Several clinically useful retinoids including 13-cis RA (Isotretinoin or Accutane®), and 4-HPR (fenretinide), differ from ATRA (Tretinoin®) and 9-cis RA (Panretin®) in their biological activity and/or toxicity. These differences in activity are thought to be primarily due to their differential ability to bind RAR or RXR and their subsequent ability to induce gene transactivation and/or transrepression. 13-cis-RA, in conjunction with IFN-γ is an effective chemotherapeutic agent for squamous cell carcinoma of the skin and cervix and is useful in the chemoprevention of secondary head and neck tumors [13]. The RAR and RXR binding and transactivating ability of 13-cis RA are controversial. Several studies have suggested that 13-cis RA possesses a moderate affinity for RARs, especially RAR-α, but possesses little to no affinity or activating potential for RXRs [8, 39–41]. However, like ATRA and 9-cis RA, in vitro conversion of 13-cis RA to other bioactive retinoids has been demonstrated [9]. 4-HPR, a synthetic retinoid, exerts potent chemopreventive action in animal models of carcinogenesis; however, its effectiveness in humans remains to be determined [42]. Like 13-cis RA, the ability of 4-HPR to bind and transactivate RARs or RXRs is a matter of controversy [43–46]. In studies examining 4-HPR interactions with these receptors, 4-HPR appears to bind to and transactivate RAR-β and RAR-γ but not RAR-α and RXRs [47]. 4-HPR does not undergo appreciable transformation to retinoids such as ATRA [42].
Several lines of evidence presented within the current report suggest that RAR-α-selective retinoids can recapitulate all of the previously observed effects of retinoids on the differentiation process of Th2 cells: (1) ATRA and 13-cis RA, which operate mainly via RARs, induces type 2 cytokine expression; (2) 4-HPR, a potential transactivator of RAR-β/γ(but not RAR-α or RXRs, exerts minimal effects on expression of IL-4, IL-5, and IL-13; (3) TTNPB, a selective activator of RARs, is capable of inducing cytokine expression similar to ATRA, 9-cis and 13-cis RA (H. Dawson and D.D. Taub, unpublished observation); (4) AM580 stimulates type 2 cytokine synthesis to the same degree as ATRA; (5) the RAR- α-selective antagonist, Ro 41–5254, inhibits the expression of Th2-like cytokines; and (6) the use of an RXR-selective agonist alone or in combination with ATRA or AM580, exerts no effect on type 2 cytokine expression. Together, these results suggest that in vitro human Th2 cell differentiation is predominantly regulated by retinoids through the RAR-α-dependent signaling pathway.
There are a very limited number of studies that have examined the role of specific RARs in the regulation of IL-4 production by rodent and human T cells. In one study, the pan RAR selective retinoid, TTNPB, was active, but far less effective, than either ATRA or 9-cis RA in inducing IL-4 synthesis from antigen-stimulated murine T cells [48]. However, the role of the RAR receptor in this regulation is unclear because the RAR-α-selective retinoid, Am-80, failed to affect IL-4 production by antigen-stimulated murine Th2 clones [23] or antigen-stimulated murine T cells [21] and in vitro antigen-stimulated murine T cell IL-4 synthesis was stimulated by 9-cis-RA and an RXR-selective retinoid, but not by ATRA or an RAR-selective retinoid [49]. The effects of receptor-selective retinoids on the production of other Th2 cytokines were not examined in these studies. The current report has directly demonstrated an RAR-α-specific effect on type 2 cytokine expression on human PBMC and T cells in the absence of accessory cells.
