A trifluoromethyl quinazoline inhibits hepatocellular carcinoma proliferation through HDAC1 binding and multi-target mechanisms

Summary

Hepatocellular carcinoma (HCC) still has limitations such as tumor resistance, recurrence, and poor prognosis. Therefore, the development of anticancer drugs with mechanisms of action is of great significance for the treatment of HCC. In this study, we identified KZL-064, a 2-trifluoromethyl-4-aminoquinazoline derivative, which markedly inhibits HCC cell proliferation, migration, and invasion, induces apoptosis and G₂/M arrest, and, in MHCC97-H xenograft mice, oral administration achieves tumor suppression comparable to that of sorafenib without significant toxicity. Through virtual screening, molecular docking, cellular thermal shift assay (CETSA), and biologic layer interferometer (BLI), we identified histone deacetylase 1 (HDAC1) as a direct binding target. Knockdown of HDAC1 partially attenuates pro-apoptotic and cell-cycle effects, yet does not impair anti-migratory/invasive activities nor completely abolish overall efficacy. Collectively, KZL-064 emerges as a promising lead, and its anti-proliferative action likely involves multi-target mechanisms engaging HDAC1 and p53-associated apoptotic signaling, offering a conceptual framework for HCC therapy.

Graphical abstract

Highlights

• •KZL-064 engages HDAC1 and inhibits HCC proliferation
• •HDAC1 knockdown partly reduces KZL-064-induced apoptosis and arrest
• •HDAC1 knockdown does not affect migration/invasion, suggesting multi-target
• •KZL-064 suppresses HCC tumor growth in vivo with no overt toxicity

Medicine; Biological sciences; Molecular mechanism of gene regulation; Cancer

Introduction

Hepatocellular carcinoma (HCC) remains one of the most lethal and rapidly progressing malignancies globally.¹ The liver’s highly vascularized microenvironment promotes aggressive tumor growth and facilitates systemic metastasis.² However, the clinical efficacy of available therapies is often limited by the complex etiology, high tumor heterogeneity, and frequent development of drug resistance associated with HCC.^2,3 Consequently, there is a pressing need to develop anticancer agents with innovative chemical scaffolds and distinct mechanisms of action to more effectively suppress HCC proliferation, metastasis, and resistance pathways.

The quinazoline scaffold is recognized as a privileged structure in medicinal chemistry, exhibiting diverse pharmacological activities and broad applications in biomedical and antitumor drug discovery. Notably, it serves as the core framework for several clinically approved tyrosine kinase inhibitors, including vandetanib—a multi-target kinase inhibitor used in the treatment of advanced thyroid cancer.^4,5 Our previous study⁶ reported the design and synthesis of a series of 2-trifluoromethyl-4-aminoquinazoline derivatives, among which several compounds demonstrated promising antitumor activity. Given the established role of histone deacetylases (HDACs) in cancer epigenetics, we hypothesized that our quinazoline derivatives might target HDACs and exert anti-HCC activity.

Histone acetylation, dynamically regulated by histone acetyltransferases (HATs) and deacetylases, modulates chromatin structure and transcription to control key cellular processes like proliferation and apoptosis.^7,8,9 HDAC1, a class I HDAC and core component of major transcriptional repressor complexes, is frequently dysregulated in cancers. Its involvement in promoting aggressive tumor phenotypes and therapy resistance provides a strong rationale for targeting HDAC1 in solid tumors, including HCC.

In this study, we demonstrated that KZL-064 could inhibit the growth of liver cancer cells both in vitro and in vivo and identified HDAC1 as its direct target. Although the targeted binding of KZL-064 to HDAC1 played a certain role in inhibiting HCC, the overall activity of this compound was not solely determined by HDAC1.

Results

Inhibitory effects of KZL-064 on HCC cell proliferation in vitro

KZL-064 (Figure 1A), a 2-trifluoromethyl-4-aminoquinazoline derivative that corresponds to compound A20 in our previously published China patent (CN114436975A)¹⁰ and to compound 9 in our recent article on trifluoromethyl quinazoline derivatives targeting Werner helicase,¹¹ was provided by collaborators and resynthesized for the present study. The batch used in all experiments was fully characterized, with a purity of >98% (determined by high-performance liquid chromatography [HPLC]), a yield of 86%, and complete ¹H-NMR, ¹³C-NMR, and high-resolution mass spectrometry (HRMS) data provided in the supplemental information (Scheme S1 and Figures S1–S3), and was evaluated for anti-HCC activity. The cytotoxic effects of KZL-064 on HCC cell lines and normal liver cell lines are shown in Figure S4. KZL-064 inhibited MHCC97-H and HUH7 cell proliferation in a time- and dose-dependent manner at low-nanomolar concentrations (Figures 1B and 1C). Treatment with 2.5–10 nM KZL-064 for 24 h induced distinct morphological changes in HCC cells, including retraction of pseudopodia and cell rounding (Figure 1D).

Figure 1.

KZL-064 inhibited the proliferation of HCC cell lines in a time- and dose-dependent manner

(A) Chemical structure of KZL-064.

(B) Viability of MHCC97-H cells treated with increasing concentrations of KZL-064 for the indicated times (n = 3).

(C) Viability of HUH7 cells treated with increasing concentrations of KZL-064 for the indicated times (n = 3).

(D) Representative images showing morphological changes in MHCC97-H and HUH7 cells after 24 h treatment with KZL-064 (2.5 or 10 nM) (n = 3).

(E) Colony formation assay of MHCC97-H and HUH7 cells treated with 2.0 nM KZL-064 for 24 h (n = 3).

