Preclinical Evaluation of the HDAC Inhibitor Chidamide in Transformed Follicular Lymphoma

The key factors leading to transformed follicular lymphoma (t-FL) include the aberrations of epigenetic modifiers as early and driving events, especially mutations in the gene encoding for histone acetyltransferase. Therefore, reversal of this phenomenon by histone deacetylase (HDAC) inhibitors is essential for the development of new treatment strategies in t-FL. Several t-FL cell lines were treated with various doses of chidamide and subjected to cell proliferation, apoptosis and cell cycle analyses with CCK-8 assay, Annexin V/PI assay and flow cytometry, respectively. Chidamide dose-dependently inhibited cell proliferation, caused G0/G1 cycle arrest and triggered apoptosis in t-FL cells. In addition, the effects of chidamide on tumor growth were evaluated in vivo in xenograft models. RNA-seq analysis revealed gene expression alterations involving the PI3K-AKT signaling pathway might account for the mechanism underlying the antitumor activity of chidamide as a single agent in t-FL. These findings provide a basis for further clinical exploration of chidamide as a promising treatment for FL.


INTRODUCTION
Follicular lymphoma (FL) is the most common indolent lymphoma, accounting for approximately 20% of all non-Hodgkin lymphoma (NHL) cases (1,2). With the use of current front-line regimens, the majority of FL patients have an initial response to therapy, with 40 to 80% demonstrating complete response (3,4). However, despite improvement in front-line treatment, conventional therapy for FL is not curative, and approximately 20% of patients still experience either refractoriness or early relapse, which occurs in the first 2 years after diagnosis and treatment by chemoimmunotherapy (5)(6)(7). Moreover, such early relapse cases are often chemo-resistant, leading to significantly shorter survival (4,8). Thus, integrating molecular targeted therapies into current treatment protocols and adjusting conventional treatment to improve survival, without compromising long-term quality of life is urgently needed in FL patients with poor prognosis.
Histone deacetylases (HDACs), which act as "epigenetic erasers", are known to catalyze the removal of acetyl groups from histones and non-histone proteins, thereby altering the transcription of oncogenes and tumor suppressor genes (9,10). Aberrant HDAC expression occurs in both solid tumors and hematological cancers, including B-cell lymphoma (11,12). Prior findings showed that dysregulation of histone acetylation contributes to lymphomagenesis, particularly in GC-derived lymphomas (13,14). FL, similar to other cancers, has recurrent alterations in genes involved in maintaining chromatin structure and transcription machinery genes (4,8). For example, somatic mutations or genomic loss in the CREBBP and EP300 genes that encode HATs lead to an imbalance between acetylation and deacetylation, and the occurrence of these mutations is associated with disease relapse and poor prognosis in FL (15)(16)(17). Perturbing the balance between histone acetylation and deacetylation, which is tightly regulated by HDACs, is one of the main mechanisms by which epigenetics may be exploited to harness chromatin remodeling (18). In this context, it raises the possibility that HDACs may serve as a potentially attractive therapeutic target in this disease.
According to previous studies, Class I HDACs 1-3 are the most important HDAC enzymes with close associations with the corresponding malignant phenotypes (19,20). Currently, multiple HDAC inhibitors have been developed, Chidamide is a noteworthy drug that may target specifically subtypes 1, 2 and 3 of Class I and subtype 10 of Class IIb HDACs and lead to increased acetylation of histones H3 and H4, resulting ultimately in the activation of gene transcription (21,22). Chidamide was first approved by the Chinese FDA for the treatment of relapsed or refractory (R/R) peripheral T cell lymphoma (PTCL) (23,24). Afterwards, more and more studies have focused on the anticancer effects of chidamide in various tumors, including hematological tumors (25)(26)(27). However, significant knowledge gaps remain, including the mechanism underlying chidamide's therapeutic effects. In addition, its clinical utility in FL is currently unclear.
This study evaluated the effects of chidamide in wellcharacterized transformed follicular lymphoma cell lines and a xenograft model of t-FL. In addition, we analyzed the molecular basis of chidamide's effects by evaluating gene expression using microarrays in cells treated with the single agent in t-FL: 1) to identify genes and pathways affected by chidamide; 2) to determine biomarkers that could be used in preclinical studies.

