Article Text

Download PDFPDF

Histone chaperone FACT complex mediates oxidative stress response to promote liver cancer progression
  1. Jialing Shen1,2,
  2. Mengnuo Chen1,2,
  3. Derek Lee1,2,
  4. Cheuk-Ting Law1,2,
  5. Lai Wei1,2,
  6. Felice Ho-Ching Tsang1,2,
  7. Don Wai-Ching Chin1,2,
  8. Carol Lai-Hung Cheng1,2,
  9. Joyce Man-Fong Lee1,2,
  10. Irene Oi-Lin Ng1,2,
  11. Carmen Chak-Lui Wong1,2,
  12. Chun-Ming Wong1,2
  1. 1State Key Laboratory of Liver Research, University of Hong Kong, Hong Kong, Hong Kong
  2. 2Department of Pathology, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong, Hong Kong
  1. Correspondence to Dr Chun-Ming Wong, State Key Laboratory of Liver Research and Department of Pathology, University of Hong Kong Li Ka Shing, Hong Kong, Hong Kong; jackwong{at}pathology.hku.hk; Dr Carmen Chak-Lui Wong, Department of Pathology, The University of Hong Kong, Hong Kong, China; carmencl{at}pathology.hku.hk

Abstract

Objective Facilitates Chromatin Transcription (FACT) complex is a histone chaperone participating in DNA repair-related and transcription-related chromatin dynamics. In this study, we investigated its oncogenic functions, underlying mechanisms and therapeutic implications in human hepatocellular carcinoma (HCC).

Design We obtained HCC and its corresponding non-tumorous liver samples from 16 patients and identified FACT complex as the most upregulated histone chaperone by RNA-Seq. We further used CRISPR-based gene activation and knockout systems to demonstrate the functions of FACT complex in HCC growth and metastasis. Functional roles and mechanistic insights of FACT complex in oxidative stress response were investigated by ChIP assay, flow cytometry, gene expression assays and 4sU-DRB transcription elongation assay. Therapeutic effect of FACT complex inhibitor, Curaxin, was tested in both in vitro and in vivo models.

Results We showed that FACT complex was remarkably upregulated in HCC and contributed to HCC progression. Importantly, we unprecedentedly revealed an indispensable role of FACT complex in NRF2-driven oxidative stress response. Oxidative stress prevented NRF2 and FACT complex from KEAP1-mediated protein ubiquitination and degradation. Stabilised NRF2 and FACT complex form a positive feedback loop; NRF2 transcriptionally activates the FACT complex, while FACT complex promotes the transcription elongation of NRF2 and its downstream antioxidant genes through facilitating rapid nucleosome disassembly for the passage of RNA polymerase. Therapeutically, Curaxin effectively suppressed HCC growth and sensitised HCC cell to sorafenib.

Conclusion In conclusion, our findings demonstrated that FACT complex is essential for the expeditious HCC oxidative stress response and is a potential therapeutic target for HCC treatment.

  • hepatocellular carcinoma
  • SUPT16H
  • SSRP1
  • FACT complex
  • ROS
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known on this subject?

  • Hepatocellular carcinoma (HCC) is a major malignancy worldwide and a leading cause of death in the world. However, current therapies for patients with HCC are unsatisfactory due to chemoresistance and drug resistance.

  • Histone chaperones control the dynamic functions of chromatin and are known to play essential roles in DNA replication, DNA damage repair and RNA transcription. Facilitates Chromatin Transcription (FACT) complex is a histone chaperone critical for nucleosome assembly and disassembly during transcription elongation. However, how deregulation of FACT complex contributes to human carcinogenesis remains elusive.

What are the new findings?

  • Deregulation of histone chaperones is a common trait of human HCC.

  • Two components of FACT complex, SUPT16H and SSRP1, are frequently upregulated in human HCC and associated with poor prognosis of patients with HCC.

  • FACT complex is regulated by NRF2 and KEAP1 at transcriptional level and protein level, respectively. FACT complex is upregulated in HCC in response to oxidative stress.

