Simultaneous determination of all-trans-, 13-cis- and 9-cis-retinoic acid, their 4-oxo metabolites and all-trans-retinol in human plasma by high-performance liquid chromatography

All-trans-retinoic acid (all-trans-RA) and 13-cis-retinoic acid (13-cis-RA), due to their effects on cell differentiation, proliferation and angiogenesis, improved treatment results in some malignancies. Pharmacokinetic studies of all-trans-RA and 13-cis-RA along with monitoring of retinoic acid metabolites may help to optimize retinoic acid therapy and to develop new effective strategies for the use of retinoic acids in cancer treatment. Therefore, we developed a HPLC method for the simultaneous determination in human plasma of the physiologically important retinoic acid isomers, all-trans-, 13-cis- and 9-cis-retinoic acid, their 4-oxo metabolites, 13-cis-4-oxoretinoic acid (13-cis-4-oxo-RA) and all-trans-4-oxoretinoic acid (all-trans-4-oxo-RA), and vitamin A (all-trans-retinol). Analysis performed on a silica gel column with UV detection at 350 nm using a binary multistep gradient composed on n-hexane, 2-propanolol and glacial acetic acid. For liquid-liquid extraction a mixture of n-hexane, dichloromethane and 2-propanolol was used. The limits of detection were 0.5 ng/ml for retinoic acids and 10 ng/ml for all-trans-retinol. The method showed good reproducibility for all components (within-day C.V.: 3.02–11.70%; day-to-day C.V.: 0.01–11.34%. Furthermore, 9-cis-4-oxoretinoic acid (9-cis-4-oxo-RA) is separated from all-trans-4-oxo-RA and 13-cis-4-oxo-RA. In case of clinical use of 9-cis-retinoic acid (9-cis-RA) the pharmacokinetics and metabolism of this retinoic acid isomer can also be examined.     

Determination of 13-cis-retinoic acid and its major metabolite, 4-oxo-13-cis-retinoic acid, in human blood by reversed-phase high-performance liquid chromatography

A high-performance liquid chromatography (HPLC) method for the quantitation of 13-cis-retinoic acid (13-cis-RA) and its major metabolite, 4-oxo-13-cis-RA, in human blood has been developed. The method includes extraction of 1 ml of blood with diethyl ether at pH 6 and the analysis of the extract by reversed-phase HPLC with solvent programming and detection at 365 nm. The quantitation ranges for 13-cis-RA and 4-oxo-13-cis-RA are 10–2000 and 50–2000 ng/ml of blood, respectively. The method also provides estimates of the concentrations of all-trans-RA and 4-oxo-all-trans-RA. The mean intra- and inter-assay variabilities for all four compounds were 6% or less. The method separates 13-cis-RA and 4-oxo-13-cis-RA from 9-cis-RA, all-trans-RA, 4-oxo-all-trans-RA, and some other possible metabolites, such as hydroxy and epoxy retinoic acids. The method has been successfully applied to the analyses of over 1200 blood samples from four 13-cis-RA clinical studies.     

Mouse alcohol dehydrogenase 4: kinetic mechanism,substrate specificity and simulation of effects of ethanol on retinoid metabolism

Mouse ADH4 (purified, recombinant) has a low catalytic efficiency for ethanol and acetaldehyde, but very high activity with longer chain alcohols and aldehydes, at pH 7.3 and temperature 37°C. The observed turnover numbers and catalytic efficiencies for the oxidation of all-trans-retinol and the reduction of all-trans-retinal and 9-cis-retinal are low relative to other substrates; 9-cis-retinal is more reactive than all-trans-retinal. The reduction of all-trans- or 9-cis-retinals coupled to the oxidation of ethanol by NAD⁺ is as efficient as the reduction with NADH. However, the Michaelis constant for ethanol is about 100 mM, which indicates that the activity would be lower at physiologically relevant concentrations of ethanol. Simulations of the oxidation of retinol to retinoic acid with mouse ADH4 and human aldehyde dehydrogenase (ALDH1), using rate constants estimated for all steps in the mechanism, suggest that ethanol (50 mM) would modestly decrease production of retinoic acid. However, if the K_m for ethanol were smaller, as for human ADH4, the rate of retinol oxidation and formation of retinoic acid would be significantly decreased during metabolism of 50 mM ethanol. These studies begin to describe quantitatively the roles of enzymes involved in the metabolism of alcohols and carbonyl compounds.     

