Retinyl esters are elevated in progeny of Retinol Dehydrogenase 11 deficient dams

Abstract

Retinol dehydrogenase 11 (RDH11) is an NADPH-dependent retinaldehyde reductase that was previously reported to function in the visual cycle. Recently, we have shown that RDH11 contributes to the maintenance of retinol levels in extraocular tissues under conditions of vitamin A deficiency or reduced vitamin A availability. RDH11 is also expressed in the embryo. Rdh11 knockout animals do not display embryonic defects and appear to develop normally to the adult stage, but the exact function of RDH11 during development is not yet known. In contrast to RDH11-null mice, animals that lack dehydrogenase/reductase 3 (DHRS3), the enzyme that functions as a retinaldehyde reductase and is essential for the maintenance of retinoid homeostasis during embryogenesis, rarely survive until birth. Here, we investigated whether inactivation of RDH11 together with DHRS3 exacerbates the severity of retinoid homeostasis disruption in embryos that lack both enzymes compared to DHRS3-null mice. The results of this study indicate that in vitamin A sufficient animals, the loss of RDH11 in addition to DHRS3 does not appear to significantly impact the total levels of retinoic acid, free retinol, or retinyl esters in Rdh11^−/−/Dhrs3^−/− embryos in comparison to Dhrs3^−/− embryos. Surprisingly, Rdh11^−/− single gene knockout embryos obtained from breeding of Rdh11^−/− dams display elevated levels of embryonic retinyl esters compared to wild type embryos. The mechanism of the maternal effect of Rdh11 status on fetal retinoid stores remains to be elucidated.

1. Introduction.

All-trans-retinoic acid (RA), the main physiologically active metabolite of vitamin A, serves as a ligand for nuclear receptors and regulates the expression of many genes important for development, differentiation, proliferation, reproduction and immune response [reviewed in ¹]. RA is synthesized from all-trans-retinol via two enzymatically controlled steps: oxidation of retinol to retinaldehyde, and oxidation of retinaldehyde to RA [reviewed in ²]. Members of the short-chain dehydrogenases/reductases (SDR) enzyme superfamily [3] have been shown to play a prominent role in the control of the first reaction, which is reversible and rate-limiting.

The biosynthesis of RA is tightly controlled. Our recent studies revealed that two members of the SDR superfamily are critically important in the maintenance of RA homeostasis. The first enzyme, retinol dehydrogenase 10 (RDH10, SDR16C4 in humans [4]) catalyzes the oxidation of all-trans-retinol to all-trans-retinaldehyde, whereas the second enzyme, dehydrogenase/reductase 3 (DHRS3, SDR16C1 in humans), acts to convert the all-trans-retinaldehyde produced by RDH10 back to retinol, but only when DHRS3 is physically bound to RDH10 [5, 6]. Both RDH10 and DHRS3 are associated with the membrane fractions of the cells. Consistent with the function of DHRS3 as a retinaldehyde reductase, E13.5 and E14.5 Dhrs3^−/− embryos have reduced levels of free retinol and retinyl ester stores, and elevated levels of RA in comparison to wild type littermates [6, 7]. However, the absence of DHRS3 does not result in a complete loss of the membrane-bound retinaldehyde reductive activity in tissues of DHRS3-null embryos, suggesting that there might be additional retinaldehyde reductases contributing to the conversion of retinaldehyde to retinol during embryogenesis [6, 7]. Whether these enzymes play a role in establishing the retinoid homeostasis in the embryo is currently unknown.

