MTAP and PRMT5 Co-Deletion Synthetic Lethality (Chr9p21.3)

MTAP (5′-methylthioadenosine phosphorylase, EC 2.4.2.28, encoded on human chromosome 9p21.3 across 8 exons, ~31 kDa monomer assembling into a catalytically active homotrimer) is the sole mammalian salvage enzyme responsible for phosphorolytically cleaving the polyamine biosynthetic byproduct 5′-deoxy-5′-methylthioadenosine (MTA) into adenine and 5-methylthio-α-D-ribose-1-phosphate (MTR-1-P). Because the MTAP gene resides only ~100 kb telomeric to the canonical tumor suppressor locus CDKN2A/CDKN2B (encoding the cell-cycle inhibitors p16INK4a, p14ARF, and p15INK4b), homozygous focal deletions of Chr9p21.3—the single most frequent somatic copy-number deletion across all human malignancies—co-delete MTAP as a collateral passenger event in 80% to 90% of CDKN2A-null tumors (accounting for ~15% of all human cancers, including ~40–50% of glioblastomas and mesotheliomas, ~25–35% of pancreatic ductal adenocarcinomas and urothelial bladder carcinomas, and ~15–20% of non-small-cell lung cancers and melanomas).

In MTAP-deficient cancer cells, loss of MTAP enzymatic activity causes intracellular MTA concentrations to rise 10-fold to 40-fold (from ~0.1–0.3 μM in normal diploid tissues to ~2–15 μM in MTAP-null tumors). Because MTA is a structural analog of the universal methyl-donor cofactor S-adenosyl-L-methionine (SAM), elevated intracellular MTA acts as a selective, endogenous competitive inhibitor (Ki ≈ 0.26 μM vs. SAM Km ≈ 1.5–2.0 μM) of protein arginine methyltransferase 5 (PRMT5). This partial baseline inhibition (~50–75% reduction in active enzyme) renders MTAP-null tumor cells vulnerable to further depletion of SAM (via MAT2A inhibition) or to second-generation MTA-cooperative PRMT5 inhibitors (BMS-986504/MRTX1719, AMG 193, TNG908, TNG462), establishing the MTAP–PRMT5 axis as a premier model of collateral synthetic lethality in clinical drug discovery.

Polyamine Biosynthesis and Stoichiometric MTA Generation

In proliferating mammalian cells, synthesis of the polyamines spermidine and spermine (which stabilize ribosomal RNA, tRNA, and chromatin structure during translation and DNA replication) requires the aminopropyl donor decarboxylated S-adenosylmethionine (dcSAM), generated from SAM by S-adenosylmethionine decarboxylase (AMD1). Spermidine synthase (SRM) and spermine synthase (SMS) transfer the aminopropyl group from dcSAM to putrescine and spermidine, respectively, releasing one molecule of MTA for every aminopropyl group transferred.

In normal MTAP-proficient cells, homotrimeric MTAP uses inorganic orthophosphate (HPO₄²⁻) to catalyze the nucleophilic attack on the C1′ glycosidic bond of MTA, yielding free adenine (which is recycled into adenosine monophosphate [AMP] via adenine phosphoribosyltransferase [APRT], supporting cellular adenosine triphosphate pools) and MTR-1-P (which is converted in five enzymatic steps via the methionine salvage pathway—including MRI1, APIP, ENOPH1, and ADI1—into 2-keto-4-methylthiobutyrate [KMTB] and transaminated back into L-methionine). Methionine adenosyltransferase 2A (MAT2A, complexed with its regulatory β-subunit MAT2B) then condenses L-methionine and ATP to regenerate SAM.

Metabolic Consequences of Homozygous Chr9p21.3 Co-Deletion

When Chr9p21.3 undergoes homozygous deletion, complete loss of MTAP blocks both adenine salvage and methionine recycling from MTA:

1. Intracellular MTA Accumulation: MTA accumulates to micromolar steady-state levels (2–15 μM) inside the MTAP-null tumor cytoplasm and nucleus, while excess MTA passively diffuses across nucleoside transporters into the tumor microenvironment—where stromal fibroblasts and endothelial cells that retain intact MTAP clear extracellular MTA.
2. De Novo Purine Synthesis Dependency: Deprived of MTAP-salvaged adenine, MTAP-null cells rely more heavily on de novo purine biosynthesis (GART, ATIC, PPAT), explaining historical observations of heightened sensitivity to antifolates (pemetrexed, methotrexate) and L-alanosine when extracellular purine salvage is restricted.

