Cancer therapy-induced dermatotoxicity as a window to understanding skin immunity

INTRODUCTION

The advent of targeted molecular therapy and immunotherapy in clinical oncology heralded explosive growth of data from the clinic that would shed new light on basic biology and accelerate research progress in molecular medicine. Conventional cancer therapies, which rely on nonspecific cytotoxic chemicals and radiation, generated limited insight into complex molecular pathways operating in cells and tissues and producing physiologic and pathologic outputs. Human biology could be approached and understood only through extrapolating data from preclinical model systems where sophisticated genetic tools for functional analysis were available. These limitations impeded the exploration of exactly how molecules and cells function in human tissues and the generation of mechanistic insights directly relevant to human health and disease. Unlike conventional oncology drugs, small-molecule inhibitors of signaling enzymes and monoclonal antibodies blocking ligand-receptor interactions—the main staples of targeted molecular therapy and immunotherapy—exert their pharmacological effects by creating a highly specific loss-of-function or gain-of-function state in vivo, akin to gene knockout and transgene expression in mice. Not only do the intended pharmacological actions of these advanced therapeutics provide meaningful insights, but their unintended adverse effects, if on-target, also reveal precious information regarding the physiologic role of the targeted molecules in the tissues where toxicity is observed.

The skin is a large organ that serves as the interface between the body and the environment. It protects against physical and chemical threats and microbial pathogens. This essential function depends on the interplay between the epidermal barrier structure and the immune system. Epidermal barrier dysfunction or impaired immunity leads to diverse forms of skin pathology. Cancer therapies are designed to suppress tumor growth or enhance antitumor immunity, but many (almost all) of those developed to date also interfere with molecular processes crucial for epithelial or immune homeostasis in the skin, producing on-target dermatotoxicity. In this review, we examine various forms of dermatotoxicity occurring in such scenarios, portray newly emerging mechanistic insights coming from their analysis, and present an updated conceptual framework of how skin immunity operates. This knowledge will also help devise novel clinical strategies for the detection, prevention, and treatment of therapy-induced dermatotoxicity in cancer patients. For detailed information on the epidemiology, clinicopathological features, and management of cancer therapy-induced dermatotoxicity, we refer readers to comprehensive expert reviews focused on these subjects (1–4).

THREE TELLING CASES IN CANCER THERAPY-INDUCED DERMATOTOXICITY

Cutaneous adverse events caused by immune checkpoint inhibitor therapy

CTLA-4, PD-1, and other immune checkpoint molecules play a crucial role in the control of adaptive immunity, restraining T cell activation in lymphoid and non-lymphoid tissues. CTLA-4 is expressed on antigen-encountered T cells and intercepts costimulatory signals from antigen-presenting cells, thereby limiting antigen-specific T cell priming (5). PD-1 is highly expressed in T cells undergoing chronic antigen stimulation in peripheral tissues and interacts with its ligands, PD-L1 and PD-L2, to induce a dysfunctional T cell state (6). Monoclonal antibodies blocking CTLA-4, PD-1, and PD-L1 are major immune checkpoint inhibitors (ICIs) used in the treatment of a growing number of cancer types. In addition, new ICIs, including a LAG-3 inhibitor, have become available for clinical use or are under development, expanding the arsenal of cancer immunotherapy (7).