Our current data appear to contradict findings of two published studies where 13-cis, ATRA, and 9-cis-RA were all equipotent in inhibiting IFN-γ production from rodent Th1 cells [25]. In addition, in vitro antigen-stimulated murine T cell IL-4 synthesis was stimulated by 9-cis-RA and an RXR-selective retinoid, but not by ATRA or an RAR-selective retinoid [49]. Our data regarding RAR-α also differ with those of another report where an experimental RAR-α-selective retinoid failed to affect the production of IFN-γ by Th1 clones and IL-4 by mouse Th2 T cell clones [23] or antigen-stimulated naïve T cells [21]. In a more recent study, ATRA, 9-cis-RA, and an RAR agonist were found to be potent inducers of cellular proliferation and IL-2 production by enriched human T cells [50]. These effects were found to occur predominantly with RAR agonists and only marginal effects being observed through the use of RXR agonists. More interestingly, while we have also observed increased activation and proliferation by human T cells upon culture with RAR agonists, no significant changes in the production of IL-2 (another type 1 cytokine) were observed post RAR agonist treatment in our hands (data not shown). In addition to authentic species and differences in biological activities of synthetic retinoid putatively acting through the same receptor, simple differences between cell purification procedures, culture conditions, activation regiment, age of donors, memory/naïve and CD4/CD8 ratios, nutritional status, species and the donor population being examined may account for differences between studies. However, we have found similar results with ATRA, 9-cis- RA and Am580 when using pig PBMCs (H. Dawson, unpublished).
One potential limitation of our study is the relatively weak antagonistic activity demonstrated by the RAR- antagonist, Ro 41–5254. This antagonist binds to RAR-α with four and two times less affinity than ATRA and AM580, respectively [51]. Previous studies have effectively utilized this compound only at a molar ratio of 1000:1 to 100:1 of antagonist to agonist [35, 51]. Using the inhibition of induction of CD38 antigen expression by ATRA as a bioassay for RAR-α-mediated activation, this compound exhibited a 10%, 40%, and 80% inhibitory activity at a molar ratio of 10:1, 100:1, and 1,000:1, respectively (data not shown). Using a 100:1 molar ratio of antagonist to agonist (AM580 or ATRA), we have also observed a 60–70% reduction in the synthesis of Th2-associated cytokines by CD3-stimulated PBMC (data not shown). There is recent evidence that Ro 41–5254 is a peroxisome proliferator-activated receptor-gamma (PPAR-γ) agonist [52]. PPAR-γ agonists can inhibit Th1 and Th2 responses. Although Ro 41–5254 exhibited little activity in this regard in our system, we cannot rule out interactions between the retinoid and PPAR pathways. Another limitation is that we cannot rule out the possibility that there was sufficient endogenous RXR-active retinoids in this system to permit RAR- and liganded RXR-mediated nuclear transactivation. However, as normal serum (similar to the serum utilized in our culture system) contains negligible concentrations of 9-cis RA but appreciable quantities of other retinoids such as ATRA (10 to 100 nM) and 13-cis RA (2.5 to 25 nM), we feel that this is unlikely.
RAR-α has at least two isoforms, RAR-α-1 and RAR-α-2, which arise as a consequence of differential splicing. Analysis of transgenic mice, which under- or overexpress RAR-α-1, has revealed limited effects on the development, phenotype and/or function of the immune system [53]. Adult mice expressing low amounts of RAR-α develop spontaneous T and B cell lymphomas and had very low amounts of the RA-inducible gene, CD38, in thymus and bone marrow [35, 54]. In vivo experiments have suggested that RAR-α agonists may exhibit inhibitory effects on T-cell mediated immunity. Systemic administration of RAR-α-selective retinoids have been shown to inhibit DTH [21, 29], the progression of experimental arthritis [23], prolong skin allograft survival [30], and inhibit Ab production in mice [29].
Our proposed model of the role of RAR-α in Th2 development has parallels in other cell types. Recent studies suggest that the differential engagement of individual RARs or RXRs may regulate the process of monocyte differentiation into macrophages or dendritic cells [55] and the induction or prevention of apoptosis of B cells and T cells [14, 56], as well as normal neutrophil differentiation [57]. Several recent studies have focused on the effects of receptor-selective retinoids on accessory cells in the regulation of Th1 or Th2-like cytokine production by T cells. 9-cis-RA and the RXR-selective retinoid, LG69, have been shown to inhibit LPS-stimulated IL-12 production from murine macrophages more effectively than the pan RAR-selective retinoid, TTNPB [39]. Similarly, 9-cis-RA was more effective than TTNPB in downregulating IL-12 protein synthesis in LPS- or KLH-stimulated mouse macrophages [48]. More recent human studies have demonstrated a direct effect of ATRA and RXR rexinoids on IL-2 receptor expression by human cutaneous T-cell lymphomas through RAR and RXR receptors (45,46). Overall, the bulk of evidence from all of these studies point to a predominant role for liganded RXRs in these processes.