(F) Migration of MHCC97-H and HUH7 cells following treatment with KZL-064 (1.0 or 2.5 nM) (n = 3).

(G) Invasion of MHCC97-H cells after treatment with KZL-064 (1.0 or 2.5 nM) (n = 3).

All scale bars represent 100 nM. Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Furthermore, KZL-064 significantly suppressed colony formation in MHCC97-H and HUH7 cells at 2 nM, without affecting the clonogenic growth of normal LX2 cells (Figure 1E). The compound also potently inhibited the migration and invasion capabilities of HCC cells at concentrations as low as 1–2 nM (Figures 1F and 1G).

KZL-064 inhibits tumor growth in a xenograft model

To evaluate its in vivo efficacy, we established a subcutaneous xenograft model using MHCC97-H cells in nude mice (Figures 2A and 2B). Oral administration of KZL-064 inhibited tumor growth in a dose-dependent manner, and at tested doses, both tumor volume (Figure 2C) and weight (Figure 2D) in the KZL-064-treated groups were comparable to those observed with sorafenib. Notably, KZL-064 treatment did not affect mouse body weight even at the highest dose tested (Figure 2E), and no significant differences in organ weight or histology were observed between treated and control groups (Figures 2F and S5). Representative H&E staining of tumor sections from each group is shown in Figure 2G, revealing decreased cellularity and increased necrotic areas in the KZL-064- and sorafenib-treated tumors compared with control. These effects were achieved at substantially lower doses than with sorafenib, highlighting the potent anti-HCC activity of KZL-064. Immunohistochemical scoring of subcutaneous tumor sections showed that KZL-064 did not affect HDAC1 expression but significantly inhibited Ki-67 and promoted p53 expression (Figures S6A–S6D).

Figure 2.

KZL-064 inhibited tumor growth in a subcutaneous HCC xenograft model in a dose-dependent manner

(A) Schematic of the experimental design for the in vivo efficacy study.

(B) Tumor tissues were collected after sacrifice (n = 5).

(C) Tumor volume curves of mice treated with control (DMSO), KZL-064 (0.5, 1.0, and 2.0 μg/kg), and sorafenib (20 mg/kg) administered every 2 days (n = 5).

(D) Tumor weight graph of mice treated with control (DMSO), KZL-064 (0.5, 1.0, and 2.0 μg/kg), and sorafenib (20 mg/kg) administered every 2 days (n = 5).

(E) Body weight curves of mice treated with control (DMSO), KZL-064 (0.5, 1.0, and 2.0 μg/kg), and sorafenib (20 mg/kg) administered every 2 days (n = 5).

(F) Weights of major organs (heart, liver, spleen, lung, and kidney) excised from mice treated with control (DMSO), KZL-064 (0.5, 1.0, and 2.0 μg/kg), and sorafenib (20 mg/kg) after sacrifice (n = 5).

(G) H&E staining of subcutaneous tumor sections from mice treated with control (DMSO), KZL-064 (0.5, 1.0, and 2.0 μg/kg), and sorafenib (20 mg/kg) (n = 5).

All scale bars, represent 20 μm. Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

KZL-064 binds to HDAC1 as a direct molecular target

To identify the molecular targets mediating the anti-proliferative effects of KZL-064 in HCC, we initially predicted 66 candidate proteins that might directly bind to KZL-064. By integrating this list with genes associated with liver hepatocellular carcinoma (LIHC) prognosis in the Cancer Genome Atlas (TCGA) database, we prioritized 8 potential target proteins (Figure 3A). Molecular docking was then performed to assess the binding activity and potential interaction sites between KZL-064 and these candidates. HDAC1 exhibited the lowest calculated binding energy, with three potential hydrogen-bonding sites identified as optimal for interaction (Figure 3B).

Figure 3.

KZL-064 directly engages HDAC1 as a binding target

(A) Flowchart for calculating the target protein KZL-064 in liver cancer prediction.

(B) Molecular docking model of KZL-064 binding to the catalytic pocket of HDAC1.

(C) CETSA shows the thermal stability of HDAC1 after treatment with DMSO and different concentrations of KZL-064 in MHCC97-H cells (n = 3).

(D) CETSA shows the thermal stability of HDAC1, HDAC2, and HDAC8 in MHCC97-H cells treated with DMSO and KZL-064 at different temperatures (n = 3).

(E) BLI measures the binding affinity of KZL-064 to recombinant HDAC1 protein in vitro.

Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

The results of cellular thermal shift assay (CETSA) indicated that under conditions of increasing concentration (Figure 3C) and temperature (Figure 3D), KZL-064 maintained the stability of HDAC1 in liver cancer cells. Furthermore, biolayer interferometry (BLI) using purified HDAC1 protein confirmed a strong, dose-dependent binding affinity between KZL-064 and HDAC1 monomers (Figure 3E). These results collectively establish HDAC1 as a direct molecular target of KZL-064.

Given the structural and functional homology among HDAC family members, we also tested whether KZL-064 interacts with other class I HDACs, namely HDAC2 and HDAC8. The results of CETSA showed that under different temperature conditions, there was no significant stabilization observed for HDAC2 or HDAC8 (Figure 3D).

HDAC1 expression in HCC associated with poor prognosis

Given that HDAC1 has been identified as the target of KZL-064, we further investigated its expression levels in liver cancer tissues and cell lines. Compared with adjacent non-tumor tissues, the expression of HDAC1 in HCC tissues was significantly increased (Figures 4A–4C). Western blot analysis further confirmed that the protein level of HDAC1 was higher in multiple HCC cell lines than in the normal liver cell line LX2 (Figure 4D). These findings were consistent with the pan-cancer genome data from TCGA (Figures S7A–S7E) and were also confirmed in the protein analysis of our clinical validation cohort (Figures 4E and S7F). Clinical information is summarized in Table S1. The aforementioned results indicate that high HDAC1 expression is an independent risk factor associated with the shortened overall survival of HCC patients.

