Retinoic acid and host-pathogen interactions: effects on inducible nitric oxide synthase in vivo

Vitamin A and its metabolite retinoic acid modulate the host response to pathogens through poorly characterized mechanisms. In vitro studies have suggested that retinoic acid decreases inducible NO synthase (NOS2, or iNOS) expression, a component of innate immunity, in several cell types stimulated with lipopolysaccharide (LPS) or cytokines. This study investigated the effect of retinoic acid on LPS-stimulated NOS2 expression in vivo. Wistar-Kyoto rats received all-trans retinoic acid (RA, 10 mg/kg) or vehicle intraperitoneally daily for 5 days followed by LPS (4 mg/kg) or saline intraperitoneally and were killed 6 h later. NOS2 activation was estimated by mRNA (RT-PCR) and protein (Western-blot) expression and plasma nitrate/nitrite accumulation. In sharp contrast to previous in vitro study reports, RA significantly enhanced NOS2 mRNA, protein expression, and plasma nitrate/nitrite concentration in LPS-injected rats but not in saline-injected rats. This was associated with increased expression of interleukin-2, interferon (IFN)-gamma and IFN regulatory factor-1 mRNAs in several organs and increased IFN-gamma plasma concentration. RA significantly increased mortality in LPS-injected rats. The NOS inhibitor aminoguanidine (50 mg/kg before LPS injection) significantly attenuated the RA-mediated increase in mortality. These results demonstrate for the first time that RA supplementation in vivo enhances activation of the LPS-triggered NOS2 pathway.     

Mouse conceptuses have a limited capacity to elevate the mRNA level of cellular retinoid binding proteins in response to teratogenic doses of retinoic acid.

In these studies, we wished to determine the effect of teratogenic doses of retinoic acid on the expression of cellular retinoic acid binding protein I (CRABP-I) mRNA, cellular retinoic acid binding protein II (CRABP-II) mRNA, cellular retinol binding protein I (CRBP-I) mRNA, and cellular retinol binding protein II (CRBP-II) mRNA in mouse conceptuses. Levels of CRABP-II mRNA and CRBP-I mRNA were modestly elevated (2.5-fold and 1.5-fold, respectively) in 9-day gestation conceptuses following treatment of dams with 100 mg/kg b.w. of retinoic acid. These levels were elevated by 6 hr following treatment and remained elevated until 48 and 24 hr, respectively. Two other retinoids, etretinate and retinoyl beta-glucuronide, also moderately elevated CRABP-II mRNA and CRBP-I mRNA levels in conceptuses. In contrast, the levels of CRABP-I mRNA in the conceptuses remained unaffected by treatment with any of these three retinoids. These results demonstrate that conceptuses have a limited capacity to elevate the cellular retinoid binding proteins mRNA levels and presumably the synthesis of their respective proteins in response to high, teratogenic doses of retinoic acid. As a result, an excess of free retinoic acid becomes available to the nuclear retinoic acid receptors, which may lead to inappropriate gene expression and eventual maldevelopment.     

Hydrolysis of 4-HPR to atRA occurs in vivo but is not required for retinamide-induced apoptosis

The retinamide, N-(4-hydroxyphenyl)retinamide (4-HPR), has shown promising anti-tumor activity, but it is unclear whether this compound is hydrolyzed to all-trans retinoic acid (atRA) and if so, whether this plays any role in its chemotherapeutic activity. To address this issue, the ability of 4-hydroxybenzylretinone (4-HBR), a carbon-linked analog of 4-HPR, to support growth in vitamin A-deficient (VAD) animals and to activate an atRA-responsive gene in vivo was compared to 4-HPR and atRA. Further, the non-hydrolyzable 4-HBR analog was used to determine whether the presence of the labile amide linkage in 4-HPR is essential for the induction of apoptosis in cultured MCF-7 breast cancer cells. Studies in VAD rats showed that 4-HPR, like atRA, supports animal growth and induces CYP26B1 mRNA expression in lung whereas 4-HBR does not. Analysis of plasma from 4-HPR- and atRA-treated VAD animals revealed the presence of atRA whereas it was not detected in plasma from animals given 4-HBR. To determine whether hydrolysis to atRA is necessary for apoptosis induced by 4-HPR in MCF-7 breast cancer cells, morphological and biochemical assays for apoptosis were performed. 4-HBR, like 4-HPR, induced apoptosis in MCF-7 cells. Apoptosis was not induced even at high concentrations of atRA, showing that 4-HPR and 4-HBR act in cells via a distinct signaling pathway. These results show that although limited hydrolysis of 4-HPR occurs in vivo, the ability to liberate atRA is not required for these 4-hydroxyphenyl retinoids to induce apoptosis in MCF-7 breast cancer cells. Thus the non-hydrolyzable analog, 4-HBR, may have significant therapeutic advantage over 4-HPR because it does not liberate atRA that can contribute to the adverse side effects of drug administration in vivo.     

