Could one genetic mutation be sufficient to bring on complex psychopathology? For decades, dominant models of mental illness-especially schizophrenia-have focused on polygenic origins, with thousands of common variants each contributing a minute increment of risk. New findings now undermine this view, showing that some loss‑of‑function mutations in the gene GRIN2A can cause a spectrum of early‑onset mental phenomena on their own and identify a targeted therapy using the amino acid L‑serine.
GRIN2A encodes the GluN2A subunit of the N‑methyl‑D‑aspartate receptor (NMDAR), a glutamate‑gated ion channel that is essential for synaptic transmission, neuroplasticity, and brain development. Null variants-those that truncate the protein or abolish its expression-reduce the density of NMDAR in neuronal membranes, impairing excitatory signaling in circuits involved in speech, memory, and emotional regulation. This mechanism aligns with the glutamatergic hypothesis of psychosis, which implicates NMDAR hypofunction in disorders such as schizophrenia.
In the largest such international cohort to date, investigators analyzed 196 individuals with pathogenic GRIN2A variants. Among 121 whose psychiatric histories were clear, 69.4% were null carriers and 30.6% were missense carriers. Among null carriers, 27.4% of null carriers mood disorders in 16%, anxiety disorders in 14%, and psychotic disorders in 10%-compared to 5.4% of missense carriers. Statistical modeling confirmed the six‑fold increased risk in null carriers, independent of age, sex, or family status.
Onset patterns were impressive. Compared with the Finnish FinRegistry database (>5 million individuals), null carriers demonstrated hazard ratios of 87 for psychotic disorders by age 12, 11.8 for mood disorders by age 11, and 5.84 for anxiety disorders by age 12. Given that childhood-onset psychosis is a very rare finding in the general population, these figures represent some of the highest risk elevations reported in psychiatric genetics. Importantly, six null carriers exhibited isolated psychiatric phenotypes in the absence of epilepsy or intellectual disability, challenging assumptions regarding the syndromic nature of neurodevelopmental disorders caused by a monogenic variant.
Insight into mechanism comes from the role of GluN2A in the assembly and signaling of NMDARs. Glycine and D‑serine serve as co‑agonists at the GluN1 subunit, but GluN2A‑containing receptors exhibit lower glycine site occupancy at rest. Null variants may therefore disproportionately diminish receptor activation. L‑serine, a non‑essential amino acid synthesized in astrocytes, is converted to D‑serine by serine racemase. D‑serine binds the glycine modulatory site with higher potency than glycine, enhancing the function of NMDARs. In addition to neurotransmission, L‑serine participates in the synthesis of sphingolipids, cell proliferation, and neuroprotection, and modulates microglial polarization to reduce neuroinflammation.
These properties form the basis for its therapeutic use. Open-label L-serine supplementation up to 500 mg/kg/day for more than a year in four null-version carriers with psychiatric symptoms resulted in significant improvements: cessation of hallucinations, remission of paranoia, and behavioral stabilization. Parallel observational series in GRIN2A/GRIN2B loss-of-function disorders report behavioral gains in 89% of treated individuals, developmental gains in 44%, and EEG or seizure improvements in 44%, without adverse effects. In contrast, in gain-of-function missense variants, L-serine exacerbated the symptoms, emphasizing the absolute requirement for genotype-guided treatment.
The trend of epilepsy offset and psychiatric onset in null carriers-mental disorders often arise after seizures remit-suggests distinct pathogenic pathways rather than secondary effects of epilepsy. This, along with the occurrence of non‑syndromic psychiatric cases, calls for reconsideration of genetic testing in early‑onset mental illness. Current guidelines rarely recommend sequencing for isolated psychiatric phenotypes, but monogenic, targetable entities like GRIN2A null disorders argue for its inclusion, particularly when onsets occur prior to adolescence.
At a molecular level, psychiatric risk differences between null versus missense variants may reflect altered electrophysiology but also C-terminal domain signaling loss that influences intracellular pathways and synaptic integration. This dimension of NMDAR biology remains poorly understood but could be critical for developing next-generation therapeutics. The combination of high‑penetrance genetic risk with well‑defined molecular pathology and a plausible, well‑tolerated intervention makes GRIN2A null variants one of the few existing examples of precision psychiatry. Although randomized, controlled trials will be required to establish efficacy, existing evidence suggests that L‑serine is a particularly promising candidate for the treatment of devastating, early‑onset mental illness arising from NMDAR hypofunction.