Immune-related adverse events (irAEs) occur in a substantial population of patients receiving all known regimens of ICI monotherapy and combination therapy. irAEs are diverse in etiology and affect a variety of organs (8). Being barrier tissues with continuous exposure to environmental immune stimuli and large immune cell pools in residence and transit, the skin and intestinal mucosa are the two most common sites where ICI-induced immunopathology is manifested. Cutaneous irAEs (cirAEs) are the most frequent type of ICI toxicity and develop early relative to other types of irAEs (3). cirAE development has been consistently shown to correlate with patient response to ICI and improved survival (9–11). ICI-induced dermatotoxicity ranges from pruritus and rash to leukoderma and alopecia and, albeit rare, also includes severe forms such as Stevens-Johnson syndrome and toxic epidermal necrolysis. Skin rashes in ICI-treated patients are heterogeneous in clinicopathological characteristics and categorized into distinct forms of dermatitis, including maculopapular, lichenoid, psoriasiform, eczematous, and bullous eruptions. Despite this broad spectrum of manifestations, ICI-induced cutaneous adversities likely share some mechanistic elements ultimately traceable to aberrant T cell activation, given that ICIs are designed to unleash T cell responses and activate T cell effector mechanisms. We revisit ICI-induced cirAEs and discuss their underlying mechanisms in upcoming sections of this review.

Cutaneous adverse events caused by macrophage-targeting cancer therapy

Macrophages exist in multiple subtypes and serve subtype-specific functions related to immunity and tissue maintenance. Tumor-associated macrophages (TAMs) contribute to both tumor growth and antitumor immunity depending on their functional state and interactions with other components of the tumor microenvironment. A variety of macrophage-targeted therapies have been developed and are tested for their efficacy against certain types of cancer. These therapies target molecular pathways essential for macrophage survival, recruitment, phagocytic activity, functional polarization, and interplay with adaptive immunity (12). Macrophage-targeting agents are often combined with ICIs and other pharmacological agents in clinical settings, making it difficult to discern macrophage-specific study outcomes. Some agents developed for macrophage-targeted therapy exert their effects on non-macrophage targets as well. A few clinical trials, however, investigated therapeutics specifically targeting macrophages and examined their performance as monotherapy, revealing therapeutic and toxic effects most likely attributable to macrophage biology.

CSF1R, the receptor for macrophage-colony stimulating factor, transduces signals essential for macrophage development and survival. Intriguingly, preclinical studies found that CSF1R inhibition produces antitumor effects mainly by shifting TAM phenotype to an antitumor state rather than achieving complete TAM depletion (13, 14). The early promise of CSF1R inhibitors has led to numerous clinical trials on a variety of tumor types, but these studies, which evaluated CSF1R inhibition as monotherapy or its combination with ICI or CD40 agonist treatment, have shown only marginal improvements in patient response (15–21). Another macrophage-targeting strategy that has entered clinical trials seeks to promote macrophage phagocytosis through disruption of the CD47-SIRPα axis (22–24). Preclinical studies have shown that a range of CD47-SIRPα signaling inhibitors can promote cancer cell clearance by phagocytosis and enhance tumor immunogenicity in mice (25–27). Macrophage-targeted therapies have been shown to cause a wide range of adverse effects including those associated with the skin. Pruritus, edema, and xerosis (dry skin) are common adverse events among cancer patients treated with antibody or small-molecule inhibitors of CSF1R (15, 18, 19, 21) and engineered CD47-blocking proteins (24). These skin manifestations may be secondary to cutaneous or systemic inflammatory responses. Alternatively, perturbation of macrophage function may contribute directly to changes in sensory nerve activity and epidermal and vascular permeability in the skin.

Cutaneous adverse events caused by EGFR/MEK inhibitor therapy

The RAS-RAF-MEK-ERK pathway is activated by EGFR and other growth factor receptors and transduces intracellular signals for cell proliferation during tissue development and regeneration. This growth factor signaling cascade is constitutively activated by mutations in a wide range of cancer types and becomes a prime target for intervention. In addition to functioning as oncogenic drivers in cancer cells, EGFR and MEK appear to serve functions essential for skin barrier integrity and antimicrobial defense. Cancer patients treated with EGFR and MEK inhibitors display similar and closely related cutaneous toxicities in the first week of treatment initiation, including papulopustular rash in the trunk and face, pruritus, xerosis, folliculitis, and paronychia (28, 29). Both antibody and small-molecule inhibitors of EGFR cause these adverse events, suggesting on-target dermatotoxicity. EGFR inhibitor-induced dermatotoxicity is dose-dependent and correlates with treatment effectiveness (28, 29). Treatment with EGFR and MEK inhibitors results in drastic changes in the skin immune landscape and altered gene expression, which could be attributed to and downstream of skin barrier dysfunction and microbial stimulation. We discuss in detail the mechanisms linking EGFR/MEK-dependent epidermal barrier function to skin immune homeostasis in the subsequent section of this review.

