SLU-PP-332: Acylhydrazone Chemistry and Pan-ERR Receptor Pharmacology

SLU-PP-332 is a small organic molecule rather than a peptide. Its structure joins 4-hydroxybenzohydrazide to 2-naphthaldehyde through a carbon-nitrogen double bond, giving the N-acylhydrazone 4-hydroxy-N-[(E)-naphthalen-2-ylmethylideneamino]benzamide. PubChem lists CID 5338394, formula C18H14N2O2, monoisotopic mass 290.105528 Da and InChIKey RNZIMBFHRXYRLL-XDHOZWIPSA-N, with the E configuration specified at the imine.

The scaffold is not new to nuclear-receptor chemistry. GSK4716, the acyl hydrazide agonist used for years as the ERRβ and ERRγ reference tool, carries the same 4-hydroxybenzohydrazide head group; PubChem records it as CID 5331325, C17H18N2O2, 282.34 g/mol. The two molecules differ at the aldehyde-derived end: GSK4716 carries a 4-isopropylphenyl ring, SLU-PP-332 carries a fused bicyclic 2-naphthyl ring in its place. That single substitution is what shifted potency onto ERRα, and it is the change the 2026 structure-activity work by Okda and colleagues set out to map systematically.

Computed descriptors describe a compact, lipophilic and only moderately polar solid. PubChem gives an XLogP of 3.7, a topological polar surface area of 61.7 square angstroms, two hydrogen-bond donors, three acceptors and only three rotatable bonds across 22 heavy atoms. Supplier certificates report 58 mg/mL solubility in dimethyl sulfoxide against 2 mg/mL in ethanol and no measurable water solubility, which is why every published cell protocol starts from a dimethyl sulfoxide stock.

SLU-PP-332 targets the estrogen-related receptor family, three orphan nuclear receptors designated ERRα (NR3B1), ERRβ (NR3B2) and ERRγ (NR3B3). Billon and colleagues reported in ACS Chemical Biology in 2023 half-maximal activation at 98, 230 and 430 nM respectively in cell-based cotransfection assays, a 4.4-fold potency margin favouring ERRα over ERRγ and a 2.3-fold margin over ERRβ.

The comparison run alongside those numbers explains why the compound was described as filling a gap. In the same assay format GSK4716 registered above 5000 nM at both ERRα and ERRβ and 1200 nM at ERRγ. Agonists with useful ERRα potency had been the missing piece of the ERR toolkit; the naphthyl acylhydrazone supplied one with a two-order-of-magnitude improvement at that isoform.

Receptor nomenclature invites a specific misreading worth heading off. Estrogen-related receptors share sequence homology with the classical estrogen receptors in the DNA-binding region, but no estrogen binds them and no endogenous ligand has been identified for any of the three. Kallen and colleagues framed ERRα as transcriptionally active without a ligand, driven instead by the supply of coactivator. Pharmacological agonism at this receptor therefore means stabilising an already-competent fold rather than switching an idle receptor on.

One honest gap remains in the public record for this molecule. The primary characterisation is against ERRα, ERRβ and ERRγ; a published counter-screen of SLU-PP-332 across the classical estrogen receptors or a wider nuclear-receptor panel was not located during this review, so selectivity outside the ERR subfamily is undocumented rather than established.

SLU-PP-332 acts through transcription rather than through an enzyme, so its readouts are messenger RNA counts. Billon and colleagues exposed C2C12 skeletal muscle cells to 10 µM for 24 hours and primary myocytes to 1 µM, then sequenced the transcriptome. DNA damage-inducible transcript 4, gene symbol Ddit4, emerged as the most strongly induced shared transcript across both systems.

The wider induced set is chemically legible. It includes Slc25a25, an ATP-magnesium and phosphate carrier of the inner mitochondrial membrane, pyruvate dehydrogenase kinase 4 (Pdk4), the circadian transcripts Per1 and Per2, the forkhead factor Foxo1, aminolevulinate synthase 2 (Alas2) and the myosin heavy chain transcript Myh6. Overlap between that set and the transcripts moved by an acute cycling bout in humans was reported at a p value below 2.3 × 10⁻¹⁴.

