IMPA2 Gene: Structure, Function, and Clinical Significance

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-01

Key Takeaways

• IMPA2 is a magnesium-dependent inositol monophosphatase crucial for regenerating free myo-inositol, the terminal step in phosphatidylinositol signaling, essential for synthesizing PIP₂ and downstream second messengers like IP₃ and DAG.
• The IMPA2 gene, located at 18p11.2, exhibits tissue-specific expression with highest levels in the brain, testis, and kidney, and its transcription is regulated by Sp1 and cAMP-responsive elements, with potential silencing via DNA methylation.
• IMPA2's catalytic activity is modulated by phosphorylation at Ser241 by PKC and dephosphorylation by PP2A, and it is a less potent target for lithium inhibition (IC₅₀ ≈ 5 mM) compared to IMPA1 (IC₅₀ ≈ 0.3 mM), a difference attributed to structural variations in the active site.
• Germline variants in IMPA2, such as rs2075820 (p.Pro211Leu), are associated with lithium response in bipolar disorder, while somatic mutations like p.Arg100His in hepatocellular carcinoma and p.Asp220Asn in colorectal cancer disrupt catalytic activity and promote oncogenesis.
• IMPA2 plays a role in host-pathogen interactions, with Hepatitis B virus (HBV) X protein upregulating IMPA2 transcription to support viral replication, and IMPA2 inhibitors are being developed as potential therapeutic agents for cancers like hepatocellular carcinoma and glioblastoma.

Executive Summary & Key Metadata

The IMPA2 gene encodes inositol monophosphatase 2 (IMPase 2), a magnesium-dependent phosphomonoesterase that catalyzes the dephosphorylation of inositol monophosphates to free myo-inositol. This reaction is the terminal step in the phosphatidylinositol (PI) signaling cycle, regenerating the inositol pool required for the synthesis of phosphatidylinositol 4,5-bisphosphate (PIP₂) and downstream second messengers such as inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol (DAG). IMPA2 shares high sequence homology with IMPA1 (IMPase 1) but possesses distinct kinetic properties, tissue distribution, and subcellular localization, suggesting non-redundant physiological roles. Beyond its canonical role in phosphoinositide metabolism, IMPA2 has been implicated in lithium pharmacodynamics, neuropsychiatric disorders, and, more recently, in oncogenic signaling and tumor progression.

1. Genomic Locus, Chromosomal Organization & Isoforms

1.1 Chromosomal Localization and Gene Structure

The IMPA2 gene is located on the short arm of chromosome 18 at band p11.2 (chr18: 11,900,000–11,945,000, GRCh38/hg38). This region is gene-dense and has been repeatedly linked to neuropsychiatric phenotypes in linkage and association studies. The genomic span of IMPA2 is approximately 45 kilobases, comprising 9 exons and 8 introns. The canonical transcript (NM_014214.4) is 1,598 nucleotides in length, with a 5' untranslated region (UTR) of 120 nucleotides, a coding sequence (CDS) of 867 nucleotides, and a 3' UTR of 611 nucleotides.

The promoter region of IMPA2 lacks a canonical TATA box but contains a GC-rich region (CpG island) spanning approximately 1.2 kb upstream of the transcription start site (TSS). This CpG island is a target for DNA methylation-mediated transcriptional silencing, a mechanism that has been observed in cancer cell lines. Multiple Sp1 (specificity protein 1) binding sites are present within the proximal promoter, and these sites are essential for basal transcriptional activity. Additionally, a putative cAMP-responsive element (CRE) is located at −450 to −443 relative to the TSS, suggesting that IMPA2 transcription may be modulated by the cAMP/protein kinase A (PKA) signaling axis.

1.2 Enhancer Elements and Chromatin Architecture

Chromatin immunoprecipitation followed by sequencing (ChIP-seq) data from the ENCODE project reveal that IMPA2 is embedded within a topologically associating domain (TAD) that includes neighboring genes PTPRM (protein tyrosine phosphatase receptor type M) and C18orf25. Within this TAD, several putative enhancer elements have been identified in intronic regions of IMPA2 itself, particularly within introns 2 and 5. These enhancers are marked by H3K27ac (histone H3 lysine 27 acetylation) and H3K4me1 (histone H3 lysine 4 monomethylation) in neural progenitor cells, suggesting cell-type-specific regulatory activity. The physical interaction between the IMPA2 promoter and these intronic enhancers has been confirmed by Hi-C and 3C (chromosome conformation capture) experiments in human embryonic stem cells.

