Overexpression of the mitochondrial translocator protein (TSPO) in the colon is a pathological feature observed in human ulcerative colitis (UC) (Ostuni et al., 2010), a form of inflammatory bowel disease (IBD). Consistent with human pathology, rodent models of UC have confirmed that TSPO upregulation in the colon is characteristic of UC (Ostuni et al., 2010). This finding is supported by studies investigating TSPO-binding drugs as radioligands for positron emission tomography (PET)-based diagnostic imaging; it has been concluded that TSPO detection using these probes provides a non-invasive method to evaluate the level and localization of colon inflammation (Bernards et al., 2015; Kurtys et al., 2017). High expression of TSPO in chronic UC has also been linked to an increased risk of colorectal cancers (Katz et al., 1988, 1990). Association between TSPO and inflammation is not only restricted to observations made in the colon; TSPO upregulation has been reported in conditions that induce neuroinflammation (Airas et al., 2015), cardiovascular inflammation (Fujimura et al., 2008), and rheumatoid arthritis (Narayan et al., 2018). Indicative of an active function and/or etiology for TSPO involvement in these pathologies, preclinical models have demonstrated that a variety of TSPO-binding drugs are efficacious as therapeutics in ameliorating inflammation. Protective effects of TSPO-targeting therapies have been reported for disorders such as multiple sclerosis (Daugherty et al., 2013), brain injury after intracerebral hemorrhage (Li et al., 2017), diabetic neuropathy (Giatti et al., 2009), autoimmune arthritis (Waterfield et al., 1999), and Alzheimer’s disease (Barron et al., 2013). As the etiology and pathophysiology of UC remains to be fully defined, TSPO has been considered as a therapeutic target and investigated for reducing both stress and colon inflammation.

Of direct relevance to therapeutic efficacy in UC, the TSPO-binding drug ONO-2952 (1-[(1S)-1-(4-chloro-2-methoxyphenyl)-5-fluoro-1,9-dihydrospiro [β-carboline-4,1’- cycloprop ane]-2(3H)-yl]ethenone) has been under development as a treatment for diarrhea-predominant irritable bowel syndrome (IBS-D) (Ono Pharma United States Inc.; NCT01844180, ClinicalTrials.gov). IBS-D patients have significantly higher levels of pro-inflammatory cytokines indicative of gut inflammation (Rana et al., 2012). ONO-2952 showed high selectivity in binding to TSPO, and could promote anti-stress effects by inducing a variety of protective responses (Mitsui, 2012; Mitsui et al., 2015). In preclinical rat models, ONO-2952 attenuated stress-induced defecation and rectal hyperalgesia (Kawahara et al., 2018). In the Phase 2 trial, ONO-2952 showed evidence for clinical efficacy with promising trends supporting its consideration as a potential treatment (Whitehead et al., 2017). Similarly, flunitrazepam, albeit a non-selective TSPO-binding drug, could protect the intestinal mucosa against DSS-induced damage (Ostuni et al., 2010). However, actions of other TSPO-binding drugs have not been consistent; PK11195 and Ro5-4864 were observed to enhance initial severity of colonic erosion, but accelerate recovery (Ostuni et al., 2010). Because the molecular function of TSPO remains to be elucidated (Selvaraj and Stocco, 2015; Selvaraj and Tu, 2016), TSPO-binding drugs cannot be confirmed as agonists or antagonists based on current understanding of target effects, potentially accounting for these differences in TSPO-targeting therapeutic efficacy.

First identified as a “receptor” showing high affinity binding to diazepam and its derivative 4’chlorodiazepam (Braestrup and Squires, 1977), the protein sequence of TSPO is highly conserved across microbes, plants and animals (Selvaraj and Stocco, 2015). Although pharmacology linked to TSPO-binding has indicated a variety of potential functions (Gavish et al., 1999), development and utilization of TSPO gene-deleted (Tspo ^−/− ) models have ruled out its involvement in key purported pharmacological functions such as induction of de novo steroidogenesis (Tu et al., 2014a; Banati et al., 2014; Morohaku et al., 2014b; Wang et al., 2016), mitochondrial permeability transition (Šileikyte et al., 2014), and heme biosynthesis (Zhao et al., 2016). Phenotypic observations in Tspo ^−/− models point to TSPO’s functional involvement in lipid/energy metabolism (Leduc et al., 2011; Thompson et al., 2013; Tu et al., 2016; Liu et al., 2017; Kim et al., 2019; Koganti and Selvaraj, 2020). While such metabolic modulation might be relevant to certain TSPO over-expressing colon cancers (Katz et al., 1990; Maaser et al., 2005; Königsrainer et al., 2007), the precise mechanism of TSPO function in this context remains to be defined (Selvaraj and Stocco, 2015; Betlazar et al., 2020).