Cell Viability Assessment
Cytotoxicity was determined with Cell Counting Kit-8 (CCK-8, APExBIO, Texas, USA). Briefly, 3×10 4 cells/well were seeded in 100ml medium in 96-well plates and treated with various concentrations of chidamide alone for 24, 36 and 48 h. The CCK-8 reagent (10ml/well) was then added and incubated for additional 2 h, after which absorbance at 450 nm was detected on a Bio-Rad microplate reader (Bio-Rad, CA, USA). Data from three independent triplicate experiments were presented as a percentage of viable cells relative to untreated controls. IC 50 values were determined with the GraphPad Prism 6 software.

FACS Analysis of the Cell Cycle and Apoptosis
Cells were treated with various concentrations of chidamide for the designated times. Cells were harvested and processed according to the manufacturer's protocols. For cell cycle analysis, propidium iodide (PI)/RNase staining buffer from BD Pharmingen (556463, New Jersey, USA) was used. Cells were then analyzed on a CytoFlex S flow cytometer (Beckman Coulter, CA, USA). Data analysis was performed with the FlowJo software (San Carlos, CA, USA). Apoptosis was measured with the Annexin V/PI apoptosis detection kit (BD Pharmingen, USA). Cells positive for Annexin V were determined to be apoptotic (28), and were located in the right quadrant of the dot plot. Statistical analysis was performed by ANOVA. P-values below 0.05 in comparison to the control group were considered significant.

Western Blot
Protein extraction, separation and immunoblotting were performed as previously described (29). Suite, USA) was used as a loading control. Blots were then detected using the hypersensitive ECL chemiluminescence kit (NCM Biotech, Suzhou, China) and the Bio-Rad ChemiDoc XRS + detection system (Bio-Rad, CA, USA).

RNA-Sequencing
DOHH2 cells were incubated with chidamide for 24h, followed by total RNA extraction with TRIzol reagent (Invitrogen, NY, USA) according to the manufacturer's instructions. Totally, 1mg total RNA with RIN above 6.5 was used for subsequent library preparation. Next generation sequencing libraries were constructed according to the manufacturer's protocol. Then, libraries with different indices were multiplexed and loaded on an Illumina HiSeq instrument according to the manufacturer's instructions (Illumina, CA, USA). Sequencing was carried out using the 2x150 bp paired-end (PE) configuration; image analysis and base calling were conducted with HiSeq Control Software (HCS) + OLB + GAPipeline-1.6 (Illumina) on the HiSeq instrument. KEGG pathway and GO analyses were performed using the R Studio approach. Differential expression analysis used the DESeq2 Bioconductor package, a model based on negative binomial distribution; adjusted p value (padj)<0.05 indicated differential expression.

In vivo Experiments
All animal procedures were performed in accordance with the guidelines of the Animal Care and Use Committee and Ethics Committee of Xiamen University. DOHH2 cells (200ml of PBS, 1×10 7 cells/mouse) were inoculated subcutaneously into the back of female CB17/Icr-Prkdcscid/IcrlcoCrl mice (approximately 14-16g of body weight, Xiamen University Laboratory Animal Center, Fujian, China). After 3 days, mice were randomly divided into two groups (8 animals per group), to receive vehicle (PBS with 0.2% methyl cellulose/0.1% Tween 80) and chidamide (10 mg/kg/d), administered by oral gavage for 3 successive weeks, respectively. Tumor size and body weight were measured every two days. Tumor volumes were calculated according to the formula V= (L × W 2 )/2 [V, volume (mm3); L, length (mm); W, width (mm)]. After euthanasia, tumor tissues were extracted and divided into two parts: one part was frozen at -80°C for protein extraction followed by Western blot, and the other was fixed with 4% paraformaldehyde for hematoxylin and eosin (H&E) staining, immunohistochemistry (IHC) and immunofluorescence (IF). The slides were incubated overnight at 4°C with primary antibodies targeting Ki67

Statistical Analyses
Statistical analyses were performed with Statistical Product and Service Solutions (SPSS) 21.0 (IBM Corp., Amronk, New York, USA), GraphPad Prism 6 (GraphPad Software, CA, USA) and Microsoft Office Excel (WA, USA). Unpaired Student's t-test was performed to compare group pairs. Multiple groups were compared by one-way ANOVA, followed by post-hoc Bonferroni test. All quantifications were performed based on at least three independent experiments. P<0.05 was considered statistically significant.