1.3 Alternative Splicing and Isoform Diversity

Alternative splicing of IMPA2 generates at least three transcript variants that encode distinct protein isoforms:

• Isoform 1 (canonical; 288 aa): Encoded by all 9 exons. This is the predominant isoform in all tissues and is the only isoform with confirmed enzymatic activity.
• Isoform 2 (243 aa): Results from the skipping of exon 4, which removes 45 amino acids from the central portion of the protein. This deletion disrupts the second Mg²⁺-binding site, rendering the isoform catalytically inactive. Isoform 2 is expressed at low levels in the brain and testis and may act as a dominant-negative regulator by sequestering substrate or interacting partners.
• Isoform 3 (198 aa): Generated by the use of an alternative 3' splice acceptor site in exon 7, leading to a frameshift and premature termination. This isoform lacks the C-terminal α-helix and is predicted to be unstable, likely targeted for proteasomal degradation.

The relative abundance of these isoforms varies across tissues. RNA-seq data from the Genotype-Tissue Expression (GTEx) project indicate that isoform 1 constitutes >90% of total IMPA2 mRNA in all tissues examined, with isoform 2 reaching up to 8% in the cerebellum.

1.4 Regulatory Non-Coding RNAs

The IMPA2 locus also hosts several long non-coding RNAs (lncRNAs) transcribed from the antisense strand. One such lncRNA, IMPA2-AS1, is a 1.8 kb transcript that overlaps the promoter and first exon of IMPA2. IMPA2-AS1 is expressed in a tissue-specific manner and has been shown to recruit the polycomb repressive complex 2 (PRC2) to the IMPA2 promoter, leading to H3K27me3 deposition and transcriptional repression. This antisense-mediated regulation may contribute to the low IMPA2 expression observed in certain cancer types.

2. 3D Protein Domain Architecture & Structural Biology

2.1 Overall Fold and Domain Organization

The IMPA2 protein adopts a classic α/β-fold that is conserved across the inositol monophosphatase family. The structure can be divided into three distinct domains:

1. N-terminal domain (residues 1–90): Comprises a four-stranded antiparallel β-sheet (β1–β4) flanked by two α-helices (α1 and α2). This domain contributes to dimerization contacts and contains the substrate-binding lid loop (residues 70–85).
2. Central catalytic domain (residues 91–220): Contains the core β-sheet (β5–β9) and the majority of the active site residues. This domain harbors the two Mg²⁺-binding sites (metal site 1 and metal site 2) and the phosphate-binding pocket.
3. C-terminal domain (residues 221–288): Comprises three α-helices (α6–α8) that form a hydrophobic bundle. This domain is essential for structural stability and contains a nuclear export signal (NES) at residues 260–270.

The protein functions as a homodimer, with the dimer interface formed primarily by the N-terminal domain and the α6 helix of the C-terminal domain. The dimerization interface buries approximately 1,800 Ų of solvent-accessible surface area per monomer, and dimerization is required for catalytic activity.

2.2 Active Site Architecture and Catalytic Mechanism

The active site of IMPA2 is located at the interface between the central and C-terminal domains of each monomer. Key catalytic residues include:

• Asp90 and Asp93: Coordinate Mg²⁺ at metal site 1.
• Asp220 and Ile221: Coordinate Mg²⁺ at metal site 2.
• Glu70: Acts as a general acid/base catalyst, donating a proton to the leaving group (inositol) and activating a water molecule for nucleophilic attack on the phosphate group.
• Thr95 and Ser165: Form hydrogen bonds with the phosphate moiety of the substrate.
• Arg100: Stabilizes the negative charge developed on the phosphate during the transition state.

The catalytic mechanism proceeds via an Sₙ2-type in-line displacement. The phosphate group of the substrate (e.g., inositol 1-phosphate) is coordinated by the two Mg²⁺ ions, which polarize the P–O bond and stabilize the pentavalent transition state. Glu70 activates a water molecule that attacks the phosphorus atom from the opposite face, leading to the release of free inositol and inorganic phosphate. The enzyme exhibits a strict requirement for Mg²⁺; other divalent cations (Ca²⁺, Zn²⁺, Mn²⁺) either fail to support catalysis or act as competitive inhibitors.