Given the considerable relevance of TSPO in the context of etiology, pathogenesis and therapeutic interventions for UC, we investigated Tspo ^−/− mice as a loss-of-function preclinical model to study the relevance of TSPO in homeostasis and TSPO-mediated mechanisms in UC. Such genetic studies provide an opportunity to delineate gene function and distinguish this from the non-specific actions prominently seen with different TSPO-binding drugs (Bordet et al., 2007; Li et al., 2008; Luc Do Rego et al., 2015; Tu et al., 2015; Singh et al., 2020). Our results identify a clear protective function for TSPO expression in colitis, deficiency of which results in exacerbated colonic inflammatory pathology.

In using Tspo ^−/− mice to investigate the mechanism and test the foundational value of TSPO-targeting pharmacological agents in UC, we uncover a functional link that corroborates TSPO involvement in the pathological basis of the disease. The DSS-induced colitis model has been shown to primarily induce epithelial barrier disruption, with inflammation only secondary to loss of epithelial integrity (Cooper et al., 1993; Laroui et al., 2012). This model yields colonic pathology that closely mimics that of human patients with UC (Wirtz et al., 2007). In this study, TSPO deficiency substantially exacerbated the inflammatory pathology associated with DSS-induced colitis. Although the impact of local versus global TSPO loss-of-function remains difficult to delineate, the pathologic findings provide conclusive evidence that effects specific to TSPO can influence inflammatory injury in colon. Physiologic TSPO function appears to offer a significant level of protection against colonic insults that might induce inflammation and injury. As TSPO expression is observed both in the intestinal epithelium and immune cells, we believe the exacerbated UC in Tspo ^−/− mice results from a combination of vulnerability in the epithelial barrier and dysregulation in specific immune responses.

The exacerbated UC phenotype in Tspo ^−/− mice became obvious as our pilot experiments to generate a colon cancer model using DSS and azoxymethane (Thaker et al., 2012) resulted in complete mortality in the Tspo ^−/− group during the second cycle of treatment (unpublished observations). During the 7-day DSS treatment in this study, we observed progressively aggravated clinical signs and body weight loss in Tspo ^−/− mice compared to Tspo ^fl/fl controls. The aggravated response in Tspo ^−/− mice was consistent with known pathological parameters of severity such as progressive, continuous ulceration of the colonic epithelium, edema, and shortened colon length (Okayasu et al., 1990). The severity of UC pathology Tspo ^−/− mice was also indicative of a degree of unrecoverable loss of intestinal structure.

As TSPO is expressed in the colonic epithelial cells, we first considered the possibility that specific loss of function in this cell type might be linked to a local predisposition to injury in Tspo ^−/− colons. Our finding that Krt18, the type I intermediate filament associated with the colonic epithelium, significantly decreases with UC in Tspo ^−/− mice was a clear indication of higher levels of colonic epithelial loss compared to Tspo ^fl/fl cohorts. Epithelial denudation has been correlated to the severity of DSS-induced UC in human patients (Corfe et al., 2015), and in our study, indicated the increased severity of UC pathology in Tspo ^−/− mice. In addition to serving as a marker, mutations in Krt18 that disrupt the intermediate filament networks have also been shown to functionally increase epithelial permeability in human colonocytes (Zupancic et al., 2014). Keratin downregulation has also shown to dramatically increase susceptibility to cell death mediated by inflammatory cytokines such as TNF⍺ (Caulin et al., 2000). Supporting our gene expression results, substantial epithelial loss could be visualized in histopathology of Tspo ^−/− colons. Excessive inflammation seen with UC in Tspo ^−/− mice compared to Tspo ^fl/fl cohorts could be a consequence of amplified disruption of the epithelial barrier.