Embryonal Carcinoma Cell Lines Stably Transfected with mRARβ2-lacZ: Sensitive System for Measuring Levels of Active Retinoids

Embryonal carcinoma cell lines (F9 EC and P19 EC) were stably transfected with 1.8 kb promoter sequence of RARβ2 coupled to the lacZ gene as a system for measuring active retinoids. These stable transfectants, designated F9-1.8 and P19-1.8, were used as reporter cell lines to investigate different retinoids for their ability to activate the reporter gene. F9-1.8 cells showed similar EC₅₀ values for the acidic retinoids all-trans retinoic acid (RA), 4-oxo RA, 9-cis RA, and 13-cis RA, in the range of 1–7 nM, while P19-1.8 cells were less sensitive. Retinal showed decreased activity compared to the RA isomers in both lines. However, P19-1.8 cells hardly showed β-gal activity after treatment with retinol, while the lacZ reporter in F9-1.8 cells was still inducible by this retinoid. In addition, the reporter system was used to investigate RA metabolism and its inhibition by P450 inhibitors. A combination of RA and liarozole showed a 10 times greater induction of the RARβ2-lacZ reporter in P19-1.8 cells, but not in F9-1.8 cells. The EC₅₀ value for 4-oxo RA, however, was not altered, indicating that metabolic conversion of RA to 4-oxo RA is the target for inhibition by liarozole in P19-1.8 cells. HPLC analysis revealed nearly complete inhibition of RA metabolism after liarozole treatment in P19-1.8 cells, resulting in higher levels of RA. Finally, the F9-1.8 cells were used to detect active retinoids during different stages of chick limb bud development, demonstrating that it is the limb bud mesenchyme which generates RA and not the epidermis, with a twofold higher level of RA in the posterior half than in the anterior half.     

Binding affinities of CRBPI and CRBPII for 9-cis-retinoids

Background

Cellular retinol binding-protein I (CRBPI) and cellular retinol binding-protein II (CRBPII) serve as intracellular retinoid chaperones that bind retinol and retinal with high affinity and facilitate substrate delivery to select enzymes that catalyze retinoic acid (RA) and retinyl ester biosynthesis. Recently, 9-cis-RA has been identified in vivo in the pancreas, where it contributes to regulating glucose-stimulated insulin secretion. In vitro, 9-cis-RA activates RXR (retinoid × receptors), which serve as therapeutic targets for treating cancer and metabolic diseases. Binding affinities and structure–function relationships have been well characterized for CRBPI and CRBPII with all-trans-retinoids, but not for 9-cis-retinoids. This study extended current knowledge by establishing binding affinities for CRBPI and CRBPII with 9-cis-retinoids.

Methods

We have determined apparent dissociation constants, K′_d, through monitoring binding of 9-cis-retinol, 9-cis-retinal, and 9-cis-RA with CRBPI and CRBPII by fluorescence spectroscopy, and analyzing the data with non-linear regression. We compared these data to the data we obtained for all-trans- and 13-cis-retinoids under identical conditions.

Results

CRBPI and CRBPII, respectively, bind 9-cis-retinol (K′_d, 11 nM and 68 nM) and 9-cis-retinal (K′_d, 8 nM and 5 nM) with high affinity. No significant 9-cis-RA binding was observed with CRBPI or CRBPII.

Conclusions

CRBPI and CRBPII bind 9-cis-retinol and 9-cis-retinal with high affinities, albeit with affinities somewhat lower than for all-trans-retinol and all-trans-retinal.

General significance

These data provide further insight into structure–binding relationships of cellular retinol binding-proteins and are consistent with a model of 9-cis-RA biosynthesis that involves chaperoned delivery of 9-cis-retinoids to enzymes that recognize retinoid binding-proteins.     

Effects of visual pigments on the organization of phospholipid multibilayer model membranes--a spin probe study

A cholestane spin probe was used to study the effects of the all-trans isomers of retinal and retinol, and the 13-cis- isomer of retinal on the degree of organization of oriented phospholipid multibilayer model membranes in the dark. Concentration-dependent effects were observed indicative of changes in the degree of organization of phosphatidyl choline, phosphatidyl ethanolamine, phosphatidyl serine, sphingomyelin, and brain lipids. In general the all-trans compounds improved the degree of order of the lipid films at low concentrations, and decreased it at high concentrations. The 13-cis retinal served only to decrease the degree of order. The magnitudes of these effects depend on lipid composition and the nature of the pigment polar residue. Chemical as well as nonbonding interactions are considered to be responsible for these various effects.     

