Patients who were diagnosed with R/R T-ALL and T-cell lymphoblastic lymphoma (T-LBL) and who received chidamide during salvage chemotherapy were recruited between May 2016 and June 2019 at the First Medical Center of the Chinese PLA General Hospital. This retrospective study was designed to evaluate the clinical efficacy and safety of combining chidamide with chemotherapy for patients with R/R T-ALL. The following inclusion criteria were used: ① patients whose diagnosis met the criteria of T-ALL or T-LBL according to the 2016 edition of the WHO classification of precursor lymphoid neoplasms; ② patients with primary refractory disease after induction therapy; ③ patients who relapsed within 12 months after first complete remission; ④ patients who were refractory to or relapsed after at least one salvage treatment; and ⑤ patients with an Eastern Cooperative Oncology Group (ECOG) score ≤ 2. Patients were excluded in this study if they had severe organ dysfunction (heart, kidney, liver, or other organs). The criteria of ETP were defined as previously (28). Our study was reviewed and approved by the Ethics Committee of the Chinese PLA General Hospital. All patients signed consent forms for the collection and use of their cells and clinical demographic data.

Chidamide (30mg per dose) was orally administered twice a week for a total of four doses (120mg in total per patient) during the first 2 weeks of the combined salvage chemotherapy. The chemotherapy regimen was selected according to the patient’s pretreatment history and physical status (see Supplementary Table S1). The choice of salvage chemotherapy was determined according to the patient’s physical status, previous treatment history and response, and the discretion of the oncologist in charge. Response assessments, including the evaluation of bone marrow, peripheral blood, and extramedullary lesions, were usually performed on the 28th day of each course of treatment or until hematopoietic recovery. Complete remission (CR) was defined as the percentage of bone marrow blasts dropped to be less than 5% and the resolution of extramedullary masses. A partial response (PR) was defined as the percentage of bone marrow blasts dropped to 6–25% and (or) more than 50% reduction of the tumor mass. Progressive disease (PD) was defined as the number of blast cells in the bone marrow or peripheral blood increased by more than 25%, more than 25% increase of the tumor mass, or the onset of a new tumor mass. No response (NR) was defined as the failure to meet the criteria of a PR or PD. The overall response rate (ORR) was defined as the CR plus PR. Adverse effects were evaluated according to the National Cancer Institute Common Toxicity Criteria.

Chidamide was obtained at no cost from Chipscreen Biosciences Ltd. (Shenzhen, China). RPMI 1640 medium, fetal bovine serum, and dimethyl sulfoxide (DMSO) were all purchased from Gibco Company (USA). All antibodies for Western blotting were purchased from Cell Signaling Technology (Danvers, MA, USA), except for TYK2 and p-STAT1, which were from Abcam (Cambridge, UK). T-ALL cell lines, including Jurkat and MOLT-4, were obtained from ATCC (Manassas, VA, USA). Primary T-ALL patient cells were isolated from bone marrow. Bone marrow samples were first diluted with an equal volume of PBS solution. The diluted mixture was spread over lymphocyte separation fluid (Lymphoprep 07801, Stemcell, USA) in a centrifuge tube and centrifuged at 2800 rpm for 25 minutes at 25 °C. Then, the mononuclear cells between the plasma layer and separation fluid were collected. Jurkat, MOLT-4, and primary cells were maintained in RPMI 1640 supplemented with 10% fetal bovine serum, 100 μg/mL penicillin, and 10 μg/mL streptomycin at 37 °C in a humidified atmosphere containing 5% CO₂.

Jurkat and MOLT-4 cells (3×10⁴ cells) were seeded in 96-well plates (each well contained 100 μL of cell suspension). Cells were incubated with different concentrations of chidamide for 24, 48, or 72 h. Each concentration was tested in 4 biological replicates (independent cell seeding and culture batches), with 3 technical replicates per well. Then, 10 μL of CCK-8 reagent (Dojindo Laboratories, Kumamoto, Japan) was added to each well and incubated for an additional 2–3 h. The OD values of each well were measured by a microplate reader at 450 nm, and data are presented as mean ± SD of biological replicates.