Figure 4.

HDAC1 was upregulated in HCC and associated with poor prognosis

(A–C) HDAC1 mRNA expression is elevated in HCC compared with adjacent normal tissues in a published cohort (n = 165 pairs) (A), Guangzhou cohort (n = 10 pairs) (B), and the TCGA-LIHC cohort (n = 150 pairs) (C).

(D) HDAC1 protein expression in SK-HEP1, HepG2, PLC5, HUH-7, MHCC97-H, and LX2 cells (n = 3).

(E) HDAC1 protein expression in HCC versus adjacent normal tissues in our clinical validation cohort (normal, n = 23; tumor, n = 96).

All scale bars represent 40 μm. Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

HDAC1 promotes HCC development

Given that bioinformatics and clinical validation cohort data indicate that HDAC1 promotes the progression of HCC, we established stable HDAC1 knockdown (short hairpin RNA-histone deacetylase 1, shHDAC1) and HDAC1 overexpression (HDAC1) MHCC97-H cell lines (Figure 5A) to explore its functional role and evaluated their phenotypes in vitro and in vivo.

Figure 5.

HDAC1 overexpression promoted HCC progression in vitro and in vivo

(A) Western blot confirming the establishment of stable HDAC1 knockdown and HDAC1-overexpressing MHCC97-H cell lines (n = 3).

(B) Viability of MHCC97-H cells after HDAC1 overexpression (n = 3).

(C) Viability of MHCC97-H cells after HDAC1 knockdown (n = 3).

(D) Schematic of xenograft animal experiments using overexpressed and knockdown HDAC1 MHCC97-H cells.

(E) Tumor tissues were collected after sacrifice (n = 5).

(F and I) Tumor growth curves of mice carrying HDAC1 overexpression (n = 5) (F) or HDAC1 knockdown (n = 5) (I).

(G and J) Tumor weight graph of the HDAC1 overexpression (G) and HDAC1-knockdown (J) groups.

(H and K) Representative H&E staining sections of tumors of the HDAC1 overexpression (H) and HDAC1-knockdown (K) groups.

All scale bars represent 50 μm. Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

HDAC1 overexpression significantly enhanced cell proliferation (Figure 5B), whereas HDAC1 knockdown led only to a brief and modest delay in cell growth (Figure 5C). In vivo, HDAC1 overexpression markedly accelerated tumor growth (Figures 5E-5H), while its depletion did not significantly alter the tumor growth rate (Figures 5E, 5I, and 5K). Regardless of HDAC1 overexpression or knockdown, KZL-064 treatment did not significantly affect mouse body weight (Figures S8A and S8C) or organ weight (Figures S8B and S8D).

Immunohistochemical scoring of subcutaneous tumor sections showed that, relative to the vector group, HDAC1 overexpression upregulated HDAC1 and Ki-67 protein levels without significantly affecting p53 expression (Figures S9A–S9D). In contrast, HDAC1 knockdown reduced HDAC1 and Ki-67 expression and concurrently elevated p53 expression (Figures S9A and S9E–S9G).

Notably, HDAC1 knockdown only slightly attenuated the anti-proliferative and pro-apoptotic effects of KZL-064, and the compound retained considerable activity in HDAC1-deficient cells. Collectively, these findings suggest that although HDAC1 contributes to the anti-cancer efficacy of KZL-064, the compound likely acts through multiple target mechanisms, with HDAC1 representing one of its key targets.

Contribution of HDAC1 to the anti-proliferative and pro-apoptotic effects of KZL-064

To investigate whether HDAC1 mediates the anti-HCC effects of KZL-064, we examined the phenotypic consequences of HDAC1 modulation. At identical concentrations and time points, HDAC1-overexpressing cells exhibited greater suppression of proliferation upon KZL-064 treatment, whereas HDAC1-knockdown cells showed only slightly reduced sensitivity (Figures 6A–6D). Notably, HDAC1-overexpressing cells displayed a higher proportion of apoptotic cells upon KZL-064 treatment, whereas HDAC1 knockdown partially attenuated this effect (Figures 6E and 6F). In contrast, KZL-064 equally suppressed the migration and invasion of shCtrl and shHDAC1 cells (Figures 6G and 6H), indicating that these anti-metastatic effects are largely independent of HDAC1. Collectively, these findings suggest that while HDAC1 contributes to the pro-apoptotic and anti-proliferative effects of KZL-064, it is not indispensable for the compound’s overall anti-cancer activity. KZL-064 induced a pronounced G₂/M phase arrest in MHCC97-H cells; notably, this arrest was significantly attenuated in shHDAC1 cells compared with shCtrl cells, indicating that HDAC1 depletion partially rescues KZL-064-induced cell cycle blockade (Figure 6I).

Figure 6.

HDAC1 served as an important contributor to the tumor-suppressive effects of KZL-064 in HCC

(A and B) Viability of HDAC1 (A) and shHDAC1 (B) MHCC97-H cells treated with KZL-064.

(C and D) Viability of HDAC1 (C) and shHDAC1 (D) MHCC97-H cells.

(E and F) Flow cytometry analysis of apoptosis (annexin V/PI staining) in HDAC1-overexpressing (n = 3) (E) and HDAC1-knockdown (n = 3) (F) cells treated with 2.5 or 10 nM KZL-064.