  • FACT complex accelerates the transcription elongation of antioxidant genes to enable cancer cells to survive under oxidative stress.

  • HCC cells are significantly sensitive to FACT inhibitor Curaxin treatment. Curaxin treatment also sensitised HCC cells to sorafenib.

How might it impact on clinical practice in the foreseeable future?

  • FACT complex is a sensitive and specific biomarker for HCC detection, prognostication and molecular classification.

  • FACT complex is indispensable for oxidative stress adaptation during HCC progression and under sorafenib treatment. Inactivation of FACT complex effectively abolishes HCC oxidative stress response and induces apoptosis. Therefore, targeting FACT complex by Curaxin, as a single treatment or in combination with sorafenib, is a novel strategy for HCC treatment.

Introduction

The high-order chromatin structure determines the genome architecture and gene expression of eukaryotic cells. The nucleosome, which consists of 146 bp of DNA wrapping around a histone octamer (one H3–H4 tetramer and two H2A–H2B dimers), is the fundamental building block of chromatin.1 This basic unit of DNA package plays an indispensable role in epigenetic regulation and is intimately related to DNA replication, DNA damage repair and gene transcription.2 Active disassembly, reassembly and repositioning of nucleosomes render spatial and temporal controls of chromatin accessibility for the actions of transcription factors and RNA polymerases. These dynamic changes are regulated by several factors, including ATP-dependent chromatin remodellers,3 histone-modifying enzymes4 and an increasing list of histone chaperones.5 6

Histone-modifying enzymes facilitate gene regulation through addition or removal of certain chemical groups, such as acetyl, methyl and ubiquitin, on the N-terminal tails of core histone proteins. These chemical modifications can be recognised by specific ‘reader’ proteins to modulate chromatin structure and transcriptional competence. Chromatin remodellers act by sliding or rejecting nucleosomes along the DNA strand to control the accessibility of DNA sequences.7 8 On the other hand, histone chaperones function by binding with histones to facilitate assembly and disassembly of nucleosomes in an ATP-independent manner.9 These chromatin regulating factors work coordinately to control number of events related to nucleosome organisation. Therefore, deregulation of the concerted action of these factors may result in defects in genome stability and gene expression, thereby promoting development of human diseases, such as cancer.10 Previous studies have mainly focused on diverse histone modifications and chromatin remodellers. However, the role of histone chaperones in human diseases, especially cancer, remains to be elucidated.

Histone chaperones are key proteins that function at multiple steps of nucleosome packaging and disassembly.6 They are defined as proteins that handle non-nucleosome histones and mediate the disposition or displacement of histone dimers in chromatin.6 Over the past few decades, histone chaperones have been demonstrated to participate in distinct steps of histone transportation,11 12 histone supply13 and nucleosome turnover.14 15 Therefore, histone chaperones are regarded as key factors in all aspects of histone dynamics, regulating gene transcription and DNA replication. Currently, histone chaperones could be classified only according to their histone substrate specificity. Most histone chaperones bind to either H2A–H2B dimer or H3–H4 tetramer,16 although some, like Facilitates Chromatin Transcription (FACT) complex, bind to both using different domains.

In human, histone chaperone FACT complex is a heterodimeric complex composed of the 140 kDa SUPT16H (SPT16 Homolog) and 80 kDa SSRP1 (Structure Specific Recognition Protein 1) subunits.17 It is initially found to destabilise nucleosomes along the DNA strand to facilitate RNA polymerase II–driven transcription and deposit histone proteins back afterwards.17 18 Therefore, traditionally FACT complex is defined as a transcriptional elongation factor and is regarded as an extremely stable histone carrier.17 Recent studies showed that FACT complex is involved in almost all chromatin-related processes, including DNA replication, damage repair, transcription initiation and transcription elongation.18 Emerging evidence showed that FACT complex is frequently deregulated in human cancers and might be a potential target for cancer therapeutics. However, how deregulation of FACT complex contributes to human carcinogenesis remains largely elusive.