Retinol dehydrogenase 11 (RDH11, SDR7C1 in humans) is an NADP(H)-dependent oxidoreductase, which belongs to a different SDR family (SDR7C) than RDH10 or DHRS3. In vitro, both human and mouse RDH11 enzymes interconvert retinoids, while mouse RDH11 also exhibits significant activity toward medium-chain lipid-derived aldehydes [8–10]. Expression of RDH11 in the retinal pigment epithelium and the fact that it is closely related to retinol dehydrogenase 12 (RDH12, SDR7C2 in humans), a photoreceptor-specific enzyme linked to congenital blindness, prompted multiple attempts to characterize RDH11 function in the eye. The results of these studies have been controversial, and the significance of RDH11 for the visual cycle or for the defense against reactive aldehydes generated under conditions of oxidative stress remains unclear [11–13]. Recently, we have conducted an extensive study to address the role of RDH11 enzyme in adult extraocular tissues. This study established that under conditions of restricted vitamin A supply the absence of RDH11 affects the in vivo levels of retinol in select adult mouse tissues, in particular in testis and liver, in a manner consistent with its function as a retinaldehyde reductase [14]. Since Rdh11 gene knockout by itself does not lead to any obvious embryonic abnormalities, but Rdh11 is known to be expressed as early as embryonic day 12 [12], we hypothesized that inactivation of Rdh11 gene together with Dhrs3 gene may exacerbate the DHRS3-null phenotype. Here, to elucidate the role of RDH11 in the embryo, we examined the effect of Rdh11 gene knockout on embryonic retinoid levels alone or in combination with Dhrs3 knockout.

2. Materials and Methods.

2.1. Animals.

Generation and genotyping of Rdh11 knockout and Dhrs3+/− mice was described previously [6, 11]. Animals were maintained on vitamin A sufficient diet (4 IU/g of vitamin A, catalog # D10012M from Research Diets, New Brunswick, NJ) in a facility approved by the Association for Assessment and Accreditation of Laboratory Animal Care. Noon of the day of a vaginal plug was taken as embryonic day 0.5 (E0.5). Embryos were collected by caesarean section at embryonic day 14.5, and the yolk sacs were taken for DNA extraction and genotyping. All animal experiments employed procedures approved by the University of Alabama Animal Care Committee, and conformed to recommendations of the American Veterinary Medical Association Panel on Euthanasia.

2.3. RNA was isolated from embryos using TRIzol reagent (Ambion, Life Technologies, Carlsbad, CA). The synthesis of cDNA and quantitative RT-PCR was performed as described in [14]. Sequences of primers are listed in Table 1.

Table 1. Primers for qRT-PCR analysis of the embryonic gene expression.

Transcript	Primer pair
Rdh13	5’ TTCGAGGACTTGAACTGGCA 3’ 5’ GTCACACCAGAGCCTTGCAG 3’
Rdh14	5’ GCATCCTGGTATTGTGCGAA 3’ 5’ TGGAAGTCTGGGCACCTTCT 3’
Dhrs3	5’ GGTCCATGGAAAAAGCTTGA 3’ 5’ CAATATGGCCGTTCTGGAGT 3’
Lrat	5’ CAAGGAACGCACTCAGAAG 3’ 5’ CCCGTCTAGGTGATTGACTA 3’
Stra6	5’ GTATTCATCCCTCTCGCCAA 3’ 5’ TGGTCCCCAAGAAGAAGATG 3’
Crpb1	5’ GACTTCAACGGGTACTGGA 3’ 5’ CTGCACGATCTCTTTGTCTG 3’
Rdh10	5’ CCTGGGCCGCCTCTTTGCTC 3’ 5’ GCGAACCATGCCTGCGGTCT 3’
Bco1	5’ TGCGAAGGATGAAGATGATG3’ 5’ AAACCATGAAGGTCCAGGTG 3’
β-Actin	5’ CCTAAGGCCAACCGTGAAAAG 3’ 5’ AGGCATACAGGGACAGCACAG 3’
Gapdh	5’ ATGTGTCCGTCGTGGATCTGA 3’ 5’ TTGAAGTCGCAGGAGACAACCT 3’
Hprt1	5’ CTTGCTCGAGATGTCATGAAGGAG 3’ 5’ TAATCCAGCAGGTCAGCAAAGAAC 3’

2.4. Statistics.

Pairwise differences in retinoid content and gene expression between groups were tested using unpaired t-test. Data were expressed as mean ± standard deviation, and visualized using GraphPad Prism version 7.04 for Windows (GraphPad Software, La Jolla, CA).

