A rogue population of immune cells, long overlooked in the complex landscape of the human gut, has been unmasked as a primary arsonist in the chronic intestinal firestorm known as Crohn’s disease. For decades, the immunology of inflammatory bowel disease has orbited around a simplified axis of helper T cells, with the pro-inflammatory Th1 and Th17 subsets cast as the central villains. But a groundbreaking study published today in Experimental & Molecular Medicine systematically dismantles that dogma, revealing a distinct lineage of CD8⁺ T cells that express the serine protease Granzyme K as the true architects of mucosal destruction. This discovery rewrites our cellular understanding of Crohn’s pathogenesis, shifting the spotlight from the adaptive immune system’s helper corps to a cytotoxic subset previously associated primarily with age-related inflammation, and in doing so, it unlocks a completely new therapeutic frontier. The research, led by a team of Korean scientists who meticulously profiled intestinal tissues from Crohn’s patients and healthy controls using multi-omics technologies, demonstrates that these Granzyme K-expressing CD8⁺ T cells are not just bystanders; they are highly activated, possess a tissue-resident memory signature that glues them to the gut lining, and, most critically, their secreted product, GzmK, triggers a cascade of inflammatory signals from intestinal epithelial cells that directly mimics the clinical pathology observed in patients.
The clinical conundrum of Crohn’s disease has always been its relapsing-remitting nature, punctuated by debilitating abdominal pain, severe diarrhea, fatigue, and weight loss, coupled with a destructive inflammation that can transmurally penetrate the entire bowel wall, leading to strictures, fistulas, and the frequent necessity of surgical resection. Current therapeutic strategies, ranging from broad-spectrum immunosuppressants like corticosteroids to more targeted biologics that neutralize tumor necrosis factor (TNF)-α or block leukocyte trafficking via α4β7 integrin, are notoriously imperfect. A significant fraction of patients exhibit primary non-response, and many more lose responsiveness over time due to immunogenicity or the activation of bypass inflammatory pathways. This therapeutic stalemate has long implied a missing piece of the pathogenic puzzle—an upstream cellular player or a parallel inflammatory circuit that remains untouched by existing drugs. The study by Lee and colleagues posits that the CD8⁺ T cell compartment, particularly a specific subset armed with GzmK, constitutes exactly that elusive missing link. Their data suggests that while helper T cells certainly contribute to the cytokine milieu, the direct tissue-damaging and epithelial-reprogramming effects of GzmK⁺ CD8⁺ T cells could be the primary event that initiates and perpetuates the vicious cycle of barrier disruption and microbial translocation.
To unravel this cellular mystery, the investigators harnessed the power of single-cell RNA sequencing (scRNA-seq) coupled with T cell receptor (TCR) sequencing, creating a high-resolution atlas of lamina propria lymphocytes isolated from endoscopically obtained biopsies of both inflamed and non-inflamed intestinal regions from Crohn’s patients, alongside samples from healthy individuals. This unbiased transcriptomic lens allowed the team to deconvolve the chaotic mixture of immune cells into discrete clusters, each defined by a unique genetic fingerprint. Among the anticipated populations of CD4⁺ Th1, Th17, and regulatory T cells, a distinct cluster of CD8⁺ T cells emerged with striking prominence in the inflamed tissues. This cluster was transcriptionally defined by high expression of GZMK, the gene encoding Granzyme K, alongside a suite of markers indicative of tissue residency such as ITGAE (CD103) and CD69, and the inflammatory chemokine receptor CXCR6. The cells were conspicuously absent of Granzyme B (GZMB), the major cytotoxic effector of classic killer T cells, suggesting their function was not primarily about inducing rapid apoptotic cell death but rather engaging in a more nuanced, protracted form of inflammatory signaling.
The functional identity of these cells was further deciphered through a focused analysis of their tissue-resident features. Tissue-resident memory T (TRM) cells are a non-recirculating population that permanently lodges in peripheral organs, providing a rapid, frontline defense against reinfection. However, when dysregulated, these cells can become potent drivers of local autoimmune and inflammatory conditions. The GzmK⁺ CD8⁺ T cells in the Crohn’s gut bore all the hallmarks of a pathogenic TRM population. Their clonal expansion, revealed by TCR sequencing, indicated a history of local antigen-driven proliferation within the intestinal microenvironment. They exhibited a core gene expression program that included the transcription factor Hobit (ZNF683), a master regulator of tissue residency, and RUNX3, a factor crucial for CD8⁺ T cell lineage commitment and retention in non-lymphoid tissues. This molecular machinery essentially programs the cells to stay put, relentlessly surveying the intestinal epithelium and, when triggered by a danger signal—perhaps a component of the dysbiotic microbiota or a dietary antigen—unleashing their inflammatory payload right at the mucosal interface.
