Mucosal surfaces represent a main entryway for pathogens to anatomic access and are common sites for cancer development. Enhancing immunity at these regions can provide better protection and improve strategies for treating these diseases. Our previous work in mouse models has shown that increasing the migration of vaccinia virus (VACV)-specific memory T cells to mucosal regions during vaccination boosts protection at these sites during VACV challenge (1). Correlative evidence also exists in non-human primate models; in rhesus macaques, the use of an attenuated cytomegalovirus (CMV) vaccine vector for simian immunodeficiency virus (SIV) increases effector T cell numbers at mucosal regions, resulting in vastly improved control and clearance of SIV following viral challenge (2, 3). In humans, clinical evidence further suggests that enhancing immune presence at mucosal sites corresponds positively with protection (4–9). Individuals with vitamin A deficiencies exhibit severely impaired mucosal immunity, resulting in increased susceptibility to infections (10–13). As the heightened immune protection seen is predominantly a result of increased effector T cell presence in the mucosal regions, identifying ways to promote T cell migration to these areas is likely to improve resistance to diseases affecting these areas (1, 14, 15).

Effector T cell trafficking to and entry into mucosal regions is governed by their expression of receptors that instigate mucosal homing, including C-C chemokine receptor type 9 (CCR9) and α₄β₇ integrin (α₄β₇) (1, 15–17). Upregulation of these mucosal homing receptors during T cell activation is dependent on signaling via retinoic acid receptor, a type II nuclear receptor that heterodimerizes with another nuclear receptor, the retinoid X receptor, to mediate transcription (12, 15, 17). Both the RAR and RXR possess α, β, and γ isotypes, with activation of the RARα/RXRα heterodimer implicated in transcription of CCR9 and the α₄ subunit of α₄β_7, via cooperation NFATc2 (12, 15, 16, 18–21). Binding of all-trans retinoic acid (ATRA), a biologically active vitamin A metabolite and the most abundant naturally occurring pan-RAR ligand, to the ligand-binding pocket (LBP) of the RAR triggers activation of the heterodimer, ultimately resulting in RAR-mediated transcription (22–24).

Like the RAR, the RXR also possesses an LBP, and ligand bound to both the RAR and RXR has been shown to enhance transcription of retinoid-dependent genes (18, 23, 25, 26). However, identification of endogenously occurring RXR ligands has remained limited. 9-cis-retinoic acid (9cRA), a naturally occurring stereoisomer of ATRA, has been reported as a high affinity RXR ligand, however its detection in vivo remains elusive (27–29). Fatty acids such as docosahexaenoic acid and phytanic acid are also capable of binding the RXR, however endogenously occurring levels are likely too low to activate the receptor under most physiologic scenarios (28, 29). The challenge to conclusively identify naturally occurring RXR ligands has led many groups to utilize synthetic agonists, which have since been coined ‘rexinoids’.

Despite studies showing that ATRA can promote the expression of mucosal T cell homing proteins and subsequent migration to mucosal sites in vivo, resulting in better protection against mucosal infection, little is known about the effect of rexinoids on effector T cells. The functional similarity seen between rexinoids and ATRA in experimental models indicates that these synthetic small molecules may be able to exert similar effects as ATRA on effector T cells, by influencing their migration to mucosal-associated regions (30, 31). The ability of rexinoids to bind the RXR suggests that they may improve the impact of endogenous ATRA on T cell mucosal-related function. Additionally, reports that some rexinoids bind to the RAR indicates they may be able to mimic the effect ATRA has on T cell activity (30, 32).

Here we assessed the ability of a panel of rexinoids ( Figure 1 ) to induce expression of CCR9 and α₄β₇ and to promote T cell migration in vitro. These rexinoids include a fluorobexarotene analog, A18, halogenated bexarotene analogs A20-A22, rexinoids A30-A41 which are described in our previous work and references therein, and rexinoids A52-A63 which are again described in our prior work and citations therein (33–36). A subset of rexinoids (A18, A20, and A41) were capable of exerting this effect independently of ATRA, suggesting these compounds can act as ATRA mimics while retaining lower toxicity and enhanced stability. Conversely, other rexinoids (A55, A56, and A57) displayed synergy with suboptimal doses of ATRA to enhance CCR9 expression. Moreover, treatment with ATRA mimics induced T cell migration in vitro towards the CCR9 ligand CCL25, while treatment with the ATRA cooperating rexinoids also resulted in improved migration. Furthermore, preliminary in vivo data suggest rexinoid treatment is accompanied by reduced toxicity compared to ATRA. Together, these data suggest that rexinoids may have potential to be used as a novel immunotherapeutic treatment modality for mucosal diseases by either replacing ATRA-based strategies or by being used in conjunction with non-toxic ATRA levels to bolster efficacy.