Induction of Normal Cardiovascular Development in the Vitamin A-Deprived Quail Embryo by Natural Retinoids

The biological activity of various natural retinoids and the time "window" when vitamin A activity is required for normal cardiovascular development were examined in vitamin A-deprived Japanese quail embryos. The administration of 1 μg of retinol at the beginning of incubation resuited in normal cardiovascular development in 97% of embryos; retinoic acid was toxic at this dose level. Treatment of embryos with 0.1 μg of all-trans-retinol or 13-cis-retinoic acid at the beginning of incubation resulted in normal cardiovascular development in 47 and 12% of embryos, respectively; administration of these retinoids at other time points attenuated the percentage of embryos with normal cardiovascular development. Didehydroretinol, 0.1 μg, and 9-cis-retinoic acid, 0.1 μg, were inactive at all time points examined; 9-cis-retinoic acid did not enhance the biological activity of all-trans-retinoic acid. All-trans-retinoic acid, 0.1 μg, administered during 22-28 hr of incubation induced normal cardiovascular development in 20-34% of embryos; biological activity was optimal when it was administered at 24 hr. All retinoids tested were inactive in establishing normal cardiovascular development when administered at 36 hr of incubation or later. The studies suggest that all-trans-retinoic acid is the biologically active form of vitamin A required for normal cardiovascular development in the avian embryo. There is a critical time point within the first 22-28 hr of quail embryogenesis when all-trans-retinoic acid initiates events that lead to normal cardiovascular development.     

Bioinorganic and bioorganic studies of liver alcohol dehydrogenase

Liver alcohol dehydrogenase (E.C.1.1.1.1) is an NAD⁺/NADH dependent enzyme with a broad substrate specificity being active on an assortment of primary and secondary alcohols. It catalyzes the reversible oxidation of a wide variety of alcohols to the corresponding aldehydes and ketones as well as the oxidation of certain aldehydes to their related carboxylic acids. Although the bioinorganic and bioorganic aspects of the enzymatic mechanism, as well as the structures of various ternary complexes, have been extensively studied, the kinetic significance of certain intermediates has not been fully evaluated. Nevertheless, the availability of computer-assisted programs for kinetic simulation and molecular modeling make it possible to describe the biochemical mechanism more completely. Although the true physiological substrates of this zinc metalloenzyme are unknown, alcohol dehydrogenase effectively catalyzes not only the interconversion of all-trans-retinol and all-trans-retinal but also the oxidation of all-trans-retinal to the corresponding retinoic acid. Retinal and related vitamin A derivatives play fundamental roles in many physiological processes, most notably the vision process. Furthermore, retinoic acid is used in dermatology as well as in the prevention and treatment of different types of cancer. The enzyme-NAD⁺-retinol complex has an apparent pK_a value of 7.2 and loses a proton rapidly. Proton inventory modeling suggests that the transition state for the hydride transfer step has a partial negative charge on the oxygen of retinoxide. Spectral evidence for an intermediate such as E-NAD⁺-retinoxide was obtained with enzyme that has cobalt(II) substituted for the active site zinc(II). Biophysical considerations of water in these biological processes coupled with the inverse solvent isotope effect lead to the conclusion that the zinc-bound alkoxide makes a strong hydrogen bond with the hydroxyl group of Ser48 and is thus activated for hydride transfer. Moderate pressure accelerates enzyme action indicative of a negative volume of activation. The data with retinol is discussed in terms of enzyme stability, mechanism, adaptation to extreme conditions, as well as water affinities of substrates and inhibitors. Our data concern all-trans, 9-cis, 11-cis, and 13-cis retinols as well as the corresponding retinals. In all cases the enzyme utilizes an approximately ordered mechanism for retinol–retinal interconversion and for retinal–retinoic acid transformation.     