Hepatocellular carcinoma (HCC), transformed from hepatocytes, is the predominant form of primary liver cancer. HCC is also the fifth most common and second most fatal cancer in the world.19 Treatments for patients with HCC are very limited and not effective. The two FDA-approved first-line drugs, sorafenib and lenvatinib, can only prolong patients’ survival by 3 months.20 21 Therefore, there is an urgent need to identify and develop novel molecular targets for HCC therapy. In this study, we aim to explore the contribution of histone chaperones in human carcinogenesis.

Herein, using HCC as a cancer model, we provide insights into the oncogenic functions of histone chaperone FACT complex with both in vitro and in vivo models. Blockade of FACT complex expression effectively suppressed cancer cell growth and metastasis. Interestingly, we found that FACT complex is a key regulator of cellular oxidative stress response. At protein level, FACT complex is regulated by KEAP1 (Kelch Like ECH Associated Protein 1)-mediated protein degradation and being stabilised under oxidative stress. At transcriptional level, expression of FACT complex is driven by NRF2 (Nuclear Factor, Erythroid 2 Like 2) in response to oxidative stress. The increased expression of FACT complex in turn facilitates rapid passage of RNA Pol II during transcriptional elongation to enable the prompt transcription of NRF2 and its downstream antioxidant genes. Through the mechanism, FACT complex together with NRF2 expedites the instant and robust antioxidant gene expression that is essential for HCC cells to survive under oxidative stress. Therapeutically, we previously showed that sorafenib treatment induced oxidative stress in HCC.22 23 Inhibition of FACT complex compromised the oxidative stress adaptation and thus sensitised cancer cells towards sorafenib treatment. In summary, our study revealed an indispensable role of FACT complex in oxidative stress adaptation; targeting FACT complex increased the vulnerability of cancer cells to oxidative stress and represents a new strategy for cancer treatment.

Results

Frequent upregulation of histone chaperone FACT complex in human HCC

To have a comprehensive understanding on the deregulation of histone chaperones in human HCC, we interrogated the expression of 21 known histone chaperones (see online supplementary figure S1A) in 16 pairs of HBV-associated HCC and their corresponding non-tumorous (NT) livers using transcriptome-sequencing (RNA-Seq; online supplementary table S1). We observed an overall upregulation of histone chaperones in HCC samples (figure 1A,B). The expression profile of histone chaperones was able to effectively segregate HCC from their corresponding NT liver samples in an unsupervised clustering analysis, suggesting that deregulation of histone chaperones is a common event in human HCC (figure 1B). Among the histone chaperones examined, we found that SUPT16H and SSRP1, the two subunits of the FACT complex, were strikingly upregulated in human HCC (p=1.68×10−5 and 1.28×10−6, respectively) (figure 1C,D and online supplementary figure S1B). Intriguingly, the expression of SUPT16H and SSRP1 perfectly correlated with each other (figure 1E). Western blotting and immunohistochemistry analyses further confirmed the increased expressions of SUPT16H and SSRP1 in human HCC at protein level (figure 1F and online supplementary figure S1C). Consistent with our findings, both SUPT16H and SSRP1 were also significantly upregulated in multiple human HCC cell lines and TCGA human HCC cohort with mixed aetiologies (online supplementary figure S1D,E and table S2). Combining TCGA and our cohorts, upregulations of SUPT16H and SSRP1 were detected in 80.3% and 87.9% of patients with HCC, respectively (online supplementary figure S1F). Apart from HCC, upregulations of SUPT16H and SSRP1 were also observed in other solid cancers, including lung and breast cancers (online supplementary figure S1G). Receiver Operating Characteristic (ROC) curve analysis further revealed that both SUPT16H and SSRP1 expression levels were highly sensitive and specific biomarkers to distinguish HCC from NT livers (figure 1G and online supplementary figure S1H). Clinicopathologically, overexpressions of SUPT16H and SSRP1 were significantly associated with higher HCC histological grades (online supplementary figure S1I). However, there was no correlation between SUPT16H and SSRP1 expression and patients’ aetiological factors, suggesting that overexpression of FACT complex is a common event of human HCC (online supplementary figures S2 and 3). Moreover, high FACT complex expression was associated with poor overall survival and disease-free survival of patients with HCC in the TCGA cohort (figure 1H). Taken together, we showed that the two subunits of histone chaperone FACT complex, SUPT16H and SSRP1, were frequently upregulated in human HCC and the upregulation of FACT complex is a potential biomarker for human HCC detection and prognostication.