Retinoic acid-induced developmental defects are mediated by RARbeta/RXR heterodimers in the pharyngeal endoderm

Fusion and hypoplasia of the first two branchial arches, a defect typically observed in retinoic acid (RA) embryopathy, is generated in cultured mouse embryos upon treatment with BMS453, a synthetic compound that exhibits retinoic acid receptor beta (RARbeta) agonistic properties in transfected cells. By contrast, no branchial arch defects are observed following treatment with synthetic retinoids that exhibit RARalpha or RARgamma agonistic properties. The BMS453-induced branchial arch defects are mediated through RAR activation, as they are similar to those generated by a selective pan-RAR agonist, are prevented by a selective pan-RAR antagonist and cannot be mimicked by exposure to a pan-RXR agonist alone. They are enhanced in the presence of a pan-RXR agonist, and cannot be generated in Rarb-null embryos. Furthermore, they are accompanied, in the morphologically altered region, by ectopic expression of Rarb and of several other direct RA target genes. Therefore, craniofacial abnormalities characteristic of the RA embryopathy are mediated through ectopic activation of RARbeta/RXR heterodimers, in which the ligand-dependent activity of RXR is subordinated to that of RARbeta. Endodermal cells lining the first two branchial arches respond to treatment with the RARbeta agonist, in contrast to neural crest cells and ectoderm, which suggests that a faulty endodermal regionalization is directly responsible for RA-induced branchial arch dysmorphologies. Additionally, we provide the first in vivo evidence that the synthetic RARbeta agonist BMS453 exhibits an antagonistic activity on the two other RAR isotypes.     

Inhibition of Nitric Oxide Synthase Expression and Activity in Macrophages by 3-Hydroxyanthranilic Acid,a Tryptophan Metabolite

Indoleamine 2,3-dioxygenase (IDO) and nitric oxide synthase (NOS) type II are induced in macrophages by interferon (IFN)-γ and lipopolysaccharide (LPS). Nitric oxide has been previously shown to inhibit IDO activity. We studied whether metabolites of tryptophan via the IDO pathway could alter NOS II activity. In RAW 264.7 cells, the phenolic antioxidant 3-hydroxyanthranilic acid (OH-AA), but not anthranilic acid, inhibited citrulline synthesis by NOS II at sub-millimolar concentrations, when added 1 h before IFN-γ and LPS. OH-AA inhibited NOS II activity in cytosolic extracts, suggesting a direct action of OH-AA on NOS II protein. Moreover, expression of NOS II mRNA and activation of the nuclear factor κB (NF-κB) in RAW 264.7 cells were decreased by a pretreatment with OH-AA, but not anthranilic acid, before addition of IFN-γ and LPS. This pretreatment also inhibited activation of NF-κB in response to TNF-α in lymphoblastoid J.Jhan5-1 cells. Finally, expression of a long terminal repeat of the human immunodeficiency virus (HIV-LTR)-driven luciferase reporter gene, controlled by NF-κB activation, was severely decreased by OH-AA or 3-hydroxykynurenine in J.Jhan5-1 cells. Other tryptophan derivatives were inactive. These data identify OH-AA as an aminophenolic tryptophan derivative inhibiting NF-κB activation and impairing both NOS II expression and activity in a millimolar concentration range.     

Regulation of the proliferation of cocultured gonocytes and Sertoli cells by retinoids,triiodothyronine, and intracellular signaling factors: differences between fetal and neonatal cells

The regulation of early fetal germ cell growth has not been studied in cell culture, probably due to the poor survival of these cells. However, cell culture is the only system in which the control of cell growth can be studied independently of the influence of secreted testicular factors, which are diluted in the medium. We successfully cultured dispersed testicular cells from 16.5-day-old rat fetuses in defined medium and compared the growth of these cells with that of cells from 3-day-old neonates. In this system, fetal gonocytes displayed low levels of mitotic activity and their numbers remained stable. In contrast, neonatal gonocytes displayed high levels of mitotic activity and increased in number, these characteristics resembling those observed in vivo. We found that retinoic acid had deleterious effects on the number of gonocytes but did not affect Sertoli cell proliferation in fetal and neonatal cell cultures. Moreover, in fetal cell cultures, the decrease in the number of gonocytes resulted from a decrease in mitotic activity, probably due to a direct effect of retinoids on fetal gonocytes. Among the selective agonists for the retinoic acid receptor (RARalpha agonist, RARbeta agonist, and RARgamma agonist) and the retinoic X receptor (pan-RXR agonist) tested, only the RARalpha agonist reproduced the effects of retinoic acid at concentrations lower than its Kd value in both fetal and neonatal cell cultures. As both RARalpha and RXRalpha are present in fetal and neonatal gonocytes, we suggest that retinoic acid exerts its effects on gonocytes via a RARalpha-RXRalpha heterodimer, with RARalpha functioning as an active partner and RXRalpha as a passive partner. In this culture system, we show for the first time that triiodothyronine (T3) inhibits testicular fetal Sertoli cell and germ cell growth. We also tested intracellular signaling factors and found that a cAMP analog increased Sertoli cell proliferation and germ cell survival in both fetal and neonatal cells whereas phorbol esters (PMA) strongly inhibited the proliferation of fetal but not of neonatal gonocytes. None of the tested factors (T3, dbcAMP, and PMA) seemed to interact with the all-trans retinoic acid pathway. Thus, fetal gonocytes and neonatal gonocytes differ in intrinsic properties, and their growth is not regulated in the same manner. Despite their low level of mitotic activity, fetal gonocytes were more sensitive to various factors than neonatal gonocytes.     

