NNMT Inhibitor Discovery: Tricyclic Compounds and Beyond

NNMT, short for nicotinamide N-methyltransferase, has emerged as one of the more intriguing drug targets in metabolic and cancer research over the past decade, and the discovery of tricyclic small-molecule inhibitors marked a turning point in making that target druggable with oral compounds. The enzyme sits at a crossroads of metabolism and epigenetics, consuming methyl groups that cells use both for energy regulation and gene silencing. Blocking it in animal models has produced striking improvements in obesity, blood sugar control, and even muscle function during aging. Yet designing inhibitors that work well inside living cells has proven surprisingly tricky, and the tricyclic series that came out of high-throughput screening represents only one chapter in a larger, still-unfolding story.

Why NNMT Matters as a Drug Target

NNMT’s day job is simple: it transfers a methyl group from the universal methyl donor SAM (S-adenosylmethionine) onto nicotinamide, producing 1-methylnicotinamide (MNA). That reaction does two things at once. First, it removes nicotinamide from the NAD salvage pathway, reducing levels of NAD+, a molecule central to cellular energy metabolism. Second, it consumes SAM, the same cofactor that cells rely on for methylating DNA and histones. When NNMT is highly active, both NAD+ availability and the cell’s capacity for epigenetic methylation can drop.

In metabolic disease, the consequences are measurable. NNMT expression runs high in the liver and white fat tissue of obese and diabetic animals, and knocking down or inhibiting the enzyme boosts energy expenditure, cuts fat accumulation, and improves insulin sensitivity.¹ Those effects hold up across multiple experimental approaches: genetic knockdown, antisense strategies, and small-molecule inhibitors all point in the same direction, with treated animals showing reduced body weight, lower white fat mass, better glucose tolerance, and normalized fasting blood sugar.²

In cancer, the picture is equally compelling. NNMT is overexpressed in a wide range of tumor types and is linked to tumor growth, spread, and the formation of a supportive tumor microenvironment.³ In esophageal squamous cell carcinoma, for instance, high NNMT expression drives metastasis by consuming methyl groups so aggressively that the cell can no longer maintain the histone marks needed to keep the adhesion molecule E-cadherin turned on, loosening cells from their neighbors and enabling spread.⁴ In breast cancer, NNMT depletion produces the opposite effect: a “methyl overflow” that ramps up silencing marks on pro-metastatic genes.⁵

The Enzyme’s Architecture and What Inhibitors Must Contend With

Understanding the crystal structure of NNMT was a prerequisite for rational drug design. The enzyme’s active site holds two substrates at once: SAM in one pocket and nicotinamide in an adjacent one, with a narrow “methyl transfer tunnel” connecting them where the actual chemistry takes place. A crystal structure resolved at 2.7 angstroms revealed the key residues that grip each substrate and showed how the methyl group is handed off.⁶ That structural picture gave drug hunters three possible points of attack: the nicotinamide pocket, the SAM-binding pocket, or both at once.

Each approach comes with trade-offs. Compounds that mimic SAM risk hitting other methyltransferases, since SAM is a cofactor for dozens of enzymes. Compounds that target only the nicotinamide pocket tend to be small and flat, which makes binding tight enough for drug-like potency a challenge. And compounds that span both pockets, the so-called bisubstrate inhibitors, can achieve remarkable potency but often struggle with cell permeability and the practical constraints of oral dosing. These tensions have driven much of the inhibitor discovery work over the past ten years.

The Tricyclic Series From High-Throughput Screening

The tricyclic inhibitor program began with a large-scale high-throughput screening campaign, testing hundreds of thousands of compounds for their ability to block NNMT. Out of that effort came a series built around a tricyclic core, a fused three-ring scaffold that binds in the nicotinamide pocket of the enzyme.⁷ The initial hits were modest in potency, but structural biology data guided a ring-enlargement strategy: by expanding one ring to better fill the nicotinamide binding pocket, the team saw improved inhibitory activity that justified a full lead-optimization campaign.⁸

Hundreds of analogs were synthesized in the optimization phase, each variation testing which chemical features mattered for potency, selectivity, and druglike properties. Co-crystal structures of optimized compounds like JBSNF-000028 and JBSNF-000107 revealed why the tricyclic scaffold works: the flat, planar molecule slides beneath a hairpin structural motif at the nicotinamide pocket and stacks between two amino acid residues, Tyr-204 and Leu-164. An amine group on a saturated ring forms hydrogen bonds with the side chain of Tyr-20 and the backbone of Leu-164, and sits within about 3.2 angstroms of the sulfur atom of SAH (the spent SAM cofactor). A fluorine on the aromatic ring adds another interaction with Ser-201.⁹ The earlier, weaker hit compounds lacked one or both of the hydrogen bonds to Tyr-20 and Leu-164, which neatly explained their lower potency.