Quantification of retinyl-β-glucuronides in rat urine by reversed-phase high-performance liquid chromatography with ultraviolet detection

A method is presented for the quantitation of the glucuronide conjugates of 4-oxo-all-trans-, 4-oxo-13-cis-, 13-cis-, 9-cis- and all-trans-retinoic acids in rat urine utilizing solid-phase extraction and gradient reversed-phase HPLC. The range of the R.S.D. (relative standard deviation) for both the inter- and intra-assay precision was 1.45,2–11.60%. The recovery of all retinoyl-β-glucuronides from rat urine ranged between 89 and 99%. The limit of detection was 0.01 μg/ml using 5 ml of rat urine. This method was applied to quantitate the amount of retinoyl-β-glucuronides produced in urine after the single and multiple oral administrations of 13-cis-, 9-cis- and all-trans-retinoic acids to rats.     

Biosynthesis of 9-cis-Retinoic Acid from 9-cis-β-Carotene in Human Intestinal Mucosa in Vitro

The role of 9-cis-β-carotene (9-cis-β-C) as a potential precursor of 9-cis-retinoic acid (9-cis-RA) has been examined in human intestinal microcosa in vitro. By using HPLC, uv spectra, and chemical derivatization analysis, both 9-cis-RA and all-trans-retinoic acid (all-trans-RA) have been identified in the postnuclear fraction of human intestinal microcosa after incubation with 9-cis-β-C at 37°C. The biosynthesis of both 9-cis-RA and all-trans-RA from 9-cis-β-C was linear with increasing concentrations of 9-cis-β-C (2-30 μM) and was linear with respect to tissue protein concentration up to 0.75 mg/ml. Retinoic acid was not detected when a boiled incubation mixture was incubated in the presence of 9-cis-β-C. The rate of synthesis of 9-cis- and all-trans-RA from 4 μM 9-cis-β-C were 16 ± 1 and 18 ± 2 pmol/hr/mg of protein, respectively. However, when 2 μM all-trans-β-C was added to the 4 μM 9-cis-β-C, the rate of all-trans-RA synthesis was increased to 38 ± 6 pmol/hr/mg of protein, whereas the rate of 9-cis-RA synthesis remained the same. These results suggest that 9-cis-RA is produced directly from 9-cis-β-C. Furthermore, incubations of either 0.1 μM 9-cis- or all-trans-retinal under the same incubation conditions showed that 9-cis-RA could also arise through oxidative conversion of 9-cis-retinal. Although only 9-cis-RA was detected when 9-cis-RA was used as the substrate, the isomerization of the all-trans-RA to 9-cis-RA cannot be ruled out, since both all-trans-RA and trace amounts of 9-cis-RA were detected when all-trans-retinal was incubated as the substrate. These data indicate that 9-cis-β-C can be a source of 9-cis-RA in the human. This conversion may have a significance in the anticarcinogenic action of β-C.     

Metabolism of 13-cis-retinoic acid: Identification of 13-cis-retinoyl-and 13-cis-4-oxoretinoyl-β-glucuronides in the bile of vitamin A-normal rats

The metabolism of 13-cis-[11-³H]retinoic acid has been examined in vitamin A-normal rats. Within 24 h after intravenous administration of the parent retinoid (15 μg/kg) to animals with biliary fistulas, 69 ± 9% of the dose was detected in the bile with 9 ± 6% being found in the urine. Analysis of the bile by reverse-phase high-pressure liquid chromatography demonstrated that the retinoic acid was being metabolized to several more polar compounds. A number of these compounds were sensitive to incubation with β-glucuronidase as evidenced by a change in their chromatographic behavior after treatment with the enzyme. Two of the metabolites have been identified as 13-cis-4-oxoretinoyl-β-glucuronide (8.1 ± 1.0% of the dose during the first 4 h after administration of the parent compound) and 13-cis-retinoyl-β-glucuronide (7.0 ± 4.4% of the dose). A comparison of the chromatographic profiles of bile from 13-cis- versus all-trans-retinoic acid-treated rats indicated a difference in their metabolism, with a greater proportion of the all-trans-retinoic acid being converted to compounds that eluted in the more polar regions of the column effluent.     

Kinetic characterization of recombinant mouse retinal dehydrogenase types 3 and 4 for retinal substrates

Background

Retinal dehydrogenases (RALDHs) catalyze the dehydrogenation of retinal into retinoic acids (RAs), which are required for embryogenesis and tissue differentiation. This study sought to determine the detailed kinetic properties of 2 mouse RALDHs, namely RALDH3 and 4, for retinal isomer substrates, to better define their specificities in RA isomer synthesis.

