HNRNPH2

HNRNPH2 is a nuclear RNA-binding protein that regulates alternative splicing of pre-mRNA by binding G-rich sequences and RNA G-quadruplexes, governing developmentally and oncogenically important splicing decisions (PMID:17567613, PMID:25122903, PMID:41867855). It recognizes G-rich enhancer elements—exemplified by the M2 enhancer of PLP/DM20 pre-mRNA in oligodendrocytes—and operates with its paralog hnRNPF to direct splice-site selection (PMID:17567613); mechanistically, it unfolds RNA G-quadruplexes to expose clustered G-rich sites, producing indirect cooperativity and switch-like control over hundreds of splicing events (PMID:41867855). Through such switches it drives oncogenic splice variants in glioblastoma, promoting the anti-apoptotic MADD isoform and a constitutively active RON kinase variant to support survival and invasion (PMID:21915099), and it shapes OPRM1 exon 2 inclusion at a G-containing intronic site (PMID:25122903). Nuclear import depends on a PY-NLS (residues 204–215, bearing an R-X2-4-P-Y motif and an adjacent DRP epitope) recognized by Karyopherin-β2/Transportin-1 (PMID:36711837). De novo missense variants clustering in this NLS reduce Karyopherin-β2 binding and cause cytoplasmic accumulation, defining a neurodevelopmental disorder; knock-in mice recapitulate motor, cognitive, and audiogenic-seizure phenotypes (PMID:27545675, PMID:37463454, PMID:36711837). HNRNPH2 expression is coupled to its paralog through splicing feedback: it modulates HNRNPH1 pre-mRNA splicing, and antisense-mediated Hnrnph2 knockdown drives compensatory Hnrnph1 upregulation that rescues disease phenotypes in mutant mice (PMID:42018666). Additional disease alleles act through distinct routes—the qRRM variant p.Arg114Trp impairs interaction with the LASR splicing complex (PMID:34907471), and the P213L variant disrupts binding to myelin-related transcripts, impairing oligodendrocyte differentiation and myelination (PMID:41443536).

1. 2007 High

   Established that hnRNPH2 is a sequence-specific splicing regulator that recognizes G-rich enhancers and selects splice sites, working synergistically with hnRNPF.

   Evidence siRNA knockdown, minigene splicing assays, and G-rich mutagenesis in differentiated oligodendrocytes (PLP/DM20)

   PMID:17567613

   Open questions at the time

   • Did not define structural basis of G-rich recognition
   • HNRNPH2-specific versus hnRNPH family contributions not separated
2. 2011 High

   Connected hnRNPH-controlled splicing switches to cancer cell phenotypes, showing causal roles in survival and invasion via specific oncogenic isoforms.

   Evidence siRNA knockdown plus isoform-specific ASO rescue/epistasis in glioblastoma cells with death and invasion readouts

   PMID:21915099

   Open questions at the time

   • Did not distinguish HNRNPH1 from HNRNPH2 contributions
   • No structural or binding-site map of the regulated exons
3. 2014 High

   Demonstrated direct G-containing sequence binding controls a clinically relevant splice variant and that a single SNP can tune hnRNPH-dependent splicing output.

   Evidence EMSA, minigene assays, siRNA/morpholino knockdown, and human postmortem brain analysis at OPRM1 rs9479757

   PMID:25122903

   Open questions at the time

   • In vivo physiological consequence of altered hMOR-1 levels not established
   • Did not resolve paralog-specific binding
4. 2016 Low

   Linked HNRNPH2 NLS variants to a human neurodevelopmental disorder, implicating disrupted nuclear localization.

   Evidence Whole-exome sequencing identifying NLS-clustered de novo variants with functional inference

   PMID:27545675

   Open questions at the time

   • No direct localization experiment reported in this study
   • Causality of mislocalization for phenotype not demonstrated
5. 2021 Medium

   Distinguished two mutational mechanisms—NLS variants impairing shuttling versus a qRRM variant impairing splicing-complex interaction.

   Evidence Nucleocytoplasmic shuttling assays, Co-IP for LASR interaction, and RNA-seq of patient fibroblasts

   PMID:34907471

   Open questions at the time

   • LASR interaction shown by Co-IP without structural detail
   • Causal chain from transcriptome changes to disease unresolved
6. 2023 High

   Identified Karyopherin-β2 as the import receptor and resolved the PY-NLS structural determinants quantitatively explaining variant mislocalization.

   Evidence Structure of Kapβ2•HNRNPH2 complex with site-directed mutagenesis and binding-affinity measurements (preprint)

   PMID:36711837

   Open questions at the time

   • Preprint, single lab
   • Did not test impact on downstream splicing in cells
7. 2023 High

   Showed in vivo that NLS mutations reduce Kapβ2 binding and recapitulate disease phenotypes, and revealed paralog compensation buffering loss of function.

   Evidence Two knock-in lines and KO mice, Co-IP for Kapβ2, behavioral/seizure phenotyping, and Hnrnph1 expression analysis

   PMID:37463454

   Open questions at the time

   • Did not resolve toxic gain-of-function versus loss-of-function definitively
   • Mechanism of failed Hnrnph1 upregulation in knock-ins unclear
8. 2025 High

   Defined the splicing-based feedback by which HNRNPH2 controls HNRNPH1 levels and exploited it therapeutically via ASO-driven paralog compensation.

   Evidence In vivo ASO knockdown (ICV), RNA-seq/RT-PCR splicing analysis, behavioral rescue, and iPSC-derived neuron validation

   PMID:42018666

   Open questions at the time

   • Long-term durability and safety of ASO rescue not established
   • Full set of co-regulated splicing targets not mapped
9. 2025 Medium

   Tied a specific patient mutation to a myelination program, linking impaired target-mRNA binding to oligodendrocyte differentiation defects.

   Evidence Hnrnph2P213L knock-in mice, RNA-IP/binding assays, oligodendrocyte differentiation assays, benztropine rescue, and behavioral testing

   PMID:41443536

   Open questions at the time

   • Single-lab study
   • Direct target transcripts incompletely catalogued
10. 2026 Medium

    Provided a biophysical mechanism—rG4 unfolding generating cooperative multi-site binding—for switch-like control of large splicing programs.

    Evidence In vivo/in vitro CLIP-based binding, rG4-unfolding assays, cooperativity modeling, and splicing analysis with patient data (preprint)

    PMID:41867855

    Open questions at the time

    • Preprint, single lab
    • Study covers HNRNPH broadly; HNRNPH2-specific cooperativity not isolated

• How HNRNPH2's splicing functions, paralog feedback, and post-translational regulation integrate to produce the disorder—and whether the mechanism is gain- or loss-of-function—remains unresolved.
• Gain- versus loss-of-function not definitively settled
• Cell-type-specific splicing target maps incomplete
• Post-translational regulation of HNRNPH2 in humans uncharacterized