(G and H) Transwell assays showing that KZL-064 (1–2 nM) similarly inhibited migration (n = 3) (G) and invasion (n = 3) (H) in control and HDAC1-knockdown cells.

(I) Cell-cycle analysis by PI staining revealed that HDAC1 knockdown attenuated KZL-064-induced G2/M arrest in MHCC97-H cells (n = 3).

All scale bars represent 100 nM. Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Effect of KZL-064 on HDAC1-associated protein complex expression

HDAC1 typically exerts its deacetylase function within multiprotein complexes. To determine whether KZL-064 impacts complex activity, we analyzed HDAC1-interacting proteins via Search Tool for the Retrieval of Interacting Genes (STRING) database-based pathway enrichment and assessed the expression of complex components in KZL-064-treated MHCC97-H cells (Figure 7A). KZL-064 treatment did not appreciably alter the levels of MBD2, YY1, MTA2, or the acetylation of histones H3 and H4 (H3ac and H4ac) (Figure 7C), suggesting that KZL-064 binding leaves the deacetylase complex functionally intact.

Figure 7.

KZL-064 binding to HDAC1 does not alter associated complex expression

(A) Functional pathway enrichment analysis for HDAC1-interacting proteins based on the STRING database.

(B) Immunofluorescence staining of HDAC1 (red) and p53 (green) with DAPI (blue) (n = 3).

(C) Western blot analysis of NuRD complex components and histone acetylation markers in MHCC97-H cells treated with KZL-064 (n = 3).

Data are represented as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Given that HDAC1 suppresses p53 activity, we evaluated the effects of KZL-064 treatment on the expression and localization of both proteins. Compared with their respective controls, HDAC1 expression was significantly downregulated in both DMSO- and KZL-064-treated HDAC1-knockdown cells, whereas p53 levels were markedly elevated only in HDAC1-knockdown cells following KZL-064 treatment (Figure 7B). These findings suggest that KZL-064-mediated p53 regulation may depend on HDAC1 expression status and that the presence of HDAC1 may partially constrain the activation of p53 by KZL-064.

Discussion

Patients with HCC typically face a poor prognosis, with a median survival of only 3–6 months after diagnosis and a 5-year survival rate of approximately 10.1% in advanced cases.^2,3 Surgical resection is often not feasible for advanced HCC¹²; instead, primary treatments include local ablation, interventional therapy, liver transplantation, and systemic drug therapy—often involving anti-angiogenic agents and PD-1 inhibitors, either alone or in combination.^13,14

We identified KZL-064, a quinazoline-based compound that directly binds to HDAC1. In vitro and in vivo results support its potential as an anti-HCC agent and underscore HDAC1 as a relevant therapeutic target for HCC.

Histone acetylation is dynamically regulated by two enzyme families: HATs and HDACs, which competitively modify lysine residues on histone tails, thereby influencing chromatin structure, gene transcription, and DNA repair.^7,9 Aberrant HDAC activity is implicated in oncogenesis,⁸ making HDACs attractive therapeutic targets. HDACs typically function within multi-protein transcriptional co-repressor complexes that suppress target gene expression.¹⁵ These complexes are involved in key cellular processes such as cell cycle progression, differentiation, apoptosis, and tumor proliferation.^16,17,18,19 While some HDAC inhibitors have shown efficacy in hematological malignancies,^20,21,22 their utility in solid tumors, including HCC, warrants further investigation.

HDAC1, a class I HDAC, primarily regulates gene expression by removing acetyl groups from histones. It is frequently overexpressed in cancers and contributes to malignant phenotypes such as sustained self-renewal, as reported in glioblastoma and low-grade glioma.²³ In line with this, our data show that HDAC1 is upregulated in HCC tissues and cell lines and serves as an independent prognostic risk factor for shorter overall survival. These findings position HDAC1 as a potential therapeutic target in HCC.

HDAC1 often operates within large repression complexes such as Sin3, NuRD (via ZEB2), and CoREST.^24,25,26 Beyond transcriptional regulation, HDAC1 modulates endothelial functions—including vasculogenesis, inflammatory signaling, redox homeostasis, and nitric oxide production—thereby linking it to atherosclerosis, aging, neurodegeneration, and other pathophysiological processes.^27,28,29 Recent studies also implicate HDAC1 in G2/M-phase progression, DNA repair through interaction with hTERT, and promotion of metastasis—for example, by repressing DUSP27, leading to ERK/MMP2 activation in nasopharyngeal carcinoma.³⁰ Our bioinformatic and proteomic analyses further confirm HDAC1 overexpression in HCC and its association with poor clinical outcomes.

To elucidate the mechanism of KZL-064, we performed a series of binding and functional studies. Molecular docking coupled with experimental validation identified HDAC1 as a direct target of KZL-064. The compound stabilized HDAC1 in CETSAs and exhibited strong binding affinity in BLI experiments. Intriguingly, despite direct binding to HDAC1, KZL-064 did not detectably alter the expression or composition of HDAC1-associated multiprotein complexes, nor did it affect total p53 protein levels. This apparent paradox may be explained by the possibility that KZL-064 interacts with a discrete region on HDAC1 that is critical for its enzymatic activity but dispensable for the structural integrity of its corepressor assemblies. Alternatively, KZL-064 binding might induce a subtle conformational shift that impairs substrate recognition without dissociating the complex. Consequently, p53 activation is associated with enhanced acetylation-dependent nuclear retention and transcriptional competence, rather than changes in p53 protein abundance or complex remodeling. These observations suggest that KZL-064 achieves functional HDAC1 engagement while preserving the scaffolding architecture of the deacetylase complex, a feature that may contribute to its favorable safety profile and distinct multi-target pharmacology.