Effect of Chidamide on t-FL Cell Viability
We evaluated the anti-proliferative activity of chidamide in four t-FL cell lines, including RL, DOHH2, SU-DHL4 and Karpas422

Chidamide Induces Caspase-Dependent Apoptosis in t-FL Cells
To further assess the antitumor effect of chidamide on t-FL cells, flow cytometry after Annexin V/PI staining was performed to examine whether chidamide induces apoptosis in t-FL cells. Four t-FL cell lines were cultured with increasing concentrations of chidamide for 24 and 48 h before apoptosis assessment. Consistent with the CCK-8 assay, dose and time-dependent induction of cell death was evident in both DOHH2 and SU-DHL4 cells, whereas RL and Karpars422 cells showed reduced, although significant induction of apoptosis ( Figure 2 and Supplementary Figure 1), further indicating that DOHH2 and SU-DHL4 cells were more sensitive to chidamide and suggesting cell line-specific differences. Thus, besides the reduced cell proliferation, increased cell death might be another factor contributing to chidamide activity in t-FL cells. We next sought to confirm the potential mechanism underlying chidamide's anti-apoptotic effect. Western blot was carried out to detect the expression of activated caspase-3 and cleaved PARP after chidamide treatment for 12, 24 and 36 h, respectively, in DOHH2 and SU-DHL4 cells. Chidamide treatment for 24 h markedly upregulated cleaved caspase-3 and cleaved PARP ( Figure 3). Taken together, these results suggested that chidamide induced apoptosis in FL cells by triggering the caspase dependent pathway.
Chidamide Induces Cell Cycle Arrest in the G0/G1 Phase Aberrant HDAC expression has previously been shown to impair a subset of genes involved in cell cycle regulation. Accordingly, inducing cell cycle arrest may be an underlying mechanism of chidamide's effect on FL cells. In the above CCK-8 assay, chidamide exhibited a potent inhibitory effect on the proliferation of all four FL cells studied. We further investigated the effect of chidamide on cell cycle distribution in these four FL cell lines.  signaling, PI3K/AKT signaling and cell cycle regulation ( Figures 5B, C). The downregulated transcripts mainly comprised HDAC, P53 and CDK2 or genes involved in the PI3K/AKT pathway ( Figure 5D). In this context, Western blot was performed to validate the target specificity of chidamide on HDAC1, 2, 3 and 10. As shown in Figure 6, exposure of DOHH2 and SU-DHL4 cells to chidamide resulted in time-dependent downregulation of HDAC1, 2, 3 and 10, thereby causing hyperacetylation of histones H3 and H4. Chidamide also upregulated p27 and downregulated phosph-CDK2 (Thr160) in a time-   Figure 3), which might be partially responsible for their lower sensitivity to chidamide. Taken together, these results suggested that the mechanism underlying the anti-proliferative activity of chidamide might involve downregulation of molecules related to the PI3K kinase pathway and cell cycle arrest in t-FL cells.

Chidamide Affects the PI3K/PDK1/AKT Signaling Pathway in t-FL Cells
Chidamide negatively regulated transcripts encoding MYC-and P53-regulated genes, as well as PI3K/AKT signaling pathway effectors, including HDAC10, AKT1, CDK2, MYC, PARPBP, PCNA, PDK1, PRDM10 and PRDM15 in DOHH2 cells ( Figure 5D). The inhibitory effect of chidamide on the PI3K/ AKT signaling pathway was further confirmed at the protein level. Western blot showed a marked reduction in the expression of PDK1 and phospho-AKT (Ser473 or Thr308) in both DOHH2 and SU-DHL4 cells after chidamide treatment for 36 h ( Figure 6). In contrast, chidamide did not affect the expression of phospho-AKT (Ser473) in RL and Karpas-422 cells (Supplementary Figure 3). Thus, we speculated that chidamide might promote apoptosis and suppress proliferation in part by disrupting the PI3K/PDK1/AKT signaling pathway in t-FL cells.

