RBM12

RBM12 is a multi-RRM-domain RNA-binding protein that functions as a post-transcriptional regulator of gene expression across diverse cellular contexts. It preferentially binds 5′ UTRs of target transcripts via its first RRM domain to suppress fetal hemoglobin expression independently of BCL11A and ZBTB7A, and represses GPCR/cAMP/PKA signaling by controlling expression of adenylyl cyclases, phosphodiesterases, and PKA subunits (PMID:35622975, PMID:37543364). In hepatocellular carcinoma, RBM12 binds JAK1 mRNA through its fourth RRM domain and recruits the translation factor EIF4A2 via its second RRM domain to enhance JAK1 translation, activating JAK1/STAT1-dependent PD-L1 expression and immune evasion (PMID:39187545). Loss-of-function truncating variants in RBM12 that remove an RNA-recognition motif co-segregate with familial psychosis (schizophrenia, schizoaffective disorder, psychotic bipolar disorder) in independent kindreds and are associated with atherosclerotic cardiovascular disease (PMID:28628109, PMID:41190437).

1. 2008 Medium

   Before RBM12's function was characterized, its genomic organization was clarified by showing that RBM12 and CPNE1 share a common promoter and 5′ exons, establishing that these two genes are co-regulated transcriptionally across vertebrates.

   Evidence RT-PCR and comparative genomics in human, mouse, and zebrafish

   PMID:18831769

   Open questions at the time

   • Functional consequences of RBM12–CPNE1 co-expression were not tested
   • No evidence that the shared promoter produces a fusion protein versus independent transcripts
2. 2017 High

   The first link between RBM12 and human disease was established when truncating mutations removing a predicted RRM domain were shown to co-segregate with familial psychosis in two independent kindreds, identifying RBM12 loss-of-function as a high-penetrance risk factor for schizophrenia-spectrum disorders.

   Evidence Whole-genome sequencing with long-range phasing and family-based segregation analysis in Icelandic and Finnish pedigrees

   PMID:28628109

   Open questions at the time

   • No biochemical or cellular mechanism linking RBM12 loss to psychosis pathophysiology was provided
   • The specific RNA targets disrupted by the truncating mutations were unknown
3. 2022 High

   A genome-wide CRISPR screen identified RBM12 as a suppressor of fetal hemoglobin expression, revealing for the first time that RBM12 acts as a post-transcriptional repressor by preferentially binding 5′ UTRs of target mRNAs, with its first RRM domain being essential for this function.

   Evidence CRISPR/Cas9 screen targeting RNA-binding proteins, shRNA depletion, eCLIP-seq, RRM domain deletion constructs in erythroid cells and SCD patient-derived cells

   PMID:35622975

   Open questions at the time

   • The specific 5′ UTR-bound transcripts responsible for HbF de-repression were not individually validated
   • Whether RBM12 acts by translational repression versus mRNA stability in erythroid cells was not resolved
   • Independence from known HbF regulators was shown but the epistatic relationship with other RNA-binding proteins was not mapped
4. 2023 High

   Connecting RBM12 to the psychosis-associated signaling pathway, loss of RBM12 was shown to cause hyperactive cAMP/PKA signaling through dysregulated expression of adenylyl cyclases, phosphodiesterases, and PKA subunits, and psychosis-linked truncating variants failed to rescue this phenotype, providing a mechanistic link between RBM12 loss-of-function and neuronal signaling dysfunction.

   Evidence siRNA/shRNA knockdown in HEK293 cells and iPSC-derived neurons with cAMP assays, PKA activity assays, RNA-seq, and rescue experiments with disease-associated truncating mutants

   PMID:37543364

   Open questions at the time

   • Whether RBM12 directly binds mRNAs encoding cAMP pathway components (as it binds 5′ UTRs in erythroid cells) was not demonstrated by CLIP
   • The specific RRM domains required for cAMP pathway regulation were not mapped
   • In vivo neuronal phenotypes of RBM12 loss-of-function were not examined
5. 2024 High

   A detailed translational mechanism was delineated in hepatocellular carcinoma, where RBM12 uses distinct RRM domains for separate functions: RRM4 binds JAK1 mRNA while RRM2 recruits EIF4A2 to enhance ribosome loading on JAK1, activating JAK1/STAT1-dependent PD-L1 transcription and immune evasion.

   Evidence RNA-IP, Co-IP, domain-specific RRM deletion mutants, and ribosome distribution assays in HCC cell lines

   PMID:39187545

   Open questions at the time

   • Whether the RBM12–EIF4A2 interaction occurs in non-cancer cell types was not tested
   • Structural basis for domain-specific RNA versus protein binding is unknown
   • In vivo tumor immune evasion phenotype upon RBM12 loss was not demonstrated
6. 2025 Medium

   Beyond neuropsychiatric disease, RBM12 loss-of-function was linked to atherosclerotic cardiovascular disease through a large exome-wide association study, broadening the clinical significance of RBM12 haploinsufficiency.

   Evidence Exome-wide association study in 434,438 UK Biobank participants

   PMID:41190437

   Open questions at the time

   • No mechanistic experiment connects RBM12 to vascular or inflammatory biology
   • Whether the cardiovascular association is mediated through the cAMP/PKA or JAK/STAT pathways is unknown
   • Replication in independent cohorts has not been reported

• Key unresolved questions include the structural basis of how individual RRM domains achieve target selectivity, whether RBM12 acts as a translational repressor versus activator depending on cellular context, and the in vivo consequences of RBM12 loss in animal models for neuronal and vascular phenotypes.
• No structural data exist for any RBM12 RRM domain or full-length protein
• No animal model phenotype for RBM12 loss-of-function has been reported
• Context-dependent switching between translational repression (erythroid, cAMP pathway) and activation (JAK1 in HCC) is mechanistically unexplained