MECHANISMS DRIVING CANCER THERAPY-INDUCED DERMATOTOXICITY

Pathogenic T cell states resulting from impaired immune tolerance

Developing T and B lymphocytes bearing receptors specific for self-antigens are clonally eliminated in a process referred to as central tolerance. Small fractions of self-reactive lymphocytes that have escaped this purge and exited generative lymphoid organs are subjected to additional control in peripheral tissues, which induces a state of unresponsiveness to antigen stimulation and attenuated effector function. CTLA-4 and PD-1 play a role in peripheral tolerance, mediating cell-autonomous as well as regulatory T cell-dependent mechanisms suppressing T cell activation. These immune checkpoint functions are essential for preventing autoimmunity and uncontrolled immune responses to foreign antigens from innocuous sources. ICIs enhance T cell priming against tumor antigens and steer tumor-infiltrating T cells (TILs) toward a state of augmented antitumor effector function, yet they also appear to break peripheral T cell tolerance and disrupt immune homeostasis, permitting self-destructive inflammatory responses (Figure 1).

Figure 1. Immune checkpoint inhibition unleashes T cell responses and drives T cell- and cytokine-mediated dermatotoxicity.

CTLA-4 inhibition enhances antigen-specific T cell priming, whereas PD-1 inhibition repositions T cells for antigen access and unlocks their effector potential. CTLA-4 and PD-1 inhibitor treatment causes cutaneous adverse events by mobilizing T cells reactive to self-antigens and commensal bacterial antigens.

Multiplex immunofluorescence imaging and spatial transcriptomics analysis have been performed of skin lesions resulting from ICI treatment (mostly PD-1 inhibitor monotherapy); this study revealed increased CD4 and CD8 T cell abundance in maculopapular, lichenoid, and bullous lesions and detected gene expression signatures of tissue-resident T cell memory (CD103, CD69) and IFNγ/TNF signaling (IFN-γ, IL-15, TNF, CXCL9/10/11) as immune features common to the three clinical types of dermatotoxicity (30). These findings are in part consistent with what was observed in patients with ICI colitis: the single-cell transcriptomics analysis performed in two independent studies captured IFNγ- and cytolytic enzyme-producing colonic CD8 T cells as the most prominently expanded T cell cluster during ICI-induced inflammation and provided evidence suggesting tissue-resident memory T cells as the precursor of these pathogenic effector T cells (31, 32). Taken together, these data from ICI dermatitis and colitis call attention to the need for deeper research into T cells committed to tissue residency and IFN-γ-dependent effector function as a major responder to ICI treatment in barrier organs and a driver of irAEs.

Immune checkpoint molecules appear to hold antigen-experienced T cells in check and allow them to silently co-exist with their cognate antigens in peripheral tissues, a condition rendered precarious by ICIs. This idea was tested in a study that used mice in which transgenic expression of an experimental antigen was induced in adult skin and subsequently led to the mobilization of antigen-specific CD8 T cells (33). The mouse line used in this study was designed to exert multiple layers of genetic and pharmacological control over the transgene such that its expression would be switched on in skin areas exposed to a topically administered chemical inducer while its leaky expression in the thymus (hence central tolerance to the antigen expressed from the transgene) would be prevented. The experimental antigen, mainly detected in the epidermis upon induction, elicited skin infiltration of CD8 T cells but did not cause overt skin pathology; these T cells expressed checkpoint inhibitor molecules but did not actively produce cytokines. This local antigen expression precipitated lichenoid dermatitis, however, when combined with administration of anti-PD-1 antibody; in this pathologic condition, antigen-specific T cells, which mostly remained in the dermal area in the absence of ICI treatment, penetrated across the dermoepidermal junction, accessed antigen-expressing epidermal cells, and underwent differentiation into fully mature effector cells producing IFN-γ, TNF, and cytolytic enzymes (33). These findings shed light on how immune checkpoint molecules keep T cells inactive upon antigen encounter in peripheral tissues and point to the T cell states and effector pathways mediating irAEs in barrier tissues. Pathogenic T cells driving cirAEs likely react to self-antigens produced from skin cells or skin microbiota-derived antigens. The source of cirAE-associated antigens as well as the type of immunity directed to them are critical determinants of the specific form of the resulting pathology.

