Structure-Guided Development of Selective HDAC4 and HDAC7 Inhibitors That Modulate Th17 Cell Transcription

Abstract

Class IIa histone deacetylases (HDACs) are attractive therapeutic targets due to their specialized regulatory functions and restricted expression patterns. Recent studies identified HDAC4 and HDAC7 as critical regulators of Th17-cell differentiation, acting through distinct transcriptional mechanisms to promote Th17 lineage commitment. Building on these findings, we designed, synthesized, and characterized a series of selective HDAC4/7 inhibitors derived from the Class IIa scaffold TMP269. Structure-guided optimization of the metal-proximal and hydrophobic-binding regions generated analogs with improved potency and enhanced isoform selectivity. Luciferase-based transcriptional assays, protein-binding studies, and molecular docking analyses revealed distinct structure–activity relationships that govern HDAC4 versus HDAC7 selectivity. Several compounds exhibited potent and selective inhibition of HDAC4- or HDAC7-dependent transcriptional activity, providing valuable chemical probes to dissect the individual functions of these enzymes in Th17 biology. These next-generation inhibitors establish a foundation for developing targeted therapies for Th17-driven inflammatory disorders, including inflammatory bowel disease.

Histone deacetylases (HDACs) remove acetyl groups from lysine residues and regulate chromatin accessibility and gene transcription. The human HDAC family comprises 18 enzymes that are classified into four groups based on sequence homology and catalytic mechanism. Class I, II, and IV HDACs are Zn²⁺-dependent hydrolases, whereas Class III sirtuins require NAD⁺ as a cofactor. Because individual HDAC isoforms exhibit distinct biological functions, isoform-selective HDAC inhibition has emerged as an important therapeutic strategy.

Several HDAC inhibitors, including vorinostat (1, SAHA), belinostat (2), panobinostat (3), and romidepsin (4). (Figure ), have been approved for the treatment of hematologic malignancies, validating HDACs as druggable therapeutic targets. , However, these pan-HDAC inhibitors frequently cause dose-limiting toxicities, including hematologic, gastrointestinal, metabolic, and cardiac adverse effects. These limitations have motivated the development of isoform-selective inhibitors with improved therapeutic windows.

1.

FDA-approved HDAC inhibitors.

Among HDAC family members, Class IIa HDACs (HDAC4, HDAC5, HDAC7, and HDAC9) have particularly attractive targets because of their specialized regulatory functions. Unlike other HDAC subclasses, they possess minimal intrinsic deacetylase activity owing to a conserved Tyr-to-His substitution within the catalytic pocket. , Instead, Class IIa HDACs function primarily as signal-responsive transcriptional regulators through interactions with transcription factors and coregulatory complexes mediated by their large N-terminal domains. ,− Dysregulation of these pathways has been implicated in cancer, cardiovascular disease, and immune disorders.

Recent studies from our laboratory identified HDAC4 and HDAC7 as key regulators of Th17-cell differentiation. HDAC4 cooperates with JunB to activate Th17 signature genes, including Il17a and Il17f, whereas HDAC7 associates with Aiolos and the SMRT/NCoR1–HDAC3 corepressor complex to repress negative regulators such as Il2. These findings establish HDAC4 and HDAC7 as attractive therapeutic targets for Th17-driven inflammatory diseases, including inflammatory bowel disease, multiple sclerosis, and rheumatoid arthritis.

Despite their biological importance, selective chemical probes capable of distinguishing HDAC4 from HDAC7 are not available. Such probes are needed to define the individual functions of these closely related enzymes and to evaluate their therapeutic potential. Here, we report the structure-guided design, synthesis, and characterization of a series of HDAC4- and HDAC7-selective inhibitors derived from the Class IIa inhibitor TMP269 (5). Functional, biochemical, and molecular modeling studies reveal distinct structure–activity relationships governing HDAC4 versus HDAC7 selectivity and identify potent chemical probes for interrogating their individual transcriptional functions.

