HNRNPH2

HNRNPH2 is an X-linked nuclear RNA-binding protein that functions as a key regulator of alternative splicing in the nervous system and oligodendrocyte lineage. It is imported into the nucleus via a PY-NLS (residues 204–215) recognized by Karyopherin-β2/Transportin-1, and disease-causing missense mutations within this NLS reduce Kap-β2 binding 70–100-fold, causing cytoplasmic mislocalization, while a qRRM2 variant (R114W) impairs interaction with the LASR splicing complex and broadly disrupts alternative splicing (PMID:34907471, PMID:37463454, PMID:36711837). HNRNPH2 autoregulates its paralog HNRNPH1 through alternative splicing of HNRNPH1 pre-mRNA, and loss of HNRNPH2 is compensated by HNRNPH1 upregulation—a compensation that fails when mutant HNRNPH2 protein is present, indicating a gain-of-function or dominant-negative disease mechanism (PMID:37463454, PMID:42018666). De novo missense variants in HNRNPH2 cause an X-linked neurodevelopmental syndrome characterized by motor and cognitive impairment and seizures, and in oligodendrocytes, disrupted HNRNPH2–target transcript binding leads to myelination defects that are pharmacologically rescuable (PMID:27545675, PMID:41443536).

1. 2016 Medium

   Identification of de novo HNRNPH2 NLS variants in unrelated females established that this X-linked RNA-binding protein is essential for neurodevelopment, raising the mechanistic question of how NLS disruption causes disease.

   Evidence Whole-exome sequencing of six independent families with neurodevelopmental phenotypes; bioinformatic prediction of variant effects

   PMID:27545675

   Open questions at the time

   • No functional validation of NLS variants at the protein level
   • Disease mechanism (loss-of-function vs. gain-of-function) was unknown
   • Specific splicing targets in the nervous system were uncharacterized
2. 2021 High

   Functional characterization of patient-derived variants revealed that NLS mutations cause cytoplasmic mislocalization whereas a qRRM2 mutation disrupts interaction with the LASR splicing complex and broadly alters alternative splicing, establishing two distinct molecular routes to disease pathology.

   Evidence Subcellular localization assays for NLS variants; co-immunoprecipitation for LASR interaction; RNA-seq of patient fibroblasts

   PMID:34907471

   Open questions at the time

   • Structural basis of NLS recognition by Karyopherin-β2 was unknown
   • Whether mislocalized protein exerts a toxic cytoplasmic function was untested
   • The relationship between HNRNPH2 and its paralog HNRNPH1 in disease compensation was unclear
3. 2023 High

   In vivo knockin mouse models demonstrated that NLS mutations reduce Kap-β2 binding and cause neurodevelopmental phenotypes, while knockout mice are phenotypically normal due to compensatory upregulation of Hnrnph1—compensation that fails in the presence of mutant protein—resolving the gain-of-function versus loss-of-function debate.

   Evidence Knockin and knockout mouse models; behavioral phenotyping; co-immunoprecipitation for Kap-β2 binding; subcellular fractionation; Western blotting for Hnrnph1 levels

   PMID:37463454

   Open questions at the time

   • Whether the toxic mechanism involves cytoplasmic aggregation or aberrant cytoplasmic RNA binding was unresolved
   • Specific neural cell types most affected were not delineated
   • Therapeutic strategy to exploit paralog compensation had not been tested
4. 2023 High

   Structural determination of the Kap-β2–HNRNPH2 NLS complex revealed a PY-NLS motif (206RPGPY210) plus a novel 211DRP213 epitope contacting Kap-β2 W373, and mutagenesis showed 70–100-fold affinity reduction for disease variants, providing an atomic-level explanation for impaired nuclear import.

   Evidence Crystal structure of Kap-β2–HNRNPH2 complex; site-directed mutagenesis; quantitative binding affinity measurements (preprint)

   PMID:36711837

   Open questions at the time

   • Preprint; awaits peer review
   • Whether the DRP epitope is functionally required in vivo has not been tested
   • Post-translational regulation of NLS accessibility is unexplored
5. 2025 High

   ASO-mediated knockdown of mutant Hnrnph2 rescued seizures, motor, and cognitive deficits in knockin mice by inducing compensatory Hnrnph1 upregulation, and mechanistic analysis showed HNRNPH2 autoregulates HNRNPH1 via alternative splicing, establishing a direct paralog cross-regulation circuit with therapeutic relevance.

   Evidence Intracerebroventricular ASO injection in neonatal/juvenile mice; behavioral and electrophysiological rescue; RT-PCR and RNA-seq of HNRNPH1 splicing in human iPSC-derived neurons

   PMID:42018666

   Open questions at the time

   • Long-term safety and durability of ASO-based strategy in humans is unknown
   • Whether additional splicing targets beyond HNRNPH1 contribute to rescue is unresolved
   • The cis-elements on HNRNPH1 pre-mRNA bound by HNRNPH2 are not mapped
6. 2025 High

   A P213L variant in the NLS/DRP epitope disrupts HNRNPH2 binding to target transcripts in oligodendrocytes, causing downregulation of myelination genes, impaired OPC differentiation, and hypomyelination—a phenotype rescued by benztropine—identifying a cell-type-specific role for HNRNPH2 in myelination.

   Evidence Hnrnph2-P213L knockin mice; RNA-binding assays; myelin histochemistry; gene expression analysis; pharmacological rescue with benztropine and behavioral testing

   PMID:41443536

   Open questions at the time

   • The specific myelination-related mRNA targets directly bound by HNRNPH2 are not comprehensively identified
   • Whether the myelination defect is independent of the splicing defect or downstream of it is unclear
   • Applicability of benztropine rescue to other HNRNPH2 variants is untested

• Key unresolved questions include the full repertoire of neural and glial HNRNPH2 splicing targets, whether cytoplasmic mutant protein forms toxic assemblies, and whether the cooperative rG4-mediated splicing regulation applies specifically to HNRNPH2 versus the broader HNRNPH family in disease-relevant tissues.
• Comprehensive CLIP-seq mapping of HNRNPH2-specific versus HNRNPH1-specific targets in neural tissue is lacking
• No structural model of HNRNPH2 qRRM–RNA or qRRM–LASR interface exists
• The contribution of cytoplasmic HNRNPH2 gain-of-function versus nuclear loss-of-function to disease is not resolved