Antigens shared between cancer cells and normal skin structures

The immune response to neoantigens, which arise from cancer-specific DNA, RNA, and protein alterations, is the mechanistic underpinning of ICI therapy, but the occurrence of vitiligo-like depigmentation (VLD) in ICI-treated melanoma patients hints at an additional contribution of antigens common to melanoma cells and normal melanocytes (Figure 2). Vitiligo is a group of autoimmune disorders where melanocytes are targeted for destruction by cytotoxic T cells. The prevailing mechanism for vitiligo posits a sequence of events that begins with innate immune sensing of stressed melanocytes, proceeds through the formation and skin infiltration of melanocyte-reactive CD8 T cells producing IFN-γ, and ends with T cell killing of melanocytes (34). This pathogenic process is amplified by IFN-γ-stimulated keratinocytes, which produce the chemokines CXCL9/10 and thereby recruit additional T cells (35, 36), and sustained by tissue-resident memory T cells, which depend on IL-15 for persistence in lesional skin (37). T cells recognizing melanocyte differentiation antigens (MDAs)—pigmentation-related proteins expressed in highly differentiated melanocytes such as Melan-A (also known as MART-1), tyrosinase, and gp100—have been detected in vitiligo patients (38, 39) as well as in melanoma patients who responded to ICI therapy (40, 41). Studies have identified T cell receptor (TCR) clonotypes shared between patient-matched melanoma and VLD lesions (42, 43). Research performed in mouse models has shown that melanoma immunotherapy, when effectively controlling tumors through neoantigen-directed immune responses, leads to an increase in MDA-specific TILs and memory T cells, immune-mediated depigmentation, and protection against newly engrafted MDA-expressing tumors with a low neoantigen load (41, 44, 45), suggesting a widened repertoire of antigens associated with T cell immunity. This epitope spreading may be attributable to either de novo formation of MDA-specific T cells or a drastic expansion of their pools from pre-existing precursors. In addition to MDAs, immune attack of melanocytes in ICI-treated patients may involve other self-antigens expressed across multiple tissues given that VLD has been reported to develop in non-melanoma cancer patients (46–48).

Figure 2. T cells and antibodies recognizing antigens shared between cancer cells and normal skin cells drive autoimmune-mediated cutaneous adverse events.

Melanoma and non-small cell lung cancer (NSCLC) may elicit autoreactive T cells and autoantibodies that attack melanocytes and keratinocytes.

Non-small cell lung cancer (NSCLC) and normal skin exhibit a substantial degree of overlap in their transcriptomes with many predicted T cell antigens co-expressed in both tissues (Figure 2). An analysis of ICI-treated NSCLC patients revealed that some of these shared antigens, including keratin-6, keratin-14, desmocollin-3, maspin (also known as serpin-B5), and LL37 (also known as cathelicidin), were associated with cirAE incidence; some TCR clonotypes of peripheral blood T cells reactive to these antigens were also detected among the TILs and lesional skin T cells of these patients (49). A study performed under a similar design of sampling and analysis found that other proteins common to NSCLC tumors and normal skin, such as BP180 (also known as the α1 chain of collagen-7), served as B cell antigens: titers of serum immunoglobulin (Ig)-G specific for BP180 were associated with the rate of cirAEs and predictive of ICI response and prolonged survival (50). Of note, although autoantibodies against BP180 are best known as a pathogenic factor of bullous pemphigoid with their ability to promote skin blister formation well established, the form of dermatotoxicity with which BP180-specific IgG titers showed a correlation in this study was lichenoid dermatitis. This finding does not necessarily indicate a causal role for BP180-specific IgG in the formation of lichenoid lesions; this autoantibody may arise as a byproduct of immune dysregulation in ICI-treated NSCLC patients, yet still has the potential to serve as a predictive marker for cirAE development.