Moving from enzymatic assays to cell-based assays revealed a recurring headache in the NNMT field: cellular potency lagged behind test-tube potency by roughly 30-fold, likely due to limited cell-membrane permeability. Still, the most active compounds did measurably reduce MNA production inside human U2OS cells, with the best performer lowering endogenous MNA levels by about 75 percent after just one hour of treatment.⁸

The Tricyclics in Living Animals

The animal data for the tricyclic series has been among the most encouraging results in the NNMT inhibitor field. JBSNF-000028, dosed orally twice daily at 50 mg/kg to mice with diet-induced obesity, produced a statistically significant reduction in body weight starting from around day 23 of treatment, without any change in how much the animals ate. Fed blood glucose dropped significantly by day 21, oral glucose tolerance normalized to the level of lean control animals by day 28, and the compound drove down plasma triglycerides, LDL cholesterol, and liver triglycerides and cholesterol.¹⁰

An earlier related compound, JBSNF-000088, showed a similar profile: significant body weight reduction throughout treatment, lower fasting blood glucose and insulin, improved glucose tolerance normalized to lean controls, and a marked improvement in a standard measure of insulin resistance.⁸ The fact that two structurally related tricyclic compounds produced overlapping benefits, each administered orally rather than by injection, reinforced confidence that the target itself was driving the outcomes rather than some off-target quirk of a single molecule.

One cautionary finding emerged from selectivity profiling. When JBSNF-000028 was tested against a broad panel of metabolic and receptor targets at a high concentration, it came up clean on everything except monoamine oxidase A (MAO-A), which it inhibited by about 90 percent.⁸ MAO-A breaks down serotonin and norepinephrine, and its inhibition is the basis for a class of antidepressants but also carries dietary and drug-interaction risks. Whether this off-target activity would matter at therapeutic doses or could be engineered out in further optimization is an open question, but it highlights why selectivity profiling remains a critical gate for any NNMT drug candidate.

Bisubstrate Inhibitors and the Potency Ceiling

While the tricyclic series was being optimized for oral dosing and metabolic disease, a parallel line of work pursued bisubstrate inhibitors that bridge both the nicotinamide and SAM pockets simultaneously. The logic is elegant: by occupying the entire active site at once, a single molecule can achieve extraordinary binding affinity. The compound NS1, a nicotinamide-SAM conjugate connected by an alkyne linker, was designed to mimic the geometry of the methyl-transfer transition state. The alkyne’s linear shape threads through NNMT’s methyl-transfer tunnel with near-perfect shape complementarity, and NS1 proved to be a subnanomolar inhibitor.¹¹

A closely related compound, LL320, used a propargyl linker and achieved a binding constant of 1.6 nanomolar, confirmed by its own co-crystal structure showing occupation of both pockets.¹² Later optimization produced compounds like II559 and II802, with binding constants around 1.2 and 1.6 nanomolar respectively, and over 5,000-fold selectivity for NNMT over closely related methyltransferases. These compounds also achieved the best cellular potency in the bisubstrate class, with cellular activity around 150 nanomolar.¹³

The bisubstrate approach produces the most potent NNMT inhibitors yet reported, but these molecules carry adenosine-like fragments that make oral bioavailability difficult to achieve. They have been invaluable as tool compounds for proving that NNMT inhibition has biological consequences in cell culture, and their selectivity data provides some of the best evidence that inhibiting NNMT specifically, rather than methyltransferases broadly, is what produces the observed effects. Whether they can be turned into clinical drug candidates is less certain.

Covalent and Mechanism-Based Approaches

A different strategy avoids the permeability problem by using small, cell-penetrant molecules that form permanent bonds with the enzyme. NNMT has an active-site cysteine residue (C159) in its SAM-binding pocket that is susceptible to electrophilic attack. Researchers exploited this with alpha-chloroacetamide compounds that achieved sub-micromolar potency in test-tube assays and excellent selectivity across the proteome in cell lysates, as measured by activity-based protein profiling.¹⁴ These covalent inhibitors also exposed an uncomfortable reality for the field: even highly selective inhibitors in cell lysates can face “target engagement challenges” in intact cells, where the compound must reach the enzyme before reacting with other cellular thiols.

An even cleverer variation exploits the enzyme’s own catalytic activity. Suicide substrates like 4-chloropyridine are designed so that when NNMT methylates their nitrogen, the resulting positive charge makes the adjacent carbon vulnerable to attack by that same C159 cysteine, permanently inactivating the enzyme from within. This mechanism was confirmed in vitro and in cells, and an alkyne-tagged version was used to selectively label NNMT, providing a tool for tracking where and when the enzyme is active.¹⁵

A related concept applies to substrate-analog inhibitors, where a compound enters the enzyme disguised as nicotinamide and gets methylated, but the methylated product is itself a potent inhibitor that stays stuck in the active site. The substrate form is uncharged and crosses cell membranes easily; the product form is charged and potent. This “on-target biotransformation” strategy combines the permeability of one species with the potency of the other.¹⁶ The catch: the charged product leaks out of cells, making the inhibition transient rather than sustained.