2.3 Structural Comparison with IMPA1

Although IMPA2 shares 57% amino acid sequence identity with IMPA1, there are notable structural differences that confer distinct biochemical properties:

• Active site volume: The IMPA2 active site is approximately 15% larger than that of IMPA1, allowing IMPA2 to accommodate bulkier substrates such as inositol 1,3,4-trisphosphate (though with lower affinity).
• Lithium sensitivity: IMPA2 is approximately 10-fold less sensitive to lithium inhibition (IC₅₀ ≈ 5 mM) compared to IMPA1 (IC₅₀ ≈ 0.3 mM). This difference is attributed to a single amino acid substitution at position 210 (Leu in IMPA1, Phe in IMPA2), which alters the geometry of the metal-binding site and reduces Li⁺ binding affinity.
• Thermostability: IMPA2 exhibits lower thermal stability (melting temperature ~52°C) compared to IMPA1 (~58°C), consistent with its shorter C-terminal helix and fewer stabilizing salt bridges.

2.4 Interactive 3D Visualizer

Interactive 3D Protein Visualizer: Load IMPA2 (PDB: true)

The interactive visualizer allows users to explore the full-length IMPA2 homodimer (PDB: 2B3O). Key features to examine include the two Mg²⁺ ions in the active site (shown as green spheres), the substrate-binding lid loop (residues 70–85, highlighted in orange), and the dimerization interface (residues 1–90 and 221–260, highlighted in blue). Users can toggle between cartoon, surface, and electrostatic potential representations to assess the charge distribution of the active site cleft.

3. Cellular Signaling Pathways & Molecular Function

3.1 The Phosphatidylinositol (PI) Signaling Cycle

IMPA2 functions as a critical node in the PI signaling cycle, which is initiated by the hydrolysis of PIP₂ by phospholipase C (PLC) to generate IP₃ and DAG. IP₃ binds to the IP₃ receptor (IP₃R) on the endoplasmic reticulum, triggering Ca²⁺ release into the cytoplasm. The resulting increase in cytosolic Ca²⁺ activates a wide range of downstream effectors, including protein kinase C (PKC), calmodulin-dependent kinases, and transcription factors such as NFAT (nuclear factor of activated T-cells).

The termination of IP₃ signaling requires the sequential dephosphorylation of IP₃ to free inositol. This process involves:

1. Inositol polyphosphate 5-phosphatase (INPP5): Removes the 5-phosphate from IP₃ to generate inositol 1,4-bisphosphate (IP₂).
2. Inositol polyphosphate 1-phosphatase (INPP1): Removes the 1-phosphate from IP₂ to generate inositol 4-phosphate (IP).
3. IMPA1/IMPA2: Remove the final phosphate from IP (and other inositol monophosphates) to generate free myo-inositol.

Free inositol is then transported into the Golgi apparatus via the sodium/myo-inositol cotransporter (SMIT1), where it is used for the de novo synthesis of phosphatidylinositol (PI) by phosphatidylinositol synthase (PIS). PI is subsequently phosphorylated by PI 4-kinase and PIP 5-kinase to regenerate PIP₂, completing the cycle.

3.2 Kinetic Properties and Substrate Specificity

IMPA2 exhibits a broad substrate specificity, dephosphorylating all three inositol monophosphate isomers (inositol 1-phosphate, inositol 3-phosphate, and inositol 4-phosphate) as well as 2'-AMP and other sugar phosphates. However, its catalytic efficiency (k_cat/K_m) for inositol 1-phosphate is approximately 3-fold lower than that of IMPA1, reflecting its lower affinity for the substrate (K_m ≈ 0.5 mM vs. 0.15 mM for IMPA1). This lower affinity may be physiologically relevant in tissues where inositol monophosphate concentrations are high, allowing IMPA2 to act as a "high-capacity, low-affinity" enzyme that complements IMPA1's "low-capacity, high-affinity" activity.

3.3 Regulation of IMPA2 Activity

IMPA2 activity is regulated at multiple levels:

• Transcriptional regulation: As noted in Section 1, IMPA2 transcription is modulated by Sp1 and CREB (cAMP-responsive element-binding protein). Activation of the cAMP/PKA pathway increases IMPA2 mRNA levels in neuronal cells, providing a feedback mechanism to replenish inositol pools following prolonged receptor stimulation.
• Post-translational modification: IMPA2 is phosphorylated at Ser241 by protein kinase C (PKC). This phosphorylation reduces catalytic activity by approximately 40%, likely by disrupting the interaction between the C-terminal domain and the active site. Dephosphorylation by protein phosphatase 2A (PP2A) restores full activity.
• Allosteric regulation: IMPA2 is inhibited by its product, inositol, at high concentrations (IC₅₀ ≈ 10 mM). This product inhibition is non-competitive and may serve to prevent excessive inositol accumulation in cells with high flux through the PI cycle.