Crucially, the team did not stop at transcriptomic inference; they validated their findings at the protein level using high-dimensional flow cytometry and immunohistochemistry, visually confirming the dense accumulation of GzmK⁺ CD103⁺ CD8⁺ T cells in the epithelial layer and the underlying lamina propria of Crohn’s lesions. The mere presence of these cells did not guarantee their culpability, so a series of elegant functional experiments were engineered to establish a causal chain. The researchers co-cultured human intestinal organoids—miniature, three-dimensional gut structures derived from primary epithelial stem cells—with recombinant Granzyme K protein. The result was nothing short of dramatic. The GzmK-exposed organoids did not simply die; instead, they adopted a profoundly inflammatory transcriptional program. The epithelial cells began to actively secrete neutrophil-recruiting chemokines such as CXCL1, CXCL5, and CXCL8 (IL-8), the latter being a potent chemoattractant that drives the massive influx of neutrophils which form the crypt abscesses, classic histological hallmarks of active Crohn’s disease. This finding redefines GzmK not as a simple pro-apoptotic executioner but as a molecular switch that reprograms epithelial cells into active participants in the inflammatory process, effectively bridging the cytotoxic T cell response to the innate immune granulocyte infiltration observed pathologically.
Digging deeper into the mechanistic basis of this epithelial reprogramming, the study elucidated the intracellular signaling pathway triggered by Granzyme K. Unlike Granzyme B, which cleaves caspases to trigger programmed cell death, GzmK proteolytic activity was shown to work in a distinct fashion, potentially entering the epithelial cell via a receptor-independent mechanism and cleaving specific intracellular substrates to activate the canonical NF-κB pathway. This pathway is the master transcriptional regulator of inflammation, and its activation led to the production of the neutrophil chemoattractants and other cytokines like IL-6 and IL-1β family members. This signaling circuit creates a self-amplifying inflammatory loop: the GzmK⁺ CD8⁺ TRM cells, positioned at the epithelial border, sense a stress signal and release GzmK, which then instructs the epithelial barrier to broadcast a “danger cry” that recruits an army of neutrophils. The arriving neutrophils degranulate, releasing reactive oxygen species and matrix metalloproteinases that further damage the epithelium, exposing more microbial stimuli, which in turn sustain the activation of the pathogenic CD8⁺ T cells. This positive feedback loop explains not only the chronicity of inflammation but also its transmural, tissue-destructive nature, as persistent neutrophilic inflammation is a primary driver of fibrosis and stricture formation.
A particularly fascinating dimension of this discovery is how it connects Crohn’s disease to the emerging field of “inflammaging.” Granzyme K-expressing CD8⁺ T cells were previously described by other labs as a dominant population that accumulates in various tissues with advancing age and contributes to chronic, low-grade sterile inflammation in the elderly. Their presence in the intestines of Crohn’s patients, many of whom are young adults, suggests a form of premature immunological aging occurring locally within the gut mucosa. The triggers that drive the accumulation of these cells in a younger cohort remain speculative but might involve chronic, subclinical stimulation from a dysregulated gut microbiome, recurrent enteric infections, or genetically determined defects in epithelial barrier function that create a persistent state of immunological stress. Whatever the initiating spark, the resulting landscape is one in which the GzmK⁺ CD8⁺ T cells become aberrantly installed as long-term residents of the intestinal tissue, operating under a biological program that is normally reserved for protecting against varicella zoster or other reactivating pathogens, but here mistakenly turned against the host’s own barrier interface.
The gene expression profiles observed open up entirely new possibilities for non-invasive diagnostics. The researchers investigated whether the transcriptomic signature of these pathogenic CD8⁺ TRM cells could be detected in the peripheral blood of patients, a pursuit that has become a holy grail in gastroenterology, where serial biopsies are burdensome. By applying deconvolution algorithms to whole-blood bulk RNA-seq datasets or looking for epigenomic footprints, they provided evidence that a circulating pool of these cells, or at least their shed mRNA and cell-free protein products, can indeed be identified, with Granzyme K itself standing out as a strong candidate biomarker. Serum levels of Granzyme K correlated robustly with endoscopic disease activity, distinguishing patients with active mucosal inflammation from those in deep remission with a degree of accuracy that could significantly improve upon C-reactive protein (CRP) or fecal calprotectin. The ability to directly monitor the activity of this pro-inflammatory CD8⁺ axis via a simple blood test would revolutionize the way clinicians track treatment response, potentially enabling the proactive titration of therapies before catastrophic tissue damage manifests as clinical symptoms.