High-affinity benzodiazepine binding sites indicative of TSPO expression have also been identified in mast cells (Miller et al., 1988). Rats deficient in mast cells fail to develop clinical signs of DSS-induced colitis, indicating a crucial role for mast cells in UC pathogenesis (Araki et al., 2000). Moreover, it has been demonstrated that mast cell activation, identified by tryptase levels, has an essential proinflammatory role in triggering DSS-induced UC (Hamilton et al., 2011). Mast cell tryptase can also act as a trigger for other mediators of inflammation (Compton et al., 1998; Huang et al., 1998), potentially contributing to UC pathology. Functionally, different drugs that bind TSPO have been shown to have varying effects on mast cell degranulation (Miller et al., 1988; Suzuki-Nishimura et al., 1989; Sano et al., 1990; Yousefi et al., 2013). Investigating whether Tspo deletion affected the mast cell response, we identified significant upregulation of murine Mcp6 (which is similar to the human mast cell hTryptase-β) in Tspo ^−/− colons. This can be considered an activation response, as Cd34 expression indicative of mature mast cells (Drew et al., 2002), albeit increased in Tspo ^−/− colons, was not significantly different between subsequent (day-4 and day-7) timepoints between the two genotypes. It has been shown that Mcp6 deficiency (Hamilton et al., 2011), or tryptase inhibition (Tremaine et al., 2002; Liu et al., 2021), could reduce UC pathology and intestinal fibrosis. In parallel, degranulation of mast cells has also been associated with the release of proinflammatory TNF⍺ (Gordon and Galli, 1991; Bischoff et al., 1999). We observed higher levels of Tnfα expression during UC in Tspo ^−/− colons, although this difference was not statistically significant. In forms of mast cell driven colitis, TNF⍺-driven effects have been demonstrated to be crucial for UC pathology (Rijnierse et al., 2006). It has also been shown that TNFα and other proinflammatory cytokines could persist in the colon at high levels for at least 14 days after DSS withdrawal (Yan et al., 2009). Therefore, protracted effects on inflammation could also be anticipated in Tspo ^−/− mice driven by both the inflammatory milieu and the degree of injury sustained.

We also observed that Cd11c levels peaked at day-4 of UC in Tspo ^−/− colons, which was significantly different and earlier than that measured for UC in Tspo ^fl/fl colons. Cd11c is expressed on the surface of monocytes, macrophages, and dendritic cells in the intestinal mucosa; these cells are important players in immune recognition and tolerance to the gut microbiota (Bain and Mowat, 2014). Increased infiltration of CD11c⁺ macrophages has been demonstrated to trigger mucosal injury in UC (Bernardo et al., 2018; Fuke et al., 2018), while reduction in CD11c⁺ macrophages ameliorated mucosal pathology (Arnold et al., 2015). Earlier up-regulation of Cd11c in Tspo ^−/− animals may reflect and/or influence the rapid onset and severity of colonic inflammation. In addition, CD11c⁺ phagocytes may also increase gut mucosal epithelial permeability (Ren et al., 2017), further contributing to colonic pathology through increased bacterial translocation across the gut barrier.

The upregulation of baseline colonic Cdx2 expression in Tspo ^−/− mice indicates that TSPO deficiency contributed to a homeostatic shift, perhaps compensating for decreased epithelial barrier function. Cdx2 gene-deleted heterozygous (Cdx2 ^+/− ) mice showed hypersensitivity to DSS-induced acute UC and increased permeability of the colonic epithelium (Calon et al., 2007). Cdx2 is known to control cell-cell adhesion and extracellular matrix composition/interactions (Lorentz et al., 1997); baseline Cdx2 upregulation may serve as a compensatory mechanism to offset increased vulnerability to epithelial permeability in Tspo ^−/− mice. High expression of TNFα, through NF-κB and p38 MAPK pathways, could downregulate Cdx2 (Coskun et al., 2014), suggesting that there might be significant cross-talk between epithelium-specific and different immune pathways.