The Heterooctameric PRMT5–WDR77 (MEP50) Methylosome

Mammalian protein arginine methyltransferases (PRMT1–PRMT9) transfer methyl groups from SAM to the guanidino nitrogen atoms of arginine residues, generating ω-N^G-monomethylarginine (MMA), asymmetric dimethylarginine (aDMA, catalyzed by Type I enzymes such as PRMT1), or symmetric dimethylarginine (sDMA, catalyzed by Type II enzymes). PRMT5 is the primary mammalian Type II methyltransferase responsible for >90% of cellular sDMA marks.

Biochemically, PRMT5 (637 residues, comprising an N-terminal TIM-barrel domain, a central Rossmann-fold SAM-binding catalytic domain with a double-E loop Glu435/Glu444, and a C-terminal β-barrel dimerization domain) assembles with the seven-bladed WD40-repeat cofactor WDR77 (MEP50) into a 430-kDa heterooctameric (PRMT5)4–(WDR77)4 methylosome. Substrate adaptor proteins—specifically pICln (CLNS1A) for spliceosomal Sm proteins, RIOK1 for ribosomal biogenesis factors, and COPR5 for histones (H4R3me2s, H3R8me2s, H2AR3me2s)—recruit distinct substrate cohorts to the PRMT5–WDR77 complex via a conserved GQF (PBW) motif.

Selective Inhibition by MTA and Retained-Intron Neoantigen Generation

Unlike Type I PRMTs (PRMT1, PRMT4/CARM1) and lysine methyltransferases, which possess roomy cofactor pockets that bind SAM with high affinity but reject MTA, the SAM-binding pocket of PRMT5 contains a unique hydrophobic sub-pocket around Trp579 and Phe327 that lacks electrostatic repulsion toward the uncharged thioether moiety of MTA. Consequently, MTA binds the PRMT5 active site with Ki ≈ 0.26 μM—roughly 6-fold tighter than the natural cofactor SAM (Km ≈ 1.5–2.0 μM) and more than 100-fold more potently than MTA inhibits any other cellular methyltransferase.

In MTAP-deleted tumor cells, roughly 50% to 75% of the PRMT5–WDR77 pool is trapped in a catalytically inactive PRMT5•MTA complex. When the remaining 25% to 50% of active PRMT5 is inhibited pharmacologically, symmetric dimethylation of the core spliceosome Sm proteins (SmB/B′ [SNRPB], SmD1 [SNRPD1], and SmD3 [SNRPD3]) on their C-terminal RG/RGG repeats collapses. Because the Tudor domain of Survival of Motor Neurons (SMN) requires sDMA-modified Sm tails to assemble heptameric Sm rings onto U1, U2, U4, U5, and U12 small nuclear RNAs (snRNAs), loss of PRMT5 activity triggers widespread failure of minor (U12) and weak-5′-splice-site alternative splicing, resulting in thousands of retained introns, premature termination codons, mitotic failure (MDM4, AURKB, CDK4 mis-splicing), and presentation of immunogenic HLA-A*02:01 splice-derived neoantigens.

Why First-Generation SAM-Competitive PRMT5 Inhibitors Failed in the Clinic

Following the 2016 discovery of MTAP–PRMT5 synthetic lethality in shRNA screens, first-generation PRMT5 inhibitors (GSK3326595/pemrametostat, JNJ-64619178/onametostat, and PF-06939999) were advanced into clinical trials. However, these molecules were either SAM-competitive or bound the apo/SAM-bound enzyme equally well in both MTAP-wild-type and MTAP-deleted cells. Because PRMT5 is an essential housekeeping gene in normal hematopoietic stem and progenitor cells (HSPCs), systemic inhibition of PRMT5 in MTAP-intact bone marrow caused dose-limiting thrombocytopenia, anemia, and neutropenia before reaching concentrations sufficient to eradicate MTAP-null solid tumors.

Second-Generation MTA-Cooperative PRMT5 Inhibitors and MAT2A Combination Strategies

To overcome bone marrow toxicity, structure-guided drug discovery exploited the differential cofactor occupancy between MTAP-null tumors (high [MTA], abundant PRMT5•MTA binary complex) and normal MTAP-intact bone marrow (low [MTA], PRMT5•SAM binary complex):