The therapeutic implications of this work are arguably its most tantalizing and immediately translational aspect. If Granzyme K released by CD8⁺ TRM cells is a master mediator of epithelial dysfunction and neutrophilic inflammation, then blocking GzmK proteolytic activity becomes a highly specific and conceptually novel therapeutic strategy. A monoclonal antibody or a small-molecule inhibitor targeting the enzymatic pocket of GzmK could theoretically extinguish the epithelial “eDanger call“ without broadly depleting the CD8⁺ T cell population, thereby preserving antiviral immunity against cytomegalovirus or influenza. Such an approach would operate completely orthogonally to anti-TNF or anti-integrin biologics, potentially offering efficacy in the substantial fraction of patients who have failed these drugs. Furthermore, if the cells themselves could be targeted, therapies such as anti-CXCR6 to block their retention or strategies to deplete the TRM compartment via anti-CD103 antibody-drug conjugates could be envisioned. The study marks the first preclinical validation confirming that Granzyme K is not merely a correlative marker but an active mechanistic driver, providing the foundational rationale needed for pharmaceutical investment in GzmK-centric drug development pipelines.
Examining the precise immunological synapse between the GzmK⁺ CD8⁺ T cell and the intestinal epithelial cell, the study used spatial transcriptomics to map the geography of this interaction in tissue sections. These maps revealed micro-anatomical niches within the crypts of inflamed tissue where GzmK-expressing T cells were found directly juxtaposed against LGR5⁺ intestinal stem cells and transient amplifying cells. This spatial intimacy is particularly alarming because it implies that the inflammation is not just occurring on the differentiated surface epithelium but is seeping into the regenerative compartments of the crypt. By exposing the stem cell compartment to a continuous protease-mediated stress, the pathogenic CD8⁺ T cells could be responsible for the aberrant epithelial restitution and crypt architectural distortion that pathologists use to diagnose chronic colitis. The failure of the mucosa to properly heal, known clinically as mucosal healing, is a critical predictor of long-term outcomes including hospitalization and surgery, and the spatial location of these GzmK⁺ T cells points to a direct mechanism by which they could prevent the resolution of tissue injury even when bulk inflammation is partially suppressed by steroids.
This research also forces a reconciliation with the relatively understudied biology of Granzyme K itself. The granzyme family of serine proteases is canonically stored in the cytotoxic granules of natural killer (NK) cells and CD8⁺ T cells. While Granzyme B has been exhaustively studied for its potent ability to activate caspases and induce rapid apoptosis, the functions of other members like Granzyme A and Granzyme K have remained shrouded in enigma. The current study provides compelling evidence that GzmK has evolved a distinct biological function specifically suited to barrier tissues: the propagation of inflammation through non-lethal epithelial activation. GzmK cleaves substrates after arginine and lysine residues, and its specific intracellular targets in epithelial cells may include sequestration proteins for the NF-κB pathway, such as IκB-α, whose degradation releases NF-κB to the nucleus. The demonstration that this protease can induce a secretory phenotype in a cell type that is structurally central to organ function, while leaving the cell alive to perform this new signaling role, classifies GzmK as part of a broader system of “protease-driven signal transduction” that is only just beginning to be appreciated in immunology.
In a broader ecological context of the gut, the study touches upon the intricate cross-talk between the immune system and the intestinal microbiome. The researchers performed an exploratory analysis correlating the abundance of the GzmK⁺ CD8⁺ T cell signature with microbial species identified through 16S ribosomal RNA sequencing of matched stool samples. Among the patients with high levels of these cells, there was a significant enrichment for Proteobacteria, specifically adherent-invasive Escherichia coli (AIEC), a pathobiont that has long been epidemiologically linked to ileal Crohn’s disease. The model that elegantly emerges is one in which a genetically susceptible host harbors a facultative intracellular pathogen, like AIEC, that invades the epithelium, providing a persistent antigenic trigger that drives the clonal expansion and tissue residency program of the GzmK⁺ CD8⁺ T cell pool. The T cells, once established, then become largely independent of the initial bacterial trigger, responding to stress signals from a barrier that is now perpetually breached, a phenomenon that could explain why antibiotic treatment often fails to fully resolve Crohn’s inflammation despite its temporary benefits, and why the disease persists as a chronic, self-sustaining cycle even when the inciting microbial insult is long gone.