Beyond conventional approaches, explainable AI (XAI) models—such as SHapley Additive exPlanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME), and the graph-based xSiGra framework^31,32—offer opportunities to dissect multi-target mechanisms. Applying these methods to transcriptomic or proteomic data from KZL-064-treated HCC models could help distinguish HDAC1-dependent from HDAC1-independent effects.

Importantly, several lines of evidence indicate that HDAC1 is not the sole or dominant target responsible for the anti-HCC activity of KZL-064, and direct enzymatic inhibition of HDAC1 by KZL-064 has not been demonstrated in a purified system. First, HDAC1 knockdown produced only modest phenotypic effects and did not abolish the anticancer activity of KZL-064. Second, the compound’s anti-migratory and anti-invasive effects were largely HDAC1 independent. Third, direct biochemical evidence demonstrating that KZL-064 inhibits HDAC1 enzymatic activity in a purified system is currently lacking, and selectivity profiling among HDAC family members remains incomplete. While CETSA suggested preferential engagement of HDAC1 over HDAC2 and HDAC8, direct enzymatic inhibition assays against a broader panel of HDAC isoforms will be required to fully define the potency and selectivity of KZL-064. Such studies, combined with structural analysis and unbiased target deconvolution approaches, are warranted in future investigations. Collectively, these observations suggest that KZL-064 operates through a multi-target mechanism, with HDAC1 representing one of several contributing targets.

Several limitations should be considered. First, direct enzymatic inhibition assays against HDAC1 and related isoforms are lacking; CETSA and BLI demonstrate binding but not inhibitory potency or selectivity, and histone acetylation changes could reflect multi-target epigenetic disruption. Second, the incomplete loss of activity upon HDAC1 knockdown suggests that additional targets contribute to the anti-HCC effects of KZL-064. Third, pharmacokinetic data to support the low-dose efficacy are not yet available. Future studies incorporating enzymatic HDAC inhibition assays, unbiased target deconvolution (e.g., chemoproteomics), and comprehensive PK/PD analysis will be essential to fully characterize the mechanism and selectivity profile of KZL-064.

In conclusion, KZL-064 is a quinazoline-based compound that directly binds to HDAC1 and inhibits HCC. Our findings support the potential relevance of HDAC1 as a therapeutic target in HCC and demonstrate that KZL-064 binds to HDAC1 and induces cell-cycle arrest and apoptosis. However, the incomplete loss of activity upon HDAC1 knockdown and the HDAC1-independent anti-migratory effects suggest that KZL-064 acts through a multi-target mechanism. These results provide a rationale for further development of KZL-064 and highlight the importance of considering multi-target pharmacology in HDAC-directed therapies for HCC.

Limitations of the study

We acknowledge several limitations in the current study. Direct biochemical evidence that KZL-064 inhibits HDAC1 enzymatic activity in a purified system is lacking; definitive confirmation will require in vitro deacetylase activity assays and comprehensive selectivity profiling against other HDAC isoforms. Furthermore, while HDAC1 contributes to the anti-proliferative and pro-apoptotic effects of KZL-064, the modest impact of HDAC1 knockdown on overall drug sensitivity indicates that additional targets are involved. To this end, definitive proof of HDAC1’s functional contribution would require rescue experiments in which HDAC1 expression is restored in knockdown cells or the use of selective HDAC1 chemical inhibitors to phenocopy the effects of KZL-064. Such experiments, along with unbiased target deconvolution approaches such as chemical proteomics, will be essential in future investigations to fully establish the causal role of HDAC1 and elucidate the molecular mechanisms underlying the anti-HCC effects of KZL-064. Additionally, the present study does not include pharmacokinetic data (e.g., plasma and tumor exposure and bioavailability) to support the efficacy observed at the low oral doses of KZL-064. Future studies incorporating comprehensive pharmacokinetic/pharmacodynamic (PK/PD) analysis will therefore be necessary to fully characterize the in vivo pharmacological profile of KZL-064. Nevertheless, the robust efficacy observed at these low doses highlights the potent anti-HCC activity of KZL-064 and supports its continued investigation.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Lei Yu (YL@hrbmu.edu.cn).

Materials availability

The compound KZL-064 used in this study was provided by the Natural Products Research Center of Guizhou Province. It is available from the lead contact upon request, but a material transfer agreement may be required.

Data and code availability

The proteomics data presented in Figure 4A and the transcriptomics data shown in Figure 4C were derived from the TCGA public database, with accession codes provided in the key resources table. The mass spectrometry proteomics data presented in Figure 4B have been deposited in the ProteomeXchange Consortium via the iProX partner repository^33,34 under accession code PXD081037, which is also listed in the key resources table. No original code was generated in this study. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported by the project of State Key Laboratory of Functions and Applications of Medicinal Plants, Guizhou Medicinal University (no. FAMP202104K), Special Funded Research Project of the Fourth Hospital of Harbin Medical University (no. HYDSYTB202206), and the Guizhou Provincial Committee Organization Department (no. QKHPTRC-GCC [2022] 034-1). We would like to thank MogoEdit (https://www.mogoedit.com) for its English editing during the preparation of this manuscript.