Non-SAM-Mimetic Bisubstrate Inhibitors and Moving Beyond Adenosine

One of the field’s more recent achievements has been the development of bisubstrate inhibitors that avoid mimicking SAM’s adenosine portion. This matters because adenosine-like fragments are notorious for off-target effects and poor drug properties. Starting from a high-throughput screening hit, structure-based optimization yielded a lead compound with over a 1,000-fold improvement in potency, from 10 micromolar down to about 8.4 nanomolar. That lead showed sub-micromolar activity in cell-based assays and strong selectivity within the methyltransferase family.¹⁷ In rodents, the compound concentrated in kidney tissue and achieved dose-dependent NNMT inhibition in a renal fibrosis model, hinting at applications beyond obesity and cancer.

Beyond Metabolic Disease and Cancer

NNMT inhibitors are now being tested in models of age-related muscle loss. In aging mice, NNMT inhibitor treatment improved grip strength and slowed the decline in quadriceps mass and whole-body lean mass. The treated animals had higher NAD+ levels in their muscles, along with increased expression of PGC1-alpha and activated AMPK, both of which are associated with mitochondrial health and exercise adaptation.¹⁸ The NAD+ connection is particularly interesting: by preventing NNMT from siphoning off nicotinamide, inhibition funnels more of it back into the NAD salvage pathway, effectively boosting the cell’s NAD+ supply without administering NAD+ precursors directly.

Separately, a compound called 5-amino-1-methylquinoline (5A1MQ) was tested in obese mice and showed dose-dependent reductions in body weight and fat mass, improved glucose tolerance and insulin sensitivity, and reduced signs of fatty liver disease including lower liver triglycerides and less inflammatory cell infiltration.¹⁹ 5A1MQ has attracted a following in the biohacking community for its purported fat-loss effects, though it remains an experimental compound without clinical trial data in humans.

What Drives NNMT Overexpression in Disease

Understanding what pushes NNMT levels up in the first place matters for both drug development and patient selection. In cancer, the transcription factor STAT3 appears to be a key switch. Stimulating liver cancer and colon cancer cells with the inflammatory signal interleukin-6 drove up NNMT expression through STAT3 activation, and knocking down STAT3 in cancer cells with constitutively high NNMT brought it back down. In colon cancer tissue samples, activated STAT3 and NNMT expression were tightly correlated.²⁰ TGF-beta-1, another signal associated with tissue remodeling and worse outcomes, was found to be overexpressed alongside NNMT in clear cell renal cell carcinoma, suggesting it may drive NNMT upregulation in that context.²¹

These upstream regulators offer a few practical implications. First, NNMT overexpression in cancer is not random but is driven by specific inflammatory and growth-factor signals, which means tumors with active IL-6/STAT3 signaling or high TGF-beta might be the ones most likely to benefit from NNMT inhibition. Second, combination strategies that target both the upstream signal and the enzyme itself could potentially hit the problem from two angles.

The Gap Between Preclinical Promise and Clinical Reality

Despite the breadth of preclinical data, no NNMT inhibitor has yet entered a pivotal human clinical trial. The field faces several challenges that help explain the gap. Cell permeability remains a persistent issue: many of the most potent inhibitors in enzymatic assays lose much of their punch once they need to cross a cell membrane, as the tricyclic series demonstrated with its 30-fold potency drop from test tube to cell. The bisubstrate inhibitors face an even steeper version of the same problem because of their large, charged structures.

Selectivity is another ongoing concern. NNMT’s active site shares features with dozens of other SAM-dependent methyltransferases, and compounds that occupy the SAM pocket risk unintended effects on DNA and histone methylation, the very epigenetic machinery that cells depend on for normal gene regulation. The tricyclic series sidesteps this partly by binding primarily in the nicotinamide pocket, but even it showed that unexpected off-target activity against MAO-A that would need to be resolved. The best-in-class bisubstrate inhibitors have demonstrated thousands-fold selectivity over related methyltransferases, but that level of selectivity profiling has only been done for a handful of the closest enzyme relatives, not the full human methyltransferase panel.

Then there is the question of which disease to pursue first. Obesity, type 2 diabetes, cancer, kidney fibrosis, and sarcopenia are all on the table. Each has different regulatory pathways, different patient populations, and different competitive landscapes. A metabolic indication might be the most straightforward path based on the animal data, but it would need to compete against GLP-1 receptor agonists and other approaches that already have extensive human evidence. An oncology indication might face a lower bar for initial approval but requires identifying which tumor types are most NNMT-dependent and whether inhibition alone is sufficient or needs to be paired with other therapies like immune checkpoint blockade.

The science of NNMT inhibition has moved rapidly from target validation through multiple generations of inhibitor chemistry, with the tricyclic series providing the first clear proof that an orally dosed small molecule can hit this enzyme hard enough to reverse metabolic dysfunction in a living animal. Whether the tricyclic scaffold or one of the newer non-SAM-mimetic bisubstrate compounds becomes the backbone of a clinical candidate remains to be seen, but the target itself has earned its place on the shortlist of metabolic enzymes worth drugging.