Ligation of retinoic acid receptor alpha regulates negative selection of thymocytes by inhibiting both DNA binding of nur77 and synthesis of bim

Negative selection refers to the selective deletion of autoreactive thymocytes. Its molecular mechanisms have not been well defined. Previous studies in our laboratory have demonstrated that retinoic acids, physiological ligands for the nuclear retinoid receptors, selectively inhibit TCR-mediated death under in vitro conditions, and the inhibition is mediated via the retinoic acid receptor (RAR) alpha. The present studies were undertaken to investigate whether ligation of RARalpha leads to inhibition of TCR-mediated death in vivo and to identify the molecular mechanisms involved. Three models of TCR-mediated death were studied: anti-CD3-mediated death of thymocytes in wild-type mice, and Ag- and bacterial superantigen-driven thymocyte death in TCR-transgenic mice expressing a receptor specific for a fragment of pigeon cytochrome c in the context of the E(k) (class II MHC) molecule. Our data demonstrate that the molecular program of both anti-CD3- and Ag-driven, but not that of superantigen-mediated apoptosis involves up-regulation of nur77, an orphan nuclear receptor, and bim, a BH3-only member of the proapoptotic bcl-2 protein family, proteins previously implicated to participate in the negative selection. Ligation of RARalpha by the synthetic agonist CD336 inhibited apoptosis, DNA binding of nur77, and synthesis of bim induced by anti-CD3 or the specific Ag, but had no effect on the superantigen-driven cell death. Our data imply that retinoids are able to inhibit negative selection in vivo as well, and they interfere with multiple steps of the T cell selection signal pathway.     

Induction of β-retinoic acid receptor mRNA by teratogenic doses of retinoids in murine fetuses

Retinoic acid is required for normal growth and development, however excessive doses are teratogenic. Recently several nuclear retinoic acid receptors (RAR) have been identified and postulated to mediate the response of retinoic acid at the gene level. We wished to determine if alpha-RAR mRNA or beta-RAR mRNA levels are modulated by teratogenic doses of retinoic acid in vivo. We have found that beta-RAR mRNA levels in 9-day-gestation mouse conceptuses are increased as early as 3 h after administration of a completely teratogenic dose of retinoic acid (100 mg/kg body weight; b.w.) and reach a maximum of approximately sixfold after 6 h of treatment. Maternal liver and maternal kidney demonstrated a similar pattern of increase in beta-RAR mRNA, however this was only approximately threefold. Retinoic acid dose-response experiments demonstrated a reduced increase of beta-RAR mRNA levels with 10 mg/kg b.w. (minimally teratogenic dose), and no increase with a more-physiological dose of 1 mg/kg b.w. in the conceptuses. beta-RAR mRNA levels were elevated in 18-day-gestation fetuses to a similar extent to that observed in the 9-day-gestation conceptuses. Therefore, the twofold difference in the extent to which beta-RAR mRNA levels increase does not occur because the fetuses are at a developmental stage that is sensitive to the teratogenic effects of retinoic acid. Finally, treatment with another teratogenic retinoid, etretinate, and a nonteratogenic retinoid, retinoyl beta-glucuronide, both resulted in increase in the level of beta-RAR mRNA in the conceptuses and the maternal tissues. Therefore, an increase in beta-RAR mRNA levels caused by treatment with retinoids does not necessarily commit a fetus to undergo an abnormal pattern of development characteristic of teratogenic retinoids.     