The clinical diversity of Crohn’s disease, with its distinct Montreal classification phenotypes—ileal, colonic, ileocolonic, and the presence or absence of structuring and penetrating behavior—has always suggested that multiple pathogenic pathways might converge under the same diagnostic umbrella. The work of Lee and colleagues remarkably stratifies this heterogeneity through the lens of the GzmK⁺ CD8⁺ TRM axis. Their data indicated that while these pathogenic cells were present across all intestinal segments, their transcriptional program and functional output were most pronounced in patients with stricturing and penetrating disease, the most severe and morbid phenotypes. The fibrosis that leads to strictures involves the activation of myofibroblasts and excessive deposition of extracellular matrix, a process known to be driven by factors like transforming growth factor-β (TGF-β) and IL-13. The GzmK⁺ CD8⁺ T cells were found to co-express metalloproteinases and factors that can proteolytically activate latent TGF-β, suggesting that these cells might simultaneously drive inflammation and the fibrotic complications that severely debilitate patients, linking active immune destruction directly to the structural sequelae that require surgical intervention.
As with all landmark translational discoveries, the path from bench to bedside is fraught with challenges. The first is target specificity—any pharmacological agent designed to inhibit Granzyme K must be exquisitely specific to avoid cross-inhibiting the vital anti-infective functions of Granzyme B or other homologous proteases within the granzyme family. The second major challenge is delivery; the pathogenic cells reside deep within the mucosa, potentially shielded from systemically circulating antibodies, which might necessitate the development of orally administered small molecules or gut-restricted peptides capable of penetrating the lamina propria. The study itself serves as a launching pad, providing the proof-of-concept human data that will encourage medicinal chemists and protein engineers to take on these challenges. It establishes the crystal structure-informed binding pockets of Granzyme K as a druggable interface, noting that because GzmK, unlike caspases, can operate on extracellular substrates and in the extracellular milieu, an inhibitor could potentially block the inflammatory signaling without even needing to enter the epithelial cells, simply by intercepting the protease in the intercellular space immediately after secretion from the T cell.
The metamorphosis of the CD8⁺ T cell from a simple executioner to a sophisticated communicator of inflammatory messages has profound implications beyond gastroenterology. The paradigm established in this study—that a tissue-resident cytotoxic cell can use a granzyme as a paracrine factor to reprogram structural cells into inflammatory amplifiers—may be a generalizable mechanism underlying other chronic inflammatory diseases of barrier tissues, including ulcerative colitis, psoriasis, atopic dermatitis, and chronic obstructive pulmonary disease. The search for GzmK⁺ CD8⁺ TRM cells in the synovial tissues of rheumatoid arthritis patients or the salivary glands of Sjögren’s syndrome patients is likely already underway in laboratories around the world. This study encourages a fundamental rethinking of what it means for a cell to be cytotoxic. Cytotoxicity, as illustrated by the GzmK⁺ population, is not exclusively a lethal act; it can be a mode of communication that fundamentally alters the behavior of the target cell without killing it, a concept that might be termed “behavior-altering cytotoxicity.” This novel mechanism, in which a T cell protease orchestrates the local immune environment by dictating epithelial chemokine output, positions the intestinal barrier as a dynamic, string-instrument-like structure, with the T cell playing the strings to produce an inflammatory concert that, once begun, is terrifyingly difficult to silence. The identification of the key musician, Granzyme K, finally offers a chance to cut the strings.
Subject of Research: The role of Granzyme K-expressing CD8⁺ tissue-resident memory T cells in driving intestinal inflammation and epithelial reprogramming in patients with Crohn’s disease, elucidated through single-cell transcriptomics, organoid co-cultures, and mechanistic signaling studies.
Article Title: Granzyme K CD8⁺ T cells with tissue-resident features promote intestinal inflammation in patients with Crohn’s disease
Article References: Lee, Y., Kim, TY., Kim, Y. et al. Granzyme K CD8⁺ T cells with tissue-resident features promote intestinal inflammation in patients with Crohn’s disease. Exp Mol Med (2026). https://doi.org/10.1038/s12276-026-01763-7
Image Credits: AI Generated
DOI: 10.1038/s12276-026-01763-7
Keywords: Crohn’s disease, Granzyme K, CD8⁺ T cells, tissue-resident memory T cells, intestinal inflammation, single-cell RNA sequencing, epithelial reprogramming, neutrophilic infiltration, mucosal immunology, inflammaging, therapeutic target, biomarker