Author contributions

S. Wang and X.C. performed the conceptualization, methodology, formal analysis, and original draft writing. G.Y. completed the compound synthesis and wrote the chemical part of this article. X. Wan and S. Wei participated in the animal experiments. T.A., X. Wang, S. Wang, and W.H. jointly completed the collection and processing of the clinical research section. J.Y. and S.C. completed the pharmacological experiment and wrote the pharmacological part of this article. D.Z. jointly completed the HDAC1 protein purification and bio-membrane interference. B.X., H.L., and L.Y. conducted the pharmacological experiment guidance and reviewed the article.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
HDAC1 Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#34589; RRID: AB_2756821
HDAC2 Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#57156; RRID: AB_2756828
HDAC8 Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#66042; RRID: AB_2895217
H3 Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#4499; RRID: AB_10544537
H3ac Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#9649; RRID: AB_823528
H4ac Rabbit Monoclonal Antibody	Cell Signaling Technology	Cat#2594; RRID: AB_2248400
MBD2 Rabbit Monoclonal Antibody	Abcam	Cat#ab188474
YY1 Rabbit Polyclonal Antibody	Proteintech	Cat#22156-1-AP; RRID: AB_11182932
MTA2 Rabbit Polyclonal Antibody	Proteintech	Cat#17554-1-AP; RRID: AB_10638601
GAPDH Rabbit Monoclonal Antibody	Huaan Biotechnology	Cat#ET1601-4
Biological samples
Human HCC and adjacent normal samples (Clinical information of HCC patients see Table S1)	4th Hospital of Harbin Medical University	N/A
Chemicals, peptides, and recombinant proteins
KZL-064	Natural Products Research Center of Guizhou Province	N/A
Sorafenib	GLPBIO	Cas#284461-73-0
D-luciferin	GLPBIO	Cas#103404-75-7
Critical commercial assays
SDS-PAGE gel preparation kit	Beyotime	P0012A
BCA protein assay kit	Beyotime	P0011
Apoptosis detection kit	Beyotime	C1055S
Cell cycle detection kit	Beyotime	C1052
Deposited data
TCGA-LIHC proteomics data (Figure 4A)	TCGA public database	https://www.cancer.gov/ccg/research/genome-sequencing/tcga
Collaborator-provided mass spectrometry proteomics data (Figure 4B)	iProX33,34	PXD081037
TCGA-LIHC proteomics data (Figure 4C)	TCGA public database	https://www.cancer.gov/ccg/research/genome-sequencing/tcga
Experimental models: Cell lines
Human: LX2 cells	ATCC	RRID: CVCL_5792
Human: HEL cells	ATCC	RRID: CVCL_0001
Human: 293T cells	ATCC	RRID: CVCL_0063 or CVCL_0045
Human: MHCC-97H cells	Procell Life Science & Technology Co., Ltd. (Wuhan, China)	RRID: CVCL_4972
Human: HEPG2 cells	Procell Life Science & Technology Co., Ltd. (Wuhan, China)	RRID: CVCL_0027
Human: HEPG2.215 cells	Procell Life Science & Technology Co., Ltd. (Wuhan, China)	RRID: CVCL_L855
Human: SK-HEP1 cells	Procell Life Science & Technology Co., Ltd. (Wuhan, China)	RRID: CVCL_0525
Human: HUH7 cells	Procell Life Science & Technology Co., Ltd. (Wuhan, China)	RRID: CVCL_0336
Experimental models: Organisms/strains
Mouse: BALB/c nude	Chongqing Tengxin Huafu Experimental Animal Sales Co., Ltd	RRID: IMSR_RJ: BALB-C-NUDE
Oligonucleotides
Plasmid sequences for shHDAC1-1 and HDAC1 see Figure S14 and Table S2	This paper	N/A
Software and algorithms
R studio 4.3.0	N/A	https://www.r-project.org/
SPSS	N/A	https://www.ibm.com/products/spss-statistics
GraphPad Prism 9.0	N/A	https://www.graphpad.com/
ImageJ	N/A	https://imagej.net/ij/
SwissTargetPrediction	N/A	https://www.swisstargetprediction.ch/
STRING database	N/A	https://cn.string-db.org/
PyMOL 2.5	N/A	https://pymol.org/
AutoDock Vina 1.2.0	N/A	https://github.com/ccsb-scripps/AutoDock-Vina
Discovery Studio Visualizer	N/A	https://discover.3ds.com/discovery-studio-visualizer-download
Other
Transwell Inserts (48-well format)	Corning (NY, USA)	CLS3460-48 EA

Experimental models and study participant details

Cell culture

The mammalian LX2, 293T, and HEL cell lines used in this study were obtained from the American Type Culture Collection (ATCC) and maintained at the Key Laboratory of Natural Products, Chinese Academy of Sciences (Guizhou Province, China). The MHCC97-H, HEPG2, HEPG2.2.15, SK-HEP1, and HUH7 cell lines were sourced from Procell Life Science & Technology Co., Ltd. (Wuhan, China). All cell lines were authenticated by short tandem repeat (STR) analysis and were confirmed to be free of mycoplasma contamination. All cells were cultured in DMEM or MEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin solution (containing 1 × 10⁵ U/L penicillin and 100 mg/L streptomycin), and incubated at 37°C under 5% CO₂ in a cell culture incubator (Thermo Fisher Scientific). The complete medium was refreshed every one to two days. Cell growth was monitored regularly under a microscope, and cells were passaged upon reaching 80–90% confluence. Cells in the logarithmic growth phase were used for subsequent experiments.

Ethics approval statement

All animal experiments in this study were approved by Guizhou Medical University (Approval No. SYXK2023-0002). The immunohistochemical experiments on HCC and adjacent normal tissues were approved by Fourth Hospital of Harbin Medical University (Approval No. 2023-37).