3. Results and Discussion.

3.1. Generation of Rdh11;Dhrs3 double knockout embryos.

According to our earlier study and the report from another group [6, 7], the levels of free and esterified retinol in Dhrs3-null embryos are reduced 3- to 4-fold, which is in agreement with the loss of retinaldehyde reductase function. However, while Rdh10 knockout embryos resorb mid-gestation due to almost complete disruption of RA biosynthesis [15], Dhrs3 knockout embryos display a less severe phenotype, with few embryos surviving until birth. The levels of RA in Dhrs3^−/− embryos are only moderately elevated (by approximately 30 %, ref. 6). The modest increase in RA levels suggests that additional retinaldehyde reductases may partially complement DHRS3 function during development, allowing embryos to progress to later stages. Cultured mouse embryonic fibroblasts (MEFs) derived from Dhrs3^−/− embryos and isolated membrane fractions retained ~50 and ~20%, respectively, of the retinaldehyde reductive activity detected in control MEFs [6], indicating the presence of additional enzyme(s) capable of converting retinaldehyde to retinol. While mouse cytosolic aldo-keto reductases can catalyze retinaldehyde reduction in vitro [16], only membrane-bound enzymes of SDR superfamily have been shown to control retinol and retinaldehyde interconversion in vivo under physiological conditions. Therefore, we focused further analysis on membrane-bound RDH11, which has been recently shown to contribute to the regulation of retinoid homeostasis in adult tissues [14].

Analysis conducted by Kanan et al. [12] detected the expression of both Rdh11 transcript and protein in developing mouse eye beginning at embryonic day 12.5. This study found that the expression pattern of RDH11 was very different from the one of photoreceptor-specific RDH12, but it did not address the expression or function of RDH11 in extraocular embryonic tissues. Analysis of Rdh11 expression data available in the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov/, gene ID 17252, expression profile based on the mouse ENCODE transcriptome data [17]) suggests that Rdh11 transcript is present in the embryo as early as embryonic day 11.5. To determine whether RDH11 is responsible for the retinaldehyde reductase activity remaining in tissues of Dhrs3^−/− embryos, we have crossed Rdh11^−/− and Dhrs3+/− strains to obtain double knockout embryos.

As Rdh11^−/− and Dhrs3+/− strains have been originally established on different backgrounds (129/SvJ and C57BL/6J, respectively), we have used double heterozygotes obtained from the original crossing of the two strains to repeat heterozygote breeding and obtain animals with a homogenous mixed genetic background. These animals were used as breeders to produce four cohorts of embryos for the analysis of in vivo retinoid content: wild type (Rdh11+/+;Dhrs3+/+), Rdh11 single knockout (Rdh11^−/−;Dhrs3+/+), Dhrs3 single knockout (Rdh11+/+;Dhrs3^−/−), and double knockout (Rdh11^−/−; Dhrs3^−/−). Double knockout E14.5 embryos were obtained at expected frequencies, and did not show earlier resorption or more severe phenotype in comparison to single Dhrs3 knockout.

3.2. RA levels.

HPLC analysis of retinoid content has confirmed the earlier observation that RA levels are elevated in DHRS3-null embryos. While in our previous study RA levels in DHRS3-null embryos were trending higher than in controls, but did not reach statistical significance, the present work showed statistically significant increase in RA levels in knockouts (p value < 0.001, Table 2, Figure 1), likely due to the larger sample sizes.

Table 2. Endogenous retinoid content in embryos.

Rdh11+/−;Dhrs3+/−, Rdh11^−/−;Dhrs3+/− and Rdh11+/+;Dhrs3+/− animals were used as breeders to generate four cohorts of embryos: WT – wild type, Rdh11 KO – Rdh11 knockout, Dhrs3 KO – Dhrs3 knockout, DKO – double knockout. Levels of retinoic acid (RA), free retinol and retinyl esters determined using HPLC and expressed as pmol/g of wet weight. SD - standard deviation, n - number of samples in the group. p- values are provided for unpaired 2-tailed t-test for pairwise comparison between indicated groups.

Figure 1. Differences in endogenous retinoid content in wild type (WT), single and double knockout (KO and DKO) embryos.

In vivo levels of RA, free and esterified retinol were determined using HPLC. Horizontal bars represent means with error bars showing standard deviation. For exact values refer to Table 2.