Identifying compounds that can favorably alter T cell migration to mucosal surfaces has the potential to improve immune responses towards diseases at these surfaces. Here we tested a panel of novel rexinoids for their ability to both influence effector T cell expression of mucosal homing markers CCR9 and α4β7 and to affect their migration towards a mucosally expressed chemokine in vitro. Our results show that many rexinoids are capable of enhancing CCR9 and α4β7 expression on responding T cells. Several rexinoids induced T cell expression of CCR9 independently, mimicking the naturally occurring biologic ATRA, while others worked synergistically with subtoxic doses of ATRA to enhance expression, indicating a potential to cooperate with vitamin A present in vivo. Furthermore, both ATRA mimicking and ATRA cooperating rexinoids were seen to improve T cell migration towards the CCR9 ligand CCL25, with some outperforming bexarotene and ATRA. These findings introduce several rexinoids that can imprint T cells with a mucosal homing phenotype and influence their migration, and may have clinical relevance in treating mucosal diseases.

In addition to CCR9 and α₄β₇, the expression of a myriad of other genes have also been shown to be under the control of RAR signaling, including those that inhibit cell cycle progression and promote apoptosis (37–42). These discoveries have led to ATRA being used clinically as an anti-cancer drug; combination treatments that include ATRA have been successful in inducing cancer remission, most notably with acute promyelocytic leukemia (APL), a disease marked by an RARα translocation (43–45). Unfortunately, these favorable results are dampened by adverse side effects attributed to ATRA usage. Various toxicities, including hepatotoxicity due to retinyl ester buildup in hepatic stellate cells (HSCs), and mucocutaneous toxicity, have been reported in cancer patients receiving ATRA treatment (46–49). Here we confirmed that mice treated with ATRA fare poorly, as evidenced by their severe weight loss, physical appearance, and higher ALT levels. An additional complication seen with ATRA use is differentiation syndrome (DS), which can be life-threatening (49–52). Surprisingly, similar adverse health effects have also been reported following the use of synthetic vitamin A derivatives such as isotretinoin and acitretin, which has led us to postulate that the toxicities seen may be due to aberrant activation of the other RAR isoforms (53–55). This is supported by the finding that RARγ deficient mice show resistance to ATRA-mediated toxicity (56). Additionally, activation of all three RAR isoforms have been shown to display teratogenic potential (57). As the ATRA cooperating rexinoids demonstrate high selectivity for the RXR, their use could avoid such toxicity. Furthermore, the widespread use of ATRA is limited due to its instability when exposed to ubiquitous elements such as ultraviolet (UV) light, ambient temperatures, and oxygen (49, 58). The improved stability of rexinoids compared to ATRA is another attractive characteristic; their long shelf life and resistance to fluctuations in temperature, UV light, or oxygen presence makes them more durable treatment options.

In animal models of lung cancer, rexinoid use has been seen to mediate similar antiproliferative and proapoptotic effects on cancer cells as is observed with ATRA (59, 60). Importantly, rexinoid treatment has been shown to be better tolerated than ATRA in both animal and human models. Clinical trial results show that bexarotene, which is currently used as a treatment modality for patients with cutaneous T cell lymphoma (CTCL), can be safely administered at dosages of 300mg/m²/day, while side effects are seen with ATRA dosages higher than 45mg/m²/day (49, 61–67). However, it is currently unknown for most rexinoids whether they are behaving as ATRA mimics or acting in synergy with ATRA. Here we have not only identified RXR ligands that act similarly to ATRA in altering mucosal homing capabilities, but we have further determined if this effect is dependent on ATRA or not. The rexinoids capable of exerting their effect independently of ATRA have the potential to replace ATRA in therapeutic settings, as they could provide a similar efficacy with a considerably reduced ability to induce toxicity. ATRA cooperating rexinoids also have potential for use in treatment settings; combining these compounds with a much lower dose of ATRA may enhance ATRA mediated effects while minimizing toxicity side effects.