Figure 1

Frequent upregulation of histone chaperone FACT complex in human hepatocellular carcinoma (HCC). (A) Expression of 21 histone chaperones in human HCC determined by RNA-sequencing. The majority of histone chaperones were increasingly expressed in human HCC. (B) The expression profile of histone chaperones completely segregated 16 human HCC from their corresponding non-tumorous (NT) tissues. (C) SSRP1 and SUPT16H were remarkably upregulated in human HCC. (D) mRNA expression levels of SUPT16H and SSRP1 were significantly upregulated in 16 pairs of human HCCs. (E) Correlation between SUPT16H and SSRP1 expression in human HCC and their corresponding NT liver samples. (F) Upregulation of SUPT16H and SSRP1 in human HCC at protein level. (G) ROC analysis revealed that SUPT16H and SSRP1 mRNA expression levels were specific and sensitive biomarkers for HCC detection. (H) Analysis of TCGA data revealed that high FACT expression was associated with poor overall and disease-free survival in patients with HCC.

Overexpression of FACT complex promoted HCC proliferation and migration

To investigate the oncogenic consequences of FACT complex overexpression in human HCC, we employed a novel CRISPR/Cas9 SAM (Synergistic Activation Mediator) system to simultaneously activate the endogenous expression of SUPT16H and SSRP1 in human HCC cell line MHCC97L (figure 2A). Overexpression of FACT complex dramatically increased HCC cell proliferation rate and colony formation ability (figure 2B,C). In addition, upregulation of FACT complex accelerated HCC cell migration in transwell assay (figure 2D). Furthermore, in vivo tumourigenicity experiments demonstrated that overexpression of FACT complex significantly promoted HCC growth (figure 2E). We noted that single activation of SUPT16H or SSRP1 could also moderately promote HCC growth and migration while the effects were less prominent when compared with double activation (online supplementary figures S4 and 5). These observations were consistent with the notion that SUPT16H and SSRP1 are both required for the function of FACT complex. Together, our findings suggested that FACT complex may function as an oncogenic complex in human HCC.

Figure 2

Overexpression of FACT complex significantly promoted hepatocellular carcinoma (HCC) growth in vitro and in vivo (A) Simultaneous activation of endogenous SUPT16H and SSRP1 in MHCC97L by CRISPR/dCas9 gene activation system. Upregulation of FACT complex was confirmed at both protein and mRNA levels. (B–D) Overexpression of FACT complex significantly promoted cell proliferation, colony formation and cell migration in MHCC97L cells. (E) Overexpression of FACT complex significantly promoted HCC tumour growth in the subcutaneous implantation nude mice model. The size of tumour formed was monitored every 2–3 days. Tumour weight was detected at the end of the experiment. Overexpression of FACT complex drastically promoted in vivo tumour growth. *p<0.05; **p<0.01; ***p<0.001. FACT OE, FACT complex overexpressing cells; NTC, non-target control.