The chidamide-treated cells were washed twice with cold PBS and fixed with cold 70% ethanol overnight. After fixation, the cells were washed with cold PBS and stained in PBS supplemented with 100 µg/mL RNase A and 100 µg/mL PI (Sigma) for 15 min. Experiments included 3 biological replicates (independent cell cultures) per treatment group, with 3 technical replicates per sample. Samples were processed on a flow cytometer with a 488-nm laser and then analyzed with a MoFlo MLS sorter (Dako, Fort Collins, CO), and cell cycle distribution data are presented as mean ± SD of biological replicates.

Cells were collected after being treated with chidamide for the indicated time. They were washed with ice-cold PBS and incubated in 200 μL of binding buffer containing 10 μL of PI and 5 μL annexin-V-FITC (Sigma) for 15 min at room temperature in the dark. Experiments were performed with 3 biological replicates (independent cell cultures), and each sample was run in 3 technical replicates. Then, the stained cells were analyzed on a flow cytometer, and data (mean ± SD of annexin V-positive cells) were analyzed by FlowJo version 10.

Jurkat and MOLT-4 cells were incubated with DMSO or chidamide (1 μmol/L) at 37 °C for 48 h and then were collected. HDAC activity was measured by a Colorimetric HDAC Activity Assay kit (BioVision, San Francisco, USA) with 3 biological replicates (independent cell treatments). Each reaction (100-μL system: 50 μg protein, 10 μL assay buffer, 5 μL colorimetric substrate) was run in 3 technical replicates. The reaction mixture was incubated at 37 °C for 1 h and then was stopped by adding 10 μL of lysine developer. The reaction system was then incubated at 37 °C for another 30 min. HDAC activity was measured on a microplate reader (SpectramMax M5) at 405 nm, and data are presented as mean ± SD of biological replicates.

Chidamide-treated (1 μmol/L for 48 h) and DMSO-treated Jurkat and MOLT-4 cells were collected from 3 biological replicates (independent treatments) per cell line and washed twice with ice-cold PBS. Total RNA was extracted using TRIzol reagent (Invitrogen, USA) following the manufacturer’s instructions. RNA sequencing was performed on a Genome Analyzer IIx (Illumina, San Diego, CA, USA). Expression levels were calculated with the RPKM method. Differential gene expression was analyzed across biological replicates, with padj < 0.05 considered statistically significant. The distribution of gene expression was analyzed.

Total cellular RNA was extracted with TRIzol according to the manufacturer’s instructions. cDNA was synthesized using a PrimeScript™ RT reagent kit (Takara). Real-time PCR (RT-PCR) assays were performed using a KAPA SYBR FAST q-PCR Master Mix (2x) kit. Primers for target genes are listed in Supplementary Table S2. For cell lines, 3 biological replicates (independent treatments) were analyzed; for primary samples, 29 T-ALL patients and 5 healthy controls served as biological replicates. Each sample was analyzed in 3 technical replicates, and relative quantification of target genes was performed using 2^-ΔΔCt. Each sample was analyzed in triplicate. Data are presented as mean ± SD of biological replicates.

HDAC3 and TYK2 plasmids and their shRNA constructs were cloned into the lentiviral vector pLKO.1-puro. HDAC3 and TYK2 plasmids and their corresponding control plasmid were purchased from Genechem. The HDAC3 shRNA plasmid (sc-35,538-SH) and TYK2 shRNA plasmid (sc-27,545-SH) were from Santa Cruz. Jurkat and MOLT4 cells were transfected with lentivirus-containing medium according to the manufacturer’s protocol. Then, cells were selected using antibiotics. They were then expanded and subjected to RT-PCR and Western blot.