The RA levels in double knockout embryos were also higher in comparison to wild type embryos (p < 0.001). Importantly, there was no further increase in RA levels in Rdh11^−/−;Dhrs3^−/− embryos in comparison to Dhrs3^−/− set (Table 2, Figure 1). Thus, in animals on vitamin A-sufficient diet the absence of RDH11 did not appear to cause further increase in RA in embryos lacking DHRS3. There was also no difference in RA content between single Rdh11 knockout embryos and wild type control group, suggesting that unlike DHRS3, RDH11 does not significantly contribute to the control of RA levels during embryonic development, at least under conditions of vitamin A sufficiency.

3.3. Levels or retinol and retinyl esters.

To determine how the Rdh11 status affects the total embryonic vitamin A content in vivo, we have analyzed the levels of both free retinol and its esterified storage form, retinyl esters (Figure 1, Table 2). Similarly to the previous report [6], the levels of free retinol in Dhrs3 single KO embryos were approximately 4-fold lower than in wild type. The levels of retinyl esters stores were also decreased. Although retinol and retinyl esters in double knockout embryos were lower than in control group, the absence of RDH11 did not lead to further decrease in retinol or retinyl esters in double knockout embryos in comparison to DHRS3-null embryos. The levels of free retinol in Rdh11 single knockout embryos in comparison to wild type controls also remained unchanged.

Intriguingly, we have observed a significant increase in retinyl ester stores in Rdh11 single knockout embryos in comparison to the control group (Figure 1, Table 2). This observation was opposite to what would be expected when a retinaldehyde reductase is inactivated, i.e., a decrease in the content of retinol and retinyl esters. We hypothesized that the observed increase in retinyl esters could be due to upregulation of other membrane-bound SDR retinaldehyde reductases, or of the retinyl ester synthesizing enzyme, lecithin-retinol acyltransferase (LRAT), in Rdh11^−/− embryos. However, quantitative RT-PCR (Figure 2) did not detect statistically significant changes in the expression of Lrat, Dhrs3, or Rdh13 and Rdh14 transcripts, which encode RDH11-related SDR enzymes known to reduce retinaldehyde in vitro [18–20]. There were also no changes in the expression of Stra6 and Crbp1, which may facilitate the uptake and esterification of retinol (Figure 2).

Figure 2. Expression of genes of retinoid homeostasis in wild type (green data points) and Rdh11 knockout (red) embryos.

Transcript levels were analyzed by qRT-PCR; relative differences were expressed as a fold of change from the average of the wild type group. No statistically significant differences between control and Rdh11 knockout groups were detected for any of the transcripts.

An additional factor that might have affected the retinyl esters content in Rdh11 single knockout versus wild type embryos is the genotype of the female breeder. As it was hardly feasible to obtain a sufficient number of embryos exclusively from the breeding of double heterozygotes due to the low expected frequency of desired phenotypes, we also used Rdh11^−/−;Dhrs3+/− and Rdh11+/+;Dhrs3+/− animals as breeders. As a result, Rdh11 single knockout embryos were predominantly collected from Rdh11^−/− ;Dhrs3+/− females, and wild type – from Rdh11+/+;Dhrs3+/− females. To determine whether the Rdh11^−/− status of female breeders was responsible for the elevated retinyl ester stores in their progeny, we conducted a follow-up experiment and crossed Rdh11+/− single heterozygote animals of mixed 129/SvJ-C57BL/6J background. HPLC analysis of retinoid content showed that wild type and Rdh11 single knockout embryos produced by Rdh11 heterozygote dams did not differ in the levels of retinyl esters, as well as free retinol and RA (Figure 3, Table 3). This analysis led us to conclude that Rdh11 status of the embryo does not affect fetal levels of retinoids, but the lack of RDH11 in the female breeder leads to the elevated embryonic retinoid stores.

Figure 3. Endogenous retinoid content in wild type (WT) and Rdh11 knockout (KO) embryos.

In vivo levels of RA, free and esterified retinol were determined using HPLC. Horizontal bars represent means with error bars showing standard deviation. For exact values refer to Table 3.

Table 3. Endogenous retinoid content in embryos obtained from Rdh11 single heterozygote females.