Knockout of either subunit of FACT complex led to the degradation of another component and suppressed HCC development

To consolidate our findings, we further employed a CRISPR/Cas9 gene editing system with two independent sgRNAs to knock out SUPT16H or SSRP1 in MHCC97L and Huh-7 cells. Successful gene knockout was confirmed by Western blotting (figure 3A and online supplementary figure S6A). Interestingly, we found that knockout of either SUPT16H or SSRP1 completely ablated the protein expression of the targeted subunit and also resulted in loss of the other untargeted subunit of the FACT complex (figure 3A and online supplementary figure S6A). However, the mRNA expression levels of the untargeted subunits remain unchanged, implying that the regulation is at the protein level (figure 3B). In addition, partial knockdown of SUPT16H or SSRP1 did not result in a drastic depletion of the untargeted subunit (online supplementary figure S7A and 8A). This observation suggested that the protein stability of FACT complex is tightly regulated and the interaction between SUPT16H and SSRP1 is important to stabilise the FACT complex. Loss of one component will lead to the protein degradation of the other (figure 3C).

Figure 3

Knockout of FACT complex suppressed hepatocellular carcinoma (HCC) growth in vitro and in vivo. (A and B) Knockout of SUPT16H or SSRP1 in MHCC97L cells by CRISPR/Cas9 gene knockout system. Knockout effect was verified at protein levels in sgSUPT16H #2, #3 cells and sgSSRP1 #1, #3 cells. Knockout of SUPT16H or SSRP1 led to the degradation of the other untargeted subunit at the protein levels without affecting their mRNA expressions. (C) A schematic diagram illustrating the degradation of FACT complex in SUPT16H or SSRP1 knockout cells. (D–F) Knockout of SUPT16H or SSRP1 significantly suppressed HCC cell proliferation, migration and colony formation in MHCC97L. (G and H) Knockout of SUPT16H suppressed HCC tumour growth and lung metastasis in vivo orthotopic implantation model as evident by bioluminescence imaging. (I) Tumour volume in the liver was measured by a calliper. **p<0.01.

We further asked whether loss of the FACT complex would affect the tumourigenicity of HCC cells. We found that depletion of FACT complex, by either knockdown or knockout of SUPT16H and SSRP1, dramatically suppressed HCC cell proliferation and colony formation ability. In addition, depletion of FACT complex substantially inhibited cell migration ability of HCC cells (figure 3D–F, online supplementary figures S6B–D, 7B–D, 9). Furthermore, orthotopic liver implantation was performed to examine the effect of FACT complex depletion in in vivo HCC tumourigenesis. We found that loss of FACT complex abolished HCC growth in the liver and also reduced the incidence of lung metastasis (figure 3G–I and online supplementary figure S10). Consistently, similar results were obtained from SUPT16H and SSRP1 stable knockdown cells, although less prominent when compared with knockout cells (online supplementary figure S11 and 12). Our knockdown, knockout and overexpression models suggested that both SSRP1 and SUPT16H are vital components of the FACT complex, which plays pivotal roles in promoting HCC growth and metastasis.