Aberrant neuroimmune interactions and sensitized pruritic neural circuits

Pruritus is common among cancer patients treated with ICIs, CSF1R inhibitors, CD47 blockers, and EGFR/MEK inhibitors. The mechanisms by which these therapeutics induce itch have yet to be understood. Itch is mediated by a subset of unmyelinated C-fibers in the skin. These pruriceptive nerves relay signals to the spinal cord and brain to evoke itch and trigger scratching (51). Immune cell- and keratinocyte-derived cytokines, derivatives of amino acids and arachidonic acids, neuropeptides, and proteases activate or sensitize pruriceptive nerves, thereby functioning as endogenous pruritogens (Figure 3).

Figure 3. Cancer therapy activates neuroimmune signaling and stimulates pruritic neural circuits.

Keratinocytes and immune cells produce various endogenous pruritogens and stimulate pruriceptive sensory neurons, a subset of dorsal root ganglion (DRG) neurons relaying itch signals to the central nervous system.

Skin biopsies and serum from patients with ICI-induced eczematous rash and bullous pemphigoid often exhibit histological and molecular features associated with pruritus, such as eosinophilia (52, 53) and elevated production of IgE (52) and IL-4/13 (54, 55); high percentages of these patients respond to omalizumab and dupilumab, antibodies neutralizing IgE and blocking IL-4Rα, respectively, showing a significant improvement in pruritis and rash (52, 53, 55, 56). IL-4Rα, a chain of the receptor that binds to both IL-4 and IL-13, is expressed in a multitude of cell types and serves signaling functions related to barrier tissue inflammation and allergic reactions. Intriguingly, IL-4Rα is also highly expressed in the subtypes of sensory neurons crucial for itch and has been shown to mediate IL-4/13-dependent sensitization of skin nerves to various pruritogens (57). There are other cytokines known to exert direct action on sensory neurons and trigger or sensitize to itch, the most notable of which are thymic stromal lymphopoietin (58), IL-31 (59), and oncostatin-M (60). Multiplex cytokine profiling has captured IL-31 in elevated amounts in the skin of ICI-treated pruritic patients (61). Whether IL-31 and other pruritogenic cytokines are produced in amounts sufficient to induce itch in cirAE settings remains to be rigorously studied.

Sensory neurons innervating the skin have been reported to express PD-1 and PD-L1; studies of neuronal PD-1 and PD-L1 signaling in mice have shown its ability to modulate the excitability of nociceptive and pruriceptive neurons and inhibit pain and itch (62–67). These studies yielded results that were incongruous with one another in mechanistic details but raised the interesting possibility that ICIs might cause pruritus by acting directly on sensory neurons and disinhibiting pruriception rather than through mechanisms secondary to inflammation and mediated by neuroactive cytokines produced from non-neuronal cells.