There also exists cross-talk between Cdx2 and Wnt signaling, regulating the proliferative compartment of normal intestinal crypts (Guo et al., 2010). Our finding that the Wnt target Ccnd1 was decreased in Tspo ^−/− colons might indicate a lower capacity for epithelial proliferation and turnover, potentially affecting regeneration after ulcerative injury. However, based on similar Lgr5 and Ascl2 levels between Tspo ^fl/fl and Tspo ^−/− colons, stem cell populations and adhesion molecule expression were not altered in Tspo ^−/− colons. Although Cd34 expression also believed to be associated with these adult stem cells (Stzepourginski et al., 2017), are higher at baseline in Tspo ^−/− colons, the differences are in all probability linked to mature mast cell numbers (consistent with higher baseline Mcp6 expression levels). Although stem cell populations are similar, given the progressive severity of UC pathology Tspo ^−/− mice, it is plausible that TSPO deficiency might also negatively impact epithelial recovery and functional restoration in the colon. This speculation is supported by pharmacology that TSPO binding drugs, PK11195 and Ro5-4864 could accelerate recovery after DSS-induced UC (Ostuni et al., 2010). Although the extent of TSPO-mediated action remains unclear, positive and negative outcomes in DSS-induced UC have been reported for relatively low-affinity TSPO binding molecules such as curcumin (Li et al., 2013; Liu et al., 2013; Guo et al., 2022) and retinoic acid (Oehlers et al., 2012; Li et al., 2013; Rampal et al., 2021).

As recent developments on TSPO function have indicated possible roles in regulation of mitochondrial fatty acid oxidation (Tu et al., 2016) and oxidative stress responses (Gatliff et al., 2014), we examined genes associated with these functions in Tspo ^−/− colons. The finding that Acadm is up-regulated in Tspo ^−/− colons is consistent with evidence that there exist subtle metabolic shifts with TSPO loss-of-function. In tissues active in lipid metabolism that show very high expression of TSPO, we have previously reported a functional up-regulation of mitochondrial fatty acid oxidation, and associated increase in gene expression of Cpt1a, Acadm, Acadl, and Hadha (Tu et al., 2016). Increases in Cpt1a, Acadl and Hadha were not observed in Tspo ^−/− colons; therefore, it is conceivable that the present study’s use of entire colons for gene expression assays, as opposed to just TSPO-expressing epithelial layers or mast cells, could have diluted any cell-type specific phenotypes. With regard to oxidative stress, direct evidence previously demonstrated that TSPO could stabilize mitochondrial architecture during inflammatory stress in colonic cells (Issop et al., 2016). We have previously observed that TSPO involvement in oxidative stress is cell-type dependent (Tu et al., 2016; Zhao et al., 2016). Although we did not observe any baseline perturbations in gene expression associated with oxidative stress responses in Tspo ^−/− colons, baseline expression may not reflect active inflammatory cell states associated with UC.

In summary, despite lack of a clear structure-function relationship for the different TSPO-mediated effects that are described in the literature (Selvaraj et al., 2015; Selvaraj and Stocco, 2015), and mechanisms that delineate TSPO pharmacology (Gavish et al., 1999; Gavish and Veenman, 2018), our findings in this Tspo ^−/− model demonstrates a protective effect for TSPO on the colonic epithelium and confirms that there might be potential for therapeutic interventions with ligands that enhance TSPO function. For example, the TSPO-binding benzodiazepine derivatives Ro5-4864 (4'-chlorodiazepam) and flunitrazepam, which were both shown to quickly suppress mast cell activation in vivo and in vitro (Yousefi et al., 2013), may be considered to be TSPO agonists. Understandably, similar actions albeit with differing efficacies, occur for these same drugs in the in vivo DSS-induced colitis model (Ostuni et al., 2010). The hypothesis that TSPO likely provides a dual therapeutic effect, involving both epithelial barrier-specific functions and immune-cell mediated pathways, is an attractive mechanism that merits more detailed investigation. Fundamentally, we uncover that TSPO deletion/loss-of-function, albeit inconsequential at a resting/healthy state, makes mice more vulnerable to insults. Future studies that dissect the precise molecular function of TSPO might help describe its physiological and pharmacological significance in UC and other inflammatory disease models.