Methods

RALDH3 and 4 were expressed in Escherichia coli as His-tagged proteins and affinity-purified. Enzyme kinetics were performed with retinal isomer substrates. The enzymatic products were analyzed by high pressure liquid chromatography.

Results

RALDH3 oxidized all-trans retinal with high catalytic efficiency (V_max/K_m = 77.9) but did not show activity for either 9-cis or 13-cis retinal substrates. On the other hand, RALDH4 was inactive for all-trans retinal substrate, exhibited high activity for 9-cis retinal oxidation (V_max/K_m = 27.4), and oxidized 13-cis retinal with lower catalytic efficiency (V_max/K_m = 8.24). β-ionone, a potent inhibitor of RALDH4 activity, suppressed 9-cis and 13-cis retinal oxidation competitively with inhibition constants of 0.60 and 0.32, respectively, but had no effect on RALDH3 activity. The divalent cation MgCl₂ activated 13-cis retinal oxidation by RALDH4 by 3-fold, did not significantly influence 9-cis retinal oxidation, and slightly activated RALDH3 activity.

Conclusions

These data extend the kinetic characterization of RALDH3 and 4, providing their specificities for retinal isomer substrates.

General significance

The kinetic characterization of RALDHs should give useful information in determining amino acid residues that are involved in the specificity for retinal isomers and on the role of these enzymes in the synthesis of RAs in specific tissues.     

Retinal-protein interactions in bacteriorhodopsin monomers,dispersed in the detergent L-1690

Bacteriorhodopsin monomer dispersed in a solution of the detergent L-1690 could maintain the specific interaction between retinal and protein in the pH range 9.0-0.0 at 25°C. λ_max of the absorbance spectrum was 550 nm at pH 9.0, 556 nm at pH 5.5, 609 nm at pH 2.1 and 570 nm at pH 0.0. Increasing the NaCl concentration in the solution promoted formation of the 609 nm product at pH 5.0-3.0 and also its transition to the 570 nm product at pH 2.5-1.0. Retinal isomer analysis gave a ratio of 13-cis- to all-trans-retinal of 53 : 47 at pH 5.5. When the pH of the solution was reduced, the relative content of all-trans-retinal increased and the ratio of 13-cis- to all-trans-retinal was 14 : 86 at pH 0.0. Illumination of the solution at pH 7.2 yielded a product containing 9-cis-retinal or 9-cis, 13-cis-retinal, which may be due to a reaction other than the photoreaction cycle.     

Configuration of the carotenoid in the reaction centers of photosynthetic bacteria. Comparison of the resonance Raman spectrum of the reaction center of Rhodopseudomonas sphaeroides G1C with those of cis-trans isomers of β-carotene

The resonance Raman spectrum of the reaction center of Rhodopseudomonas sphaeroides G1C as well as those of the cis-trans isomers of β-carotene (all-trans, 9-cis, 13-cis, 15-cis and 9-cis, 13-cis- (or 9-cis, 13′-cis)) have been recorded at liquid N₂ temperature by use of the 457.9, 488.0 and 514.5 nm excitation lines. Comparison of the spectra indicated that the carotenoid in the reaction center takes the 15-cis configuration.     

Membrane phospholipids and the dark side of vision

The key step in the visual pigment regeneration process is an enzyme-catalyzedtrans tocis retinoid isomerization reaction. This reaction is of substantial general interest, because it requires the input of metabolic energy. The energy is needed because the 11-cis-retinoid reaction products are approximately 4kcal/mol higher in energy than their all-trans congeners. In the retinal pigment epithelium a novel enzymatic system has been discovered which is capable of converting all-trans-retinol into all-trans retinyl esters, by means of a lecithin retinol acyl transferase (LRAT), followed by the direct processing of the ester into 11-cis-retinol. In this process the free energy of hydrolysis of a retinyl ester, estimated to be approximately –5kcal/mol, is coupled to the endothermic (+4kcal/mol) isomerization reaction, resulting in an overall exothermic process. The overall process is analogous to ATP-dependent group transfer reactions, but here the energy is provided by the membrane phospholipids. This process illustrates a new role for membranes: they can serve as an energy source.     