After cells were collected and washed twice with ice-cold PBS, they were lysed in RIPA buffer that contained phosphatase inhibitors and protease inhibitors. Protein lysates were prepared from 3 biological replicates (independent cell cultures or patient samples). Equal amounts of proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). The PVDF membranes were blocked with 5% non-fat dry milk for 1 h at room temperature and then were incubated overnight at 4 °C with primary antibodies. Each membrane was incubated with a secondary antibody according to the manufacturer’s instructions. Then, target proteins were detected with enhanced chemiluminescence kits (Millipore, Billerica, MA, USA). Blots were repeated twice to confirm reproducibility, and representative images are shown. For co-immunoprecipitation, about 0.5–1 mg of protein lysate from 3 biological replicates was immunoprecipitated by TYK2 and HDAC3 antibodies using co-immunoprecipitation reagents (ThermoFisher, #88804) (reciprocal IP strategy). The immune complex was subjected to immunoblotting with antibodies against HDAC3, TYK2, and p-TYK2.

Xenograft experiments were established in NOD/SCID immunodeficient mice as previously described (29). Athymic mice were maintained in laminar flow air conditioning cabinets and fed with a standard diet of laboratory rodent food and water under specific pathogen-free conditions. Jurkat cells (1×10⁷ cells in 200 μL of PBS) were injected subcutaneously into the lateral flanks of mice. Tumor volumes and weights were measured every other day by 2 independent observers (technical replicates). When tumors approached 50 mm³, mice were randomly assigned to 4 groups (n=6 mice per group, biological replicates): control (PBS + normal saline), chidamide (5 mg/kg, intragastrically, 3 times/week for 2 weeks), deucravacitinib (a highly selective inhibitor of tyrosine kinase 2, MedChemExpress, China, BMS-986165) (30 mg/kg, intragastrically, twice daily for 2 weeks), or combination therapy. The mice were sacrificed on the 21st day after inoculation via intraperitoneal injection of pentobarbital sodium at a dose of 100 mg/kg (injection volume: 0.1 mL/10g body weight), which induces rapid and painless euthanasia. Tumors were measured and collected for further analysis. All protocols were approved by the Experimental Animal Ethics Committee of the Chinese PLA General Hospital. All animal experiments were performed in accordance with the international and institutional rules of the Institutional Animal Care and Use Committee (IACUC). Data are presented as mean ± SD of tumor volume/weight per group.

Data were expressed as mean ± standard deviation (SD). Patient survival analysis was calculated using the Kaplan-Meier method with the log-rank (Mantel Cox) test. Statistical analysis was performed using SPSS 25.0 software and GraphPad Prism 9.0. The Mann–Whitney test or Student’s t test was used to perform comparisons when necessary. Differences were considered statistically significant when a two-tailed P-value was < 0.05.

To evaluate the clinical efficacy of the combination therapy of chidamide and chemotherapy, 28 patients with R/R T-ALL received salvage chemotherapy and chidamide between May 2016 and June 2019 at the Chinese PLA General Hospital. The clinical characteristics of these patients are shown in Table 1. Nineteen patients were male and nine were female. The median age of these patients was 26 (range, 13–63) years. Six patients were diagnosed with T-LBL, and the remaining 16 patients were diagnosed with T-ALL. Ten patients met the criteria of the ETP immunophenotype. Before salvage treatment, 12 (43%) patients had relapsed, and 16 (57%) patients had refractory disease. The previous treatments of these 28 patients are listed in Supplementary Table S2.