Pruritus is common and often the most frequent type of adverse event among cancer patients treated with macrophage-targeting agents, including CSF1R inhibitors (15, 18, 19, 21) and CD47 antagonists (24). The dorsal root ganglion (DRG) contains the somata of pruriceptors and other sensory neurons innervating the skin as well as immune cells including macrophages. Subsets of DRG- and skin-resident macrophages patrol neuronal somata and nerve fibers for surveillance and contribute to their regeneration after tissue injury (68, 69). Furthermore, these macrophages have been shown to modulate neuronal gene expression and metabolism to set the threshold for and terminate sensory stimulation (70, 71) as well as produce IL-31 and other pruritogens (72–74). CSF1R inhibitors, CD47 blockers, and other macrophage-targeting agents may cause pruritus by depleting or functionally disrupting macrophages in the DRG and skin. A better understanding of the nature of these sensory anomalies requires studies that delve deeper into the mechanisms governing macrophage-neuron interactions and macrophage-dependent maintenance and functional modulation of pruriceptive neurons.

Epithelial barrier dysfunction and immune dysregulation linked to skin and gut bacteria

Acneiform eruptions and bacterial superinfection frequently occur in cancer patients treated with EGFR and MEK inhibitors (28, 29). Staphylococcus aureus is one of the bacteria frequently isolated from EGFR/MEK inhibitor-induced skin lesions (75–77). Studies that generated and characterized mice with keratinocyte-restricted EGFR gene ablation found that these mice exhibited increased epidermal permeability (as extrapolated from a higher rate of transepidermal water loss) and a signature of skin gene expression indicative of a defective skin barrier (77, 78). Skin barrier dysfunction in mice lacking epidermal EGFR expression was accompanied by the influx of mast cells, macrophages, eosinophils, neutrophils, and T cells in lesional skin (Figure 4); among these immune cell types, macrophage function and mast cell degranulation were found to contribute to skin pathology, as macrophage depletion and inhibition of histamine signaling with pharmacological agents suppressed abnormal epidermal differentiation and T cell infiltration in skin lesions, respectively (78, 79). The outgrowth of bacteria in the skin of EGFR/MEK inhibitor-treated cancer patients likely activates microbial sensor-dependent innate immune signaling. Bacterial stimuli and EGFR/MEK inhibition have indeed been shown to act together to reshape keratinocyte gene expression, induce epithelial cytokines such as IL-36γ, and promote cutaneous neutrophilia (80).

Figure 4. EGFR/MEK inhibitors disrupts the skin barrier function and impairs antibacterial defense.

Intracellular signaling downstream of EGFR is essential for skin homeostasis and immune defense. Inhibition of this signaling results in barrier dysfunction, hair follicle infection, and immunopathology. Staphylococcus aureus is the common etiologic agent associated with EGFR/MEK inhibitor-induced cutaneous adverse events.

Commensal bacteria in healthy human skin perform functions beneficial to the host, such as controlling pathogenic microbial growth and training the skin immune system. T cells primed to respond to commensal bacterial antigens or innate-like T cells hard-wired to sense microbial metabolites are recruited to barrier tissues inhabited by the antigen/metabolite-producing microbes, yet these T cells persist there without triggering inflammation and pathology, a state referred to as homeostatic immunity (81). Immune checkpoint molecules are thought to play an essential role in keeping commensal-specific T cells locked in a poised state for activation and cytokine production and preventing their activation in the steady-state skin with an intact epidermal barrier. This idea was supported by a study that investigated mice whose skin was colonized with Staphylococcus epidermidis, a commensal inhabiting the human skin. S. epidermidis-colonized mice did not display perceptible skin pathology but developed severe dermatitis when administered anti-CTLA-4 antibody, which by itself did not induce inflammation (82). Skin disease resulting from S. epidermidis colonization in the absence of CTLA-4 function was driven by both CD4 and CD8 T cells and depended on IL-17A but not IFN-γ. Similar commensal-CTLA-4 inhibitor cooperation has been observed in mice harboring the gut microbiota from wild-caught mice; these mice developed colitis in a manner dependent on IFN-γ-producing CD4 T cells but not CD8 T cells or IL-17A/F-producing T cells (83). These findings show that ICIs can break T cell tolerance in barrier tissues and unleash inflammatory responses driven by T cells specific to self-antigens or antigens derived from innocuous commensals.