Developing isoform-selective Class IIa HDAC inhibitors requires optimization of both the conserved catalytic pocket and the structural features unique to individual family members. TMP269 (5), a trifluoromethyloxadiazole (TFMO)-based inhibitor, is a first-generation Class IIa-selective ligand with demonstrated biological activity. Crystal structures of the HDAC7–TMP269 complex revealed that the TFMO group occupies the zinc-proximal region through interactions with conserved residues H709, D707 and D801 without directly coordinating the catalytic Zn²⁺ ion³ (Figure a). The tetrahydropyran ring is positioned within a pocket defined by three variable loops (A, B, C) (Figure b), which adopt distinct conformations in HDAC4 and HDAC7 despite their otherwise high structural similarity. These structural differences suggested opportunities to improve potency and isoform selectivity through optimization of zinc-proximal interactions and hydrophobic contacts.

2.

Structure-guided design of HDAC4- and HDAC7-selective inhibitors. (a) Structural features of the TMP269-binding pocket in HDAC4 and HDAC7, illustrated using the crystal structure of HDAC7 (PDB: 3ZNR). Residues unique to Class IIa HDACs are shown in red, whereas residues conserved only between HDAC4 and HDAC7 are colored in blue. (b) Sequence alignment of loops A, B, and C in Class IIa HDACs, colored as in panel a . (c) Design strategy for next-generation HDAC4- and HDAC7-selective inhibitors.

To strengthen interactions within the catalytic channel, we replaced the TFMO group with a trifluoromethyl-1H-pyrazole (TFPO) moiety and redesigned the tetrahydropyran region to introduce additional hydrogen-bonding and hydrophobic interactions (Figure c). These modifications were intended to engage residues R540, T625, and A807 in HDAC7, and corresponding residues S539, I628, and P806 in HDAC4. This structure-guided strategy generated a focused library of next-generation HDAC4/7 inhibitors for evaluation of potency, selectivity, and transcriptional activity.

The synthesis of analogs 15a–o is outlined in Scheme . Cyclization of commercially available 2-bromo-1-arylethan-1-one (6a–e) with 2-cyanoethanethioamide (7) afforded thiazoles 8a–e, which were alkylated to give 9a–g. Reduction of 9a–g with LiAlH₄ yielded amines 10a–g (Scheme a). In parallel, substituted ethyl 3-bromobenzoate (11a–f) were converted to trifluoromethylpyrazol esters 13a–f and hydrolyzed to acids 14a–f (Scheme b). Amide coupling of 10a–g with 14a–f furnished the target compounds 15a–o (Scheme c).

1.

^a Reagents and Conditions: (a) TEA, THF, rt, 12 h; (b) NaH, THF, 0 °C-rt; (c) 1 M LiAlH₄, THF. rt, 3 h; (d) Pd­(PPh₃)₄, Na₂CO₃, Dioxane, 80 °C, 12 h; (e) LiOH, H₂O:MeOH:THF, rt, 4 h; (f) HATU, DIPEA, DMF, rt, 12 h.

The structure–activity relationships (SARs) of compounds 15a–o were evaluated using two complementary luciferase reporter assays that measure HDAC4-dependent transcriptional activation of Il17a with JunB, or HDAC7-dependent transcriptional repression of Il2 with Aiolos in Jurkat T cells. Activities were normalized to TMP269 (5) and docking scores for HDAC4 and HDAC7 are summarized in Table .

1. Structure-Activity Relationships of HDAC4- and HDAC7-Selective Inhibitors.

Among compounds 15b–k, which retain the pyran C-ring, substitutions on the A and B rings strongly influenced potency and isoform selectivity. For HDAC4, meta-substitution on the A-ring proved optimal. Compound 15e (3-OMe) exhibited the strongest HDAC4 inhibition (1.87-fold relative activity) and a favorable docking score (−7.06 kcal/mol), while showing little activity toward HDAC7. Docking analysis suggested that the meta-methoxy group stabilizes the pyrazole core and promotes productive positioning of the amide linker within the zinc-proximal region. In contrast, the para-methoxy analog 15f displayed substantially reduced HDAC4 selectivity (1.22-fold activity), indicating that meta-substitution provides a more favorable steric and electronic environment for HDAC4 recognition.