Positive feedback regulatory loop between NRF2 and FACT complex

Next, we investigated the possible molecular mechanisms that drive the upregulation of FACT complex in HCC. In the expression analysis, we noted that the expression levels of SUPT16H and SSRP1 were strictly correlated with each other in human HCC and NT livers (figure 1E), suggesting that they might be co-regulated by the same set of transcriptional factors. We therefore analysed the promoter sequences of SUPT16H and SSRP1 genes and public available transcription factors ChIP data. In silico prediction showed that SUPT16H and SSRP1 shared more than half (95 out of 156 for SUPT16H; 95 out of 138 for SSRP1) of common transcription factors (online supplementary figures S13). Among the 95 candidates, we are particularly interested in NRF2 and BACH1 (BTB Domain and CNC Homolog 1). NRF2 is a transcription factor and a central regulator of oxidative stress response pathway through binding the ARE (antioxidant response element) in the promoters of many antioxidant genes to activate their transcription.24 BACH1 is a transcription repressor that competes with NRF2 for the ARE binding sites and therefore negatively regulates the expression of antioxidant genes.25 We speculated that both SUPT16H and SSRP1 are transcriptional targets of NRF2 and BACH1. In supporting this notion, we identified putative ARE binding sequence (ATAC—AGCA) in the promoter regions of SUPT16H and SSRP1 genes (figure 4A). The binding of NRF2 and BACH1 to the promoter regions of SUPT16H and SSRP1 was further confirmed by ChIP assay, which showed a higher NRF2 binding than BACH1 binding in MHCC97L (figure 4B). To validate the regulatory functions of NRF2 to SUPT16H and SSRP1 expression, we generated a NRF2 stable knockdown cell line. We showed that the SUPT16H and SSRP1 expression, at both mRNA and protein levels, were significantly reduced upon NRF2 knockdown (figure 4C,F and online supplementary figure S14). We also found that treatment of SFN (sulforaphane), a NRF2 activator,26 effectively activated the expression of SUP16H and SSRP1 in a dose-dependent manner (figure 4D). In addition, we showed that the expression of NRF2, SUPT16H and SSRP1 was gradually elevated on challenge of the HCC cells with increasing dose of an oxidative stress–inducing agent tBHP (tert-butyl hydroperoxide solution; figure 4E). These findings suggested that oxidative stress induced FACT complex expression through NRF2. Moreover, loss of FACT complex caused internal reactive oxygen species (ROS) accumulation in HCC cells (figure 4G). Proliferation of FACT complex depleted cells could be partially rescued by antioxidant N-acetyl-cysteine (figure 4H), suggesting FACT complex is involved in oxidative stress response. Interestingly, we found that depletion of FACT complex drastically suppressed the SFN and tBPH induced NRF2, SUPT16H and SSRP1 expressions in HCC cells (figures 4I, 5A). These data converged to show that FACT complex is transcriptionally regulated by NRF2 and feed-forward to augment NRF2 mRNA expression under oxidative stress.

Figure 4

NRF2 regulates FACT complex expression at transcriptional level. (A) The promoter regions of SUPT16H and SSRP1 contain an ARE consensus sequence. (B) ChIP assay using NRF2, BACH1 and IgG antibodies in MHCC97L cells confirmed the binding of NRF2 and BACH1 to both SUPT16H and SSRP1 promoters. (C) mRNA expression of SUPT16H and SSRP1 decreased in NRF2 knockdown cells. (D–E) Treatment of oxidising agent SFN and tBHP induced SUPT16H and SSRP1 expression in a dose-dependent manner. (F) Decreased expression of SUPT16H and SSRP1 protein in NRF2 knockdown cells before and after tBHP treatment (100μM for 6hrs). (G) Knockout of SUPT16H increased intercellular ROS in MHCC97L cells. (H) Antioxidant NAC rescued the proliferation ability of SUPT16H knockout cells. (I) NRF2 expression induction by SFN treatment (2.5μM for 16 hrs) was blocked upon FACT complex depletion.

Figure 5

KEAP1 regulates SSRP1 protein degradation and stabilisation. (A) SSRP1 protein expression was lost in SUPT16H knockout cells. SSRP1 protein can be stabilised by proteasome inhibitor MG132. (B) SSRP1 stabilisation by MG132 was also confirmed by immunofluorescence staining. SSRP1 expression transferred from cytoplasm into nucleus after treated with MG132. (C) Similarly, SUPT16H protein expression was lost in SSRP1 knockout cells and the degradation of SUPT16H can be inhibited by proteasome inhibitor MG132. (D) SUPT16H stabilisation by MG132 was also confirmed by immunofluorescence staining. SUPT16H mainly expressed in the nucleus. (E) Ubiquitin pull-down assay confirmed the SSRP1 ubiquitination in wildtype cells. Ubiquitinated SSRP1 largely increased on SUPT16H knockout. (F–G) SSRP1 protein degradation is largely blocked by H2O2 and SFN. Cells were treated with 2.5μM SFN for 16 hrs and 20μM H2O2 for 4 hrs (H) Knockout of KEAP1 increased SSRP1 expression in wildtype MHCC97L cells and prevented SSRP1 from degradation when SUPT16H was knocked out. (I) SSRP1 ubiquitination decreased in KEAP1 knockout cells. (J) SSRP1 was stabilised on KEAP1 knockout and translocated from cytoplasm into nucleus in KEAP1-deficient cells.