It is well-established that effector T cell infiltration into affected tissues positively correlates with protection from viral infection and tumor regression, therefore identifying methods that can specifically impact their migratory ability may improve immune responses in these microenvironments (1, 14, 68, 69). Our discovery of several rexinoids that favorably modulate T cell mucosal homing abilities in vitro indicates that they may be useful as an adjuvant during vaccination towards viruses that infect mucosal surfaces, and in immunotherapies targeting tumors that form at mucosal sites. We have previously shown in mouse models that ATRA has the potential to function as an adjuvant; i.p. injection of ATRA during vaccination increases the number of virus specific T cells to mucosal regions and boosts protection during viral challenge (1). However, this treatment is physically taxing to the mice, resulting in weight loss and inflammation at the injection site. Changing the route of delivery may improve tolerability, however the tradeoff is a reduction in ATRA bioavailability. Our preliminary in vivo work has shown that mice are not subject to the same physical discomforts following rexinoid treatment delivered via i.p. injection, as observed by their minimal weight loss during treatment ( Figure 6A ). Furthermore, high dose rexinoid delivered orally was well-tolerated, which could compensate for reduced bioavailability when delivered a more preferable route, unlike high dose ATRA ( Figure 6B ). Thus, administering either the ATRA mimicking or ATRA cooperating rexinoids via the same route as ATRA may result in similar immune modulating activity, without the associated pathology.

Adoptive cell transfer (ACT) and immune checkpoint blockade (ICB) are immunotherapies currently showing great promise as cancer treatment modalities (70–73). The ability of our rexinoids to modulate T cell migration suggests that their use in tandem with either ACT or ICB therapy may enhance the efficacy of these treatments by directing more effector T cells to tumors at mucosal sites. With ACT, the treatment of ex vivo expanded tumor-specific T cells with rexinoids prior to re-infusion can result in more T cells effectively homing to the mucosal tumor, which would result in tumor reduction and possible elimination while avoiding the majority of toxicity issues associated with ATRA use in vivo. ICB therapy using a combination of PD-L1 and CTLA-4 blocking antibodies has been shown to reverse tumor-specific effector T cell exhaustion and increase the number of tumor-infiltrating lymphocytes (TILs) present, resulting in improved anti-tumor immune responses (68, 72, 74). Inhibitory interactions between TILs and tumor cells are blocked by anti-PD-L1, while the use of anti-CTLA-4 likely both promotes the activation of new tumor-specific T cells and overcomes regulatory T cell inhibitory pathways. Although promising, this approach currently displays limited efficacy in a subset of patients (69, 75, 76). This may be due to the newly activated T cells ineffectively migrating to the tumor site, resulting in the current TILs becoming overwhelmed, and subsequent re-loss of function. Coupling this ICB approach with our identified rexinoids may ameliorate treatment efficacy towards mucosal cancers by better directing the migration of newly activated tumor-specific T cells to these sites. This would result in larger numbers of functional effector T cells present in the mucosal tumors, resulting in improved cancer control and patient survival.

While this work focuses on immune function resulting from interactions between the RXR and RAR, it should be noted that the RXR is promiscuous. It is an essential partner for a multitude of other receptors, all of which require heterodimeric formation with the RXR to exert their function (26, 77). Rexinoids that did not affect RAR/RXR mediated transcription in terms of CCR9 and α₄β₇ expression may play a role in mediating expression of non-immune RAR/RXR dependent genes, or may influence the expression of genes under the control of other RXR heterodimers. The potential of rexinoid treatment to beneficially regulate a variety of biological processes is an exciting and growing research area.

The authors acknowledge that the data presented in this study must be deposited and made publicly available in an acceptable repository, prior to publication.

All animal studies were reviewed and approved by Arizona State University Institutional Animal Care and Use Committee (IACUC #19-1676R).

KM performed experiments, statistical analyses, and wrote the first draft of the manuscript. PM, CW, and PJ contributed to study design and rexinoid synthesis. CW and PJ wrote sections of the manuscript. MM wrote sections of the manuscript and performed statistical analyses. HD performed experiments. JB designed and implemented the study, and wrote sections of the manuscript. All authors read, revised, and approved the manuscript prior to submission.

This work was supported in part by funds received from the NIH (NIH SBIR AI089290-01 to JB, NIH R01 AI110720-03 to JB, NIH 1R15CA139364-01A2 to CW, co-Is: PM and PJ, and NIH 1R15CA249617-01 to CW and PJ, co-I: PM) and Arizona State University (ASU GPSA and SOLS RTI funds to KM, Barrett, The Honors College Honors Thesis Reimbursement funds to HD). Patent applications covering the technologies described in this work have been applied for on behalf of the Arizona Board of Regents. This content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.