Chidamide Has Antitumor Activity in a FL Tumor Xenograft Model
Finally, the anti-tumor activity of chidamide was examined in a CB17/Icr-Prkdcscid/IcrlcoCrl mouse xenograft model bearing DOHH2 cells. In this study, DOHH2 cells (10 7 ) were injected subcutaneously into the back of mice, which were randomly divided into the vehicle control and chidamide groups. Vehicle or chidamide (10 mg/kg/day) was orally given continuously daily for three weeks ( Figure 7A). After chidamide treatment for 13  days, compared with the control group, mouse tumors showed obvious growth inhibition with no fatal toxicity ( Figures 7B, C). Although a temporary body weight loss was observed at the beginning of chidamide administration, it was recovered after a short period of time. Tumors were collected from 5 mice randomly selected per group at the study endpoint (Day 22). As shown in Figure 7D, chidamide treatment resulted in a marked reduction of tumor burden, reflected by decreased volume and weight of tumor masses, compared with the vehicle control ( Figures 7E, F). Besides, chidamide significantly prolonged survival in the treatment group compared with vehicle treated animals ( Figure 7G). The tumor tissues obtained from chidamide treated animals displayed obvious nuclear shrinkage as shown by H&E staining ( Figure 8A). Apoptosis in tumor tissues was detected by the TUNEL assay. The number of apoptotic cells was markedly increased in chidamide treated tumors ( Figures 8B, C), in agreement with in vitro findings. Immunohistochemical staining revealed that Ki-67 and PCNA levels were decreased in tumor tissues from chidamide treated mice (Figures 8D-F). Compared with the vehicle control group, chidamide treated mice showed significantly reduced p-PDK1 expression, and markedly increased expression of P27, cleaved caspase-3 and cleaved PARP ( Figure 7H). These results were consistent with in vitro findings. Taken together, these data suggested that chidamide effectively inhibited t-FL tumorigenesis and development in vivo.