During the study period, 28 patients received at least one course of salvage therapy. The salvage reinduction regimen of each patient and their response are shown in Supplementary Table S3. Sixteen patients achieved CR and three patients achieved PR. The overall response and CR rate were 68% and 57%, respectively. The CR rate for patients with ETP-ALL was 70% (7/10). Twenty-one patients underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT). Sixteen patients were in CR before allo-HSCT. Five patients with progressive disease received allo-HSCT as salvage therapy. Median follow-up for survivors was 56.5 (32–70) months. The OS rates at 1 and 3 years were 56.3% and 37.5%, respectively (Figure 1A). The progression-free survival rates at 1 and 3 years were 42% and 27%, respectively (Figure 1B). The occurrence of adverse events is listed in Supplementary Table S4. All patients developed Grade III/IV bone marrow suppression. Febrile neutropenia was the most frequent event, which occurred in 14 cases (50%). Severe septic shock occurred in two cases. Other adverse events (more than one case) included grade I drug-induced liver injury (five cases), oral mucositis (five cases), diarrhea (three cases), hypofibrinogen (three cases), and pneumonitis (three cases).

The CR rates of salvage chemotherapy in R/R T-ALL range from 30% to 50% (30). Therefore, improving the CR rate is an extremely critical step for this group of patients. The findings in this study, in which the CR rate was as high as 57% with the combined regimen, prompted us to investigate the underlying mechanism of chidamide-based therapy in T-ALL. The cell viability of both cell lines was significantly inhibited by chidamide in a dose- and time-dependent manner (Figure 2A). Then, we determined whether chidamide induced apoptosis in T-ALL cells. Compared with control cells, the percentage of apoptotic cells increased when the concentration of chidamide increased (Figure 2B, Supplementary Figures S1A, B). Furthermore, we tested whether chidamide induced cell cycle arrest in T-ALL cells. Consistent with a previous study (31), chidamide treatment increased the percentage of cells in the G0/G1 phase while decreasing the percentage of cells in the S phase (Figure 2C, Supplementary Figure S1C). The effect of chidamide on apoptosis and the cell cycle was also concentration- and time-dependent. Taken together, these data indicated that chidamide significantly inhibited the proliferation of T-ALL cells in vitro.

Given that the pleiotropic effects of chidamide and its relevant molecular mechanism remained to be clarified, the global HDAC activity in both cell lines was evaluated. Upon treatment with chidamide, the total cellular deacetylase activity was significantly inhibited (Figure 2D). This was consistent with a previous study on T-ALL (27). The antitumor effect of HDACIs was initially considered to be dependent on their effect on histone acetylation. However, increasing evidence indicated that the biological effects of HDACIs are various (32). Changes in expression levels of HDACs may also be necessary for the biological effects of HDACIs (32–34). Therefore, we also checked the expression level of HDAC1, HDAC2, HDAC3, and HDAC10 upon treatment with chidamide (Figure 2E, Supplementary Figure S2A). The expression of HDAC3, but not HDAC1, HDAC2, or HDAC10, was dramatically downregulated after treatment with chidamide in both primary T-ALL specimens and cell lines. These findings suggested that the inhibitory effect of chidamide on T-ALL cells may be dependent on HDAC3.

To determine whether chidamide-induced cell arrest of T-ALL cells was dependent on HDAC3, we first evaluated the effect of HDAC3 silencing on the viability of T-ALL cells. Next, we explored whether HDAC3 overexpression rescued the inhibitory effect of chidamide on T-ALL cells. First, HDAC3 was overexpressed in both cell lines (Figure 2F, annotated as OE-HDAC3). Overexpression of HDAC3 was found to rescue the decreased cell viability (Figure 2G), decrease the percentage of apoptotic cells (Figure 2H, Supplementary Figures S2B, C), and cause cell cycle arrest (Supplementary Figures S2D–F) induced by chidamide. Then, the expression of HDAC3 was knocked down in MOLT-4 cells (Figure 2I). The cell proliferation was significantly suppressed in HDAC3 knockdown cells (Figure 2J). The percentage of apoptotic cells in the HDAC3 knockdown cells was higher than that in control cells (Figure 2K, Supplementary Figure S2G). Moreover, HDAC3 silencing was found to reduce the proportion of cells in the S phase and induced cell cycle arrest in the G0/G1 phase (Supplementary Figures S2H–J). Altogether, these data suggested that the inhibitory effect of chidamide on T-ALL cells was at least partially dependent on HDAC3.