The gut microbiota, the largest microbial community in the human body, exerts local and systemic influences on physiology and pathology by interacting with the immune and neuroendocrine systems. Gut microbiota composition affects not only response to ICIs but also irAE incidence. Studies have captured several distinct taxa of the baseline gut microbiota as associated with irAEs although there is between-cohort and inter-study heterogeneity in the data and conclusions (84–87). These irAE-associated bacteria were identified at different taxonomic levels and included Lachnospiraceae (family), Streptococcus spp. (genus), and Bacteroides intestinalis (species). One of these studies detected specific taxa specifically associated with different types of irAEs including ICI dermatitis (87). Whereas gut bacteria appear to contribute to ICI colitis through inducing IL-1β and other pathogenic cytokines in the intestinal mucosa and altering the local immune landscape, it is unclear how they influence cirAE development and exert other systemic actions. Well-powered prospective studies are needed to generate mechanistic insights into this gut-skin axis of host-microbe interactions.

KNOWLEDGE GAPS AND FUTURE DIRECTIONS

Despite the recent progress in understanding the mechanistic underpinnings of cancer therapy-induced dermatotoxicity, several research gaps in the field remain to be closed. Advanced research tools for non-invasive specimen collection, single-cell molecular profiling, and spatial imaging should be leveraged to analyze skin lesions from patients developing cirAEs. Comprehensive and unbiased data from such analyses will uncover as-yet-unidentified molecular processes causally linked to cirAE mechanisms. Preclinical models that closely resemble pruritus, rash, VLD, and other manifestations of dermatotoxicity in humans will propel research into the underlying molecular mechanisms and help discover targets for the management of cutaneous adverse events. Severe dermatotoxicity often results in therapy interruption and discontinuation, preventing otherwise promising cancer therapeutics from realizing their full potential. Effective treatments tailored to individual forms of dermatotoxicity will improve toxicity management and maximize patient benefit. Another important question in the areas of targeted molecular therapy and immunotherapy is whether clinical strategies could be devised to suppress the toxicity of therapeutic agents while preserving their efficacy. A preclinical study showed that although anti-CTLA-4 therapy promoted both antitumor immune responses and intestinal inflammatory damage in mice, the latter (ICI colitis) depended on immunoglobulin Fcγ receptor signaling whereas this signaling function is dispensable for the former (tumor eradication); anti-CTLA-4 nanobodies, which lacked an Fc domain, could circumvent intestinal toxicity without sacrificing their antitumor effects (83). The management of cirAEs, and irAEs in general, in ICI-treated patients requires such strategies for dissociating antitumor immunity and treatment toxicity.

The use of targeted molecular therapy and immunotherapy for cancer treatment will continue to grow. Their armamentarium is expanding with new small-molecule therapeutics and biologics entering the clinical space. This progress will certainly push cancer treatment forward; it will also generate massive amounts of data that deepen our understanding of human biology and enable the development of more powerful technologies for improved patient care.

KEY POINTS.

• Therapy-induced dermatotoxicity reveals mechanisms governing skin immunity.

• Immune checkpoint inhibitors unleash T cell responses and trigger cytokine-driven dermatitis.

• Various cancer therapies alter skin neuroimmune interactions and cause itch.

• EGFR/MEK inhibitors disrupt the epidermal barrier and promote skin infection.

SYNOPSIS.

Pruritus, rash, and various other forms of dermatotoxicity are the most frequent adverse events among cancer patients receiving targeted molecular therapy and immunotherapy. Immune checkpoint inhibitors, macrophage-targeting agents, and EGFR/MEK inhibitors not only exert antitumor effects but also interfere with molecular pathways essential for skin immune homeostasis. Studying cancer therapy-induced dermatotoxicity helps us identify molecular mechanisms governing skin immunity and deepen our understanding of human biology. This review summarizes new mechanistic insights emerging from the analysis of cutaneous adverse events and discusses knowledge gaps that remain to be closed by future research.