Method details

Cell proliferation

Cells were seeded in 96-well plates at a density of 3,500 cells/well and cultured for 24 h under standard conditions (37°C, 5% CO₂). Following treatment with KZL-064 at the indicated concentrations (within the 1–20 nM range) according to the experimental design, 20 μL of 0.5% MTT solution was added to each well at designated time points, followed by a 4 h incubation at 37°C. The supernatant was then carefully removed, and 150 μL of dimethyl sulfoxide (DMSO) was added to dissolve the formazan crystals. Absorbance at 490 nm was measured using a Bio-Tek microplate reader (VE, USA) to calculate the cell proliferation inhibition rate or to generate proliferation curves.

Cell migration and invasion

Cells were cultured in complete growth medium (DMEM) and incubated at 37°C under 5% CO₂ for 24 h. Subsequently, they were trypsinized, resuspended in serum-free medium at a density of 5 × 10⁵ cells/mL, and 100 μL of the cell suspension was seeded into the upper chamber of a Transwell insert pre-coated with Matrigel for the invasion assay. The lower chamber was filled with 600 μL of complete medium containing 10% FBS and KZL-064 at the indicated concentrations (1–10 nM), with 0.1% DMSO serving as the vehicle control. Cells were allowed to migrate or invade for 24 h at 37°C in 5% CO₂. After incubation, non-migrated/non-invaded cells on the upper surface of the membrane were gently removed with a cotton swab. Cells that had migrated or invaded to the lower surface were fixed with 4% paraformaldehyde and stained with crystal violet. The number of stained cells was quantified under a microscope from several randomly selected fields.

Colony formation

Cells were seeded in 6-well plates at a density of 1,000 cells per well and incubated at 37°C under 5% CO₂ for 24 h. Following this, cells were treated with KZL-064 at the indicated concentrations (2.0 nM) or other interventions according to the experimental design. Once colony formation was observed, the culture was terminated and the medium was removed. The cells were washed twice with phosphate-buffered saline (PBS), fixed with 1 mL of 4% paraformaldehyde per well for 30 min at room temperature, and then washed again twice with PBS. Subsequently, each well was stained with 1 mL of crystal violet solution for 15 min at room temperature. After staining, the cells were gently rinsed several times with PBS, air-dried, and imaged using a digital camera.

Cell cycle and apoptosis analysis

MHCC97-H-shHDAC1 and MHCC97-H-WT hepatocellular carcinoma cells were seeded in 6-well plates at a density of 3 × 10⁵ cells per well. After adherence, cells were treated with KZL-064 at the indicated concentrations (2.5–10 nM) or vehicle control (0.1% DMSO) for 24 h.

Cell cycle analysis was performed by flow cytometry. Briefly, treated cells were harvested, fixed in 70% ice-cold ethanol overnight at 4°C, washed twice with PBS, and centrifuged at 1000 rpm for 5 min at room temperature. Cell pellets were resuspended in 500 μL PBS and treated with 5 μL RNase A (10 μg/mL) at 37.5°C for 30 min. Subsequently, cells were stained with 25 μL propidium iodide (PI, 1 mg/mL) and kept on ice until analysis.

Apoptosis detection was carried out using an annexin V-FITC/PI double-staining kit according to the manufacturer’s protocol (Beyotime, Shanghai, China). After treatment, cells were collected, washed twice with PBS, and stained with annexin V-FITC and PI. Samples were analyzed using a flow cytometer (Miltenyi Biotec), and data were processed with FlowJo software.

Cell transfection

The shCtrl, shHDAC1, vector, and HDAC1 plasmids were constructed by Chongqing Leqin Biotechnology Co., Ltd. (Chongqing, China). Plasmid sequences and related details are provided in the Figure S14 and Table S2. For lentivirus production, plasmids were transfected into 293T cells using Neofact transfection reagent (ZeroClient Biotechnology Co., Ltd., Beijing, China). Lentiviruses were then harvested and used to infect hepatocellular carcinoma cells in the presence of polybrene, following the manufacturer’s protocol.

Western blot

Antibodies were obtained from the following commercial sources: HDAC1 (#34589S; 1:5,000), HDAC2 (#57156S; 1:2,000), HDAC8 (#66042S; 1:2,000), Histone H3 (#4499S; 1:10,000), Histone H4 (#13919S; 1:5,000), Acetyl-Histone H3 (Lys27) (#8173S; 1:5,000), and Acetyl-Histone H4 (#13534S; 1:2,000) were from Cell Signaling Technology (USA). MBD2 (ab188474; 1:2,000) was from Abcam (UK). Antibodies against MTA2, p53, YY1, HDAC1, and mouse secondary antibodies were from Proteintech (China). Antibodies against BCL2, BAX, GAPDH, and rabbit secondary antibodies were from Huaan Biotechnology (China). All antibodies were used according to the manufacturers’ recommended protocols.

HDAC1 protein purification

The full-length HDAC1 gene (GenBank: BC000301.2) was synthesized and cloned into the pET-28a expression vector using NcoI/XhoI restriction sites. The recombinant plasmid was then transformed into Escherichia coli BL21(DE3) for protein expression. Bacterial cells were harvested, resuspended in lysis buffer A (50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 2.5 mM DTT) supplemented with 10 mM imidazole and 1 mM PMSF, and lysed by sonication. After centrifugation, the supernatant was loaded onto a Ni-NTA affinity column. The column was washed sequentially with buffer A containing 30 mM and 60 mM imidazole, and the target protein was eluted with buffer A containing 300 mM imidazole. Protein purity was evaluated by SDS-PAGE.