To further explore the underlying mechanism of the inhibitory effect of chidamide on T-ALL cells, we performed RNA sequencing (RNA-seq) on chidamide-treated Jurkat and MOLT-4 cells. The heatmap of differential gene expression levels across chidamide treatment group was shown in Supplementary Figure S3A. Compared with untreated Jurkat cells, 2331 upregulated genes and 935 downregulated genes were identified in chidamide-treated Jurkat cells. There were 2478 upregulated genes and 806 downregulated genes in chidamide-treated MOLT-4 cells compared to untreated cells (Figure 3A). And then, further Gene Ontology (GO) was performed. Differentially expressed genes are significantly enriched in signaling pathways of leukocyte migration/immune chemotaxis, cytokine receptor activity, and chemokine receptor activity (Supplementary Figure S3B). While performing Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses, multiple signaling pathways—including the JAK-STAT signaling pathway—were significantly inhibited following chidamide treatment (Figure 3B). Within the JAK-STAT signaling pathway, the expression of TYK2, STAT1, and BCL2 was significantly different (Figure 3C). HDAC3, but not HDAC1, HDAC2, or HDAC10, was also found to be one of the most differentially expressed genes in both cell lines upon treatment with chidamide (Figure 3C).

To verify these results, we reviewed the differential expression of genes within the JAK-STAT signaling pathway in both cell lines and primary T-ALL samples. The mRNA and protein levels of TYK2, but not of JAK1-3, were both significantly inhibited upon chidamide treatment (Figures 3D, E). The expression of STAT1, phosphorylated STAT1, and BCL2 was also clearly inhibited (Figures 3D, E, Supplementary Figure S3C). These data were in accordance with the RNA-seq results. HDAC has been reported to affect the growth of tumor cells by regulating the transcription of many oncogenes or tumor suppressor genes (35). More importantly, the alteration of HDAC expression can protect tumor cells from genotoxic drugs (36). The expression of HDAC1, HDAC2, HDAC3, and HDAC10 genes upon chidamide treatment was examined (Figure 2E). The protein level of HDAC3 was found to be significantly inhibited by chidamide. Given that the TYK2-STAT1-BCL2 signaling pathway has been reported to be constitutively activated in T-ALL cells (37), we also examined the expression level of the TYK2-STAT1-BCL2 signaling pathway in primary T-ALL cells. Compared with healthy controls, the mRNA level of TYK2 and STAT1 was also found to be upregulated (Figure 3F). The expression level of HDAC3 was found to be higher in primary T-ALL cells (Figure 3F). To further address the limitation of sample size, we analyzed an independent public transcriptomic dataset (GSE7186). Consistent expression patterns of TYK2-STAT1-BCL2 axis-related genes were observed in T-ALL samples (Supplementary Figure S3D). All these results suggested that the TYK2-STAT1-BCL2 signaling pathway and HDAC3 were activated in T-ALL cells and that they could be inhibited by chidamide.

First, we examined whether the growth of T-ALL cells was dependent on the TYK2-STAT1-BCL2 signaling pathway. TYK2 was knocked down in Jurkat and MOLT-4 cells. The cell proliferation of both cell lines was significantly suppressed in TYK2 knockdown cells (Figure 4A). The percentage of apoptotic cells in TYK2 knockdown cells was higher than that in the control cells (Figures 4B, C). Moreover, TYK2 silencing was found to reduce the proportion of cells in the S phase and induced cell cycle arrest in the G0/G1 phase (Figures 4D, E, Supplementary Figure S4A). We also examined the influence of TYK2 knockdown on its downstream signaling pathway. The expression of STAT1, phosphorylated STAT1, and BCL2 was significantly inhibited in TYK2 knockdown cells (Figures 4F–I, Supplementary Figures S4B–E). Taken together, our results consistently demonstrated that the survival of T-ALL cells was dependent on the TYK2-STAT1-BCL2 signaling pathway (37).