1. MTA-Cooperative PRMT5 Inhibitors (BMS-986504/MRTX1719, AMG 193, TNG908, TNG462, AZD3470): Discovered via fragment screening against the PRMT5•MTA co-crystal structure, these uncompetitive/cooperative small molecules occupy the arginine substrate channel immediately adjacent to bound MTA. While a bulky sulfonium methyl group on SAM sterically clashes with the drug, the smaller neutral methylthio group of MTA forms favorable van der Waals interactions and induces a conformational flip of Glu435/Trp579 that locks the inhibitor into a stable PRMT5•MTA•Inhibitor ternary complex. This confers a 30-fold to >100-fold selectivity window for MTAP-deleted tumor cells over MTAP-wild-type hematopoietic cells, producing confirmed partial and complete clinical responses across MTAP-deleted lung, pancreatic, biliary, and mesothelioma patients without severe myelosuppression.
2. Allosteric MAT2A Inhibitors (AG-270, IDE397/GSK4362676, S095035): These allosteric inhibitors bind the dimer interface of MAT2A to block SAM biosynthesis (~80% reduction in intracellular [SAM]). In MTAP-null cells where [MTA] is already high, lowering [SAM] collapses the [SAM]:[MTA] ratio below the threshold needed for PRMT5 catalysis; moreover, combining an allosteric MAT2A inhibitor (IDE397) with an MTA-cooperative PRMT5 inhibitor (AMG 193) synergistically drives nearly 100% of cellular PRMT5 into the druggable PRMT5•MTA state.

Why is MTAP co-deleted in 80–90% of CDKN2A-null cancers if MTAP loss creates a metabolic vulnerability?

On human chromosome 9p21.3, the 3′ end of MTAP lies less than 100 kb telomeric to CDKN2A (p16INK4a/p14ARF). Homozygous deletions of Chr9p21.3 arise through illegitimate recombination or non-homologous end joining across flanking fragile sites and repetitive retrotransposon elements that span several hundred kilobases to several megabases. Because the selective oncogenic advantage of simultaneously eliminating both the retinoblastoma (RB1) pathway checkpoint (p16INK4a) and the p53 (TP53) stabilization checkpoint (p14ARF) is strong, MTAP is swept along as a chromosomal passenger deletion; under baseline physiological conditions without a PRMT5 or MAT2A drug present, the remaining 25–50% of PRMT5 activity is sufficient to keep the tumor cell alive.

Why can stromal MTA clearance in the tumor microenvironment blunt MAT2A monotherapy, and how do MTA-cooperative PRMT5 inhibitors overcome it?

Because MTA is a neutral, membrane-permeable nucleoside, some intracellular MTA generated by MTAP-null carcinoma cells diffuses out into the interstitial space, where MTAP-proficient cancer-associated fibroblasts (CAFs) and macrophages phosphorolytically metabolize it. In certain stroma-rich tumors, this paracrine sink keeps tumor intracellular [MTA] at moderate levels (~1–3 μM). Whereas MAT2A monotherapy relies solely on endogenous MTA outcompeting SAM, MTA-cooperative PRMT5 inhibitors (BMS-986504, AMG 193, TNG462) act as thermodynamic "molecular glues" that trap MTA inside the PRMT5 binding pocket (Kd in the picomolar-to-low-nanomolar range), preventing MTA efflux and driving complete PRMT5 inactivation even at modest [MTA].

Why does PRMT5 inhibition selectively disrupt minor (U12) introns and weak 5′-splice-site introns rather than all pre-mRNA splicing?

Symmetric dimethylation (sDMA) of SmB/B′, SmD1, and SmD3 by PRMT5 accelerates SMN-assisted snRNP biogenesis and stabilizes protein–protein interactions within the catalytic spliceosome. Canonical high-strength U2-type introns with consensus 5′-GTAAGT donor sites require lower snRNP concentrations for rapid co-transcriptional excision, whereas slowly spliced U12-type minor introns and "detained" U2 introns with non-canonical 5′/3′ splice sites (MDM4, ATM, FANCA, KAT5/TIP60, AURKB) are kinetically sensitive to any drop in mature snRNP stoichiometry, causing selective intron retention and DNA damage response failure.

1. Kryukov, G. V., Wilson, F. H., Ruth, J. R., et al. (2016). "MTAP deletion confers enhanced dependency on the PRMT5 arginine methyltransferase in cancer cells." Science, 351(6278), 1214–1218. https://doi.org/10.1126/science.aad5214
2. Mavrakis, K. J., McDonald, E. R., Schlabach, M. R., et al. (2016). "Disordered methionine metabolism in MTAP/CDKN2A-deleted cancers leads to dependence on PRMT5." Science, 351(6278), 1208–1213. https://doi.org/10.1126/science.aad5944
3. Marjon, K., Cameron, M. J., Quang, P., et al. (2016). "MTAP deletions in cancer create vulnerability to targeting of the MAT2A/PRMT5/RIOK1 axis." Cell Reports, 15(3), 574–587. https://doi.org/10.1016/j.celrep.2016.03.043
4. Engstrom, L. D., Aranda, R., Waters, L., et al. (2023). "MRTX1719 is an MTA-cooperative PRMT5 inhibitor that exhibits synthetic lethality in preclinical models and early clinical studies of MTAP-deleted cancer." Cancer Discovery, 13(11), 2412–2431. https://doi.org/10.1158/2159-8290.CD-23-0669