Functional mitochondrial readouts accompanied the transcript work. C2C12 cells exposed to the compound showed an increase in maximum mitochondrial respiration relative to vehicle in extracellular flux measurement, together with stronger MitoTracker Red staining consistent with greater mitochondrial content. Xu and colleagues later reported the same direction of travel in neonatal rat ventricular myocytes at 10 µM, drawn from a 10 mM dimethyl sulfoxide stock, where oxidative capacity and palmitate-driven oxygen consumption both rose.

Coactivator supply is the hinge in this mechanism. The ERRα structure solved with a PGC-1α fragment showed the coactivator binding through an inverted LXXLL motif, an LLXYL sequence, rather than the canonical arrangement seen elsewhere in the nuclear-receptor family. Bonanni and colleagues reported in Frontiers in Physiology in 2025 that primary human myoblasts exposed to SLU-PP-332 shifted exactly that axis, with NOX4 protein falling and SIRT1, PGC-1α, ERRα, FNDC5, Akt and Bcl-2 rising.

SLU-PP-332 has to enter a pocket that structural biology described as essentially full. Kallen and colleagues solved the ERRα ligand-binding domain bound to a PGC-1α peptide at 2.5 angstrom resolution in 2004, deposited as PDB entry 1XB7, and reported the putative cavity almost completely occupied by side chains, the bulky Phe328 prominent among them.

That observation carries a hard consequence for medicinal chemistry. The authors concluded that a ligand larger than roughly four carbon atoms could occupy the ERRα pocket only if the receptor underwent a major conformational change. A 22-heavy-atom naphthyl acylhydrazone is far past that threshold, which is the structural reason ERRα agonists remained scarce for two decades while ERRβ and ERRγ tools appeared.

The ERRγ structures supply the binding-mode template. Wang and colleagues published ERRγ ligand-binding domain crystal structures in three functional states in 2006, including the agonist-bound form with GSK4716 and a RIP140 peptide at 2.60 angstrom resolution, deposited as PDB entry 2GPP. That work reported an unexpected rearrangement of the phenol-binding residues on agonist engagement and, unusually, activation without major rearrangement or marked stabilisation of the C-terminal helix.

Later chemistry built directly on that coordinate set. Xu and colleagues described designing the second-generation pan-ERR agonist SLU-PP-915 from the GSK4716-bound ERRγ structure, replacing the central hydrazide with five-membered heterocycles to yield a series of disubstituted thiophenes. Okda and colleagues then published the first systematic structure-activity relationship analysis of the SLU-PP-332 scaffold itself in 2026, combining synthesis, functional assays, gene-expression profiling, docking and molecular dynamics, and the 2026 review by Zhao and colleagues in ACS Pharmacology and Translational Science collects the ERR structural corpus in one place.

SLU-PP-332 engages all three isoforms, so which one carries a given response is a question about tissue rather than about the compound. In skeletal muscle Billon and colleagues reported the acute transcriptional program as ERRα-dependent. In cardiomyocytes Xu and colleagues reached the opposite assignment, with ERRγ carrying essentially the entire response.

The cardiac dependency experiment is unusually clean. Small interfering RNA knockdown of ERRγ in neonatal rat ventricular myocytes abolished compound-driven induction across the measured gene set, the single exception being Cox6a2 induction by SLU-PP-915. A cardiac-specific ERRγ knockout removed the in vivo readout entirely, which is the genetic control that separates on-target activity from off-target chemistry.

Kidney work adds a third tissue context. Wang and colleagues reported in The American Journal of Pathology in 2023 that ERR transcripts decline in both aging human and aging mouse kidney and are preserved by lifelong caloric restriction. In 21-month-old male C57BL/6 mice given 25 mg/kg per day by the intraperitoneal route for 8 weeks, the mitochondrial to nuclear DNA ratio and the quantity of assembled respiratory complexes both recovered, while STAT3 phosphorylation at Tyr705 and expression of the senescence marker p21 (Cdkn1a) fell.

Exposure figures put those results on a scale that can be checked against the potency numbers. Rodent work reported roughly 0.6 µM in skeletal muscle and 0.2 µM in plasma following 30 mg/kg by the intraperitoneal route, tissue concentrations that sit several-fold above the 98 nM ERRα EC50 and near the 230 and 430 nM values for the other two isoforms. Pan-isoform engagement at those exposures is therefore expected rather than incidental.