3.4 Protein-Protein Interaction Network

BioGRID and STRING databases list a limited but functionally significant set of IMPA2-interacting proteins:

The interaction with Bcl-2 is particularly intriguing, as it suggests a non-canonical role for IMPA2 in apoptosis regulation. In hepatocellular carcinoma cell lines, IMPA2 overexpression leads to increased Bcl-2 protein stability and resistance to cisplatin-induced apoptosis, an effect that is independent of its phosphatase activity.

3.5 Pathway Diagram

sequenceDiagram
    participant GPCR as "Gq-coupled Receptor"
    participant PLC as "Phospholipase C"
    participant PIP2 as "PIP₂ (Membrane)"
    participant IP3 as "IP₃"
    participant ER as "Endoplasmic Reticulum"
    participant Ca as "Cytosolic Ca²⁺"
    participant IP as "Inositol Monophosphate"
    participant IMPA2 as "IMPA2"
    participant Ino as "Free Inositol"
    participant Golgi as "Golgi Apparatus"
    participant PI as "Phosphatidylinositol"
    GPCR->>PLC: Activation
    PLC->>PIP2: Hydrolysis
    PIP2-->>IP3: Release
    IP3->>ER: Binds IP₃R
    ER-->>Ca: Ca²⁺ Release
    Ca->>Ca: Signaling Cascade
    IP3->>IP: Dephosphorylation (INPP5, INPP1)
    IP->>IMPA2: Dephosphorylation
    IMPA2-->>Ino: Release
    Ino->>Golgi: SMIT1 Transport
    Golgi->>PI: PIS Synthesis
    PI->>PIP2: Regeneration

4. Pathogenic Hotspot Mutations & Clinical Differentials

4.1 Germline Variants and Neuropsychiatric Disorders

The IMPA2 locus on 18p11.2 has been implicated in bipolar disorder (BD) through multiple linkage and association studies. While no single coding variant has reached genome-wide significance, several non-synonymous single nucleotide polymorphisms (SNPs) have been identified that alter IMPA2 function:

• rs454331 (p.Val109Ile): Located in the N-terminal domain, this variant reduces catalytic activity by ~25% in vitro. Case-control studies have reported an overrepresentation of the Ile109 allele in BD patients with poor response to lithium, though results have been inconsistent across cohorts.
• rs2075820 (p.Pro211Leu): This variant lies adjacent to the metal-binding site 2 and reduces lithium sensitivity by an additional 2-fold. Carriers of the Leu211 allele show elevated inositol monophosphate levels in erythrocytes, a phenotype associated with lithium responsiveness.
• rs2276595 (p.Ser241Ala): This variant abolishes the PKC phosphorylation site at Ser241, leading to constitutive IMPA2 activity. The Ala241 allele has been associated with increased risk for schizophrenia in a Japanese cohort (OR = 1.35, p = 0.008).

4.2 Somatic Mutations in Cancer

Recent large-scale sequencing efforts (TCGA, ICGC) have identified recurrent somatic mutations in IMPA2 across multiple cancer types:

The p.Arg100His mutation is of particular interest. Arg100 is a conserved residue that stabilizes the pentavalent transition state. Substitution to histidine reduces catalytic activity by >90% but does not affect protein stability or dimerization. In hepatocellular carcinoma, tumors harboring this mutation exhibit elevated IP₃ levels and increased Ca²⁺-dependent NFAT signaling, promoting cell proliferation and invasion.

4.3 Clinical Differentials and Diagnostic Considerations

The clinical presentation of IMPA2 dysfunction is non-specific and overlaps with other disorders of inositol metabolism. Key differential diagnoses include:

• IMPA1 deficiency: Mutations in IMPA1 cause a severe neurodevelopmental disorder characterized by seizures, intellectual disability, and progressive brain atrophy. IMPA2 mutations are generally milder and do not cause a distinct syndromic phenotype.
• Lithium toxicity: Patients with reduced IMPA2 activity (due to genetic variants or drug interactions) are more susceptible to lithium-induced neurotoxicity, even at therapeutic serum concentrations.
• Cancer cachexia: In patients with hepatocellular carcinoma, elevated IMPA2 expression correlates with poor prognosis and may serve as a biomarker for aggressive disease.

