Finding the TRAIL to escape granzyme B–mediated liver injury in PSC

Primary sclerosing cholangitis (PSC) is a chronic liver disease characterized by progressive inflammation and fibrosis of the intrahepatic and extrahepatic bile ducts, leading to narrowing and fibro-obliteration of these ducts.^[1] The progressive nature and lack of effective pharmacotherapy make PSC a common indication for liver transplant. Thus, there is a critical unmet need for safe and effective therapeutics that can slow or reverse PSC progression. The etiopathogenesis of PSC is still largely unknown. Although PSC is considered an immune-mediated biliary disease, and possibly, an auto-immune disease due to its strong association with several human leukocyte antigen haplotypes, immunosuppressive therapies have not been successful in patients with PSC. Nonetheless, this association strongly points to the adaptive immune system’s involvement in the pathogenesis of PSC. Although biliary immunology can be considered still in its infancy, there has been a surge in studies to interrogate the intrahepatic immune landscape in PSC.^[2] In that regard, it has been elegantly shown that cytobrush sampling of bile duct cells from patients with PSC contain increased populations of T cells compared to patients without PSC.^[3,4] However, the function of CD8+ T cells and natural killer (NK) cells in PSC pathogenesis is still unclear, as most of the studies thus far have focused on the role of CD4+ T cells, despite evidence implicating CD8+ T cells and NK cells in mouse and human sclerosing cholangitis. Indeed, interferon-gamma (IFNγ) serum levels are elevated in patients with PSC, and IFNγ, which is mainly produced by CD8+ T cells and NK cells, has been shown to switch their phenotype toward a more cytotoxic one in Mdr2^−/− mice, a mouse model of sclerosing cholangitis. In this model, genetic ablation of Ifng (the gene encoding IFNγ) resulted in reduced CD8+ T-cell–mediated and NK-cell–mediated cytotoxicity and ameliorated liver fibrosis^[5].

In an elegant study featured in the current issue of Hepatology, Kellerer et al^[6] continued to investigate the immunopathogenesis of PSC by subjecting CD45+ leukocytes from the livers of Mdr2^−/− mice to scRNA-seq (single-cell RNA sequencing) and CITE-seq (cellular indexing of transcriptomes and epitopes by sequencing). The analyses led to the identification of a cluster of CD8+ effector memory T cells (CD8+ T_EM) and a subset of CD49b^low NK cells expressing several genes associated with cytotoxicity, including Gzmb (encoding the cytotoxic protease granzyme B, GzmB) and Tnfsf10 (encoding the death ligand tumor necrosis factor–related apoptosis-inducing ligand, or TRAIL). Both clusters also expressed genes associated with tissue residency and were low in genes linked to tissue egress, suggesting they are tissue-residing cells potentially important in the early stages of the disease. Protein expression of GzmB and TRAIL was further confirmed only in subsets of CD8+ T cells and NK cells that were elevated in Mdr2^−/− compared to wild-type mouse livers. More importantly, similar hepatic cytotoxic subsets were also found to be present in human livers of patients with PSC from a previously published database.^[7] Based on these findings, the authors went on to investigate the role of TRAIL-mediated and GzmB-mediated lymphocyte cytotoxicity in the pathogenesis of PSC.

GzmB is a serine protease primarily produced by cytotoxic T cells and NK cells, with a well-established role in the induction of apoptosis in tumor or viral infected cells. GzmB is delivered into the target cells following the release of lytic granules containing GzmB and the pore-forming protein perforin by the immune effector cells; once in the cytosol, GzmB cleaves and activates several apoptosis mediators (ie, Bid, caspase 3), resulting in cell death. Previous studies have proposed a role for GzmB in NK and CD8+ T-cell–mediated cholangiocyte lysis in a mouse model of biliary atresia.^[8] Similarly, the death ligand TRAIL is also preferentially expressed in T cells and NK cells, although it can be induced in other immune cells, including NK T cells, dendritic cells, and macrophages, in pathologic conditions. The binding of TRAIL to one of its cognate death receptors (TRAIL-R1 or DR4, and TRAIL-R2 or DR5 in humans; mDR5 in mice) triggers a caspase-mediated signaling cascade culminating in apoptosis. Interestingly, germline deletion of Tnfrsf10b (encoding mDR5) in Mdr2^−/− mice, or conditional knockout of Tnfsf10 in myeloid cells in mice fed a 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) diet (another model of cholestasis-induced liver injury) are associated with increased ductular inflammation and fibrosis, suggesting the TRAIL/TRAIL-R axis is involved in restraining the cholestatic injury.^[9,10]