Induced-fit docking suggested that this selectivity arises from conformational differences in loops A and B, which affect the positioning of F812 in HDAC4 and F679 in HDAC7 (Figure a). The larger pocket surrounding F812 accommodates the bulkier ligand 15e and enables a water-mediated hydrogen bond between the methoxy oxygen and G975. In contrast, the more constrained HDAC7 pocket prevents formation of this interaction because of steric clash, thereby reducing binding affinity.

3.

Cellular assays for HDAC4- and HDAC7 activity. (a) Induced-fit docking models of 15e bound to HDAC4 and HDAC7, generated using the HDAC7 crystal structure (PDB: 3ZNR). (b) Co-immunoprecipitation analysis of HDAC7-Aiolos and HDAC4-JunB interactions in Th17 cells following 24 h treatment with vehicle, TMP269, 15e, or 15i (5 μM each). Data are representative of three independent experiments.

A distinct SAR trend emerged for HDAC7. Meta-substitution on the B-ring was critical for activity, with compound 15k (3-Cl) displaying the strongest HDAC7 inhibition (2.00-fold activity) and the most favorable docking score (−7.70 kcal/mol). Docking models suggest that the meta-chloro substituent optimally positions the aromatic ring within a hydrophobic channel adjacent to the catalytic pocket. Notably, relocation of the chloro group to the para position (15j) largely abolished HDAC7 activity despite only a modest reduction in docking score (−6.79 kcal/mol), highlighting the sensitivity of HDAC7 recognition to subtle geometric changes.

Modifications of the C-ring further defined the structural requirements of activity. Replacement of the pyran ring with a alkyl linker (15a, 15l–o) resulted in substantially weaker docking scores and minimal cellular activity, indicating that the pyran scaffold is required to maintain the bioactive conformation necessary for productive target engagement.

Overall, the SAR and docking studies demonstrate that small positional changes in the A and B rings exert profound effects on HDAC4/7 selectivity. Docking scores correlated closely with HDAC7 activity but less consistently with HDAC4 inhibition, suggesting that distinct structural determinants govern ligand recognition by the two isoforms. These findings establish key design principles for achieving isoform-selective Class IIa HDAC inhibition.

We next examined whether the most selective compounds could disrupt physiologically relevant HDAC4 and HDAC7 transcriptional complexes. Co-immunoprecipitation and immunoblot analyses showed that 15e effectively disrupted the HDAC4-JunB interactions in Th17 cells, with activity comparable to TMP269, whereas 15i was inactive (Figure b). Conversely, 15i selectively disrupted HDAC7 binding to Aiolos, whereas 15e had no detectable effect. These results confirm the cellular activity and isoform selectivity of the newly developed HDAC4- and HDAC7-selective inhibitors.

In summary, we developed a focused series of next-generation Class IIa HDAC inhibitors derived from the TMP269 scaffold and defined the structural determinants governing HDAC4 and HDAC7 selectivity. Integrated SAR and molecular docking analyses identified meta-substitution on the A ring as a key driver of HDAC4 selectivity, exemplified by compound 15e, whereas meta-substitution on the B ring was critical for HDAC7 selectivity, as demonstrated by compound 15i. The pyran C ring and amide linker were essential for maintaining the ligand geometry required for productive engagement of the catalytic channel. These compounds provide valuable chemical probes for dissecting the distinct transcriptional functions of HDAC4 and HDAC7 and establish a framework for developing selective therapeutics for Th17-driven inflammatory diseases.

Glossary

Abbreviations

HDAC

histone deacetylase

SAR

structure–activity relationship

TFMO

trifluoromethyl-oxadiazole

TFPO

trifluoromethylpyrazole

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.6c00328.

• Experimental methods, chemical synthesis, and details of compound analysis (PDF)

M.-M.Z., B.K.G., and K.L.C. conceived and designed the study. B.K.G., N.Z., A.J., J.M., A.L., J.S., R.S., and K.L.C. conducted experiments, discussed and interpreted the data with B.K.G., K.L.C. and M.-M.Z. B.K.G., K.L.C. and M.-M.Z. wrote the manuscript with the input from all coauthors.

Safety Statement: No unexpected or unusually high safety hazards were encountered.

The authors declare no competing financial interest.