SLU-PP-332 became an analytical target in 2026, when two independent laboratories published in vitro metabolite maps for doping-control screening. Möller, Krug and Thevis reported the protonated molecule at m/z 291.1142 and a major product ion at m/z 121.0288, an accurate mass matching the elemental composition C7H5O2, using positive heated electrospray at 3000 volts.

Their incubation conditions are reproducible in any metabolism laboratory. Substrate at 200 µM was held at 37 degrees Celsius for 24 hours with pooled human liver microsomes and S9 fraction at 20 mg/mL protein, an NADPH regenerating system for phase I and UDP-glucuronic acid, saccharic acid lactone and 3-phosphoadenosine-5-phosphosulfate for phase II. Nine metabolites resulted: monohydroxylation at m/z 307.1077, bis-hydroxylation at m/z 323.1026, bis-hydroxylation with reduction at m/z 325.1183, a glucuronide at m/z 467.1449, a sulfate at m/z 371.0696 and a hydroxy-sulfate at m/z 387.0645. Assignments were held to a mass error tolerance of 5 parts per million or better.

Avliyakulov and colleagues, working with pooled human liver S9 fractions and LC-MS with high-resolution detection, resolved a larger map of 22 metabolites: five monohydroxylated, three dihydroxylated, four reduced dihydroxylated species, direct glucuronide and sulfate conjugates of the parent, and a further group of conjugated hydroxy metabolites. Eight of the 22, designated M1, M7, M9, M10, M13, M14, M19 and M20, were nominated as the most abundant and the most useful for screening. The authors noted that full structural elucidation of the metabolites remains outstanding.

A structural contrast between the two SLU compounds emerges from these datasets. The SLU-PP-332 metabolite set is dominated by ring hydroxylation and phenol conjugation, with the acylhydrazone linkage surviving phase I intact. SLU-PP-915, built on a thiophene rather than a hydrazide core, instead produced an amide hydrolysis product at m/z 247.0242 and a hydrolysis-plus-oxidation species at m/z 219.0121 among its seven phase I metabolites, and three of those were synthesised independently and confirmed by nuclear magnetic resonance.

SLU-PP-332 is supplied as a solid and characterised on supplier certificates at 99.5 percent purity. Reported solubility figures are 58 mg/mL in dimethyl sulfoxide and 2 mg/mL in ethanol, with no measurable water solubility, and powder storage is stated at minus 20 degrees Celsius. Published cell protocols work from a dimethyl sulfoxide stock, commonly 10 mM, diluted into medium at 1 to 10 µM.

The acylhydrazone linkage is the chemically interesting part of the handling profile. As a class, acylhydrazones are hydrolytically robust at neutral and basic pH and degrade under acidic aqueous conditions, and they undergo E to Z photoisomerisation on ultraviolet irradiation. Fernández-Palacios and colleagues quantified that behaviour across a seven-compound acylhydrazone series in 2023, recording for one congener 24 percent conversion to the Z isomer after 5 minutes of ultraviolet exposure in acetonitrile, rising to 78 percent at 45 minutes, with thermal reversion on heating to 150 degrees Celsius. Neither result was generated on SLU-PP-332 itself, so the practical inference is procedural: amber glass, no acidic aqueous holds, and identity confirmation after any handling that involves light or low pH.

Identity and purity confirmation now has a published method to copy. Möller, Krug and Thevis separated the compound on a Poroshell 120 EC C18 column of 3.0 by 50 mm and 2.7 µm particle size with a Nucleoshell RP 18 Plus guard, using 0.1 percent formic acid in water and in acetonitrile across a 0 to 100 percent organic gradient over 10 minutes at 0.3 mL/min, with 10 µL loaded. Detection used a heated electrospray source at 3000 volts positive or minus 2600 volts negative, full scan at 60,000 FWHM resolution and product ion scans at 30,000 FWHM. The parent appears at m/z 291.1142 with a major product ion at m/z 121.0288.

SLU-PP-332 holds no marketing authorisation in any jurisdiction and is not a medicine. Material is supplied for laboratory research use only, not for human or veterinary use, and should be handled by trained personnel under institutional chemical hygiene procedures with the supplier certificate of analysis on file.