Lipopolysaccharide-induced increase of prostaglandin E(2) is mediated by inducible nitric oxide synthase activation of the constitutive cyclooxygenase and induction of membrane-associated prostaglandin E synthase

NO produced by the inducible NO synthase (NOS2) and prostanoids generated by the cyclooxygenase (COX) isoforms and terminal prostanoid synthases are major components of the host innate immune and inflammatory response. Evidence exists that pharmacological manipulation of one pathway could result in cross-modulation of the other, but the sense, amplitude, and relevance of these interactions are controversial, especially in vivo. Administration of 6 mg/kg LPS to rats i.p. resulted 6 h later in induction of NOS2 and the membrane-associated PGE synthase (mPGES) expression, and decreased constitutive COX (COX-1) expression. Low level inducible COX (COX-2) mRNA with absent COX-2 protein expression was observed. The NOS2 inhibitor aminoguanidine (50 and 100 mg/kg i.p.) dose dependently decreased both NO and prostanoid production. The LPS-induced increase in PGE(2) concentration was mediated by NOS2-derived NO-dependent activation of COX-1 pathway and by induction of mPGES. Despite absent COX-2 protein, SC-236, a putative COX-2-specific inhibitor, decreased mPGES RNA expression and PGE(2) concentration. Ketoprofen, a nonspecific COX inhibitor, and SC-236 had no effect on the NOS2 pathway. Our results suggest that in a model of systemic inflammation characterized by the absence of COX-2 protein expression, NOS2-derived NO activates COX-1 pathway, and inhibitors of COX isoforms have no effect on NOS2 or NOS3 (endothelial NOS) pathways. These results could explain, at least in part, the deleterious effects of NOS2 inhibitors in some experimental and clinical settings, and could imply that there is a major conceptual limitation to the use of NOS2 inhibitors during systemic inflammation.     

Negative regulation of LPS-stimulated expression of inducible nitric oxide synthase by AP-1 in macrophage cell line J774A.1.

The level of NOS II mRNA was markedly increased during 24 h lipopolysaccharide (LPS) stimulation, but showed no further increase thereafter. On the other hand, the level of NOS II mRNA in J774A.1 cells transfected with an expression vector containing the rat csk cDNA (J.Csk) was significantly increased during 3 h LPS stimulation, but rather decreased thereafter. Although no significant difference was observed in the activation of NF-kappaB by LPS among parental J774A.1, J774A.1 transfected with promoterless vector (J.pBK), and J.Csk cells, activity of c-Jun N-terminal kinase (JNK) and nuclear translocation of nuclear factor activator protein-1 (AP-1) were markedly upregulated in the J.Csk cells. Then luciferase reporter vectors containing NOS II promoter with mutations in two AP-1-like sites (U site, -1126 approximately -1120; L site, -524 approximately -518) were transiently transfected in J774A.1 cells. The promoter activity following LPS stimulation for 24 h was significantly increased by mutation at the L site, but not by mutation at the U site, suggesting that NOS II expression is negatively regulated, at least in part, through the AP-1-like L site in response to LPS.     

Identification and characterization of the retinoic acid response elements in the human RIG1 gene promoter

The expression of retinoic acid-induced gene 1 (RIG1), a class II tumor suppressor gene, is induced in cells treated with retinoids. RIG1 has been shown to express ubiquitously and the increased expression of this gene appears to suppress cell proliferation. Recent studies also demonstrated that this gene may play an important role in cell differentiation and the progression of cancer. In spite of the remarkable regulatory role of this protein, the molecular mechanism of RIG1 expression induced by retinoids remains to be clarified. The present study was designed to study the molecular mechanism underlying the all-trans retinoic acid (atRA)-mediated induction of RIG1 gene expression. Polymerase chain reaction was used to generate a total of 10 luciferase constructs that contain various fragments of the RIG1 5'-genomic region. These constructs were then transfected into human gastric cancer SC-M1 and breast cancer T47D cells for transactivation analysis. atRA exhibited a significant induction in luciferase activity only through the -4910/-5509 fragment of the 5'-genomic region of RIG1 gene relative to the translation initiation site. Further analysis of this promoter fragment indicated that the primary atRA response region is located in between -5048 and -5403 of the RIG1 gene. Within this region, a direct repeat sequence with five nucleotide spacing, 5'-TGACCTctattTGCCCT-3' (DR5, -5243/-5259), and an inverted repeat sequence with six nucleotide spacing, 5'-AGGCCAtggtaaTGGCCT-3' (IR6, -5323/-5340), were identified. Deletion and mutation of the DR5, but not the IR6 element, abolished the atRA-mediated activity. Electrophoretic mobility shift assays with nuclear extract from atRA-treated cells indicated the binding of retinoic acid receptor (RAR) and retinoid X receptor (RXR) heterodimers specifically to this response element. In addition to the functional DR5, the region contains many other potential sequence elements that are required to maximize the atRA-mediated induction. Taken together, we have identified and characterized the functional atRA response element that is responsible for the atRA-mediated induction of RIG1 gene.