Stabilisation of SSRP1 degradation by oxidative stress and KEAP1 eliminate

As mentioned earlier, knocking out of either subunit of FACT complex led to complete loss of the entire FACT complex. We reasoned that this is likely due to protein degradation. To test this hypothesis, we first treated knockout cells with MG132, a specific proteasome inhibitor that can reduce the degradation of ubiquitin-conjugated proteins.27 Results showed that the protein of untargeted subunit was remarkably stabilised upon MG132 treatment, which was confirmed by both Western blot and immunofluorescence staining (figure 6A–D). Next, we performed polyubiquitin pull-down assay to confirm if the stability of SUPT16H and SSRP1 is regulated by protein ubiquitination. We found that more ubiquitinated SSRP1 was detected in SUPT16H knockout cells compared with wildtype cells (figure 6E), which further indicated that loss of untargeted SSRP1 proteins was a ubiquitin–proteasome system–regulated process.

Figure 6

FACT complex is indispensable for oxidative stress response. (A) tBHP treatment (100μM for 6hrs) induced the expression of NRF2 and its downstream antioxidant genes NQO1, TXNRD1 and TKT expression in wildtype MHCC97L cells. Knockout of SUPT16H abolished the induction of these genes. (B) Knockout of SUPT16H reduced cell survival of MHCC97L under oxidative stress induced by tBHP treatment. (C) Overexpression of FACT complex sensitised the response of MHCC97L towards tBHP treatment–induced antioxidant gene expression. Cells were treated with 50μM tBHP for 4 hrs (D) Curaxin treatment diminished the induction of NRF2, NQO1, TXNRD1 and TKT expression on oxidative stress. Cells were treated with 100 μM tBPH and/or 100nM Curaxin for 24 hrs. (E) Curaxin treatment also inhibited the expression of NRF2, NQO1, TXNRD1 and TKT in in vivo subcutaneous tumour xenografts

KEAP1 and NRF2 are the master regulators of cellular oxidative stress response. KEAP1 is an E3-ligase of NRF2, which mediates NRF2 ubiquitination and proteasomal degradation. However, oxidative stress inactivates KEAP1 activity and allows NRF2 to escape from protein degradation.28 Considering the functions of FACT complex in oxidative stress response, we asked whether KEAP1 may also regulate FACT complex protein stability in a similar fashion as it regulates NRF2. We treated wildtype and SUPT16H knockout cells with tBHP, H2O2 and SFN. We demonstrated that the protein degradation of SSRP1 in SUP16H knockdown cells could be moderately but significantly rescued by oxidative stress-inducing agent H2O2 and tBHP (figure 6F and online supplementary figures 15, 19). In addition, SFN notably stabilised SSRP1 protein expression in SUPT16H knockout cells (figure 6G). In view of the function of SFN in blocking KEAP1 activity, we further hypothesised that E3 ligase KEAP1 may play an essential role in regulating SSRP1 protein ubiquitination. To this end, we used CRISPR Alt-R system to knock out KEAP1 in wildtype and SUPT16H knockout cells and evaluated the effects of KEAP1 knockout in SSRP1 ubiquitination and degradation. In wildtype cells, knockout of KEAP1 led to increased SSRP1 protein expression (figure 6H) and decreased ubiquitin level of SSRP1 (figure 6I). Interestingly, we found that SSRP1 protein translocated from cytoplasm into the nucleus upon KEAP1 knockout (figure 6J), as NRF2 does. More dramatically, knockout of KEAP1 rescued the SSRP1 protein expression in SUPT16H knockout cells (figure 6H). Taken together, SSRP1 ubiquitination in both wildtype and SUPT16H knockout cells was regulated by E3 ligase KEAP1. To conclude, the above data showed that NRF2 and FACT complex are regulated by KEAP1-mediated protein degradation.

