer34 have established guidelines and recommendations for the management of IrAEs. However, it must be highlighted that these guidelines and recommendations are primarily based on clinical recommendations from experience gained from clinical trials, daily clinical practice and general clinical consensus. Although these management recommendations are widely accepted, the level of evidence is low, and currently, no prospective, randomized trials have been undertaken to evaluate whether one treatment strategy is superior to another34.

Early recognition of IrAEs and pro-active management by clinicians remain critical to lower the morbidity and mortality associated with co-inhibitory ICM-targeted immunotherapy in cancer patients. IrAEs are usually manageable low-grade, particularly with single-agent co-inhibitory ICM-targeted mAbs. However, it is essential to point out that reporting of IrAEs outside of a clinical trial setting remains suboptimal15. A multidisciplinary team of specialists is required to manage patients diagnosed with severe IrAEs.

Anti-CTLA-4- and anti-PD-1- or anti-PDL-1-targeted mAbs have different mechanisms of action, and this has been clearly documented in the numerous clinical trials examining combination therapies, which have shown that the occurrence of grade 3 and grade 4 IrAEs in cancer patients receiving the combination of ipilimumab and nivolumab was approximately 50 per cent. This was significantly higher than either agent administered individually and resulted in treatment disruption and terminations in approximately one-third of patients35.

Clinicians, nurses and healthcare professionals who administer anti-cancer immunotherapeutic antibodies should be aware of possible new toxicities and side effects that can be severe and unpredictable. In addition, these antibodies are being increasingly administered in combination with chemotherapy, radiation therapy, targeted therapies or other agents; therefore, the incidence and severity of these toxicities may change in the future. These variations in side effect patterns will require a regular update of the various guidelines and recommendations to attain better IrAEs management. In addition and somewhat paradoxical, there is a growing body of evidence that patients who developed IrAEs are associated with a better outcome36. Biomarkers to individualize and categorize patients at risk of severe IrAEs would be clinically helpful.

Finally, co-inhibitory ICM-based treatments undoubtedly represent a significant breakthrough in the treatment of cancer patients. Such advances in treatment, however, come at high financial costs. The estimated cost of one year of co-inhibitory ICM-based treatment is approximately 100,000 USD37. Financial toxicity, therefore, remains the main barrier to access to care for our patients requiring co-inhibitory ICM-based treatment.