To elucidate the role of CD8+ T cells, Kellerer and colleagues first treated Mdr2^−/− mice with an anti-CD8α antibody, which depleted the CD8+ T cells without altering CD4+ T cells or NK cells. Interestingly, reduced GzmB and TRAIL expression by NK cells was observed along with decreased cholangiocyte apoptosis, liver injury, and fibrosis in the absence of CD8+ T cells. They then proceeded to assess the individual contribution of GzmB and TRAIL by generating Mdr2^−/− × Gzmb^−/− and Mdr2^−/− × Tnfsf10^−/− mice. The absence of GzmB reduced liver injury, fibrosis, and cholangiocyte apoptosis without affecting the frequencies of hepatic NK, CD4+ and CD8+ T cells, or their expression of TRAIL or IFNγ. In contrast, the deletion of TRAIL in Mdr2^−/− mice resulted in the expansion of T-cell and NK cell populations, pointing to a fundamental role of TRAIL in immune cell apoptosis. Comparative transcriptome analysis also revealed a change in the hepatic CD8+ T_EM cell profile, with upregulation of inflammatory, cytotoxic, and tissue residency genes, as well as enhanced activation, in the absence of TRAIL. Consistent with enhanced inflammatory and cytotoxic immune response, TRAIL deficiency also led to increased cholangiocyte apoptosis and hepatic fibrosis, in line with previous studies showing exacerbation of inflammation and fibrosis in Mdr2^−/− mice lacking mDR5.^[9] Finally, the transfer of CD8+ T cells from Tnfsf10^−/− mice into immunodeficient Mdr2^−/− mice (Mdr2^−/−xRag1^−/−) resulted in enhanced production of IFNγ with no change in GzmB, increased liver injury, augmented cholangiocyte apoptosis, and worsened fibrosis. The authors concluded that CD8+ T-cell–derived TRAIL limits the IFNγ-mediated inflammatory lymphocyte response that eventually regulates liver injury and fibrosis in PSC.

This study is an important contribution to the field of biliary immunology and provides another piece of the puzzle of PSC pathogenesis. The relevant findings of the study are (a) the identification of subsets of tissue-resident, cytotoxic CD8+ T_EM and NK cells as pivotal players in the development of mouse sclerosing cholangitis and human PSC; and (b) the opposing functions of the cytotoxic effectors GzmB and TRAIL expressed by these subsets of lymphocytes (Figure 1). While GzmB is mostly implicated in mediating the lysis of target cells, the role of TRAIL in regulating cell death and immune functions is far more complicated. Not unexpectedly, GzmB was found to induce cholangiocyte apoptosis, therefore promoting biliary injury, but had no effect on immune cell death. On the other hand, TRAIL regulated T-cell activation and survival, as well as lymphocyte cytotoxicity, thus restraining tissue injury by reducing inflammation. Both TRAIL and GzmB are expressed in response to the injury; however, TRAIL may be necessary to dampen the immune response to avoid excessive damage. The question remains as to why in chronic cholestatic conditions, expression of TRAIL is not sufficient to prevent biliary injury. TRAIL is also expressed by other cells of the immune system, including macrophages, and conditional deletion of TRAIL on myeloid cells (macrophages in particular) worsens liver injury and fibrosis in a mouse model of cholestasis. Thus, the protective role of TRAIL against cholestatic liver injury seems not to be restricted to the one expressed by lymphocytes. Whether TRAIL on lymphocytes has a more profound immunomodulatory effect on the cholestatic microenvironment than TRAIL expressed on macrophages, or possibly even has distinct target cells or been implicated in different types of cholestatic liver injury (chronic vs. acute), remains to be determined. Moreover, immune cells are not the only target of TRAIL-killing. Activated HSC, unlike quiescent HSC, are exquisitely sensitive to TRAIL-induced apoptosis, and several studies have explored the possibility of exploiting this sensitivity for antifibrotic therapies. In the case of fibrotic diseases such as PSC, the role of immune cell–derived TRAIL likely extends beyond sole modulation of the immune response and into direct antifibrotic effect through the killing of activated HSC.

FIGURE 1.

Opposing effects of TRAIL and granzyme B in sclerosing cholangitis. In response to IFNγ, cytotoxic resident memory T cells and NK cells upregulate GzmB and TRAIL. GzmB induces cholangiocyte apoptosis, leading to enhanced cholestatic liver injury and fibrosis. This effect is counterbalanced by TRAIL, which acts in a feedback loop to restrain lymphocyte activation and expansion, thereby limiting inflammation and liver damage. Figure created with Biorender.com. Abbreviations: GzmB, granzyme B; IFNγ, interferon gamma; NK, natural killer; TRAIL, tumor necrosis factor–related apoptosis-inducing ligand.

Abbreviations:

CD8+T_EM

CD8+ effector memory T cells

GzmB

granzyme B

IFNγ

interferon gamma

NK

natural killer

PSC

primary sclerosing cholangitis

TRAIL

tumor necrosis factor–related apoptosis-inducing ligand