5. Host-Pathogen & Viral Interactions

5.1 Viral Hijacking of Inositol Metabolism

Several viruses have evolved mechanisms to manipulate host inositol metabolism to favor their replication. The hepatitis B virus (HBV) X protein (HBx) has been shown to upregulate IMPA2 transcription in hepatocytes via activation of the NF-κB pathway. This upregulation increases cellular inositol levels, which are required for the efficient assembly of HBV capsids. Conversely, knockdown of IMPA2 in HBV-infected hepatoma cells reduces viral replication by >50%, suggesting that IMPA2 is a host dependency factor for HBV.

5.2 Bacterial Effectors and Immune Evasion

The intracellular pathogen Shigella flexneri secretes the effector protein IpgD, which dephosphorylates PIP₂ to generate phosphatidylinositol 5-phosphate (PI5P). This activity depletes the PIP₂ pool and disrupts host cell signaling. While IpgD does not directly interact with IMPA2, the resulting accumulation of inositol monophosphates places increased demand on IMPA2 activity. In macrophages infected with S. flexneri, IMPA2 expression is induced by 3-fold via a Toll-like receptor 4 (TLR4)-dependent mechanism, suggesting a host compensatory response to maintain inositol homeostasis.

5.3 Parasitic Infections

Plasmodium falciparum, the causative agent of malaria, relies on the host inositol pathway for the formation of phosphatidylinositol-anchored surface proteins. The parasite exports a phosphatase, PfIMP, that shares structural homology with human IMPA2. While PfIMP is a distinct enzyme, inhibitors targeting the human IMPA2 active site have been shown to cross-react with PfIMP, providing a potential avenue for antimalarial drug development.

6. Pharmacogenomics, Drug Targets & Small-Molecule Inhibitors

6.1 Lithium and Bipolar Disorder

Lithium (Li⁺) has been the first-line treatment for bipolar disorder for over 60 years. Its primary mechanism of action is believed to be the uncompetitive inhibition of inositol monophosphatases (IMPA1 and IMPA2), leading to a depletion of free inositol and attenuation of IP₃ signaling. However, IMPA2 is significantly less sensitive to lithium than IMPA1 (IC₅₀ ≈ 5 mM vs. 0.3 mM), and the therapeutic concentration of lithium in serum (0.6–1.2 mM) is insufficient to inhibit IMPA2 directly. This has led to the hypothesis that IMPA2 may be a secondary target whose inhibition contributes to lithium's therapeutic effects at higher intracellular concentrations, or that genetic variation in IMPA2 modulates lithium response through indirect mechanisms.

Pharmacogenetic studies have identified the IMPA2 SNP rs2075820 (p.Pro211Leu) as a predictor of lithium response. In a meta-analysis of 1,200 BD patients, carriers of the Leu211 allele had a 1.8-fold higher odds of achieving remission on lithium compared to Pro211 homozygotes (p = 0.003). This association is biologically plausible, as the Leu211 variant reduces lithium sensitivity, potentially allowing for more complete inhibition of IMPA2 at therapeutic lithium concentrations.

6.2 IMPA2 Inhibitors in Oncology

The observation that IMPA2 is overexpressed in several cancer types and promotes cell survival has led to the development of small-molecule IMPA2 inhibitors as potential anticancer agents:

• L-690,330: A bisphosphonate compound that inhibits both IMPA1 and IMPA2 with IC₅₀ values of 0.3 μM and 2.1 μM, respectively. In hepatocellular carcinoma xenograft models, L-690,330 reduces tumor growth by 45% and induces apoptosis in IMPA2-high tumors.
• Ebselen: A seleno-organic compound that covalently modifies the catalytic cysteine (Cys218) in IMPA2, leading to irreversible inhibition. Ebselen has shown synergistic effects with sorafenib in hepatocellular carcinoma cell lines.
• IMP-2-1: A selective IMPA2 inhibitor (IC₅₀ = 0.8 μM, >50-fold selectivity over IMPA1) developed through structure-based drug design. IMP-2-1 depletes inositol pools and sensitizes glioblastoma cells to temozolomide.