Rdh11+/−;Dhrs3+/+ animals were used as breeders. Levels of retinoic acid (RA), free retinol and retinyl esters determined using HPLC and expressed as pmol/g of wet weight. WT – wild type, KO – knockout, SD - standard deviation, n - number of samples in the group. p-values are provided for unpaired t-test between WT and KO groups.

		WT	Rdh11 KO	p-value
RA	n	9	6
	Mean ± SD	13.1 ± 3.3	12.7 ± 3.6	0.828
free retinol	n	9	7
	Mean ± SD	174 ± 30	168 ± 29	0.73
retinyl esters	n	9	7
	Mean ± SD	356 ± 204	401 ± 174	0.978

The mechanism that leads to the increase in fetal retinyl esters is currently unclear. Retinol bound to the serum retinol binding protein 4 is the major source of vitamin A for fetal development, and it is delivered through maternal circulation [21, 22]. Retinol crosses maternal-fetal barrier and is used by the developing tissues or is esterified to form embryonic retinyl ester stores. Under conditions of high dietary vitamin A intake, retinyl esters in circulating chylomicrons represent another major form of vitamin A supply for the embryo. It has been shown that under conditions of vitamin A sufficiency retinyl esters delivered in chylomicrons are largely responsible for the accumulation of fetal retinyl ester stores [21, 22]. While the embryonic supply ultimately depends on maternal vitamin A in circulation, our previous study has shown that systemic levels of retinoids in vitamin A-sufficient Rdh11 knockout animals are not different from wild type mice [14], and thus cannot explain the observed differences in fetal retinyl esters. At this point, it is not clear whether the differences in maternal Rdh11 status result in altered transfer of retinyl esters across the placenta, or affect the availability of transferred free retinol for in situ esterification in the embryo. While further investigation of this effect was beyond the scope of this study, it is interesting to note that according to the NCBI database of mouse expressed sequence tags (ESTs), Rdh11 is a maternal transcript present in the oocyte and in the zygote, while according to RNA sequencing-based mouse ENCODE transcriptome data it is also expressed in the ovary and placenta. It has been proposed that placenta may serve as temporary storage site of retinoid stores for the embryo, before the embryonic liver is formed [23]; therefore the role of RDH11 in female reproductive system merits further investigation.

The present findings are consistent with our previous conclusions that while DHRS3 directly regulates the output of RA via control of the production of its immediate precursor retinaldehyde, RDH11 appears to be important for the maintenance of tissue levels of retinol available for RA biosynthesis. The lower levels of retinol in testis of RDH11-deficient animals resulted in a mildly downregulated RA signaling [14], not in the increase of RA levels, as is the case with Dhrs3 knockout [6, 7 and present study]. It is important to note that the in vivo effects of RDH11 were detectable only under conditions of vitamin A deficiency or restricted supply [14]. There is a possibility that, if placed on vitamin A deficient diet during gestation, double knockout RDH11/DHRS3 embryos would have displayed a smaller than 30% increase in RA due to reduced retinol availability in the absence of RDH11, and survived longer. The results of the current study confirm that although DHRS3 and RDH11 enzymes catalyze the same chemical reaction in vivo, the reduction of retinaldehyde to retinol, their physiological functions are remarkably different and not overlapping.

Conclusions.

Our recent report on mouse RDH11 [14] has proven that in addition to RDH10 and DHRS3, other SDRs, which are expressed in a tissue-specific manner, can contribute to the regulation of vitamin A status in vivo in a specific set of tissues. The results of this study show that unlike DHRS3, RDH11 does not play a significant role in the maintenance of RA homeostasis in fetal tissues under conditions of sufficient vitamin A supply. However, an unexpected observation that maternal Rdh11 status has an impact on fetal retinoid stores suggests a potential new aspect in RDH11 contribution to retinoid metabolism. Future studies should uncover the mechanism underlying this effect.

The loss of RDH11 does not exacerbate the phenotype of Dhrs3^−/− embryos.

RDH11 does not affect retinoic acid homeostasis in vitamin A-sufficient embryos.

Retinyl esters are elevated in the progeny of Rdh11 knockout dams on chow diet.