DHRS3

DHRS3 is a microsomal, integral-membrane short-chain dehydrogenase/reductase that uses NADPH to reduce all-trans-retinal to retinol, thereby constraining the synthesis of retinoic acid and serving as a central buffer of retinoid homeostasis during development (PMID:24005908, PMID:25451588). The purified enzyme prefers NADPH over NADH and, beyond retinaldehyde, also reduces androstenedione, estrone, and xenobiotic substrates (PMID:25451588). Functionally, DHRS3 acts in opposition to the RA-generating machinery (Aldh1a2/Rdh10): its knockdown phenocopies excess RA while its overexpression counteracts retinal-driven RA signaling (PMID:24045938), and in cell models it promotes accumulation of retinyl ester storage forms of retinol (PMID:11861404). This activity is embedded in negative-feedback control of RA: DHRS3 transcription is directly induced by retinoic acid through RAR/RXR cis-regulatory elements (PMID:39420244), and it is additionally a transcriptional target of p53 and TAp63γ acting through distinct promoter response elements following DNA damage (PMID:20543567, PMID:21659514). DHRS3 localizes to the ER and to lipid droplets, including ER–lipid-droplet interfaces juxtaposed to mitochondria where it supports NADPH/redox buffering and limits ROS (PMID:21659514, PMID:41579973). Loss-of-function in mice produces excess ATRA with cardiac outflow tract, septal, skeletal, and palate defects (PMID:24005908), and biallelic catalytically impairing variants in humans cause a congenital syndrome of coronal craniosynostosis, congenital heart disease, and scoliosis with reduced plasma retinol and elevated RA (PMID:40519748).

1. 2002 Medium

   Established that DHRS3/retSDR1 acts on retinoid metabolism by generating retinol storage forms and that its expression is responsive to retinoic acid, first linking the enzyme to vitamin A handling.

   Evidence Overexpression of retSDR1 in SK-N-AS neuroblastoma cells with retinyl ester measurement and RA-induction expression analysis

   PMID:11861404

   Open questions at the time

   • Overexpression-only, no direct enzymatic kinetics
   • Direction of reaction (retinal→retinol vs storage) inferred from metabolite pools
   • No in vivo validation
2. 2010 High

   Identified the transcriptional control of DHRS3 by tumor-suppressor transcription factors, placing the gene downstream of p53/TAp63γ stress and developmental signaling.

   Evidence Promoter reporter assays, in vitro binding, ChIP, response-element mutagenesis, and DNA damage induction

   PMID:20543567

   Open questions at the time

   • Does not establish the biochemical function of the DHRS3 protein itself
   • Link between p53/p63 induction and retinoid outcome not measured
3. 2011 Medium

   Defined DHRS3 subcellular localization to the ER and lipid droplets via an N-terminal targeting signal, connecting its activity to lipid-storage organelles and confirming it as a p53 target.

   Evidence Subcellular fractionation, confocal co-localization, p53 microarray target identification and activation experiments

   PMID:21659514

   Open questions at the time

   • Mechanistic role of LD localization in catalysis not resolved
   • Single lab
   • Topology not yet determined
4. 2012 Medium

   Showed DHRS3 induction is RAR-α-selective and that it is a microsomal protein, refining the receptor subtype driving its feedback regulation.

   Evidence Microarray/RT-PCR, in vitro transcription-translation, RAR-α-selective agonist (Am580), and rat tissue/LPS/RA treatments in vivo

   PMID:22790594

   Open questions at the time

   • Two isoforms detected but functional difference unknown
   • Cis-elements not yet mapped
5. 2013 High

   Demonstrated in vivo that DHRS3 is a retinaldehyde reductase essential for limiting embryonic ATRA, resolving the physiological consequence of its enzymatic activity.

   Evidence Dhrs3-knockout mouse with quantitative retinoid metabolite measurement and RA gene-expression analysis; reciprocal Xenopus gain/loss-of-function antagonizing Aldh1a2/Rdh10

   PMID:24005908 PMID:24045938

   Open questions at the time

   • Direct in vitro enzymatic reconstitution not yet performed
   • Cofactor preference inferred not measured
6. 2014 High

   Biochemically defined DHRS3 as an integral-membrane, NADPH-preferring reductase with a broad substrate range, providing the first reconstitution of purified enzyme.

   Evidence Recombinant expression, membrane topology determination, NADPH/NADH cofactor assays, multi-substrate enzymatic assays, purification and reconstitution

   PMID:25451588

   Open questions at the time

   • No structural model
   • Relative physiological importance of non-retinoid substrates unknown
7. 2018 Medium

   Extended DHRS3 regulation to the post-transcriptional level and to a new cellular outcome, showing miR-223 represses DHRS3 to control osteoblast differentiation.

   Evidence Dual-luciferase target validation, miR-223 mimic/inhibitor, DHRS3 overexpression and rescue, differentiation marker assays in human MSCs

   PMID:29794437

   Open questions at the time

   • Whether retinoid metabolism mediates the osteogenic effect not established
   • Single lab
8. 2024 Medium

   Mapped the direct RAR/RXR cis-regulatory feedback loop and added a tumor-context role for DHRS3 in driving melanoma cell-state plasticity.

   Evidence Reporter assays with cis-element mapping and vitamin-A manipulation in mice; lipid-droplet envelope proteomics and overexpression in melanoma cells

   PMID:39420244 PMID:39479752

   Open questions at the time

   • RAR/RXR element mapping in mouse; human conservation not addressed
   • Melanoma role from gain-of-function only
9. 2025 High

   Established DHRS3 as a Mendelian disease gene, showing biallelic catalytically impairing variants cause a congenital craniofacial/cardiac/skeletal syndrome through excess RA.

   Evidence Cell transfection of mutant constructs with retinoid metabolite quantification, in vitro activity assays for multiple alleles, and patient plasma retinoid measurements

   PMID:40519748

   Open questions at the time

   • Genotype–phenotype correlation across variants incomplete
   • No structural rationale for specific missense effects
10. 2026 Medium

    Connected DHRS3 to redox buffering and organelle contact sites, showing m6A/YTHDF2 stabilization and LRAT-dependent ER-LD-mitochondria positioning underlie its radioprotective NADPH/ROS function, and identified an Nrf2 physical and transcriptional axis.

    Evidence MeRIP-seq/reporter for m6A; YTHDF2 and LRAT perturbation with ROS, NADP+/NADPH and clonogenic assays; spatial imaging; Co-IP/GST pull-down/SPR/CETSA/ChIP for Nrf2 interaction

    PMID:41579973 PMID:41993611

    Open questions at the time

    • Functional consequence of the DHRS3–Nrf2 physical interaction unresolved
    • Single lab for each axis
    • Mechanism linking NADPH buffering to retinoid catalysis not fully integrated

• How DHRS3's multiple regulatory inputs (RA/RAR-RXR feedback, p53/p63, miR-223, m6A/YTHDF2, Nrf2) and its dual retinoid-reductase and NADPH/redox-buffering roles are integrated in specific tissues remains unresolved.
• No structural model of the enzyme
• Tissue-specific weighting of regulatory inputs unknown
• Direct mechanistic role of the Nrf2 interaction undefined