Next, we explored whether chidamide-induced cell arrest of T-ALL cells was dependent on TYK2. As expected, overexpression of TYK2 (annotated as OE-TYK2) was found to rescue the decreased cell viability (Figure 5A), decrease the percentage of apoptotic cells (Figures 5B, C, Supplementary Figure S5A), and alleviate cell cycle arrest (Figures 5D–F, Supplementary Figure S5B) induced by chidamide. The expression of STAT1, phosphorylated-STAT1, and BCL2 was found to be upregulated in cells with TYK2 overexpression (Figures 5G–J, Supplementary Figures S5C–E). Altogether, our data supported that the inhibitory effect of chidamide on T-ALL cells was at least partially in a TYK2-dependent manner.

Given that chidamide inhibited the growth of T-ALL cells and that overexpression of HDAC3 or TYK2 at least partially rescued the inhibitory effect of chidamide, we sought to explore the impact of HDAC3 expression on the TYK2-STAT1-BCL2 signaling pathway and the interaction between HDAC3 and TYK2. We found that HDAC3 silencing decreased the expression of TYK2, STAT1, phosphorylated STAT1, and BCL2 (Figures 6A–D). Similarly, overexpression of HDAC3 led to the upregulation and activation of the TYK2-STAT1-BCL2 signaling pathway (Figures 6E–H, Supplementary Figures S6A–D). To further explore the relationship between HDAC3 and TYK2, we examined whether HDAC3 is associated with TYK2 in T-ALL cells. Co-immunoprecipitation assays revealed an association between endogenous HDAC3 and TYK2 in Jurkat cells (Figure 6I), which was further confirmed in MOLT-4 cells (Supplementary Figure S6E). To corroborate this finding in an orthogonal system, co-immunoprecipitation was also performed in HEK293T cells co-expressing HDAC3 and TYK2, again supporting their association (Figure 6J). Notably, increased HDAC3 abundance correlated with reduced acetylation of TYK2 and enhanced TYK2 phosphorylation, suggesting that HDAC3 is linked to modulation of TYK2 post-translational status. Consistent with this association, perturbation of HDAC3 expression altered downstream TYK2-STAT1-BCL2 signaling activity in T-ALL cells. Taken together, these results suggested that HDAC3 is associated with TYK2 and contributes to activation of the TYK2-STAT1-BCL2 signaling pathway.

To further validate whether there was a synergistic inhibitory effect between chidamide and TYK2 inhibition in T-ALL cells, deucravacitinib, a highly selective inhibitor of TYK2, was used. The proliferation of T-ALL cells was profoundly inhibited by the combination treatment of chidamide and deucravacitinib (Figure 7A). Using the Chou-Talalay method, we found that 0.3 μM chidamide and 4 μM Deucravacitinib inhibited cell growth by more than 50% in Jurkat cells (Combination Index 0.6). Compared with a single reagent, the percentage of apoptotic cells was significantly higher in the cells treated with the combination treatment of chidamide and deucravacitinib (Figure 7B and Supplementary Figure S7A). Then, a xenograft mouse model was established in NOD/SCID mice by subcutaneously inoculating the mice with wild-type Jurkat cells (total = 24) (Figures 7C–G). The mice were randomized into four groups (n=6). They were treated with chidamide, deucravacitinib, chidamide with deucravacitinib, or a PBS control. The volume and weight of the tumors in the combination group were less than those of tumors in the other groups (Figures 7C–F). Moreover, the expression of phosphorylated STAT1, BCL2, and Ki-67 in the combination group was lower than that in the other groups (Figure 7G, Supplementary Figure S7B, C). These results indicated that chidamide and a TYK2 inhibitor synergistically inhibited the growth of T-ALL cells.