Simultaneous analysis of retinol,all-trans- and 13-cis-retinoic acid and 13-cis-4-oxoretinoic acid in plasma by liquid chromatography using on-column concentration after single-phase fluid extraction

A reversed-phase high-performance liquid chromatographic method for the simultaneous analysis of retinol, all-trans-retinoic acid, 13-cis-retinoic acid and 13-cis-4-oxoretinoic acid in human plasma and cell culture medium is described. Sample preparation involves precipitation of proteins and extraction of retinoids with 60% acetonitrile. After centrifugation, the acetonitrile content of the supernatant is reduced to 45%, allowing on-column concentration of analytes. Injection volumes up to 2.0 ml (equivalent to 0.525 ml of sample) can be used without compromising chromatographic resolution of all-trans-retinoic acid and 13-cis-retinoic acid. Retinoids were stable in this extract and showed no isomerization when stored in the dark in a cooled autosampler, allowing automated analysis of large series of samples. Recoveries from spiked plasma samples were between 95 and 103%. Although no internal standard was used, the inter-assay precision for all retinoids was better than 6% and 4% at concentrations of 30 nM and 100 nM, respectively. The method is a valuable tool for the study of cellular metabolism of all-trans-retinoic acid, as polar metabolites of this compound can be detected with high sensitivity in cell culture media.     

Isocratic reversed-phase liquid chromatographic of all-trans-retinoic acid and its major metabolites in new potential supplementary test systems for developmental toxicology

An isocratic reversed-phase high-performance liquid chromatographic procedure for the determination of all-trans-retinoic acid (all-trans-RA) and its metabolites, all-trans-4-oxo-RA, 5,6-epoxy-RA, 9-cis-RA and13-cis-RA, in mouse plasma and embryo and in new in vitro potential test systems for development toxicology has been developed. These compounds, their biological precursor retinol (vitamin A) and the internal standard were resolved on a Spherisorb ODS-2 (5 μm) column (250×4.6 mm I.D.) with acetonitrile-water-methanol-n-butyl alcohol (56:37:4:3, v/v) containing 100 mM ammonium acetate and 70 mM acetic acid as the elution system with a total run time of 23 min. The assay was linear over a wide range, with a lower limit of quantitation of 50 ng/ml or 10 ng/ml of protein for all-trans-RA, 13-cis-RA and 9-cis-RA and of 25 ng/ml or 5 ng/ml protein for the 4-oxo- and 5,6-epoxy-metabolites. At these concentrations, intra-assay coefficients of variation (C.V.) of the retinoids were 3–9%. Mean intra-assay C.V. averaged 5–7% in the tissues studied. Its use is discussed for RA measurements in some of the new test systems — Drosophila melanogaster, sea urchin embryos and cultured human keratinocytes — that have to be evaluated in toxicological testing, supplementary to standard assays in mammals.     

Simultaneous determination of all-trans-, 13-cis-, 9-cis-retinoic acid and their 4-oxo-metabolites in plasma by high-performance liquid chromatography

A gradient reversed-phase high-performance liquid chromatographic technique is described for the easy separation and quantification of some retinoids; all-trans-retinoic acid, 13-cis-retinoic acid, 9-cis-retinoic acid and their corresponding 4-oxometabolites, in plasma. The method involved a diethyl ether-ethyl acetate (50:50, v/v) mixture extraction at pH 7 with acitretin and 13-cis-acitretin as internal standards. A Nova-Pak C₁₈ steel cartridge column was used. The mobile phase was methanol-acetonitrile (65:35, v/v) and 5% tetrahydrofuran (solvent A) and 2% aqueous acetic acid (solvent B) at 1 ml/min. The gradient composition was (only the percentages of solvent B are mentioned): I, 25% solvent B at the time of injection; II, 12% solvent B at 11 min until 30 min; III, 25% solvent B and maintenance of 25% solvent B for 10 min until a new injection. Total time between injections was 40 min. Detection was by absorbance at 350 nm. The precision calculated for plasma concentrations ranging from 2 to 250 ng/ml was better than 15% and the accuracy was less than 12%. The linearity of the method was in the range of 2 to 400 ng/ml of plasma. The limit of quantification was 2 ng/ml for each of the compounds. The HPLC method was applied to plasma specimens collected from animals receiving single dose administrations of all-trans-retinoic acid, 13-cis-retinoic acid and 9-cis-retinoic acid.