DISCUSSION
Transformed follicular lymphoma (t-FL) is considered a disease predominantly caused by several epigenetic aberrations (e.g., mutations affecting the epigenetic modifiers KMT2D, EZH2,  CREBBP and MEF2B) rather than sequential acquisition of genetic aberrations (8,18). Thus, targeting enzymes involved in the regulation of DNA methylation and histone modifications might be critical for developing more effective treatment strategies in t-FL. In this regard, histone deacetylases (HDACs) represent the most widely studied therapeutic targets, with HDAC inhibitors (HDACi) including chidamide, vorinostat and romidepsin, approved for use in the therapy of cutaneous T-cell lymphoma (23,30). However, clinical practice has not yet delivered desirable results by applying HDACi in the treatment of t-FL. Furthermore, the biological effects of HDACi in t-FL remain unclear, as well as the identification of response mechanisms. Taking these observations into consideration, we evaluated the activity and underlying mechanism of a benzamide-type selective HDAC inhibitor, chidamide, in preclinical models of t-FL cells.
HDACs have been demonstrated to play a crucial role in the pathogenesis of lymphoma (31,32). Gil et al. reported that aberrant expression of HDAC9 could lead to lymphoproliferative disorders, including germinal center (GC) and post-GC lymphomas (33). In follicular lymphoma (FL) and diffuse large B cell lymphoma (DLBCL), Bcl6 recruits HDAC3 to repress transcription and trigger B cell lymphoma (34). All these data strongly suggest that HDACs are promising therapeutic targets for GC lymphomas. Furthermore, HDAC inhibitors have been reported to induce Bcl6 downregulation in GC lymphomas including DLBCL (35). Herein, we demonstrated that chidamide had an anticancer effect as a single agent in several t-FL cell lines. At clinically achievable concentrations, chidamide showed cell type-and dose-dependent cytotoxicity in t-FL cells at 24 h, with IC50 values ranging from 4.5µM to 30µM, and varied responses to chidamide among different cells. In DOHH2 and SU-DHL4 cells, which were relatively more sensitive to chidamide, a higher degree of apoptosis induced by chidamide was found as strongly evidenced by the activation of caspase-3. This is likely due to the cell origins or unknown differences of their genetic differentiation. Since chidamide inhibited cell proliferation and induced apoptosis in t-FL cells, we further investigated whether chidamide regulates cell cycle progression, which is one of the main mechanisms by which HDACi induce tumor cell death. In this context, previous studies have shown that G1 arrest appears to be a common response to chidamide in various tumor cells (36)(37)(38). Thus, cell cycle regulators, including cyclins and CDK inhibitors (e.g., p21 and p27), may be tightly controlled by chidamide (27,36,39). In general, p27 is known to control G1 length and cell cycle exit by inhibiting the kinase activity of CDK2 bound to cyclin E, thereby causing the dephosphorylation of retinoblastoma protein (Rb), which blocks E2F activity in the transcription of genes required for G1/S transition (40,41). In this study, microarray analysis of chidamide-treated t-FL cells pointed to cell cycle arrest in the G1 phase. Upon chidamide treatment, CDK2 was found to be specifically inhibited in DOHH2 and SU-DHL4 cells accompanied by p27 activation. However, obvious changes of p27 expression were not observed in RL and Karpas-422 cells, in disagreement with the antiproliferative phenotype as well as G1 arrest in these cells, indicating that p27-mediated inhibition of CDK2 might partially contribute to chidamide-induced cell cycle arrest. On the other hand, we showed that chidamide treatment simultaneously caused the accumulation of histone H3 acetylation and the activation of p27 and cell cycle exit in DOHH2 and SU-DHL4 cells. The present findings indicate that the mechanism by which chidamide induces G1 arrest by inhibiting HDACs is cell-specific.
The integration of gene expression profiling and sensitivity in cancer cells allowed the identification of functional pathways that might predict the response to chidamide (27,42). DOHH2 cells with the highest sensitivity to chidamide had high expression levels of genes involved in the PI3K/PDK1/AKT pathway, which is regulated in several human carcinomas, including lymphoid malignancies (43,44). In addition, an important biological effect of chidamide, both in vitro and in vivo, was the downregulation of the PI3K/AKT pathway. Several studies have indicated that PI3K/AKT pathway downregulation is a relevant mechanism of chidamide's effects in various cancer cell lines (27,45,46). The PI3K/AKT pathway was shown to be activated in human cancers by oncogenic mutations of the PIK3CA gene encoding the catalytic subunit p110a (47). In lymphoid malignancies, including DLBCL (48), MCL (49) and FL (50), hyperactivation of AKT is due to its enhanced activation/phosphorylation at Serine 473 (Ser473). In accordance with apoptosis and caspase-3 activation observed at 24 to 48 h, AKT phosphorylation at ser473 was inhibited in both DOHH2 and SU-DHL4 cells by chidamide treatment. Taken together, these findings suggested that chidamide modulated PI3K/AKT signaling, which is known to be involved in cell proliferation, cell cycle regulation, apoptosis and tumor development ( Figure 9).
As the activity of chidamide, as monotherapy at the doses tested in this study was modest, understanding potential biomarkers that are predictive of response is very important for the design of future clinical trials (51). In addition, emerging data from in vitro studies indicate that HDAC inhibitors, such as chidamide, may have improved activity when used in combination therapy (52)(53)(54). To this end, several recent studies have provided a strong preclinical rationale for combination with chemotherapy, immunotherapy, or molecular targeted therapy, paving the way for possible studies in selected populations (53)(54)(55). In our previous work, we revealed that  or MLL-menin inhibitor (56) has a synergistic inhibitory effect on acute myeloid leukemia cells when combined with chidamide. Combination with a demethylating agent also showed benefit in diffuse large B cell lymphomas (DLBCLs) (53). In addition, chidamide could increase PD-L1 expression in the tumor microenvironment, and preclinical studies have demonstrated synergy between chidamide, and PD-1 blockade in solid tumors (55,57). A single-arm-phase II study is therefore currently ongoing to evaluate the activity of chidamide in combination with sintilimab in relapsed or refractory peripheral T-cell lymphomas (58). It is possible that chidamide administration in the context of these combination strategies could further enhance the killing of tumor cells.
In summary, this study demonstrated the importance of HDACs in the progression and transformation of FL, and provided a critical link between epigenetic changes and increased FL aggressiveness. We also provided evidence that chidamide exerts anticancer effects by inducing G1 arrest and apoptosis via PI3K/PDK1/AKT signaling pathway inactivation. Furthermore, in t-FL cells relatively sensitive to chidamideinduced apoptosis, chidamide caused significant changes in the FIGURE 9 | Illustration of the mechanism underlying chidamide induced follicular lymphoma growth inhibition. Chidamide exerts antitumor activity toward FL both in vitro and in vivo, potentially by targeting the PDK1-Akt-P27-CDK2 pathway and triggering caspase dependent apoptosis, finally markedly blocking cells in the G0/G1 phase. transcriptome profile, providing a compelling rationale for chidamide as an effective single-agent in aggressive FL. The data presented here provide the basis for further exploration of chidamide in combination therapies.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: SAMN22127095.

ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Care and Use Committee and Ethics Committee of Xiamen University. Written informed consent was obtained from the owners for the participation of their animals in this study.  Supplementary Table 1 | IC50 values of chidamide in FL cell lines.