Bio-membrane interference

Recombinant HDAC1 was biotinylated and purified to remove free biotin. A streptavidin-coated biosensor was pre-hydrated in assay buffer for 20 min, followed by loading with biotinylated HDAC1. Binding kinetics were measured on an Octet96 instrument under standard conditions: 60 s baseline equilibration, 60 s association, and 60 s dissociation. The dissociation constant (Kd) was calculated by fitting the real-time binding curves using the instrument’s analysis software.

Cellular thermal shift assay (CETSA)

MHCC97-H cells in logarithmic growth phase were seeded in 6-well plates at a density of 3 × 10⁵ cells/mL and treated with 1–10 nM KZL-064 for 1 h. Cells were then harvested and subjected to a CETSA. Briefly, cell pellets were heated at temperatures ranging from 53°C to 63°C for 3 min, followed by rapid cooling in liquid nitrogen; this cycle was repeated three times to promote drug-target interaction. Total protein was subsequently extracted, and HDAC1 levels were analyzed by western blotting.

Drug target prediction and bioinformatics analysis of liver cancer prognosis

Molecular structures and Simplified Molecular-Input Line-Entry System (SMILES) notations of the compounds were generated using the SwissTargetPrediction tool.³⁵ Target prediction was performed based on 2D/3D structural similarity to known bioactive compounds in the ChEMBL database.^36,37 Key physicochemical properties—including predicted oral bioavailability, logP, and numbers of hydrogen-bond donors and acceptors—were also evaluated.

RNA-seq data and corresponding clinical information for patients with liver cancer were downloaded from TCGA database.³⁸ Kaplan-Meier survival curves were plotted, and p-values, hazard ratios (HR) with 95% confidence intervals (CI) were calculated using R software to compare survival between groups stratified by candidate gene expression. Statistical significance was defined as p < 0.05.

Molecular docking

Protein structures were retrieved from the Protein DataBank (PDB). Prior to docking, protein structures were prepared using PyMOL 2.5. Molecular docking was then performed with AutoDock Vina 1.2.0. The resulting docking poses were visualized and analyzed using PyMOL 2.5 and Discovery Studio Visualizer.

Animal experiments

Four-week-old male BALB/c nude mice were purchased from Chongqing Tengxin Huafu Experimental Animal Sales Co., Ltd. All animal experiments were conducted in accordance with the guidelines of the Chinese Association for Laboratory Animal Science and approved by the relevant institutional ethics committee.

MHCC97-H cells (1 × 10⁶) were subcutaneously injected into the dorsal flank of each mouse. After three days, mice were randomly assigned to five groups (n = 5): vehicle control (0.1% DMSO), sorafenib (20 mg/kg, positive control), and KZL-064 at low (0.5 μg/kg), medium (1.0 μg/kg), or high (2.0 μg/kg) doses. The oral doses of KZL-064 (0.5, 1.0, and 2.0 μg/kg) were selected based on in vitro potency data (IC₅₀ values of 2–10 nM against MHCC97-H cell) and a preliminary dose-finding study in tumor-bearing mice, which identified this range as effective and well tolerated. The dose of sorafenib (20 mg/kg) was chosen based on published literature establishing this as an effective and well-tolerated dose in murine HCC xenograft models.^39,40 Treatments were administered orally every two days for 14 days. Tumor volume was measured every two days using a caliper and calculated as V = (L × W²)/2, where L is the longest diameter and W is the perpendicular shorter diameter. Body weight was recorded on the same schedule. At the endpoint, mice were euthanized, and xenograft tumors were excised and weighed.

For studies involving HDAC1 knockdown or overexpression, mice were randomly divided into four groups (n = 5) after subcutaneous inoculation of genetically modified MHCC97-H cells: shHDAC1, shCtrl, HDAC1, and vector. Tumor volume and body weight were monitored every two days for 14 days, after which mice were euthanized and tumors were harvested and weighed.

Hematoxylin and eosin (HE) staining

Tumor tissues were fixed in 4% paraformaldehyde, dehydrated through a graded ethanol series (70%, 80%, 90%, and 100%), cleared in xylene, embedded in paraffin, and sectioned at 5 μm thickness. Sections were then stained with hematoxylin for 3–5 min, rinsed under running water, counterstained with eosin, dehydrated through graded ethanol (85% and 95%), cleared in xylene, and mounted with neutral resin. Stained sections were examined under a light microscope, and representative images were acquired for analysis.

Immunohistochemistry

Tumor samples from human subjects (including tumor and adjacent normal tissues) and from mouse xenografts were paraffin-embedded and sectioned. Following antigen retrieval, endogenous peroxidase activity was blocked. Sections were then incubated with the indicated primary antibodies, followed by appropriate secondary antibodies. Immunoreactivity was visualized using a DAB detection kit (Servicebio, G1212). Stained sections were examined and imaged under a light microscope (Nikon, Japan).

Quantification and statistical analysis

Statistical analysis

Statistical analysis was performed using SPSS (version 26.0) and GraphPad Prism (version 9.0). Data from at least three independent experiments are presented as mean ± standard deviation (SD). Differences between two groups were assessed using the unpaired two-tailed Student’s t test, whereas comparisons among multiple groups were performed by one-way analysis of variance (ANOVA) followed by appropriate post-hoc tests. The correlation between HDAC1 gene copy number and its mRNA expression in clinical samples was evaluated using Pearson’s correlation coefficient. Kaplan-Meier survival curves were plotted based on HDAC1 expression levels, and differences in overall survival were compared using the log rank test. p < 0.05 was considered statistically significant.