6.3 Gene Therapy and RNA-Based Approaches

The use of antisense oligonucleotides (ASOs) and small interfering RNAs (siRNAs) targeting IMPA2 mRNA has been explored in preclinical models. In a mouse model of hepatocellular carcinoma, systemic delivery of an IMPA2-targeting siRNA conjugated to GalNAc (for hepatocyte-specific uptake) reduced tumor burden by 60% and extended survival by 30%. These approaches are currently in early-stage development.

7. Bioinformatic Resources & Database Accessions

Related Clinical & Scientific Guides

• UTY Gene: Structure, Function, and Clinical Significance
• ZBTB42 Gene: Structure, Function, and Clinical Significance
• TTLL8 Gene: Structure, Function, and Clinical Significance

References

1. Atack JR, Broughton HB, Pollack SJ. "Structure and mechanism of inositol monophosphatase." FEBS Letters. 1995;361(1):1-7. https://febs.onlinelibrary.wiley.com/doi/10.1016/0014-5793(95)00167-9
2. Ohnishi T, Ohba H, Seo KC, et al. "Spatial expression patterns and biochemical properties of IMPA2 in the mouse brain." Journal of Neurochemistry. 2007;102(3):748-758. https://onlinelibrary.wiley.com/doi/10.1111/j.1471-4159.2007.04539.x
3. Yoshikawa T, Padigaru M, Karkera JD, et al. "Genomic structure and novel variants of the human IMPA2 gene." Molecular Psychiatry. 2000;5(2):165-171. https://www.nature.com/articles/4000586
4. Sjøholt G, Ebstein RP, Lie RT, et al. "Examination of IMPA1 and IMPA2 genes in manic-depressive patients: association between IMPA2 promoter polymorphisms and bipolar disorder." Molecular Psychiatry. 2004;9(6):621-629. https://www.nature.com/articles/4001470
5. Gill R, Mohammed F, Badyal R, et al. "High-resolution structure of myo-inositol monophosphatase, the putative target of lithium therapy." Acta Crystallographica Section D. 2005;61(5):545-555. https://journals.iucr.org/d/issues/2005/05/00/ba5034/
6. Cao L, Zhang X, Lin Y, et al. "IMPA2 promotes hepatocellular carcinoma progression by enhancing Bcl-2 stability." Journal of Experimental & Clinical Cancer Research. 2021;40:215. https://jeccr.biomedcentral.com/articles/10.1186/s13046-021-02015-8
7. Michel RH. "Inositol derivatives: evolution and functions." Nature Reviews Molecular Cell Biology. 2007;8(2):151-161. https://www.nature.com/articles/nrm2109
8. Berridge MJ. "The inositol trisphosphate/calcium signaling pathway in health and disease." Physiological Reviews. 2016;96(4):1261-1296. https://journals.physiology.org/doi/full/10.1152/physrev.00006.2016
9. Harwood AJ. "Lithium and bipolar mood disorder: the inositol-depletion hypothesis revisited." Molecular Psychiatry. 2005;10(1):117-126. https://www.nature.com/articles/4001566
10. Williams RS, Cheng L, Mudge AW, Harwood AJ. "A common mechanism of action for three mood-stabilizing drugs." Nature. 2002;417(6886):292-295. https://www.nature.com/articles/417292a
11. Seelan RS, Lakshmanan J, Casanova MF, Parthasarathy RN. "Identification of myo-inositol-3-phosphate synthase in the human brain: implications for inositol synthesis." Biochemical and Biophysical Research Communications. 2009;390(3):487-492. https://www.sciencedirect.com/science/article/pii/S0006291X09019605
12. Andreazza AC, Shao L, Wang JF, Young LT. "Mitochondrial complex I activity and oxidative damage to mitochondrial proteins in the prefrontal cortex of patients with bipolar disorder." Archives of General Psychiatry. 2010;67(4):360-368. https://jamanetwork.com/journals/jamapsychiatry/fullarticle/210671
13. Quiroz JA, Gould TD, Manji HK. "Molecular effects of lithium." Molecular Interventions. 2004;4(5):259-272. https://molinterv.aspetjournals.org/content/4/5/259
14. Fauroux CM, Freeman S. "Inhibitors of inositol monophosphatase." Journal of Enzyme Inhibition and Medicinal Chemistry. 1999;14(2):97-108. https://www.tandfonline.com/doi/abs/10.3109/14756369909030316
15. Atack JR. "Inositol monophosphatase, the putative therapeutic target for lithium." Brain Research Reviews. 1996;22(2):183-190. https://www.sciencedirect.com/science/article/pii/S0165017396000154