Role of FACT complex in oxidative stress response

We further asked whether FACT complex plays a role in oxidative stress response and loss of FACT complex would have impact on the expression of antioxidant genes. We demonstrated that depletion of FACT complex drastically abolished the induction of NRF2, and its downstream antioxidant genes NQO1 (NADPH Quinone Dehydrogenase 1), TXNRD1 (Thioredoxin Reductase 1) and TKT (Transketolase; figure 5A), which are essential for cancer cells to survive under oxidative stress. Consequently, depletion of FACT complex created a vulnerability towards oxidative stress, and treatment of tBHP caused excessive growth reduction in FACT complex depletion cells (figure 5B). On the other hand, overexpression of FACT complex significantly augmented the expression of antioxidant genes on the challenge of tBHP treatment (figure 5C). In line with the above observation, we further demonstrated that treatment of Curaxin (CBL0137), a FACT complex inhibitor, reduced the oxidative stress–induced NRF2, NQO1, TXNRD1 and TKT expression in in vitro cell culture treated with tBHP (figure 5D) as well as in vivo subcutaneously implanted tumour growth in nude mice (figure 5E). Collectively, our findings suggested that FACT complex plays an indispensable role in oxidative stress response by regulating the expression of antioxidant genes.

Loss of FACT complex reduced transcription elongation rate of NRF2 and antioxidant genes in HCC

Mechanistically, FACT complex functions as a histone chaperone that facilitates nucleosome assembly and disassembly. We reasoned that FACT complex may play a critical role to modulate the transcription elongation of antioxidant genes under oxidative stress response by enabling the rapid passage of RNA polymerase II through the chromatin for expeditious mRNA transcription (figure 7A). To further validate this notion, we performed 4sU-DRB transcription elongation assay to assess the effect of FACT complex on the transcriptional elongation of antioxidant genes (online supplementary figure S16). Within a short incubation time (0, 4 and 8 min), the amount of 4sU-labelled RNA in the middle of the target (~1 to 4 kb from the transcription start site), which directly reflects the transcription elongation rate, was determined. Here, we demonstrated that knockout of FACT complex and Curaxin treatment substantially reduced 4sU incorporation rate of NRF2 at a distance-dependent manner, implying a deaccelerated transcription elongation rate (figure 7B). Moreover, knockout of FACT complex and Curaxin treatment also decelerated the transcription elongation rate of other antioxidant genes, such as TKT, NQO1, and TNXRD1 (figure 7C,D). Taken together, when FACT complex is absent, RNA polymerase II passage is blocked and transcription elongation rate in HCC cells was drastically impeded (figure 7A). In response to oxidative stress, cancer cells elicit rapid transcription of antioxidant genes to cope with the stress. When FACT complex is depleted, transcription elongation rate of those antioxidant genes would be limited. Cancer cells are therefore no longer able to survive oxidative stress, consequently leading to excessive cell death.

Figure 7

Loss of FACT complex reduced transcription elongation rate of NRF2 and antioxidant genes in hepatocellular carcinoma. (A) A schematic diagram illustrating the function of FACT complex in nucleosome disassembly for passage of polymerase ll during transcription elongation. (B–C) Transcriptional elongation rate was determined by 4sU-DRB assay. Oxidative stress response genes NEF2, NQO1, TXNRD1 and TKT showed decreased transcription elongation rate on SUPT16H knockout or Curaxin treatment.

Inactivation of FACT complex sensitised the effect of sorafenib treatment

After confirming the oncogenic function of FACT complex, we tested the therapeutic potential of Curaxin that binds to DNA, thereby trapping FACT complex in chromatin.29 We showed that HCC cells were highly sensitive to Curaxin treatment, whereas immortalised hepatocyte line MIHA was more tolerant to Curaxin compared with HCC cell lines (figure 8A,B), indicating that Curaxin selectively suppressed cancer cell growth. We further examined the in vivo effect of Curaxin on HCC tumourigenesis. Mice were subcutaneously injected with MHCC97L cells and treated with Curaxin on a 5-day-on/2-day-off pace. After 2 weeks, the group of mice under Curaxin treatment showed significant reduction in tumour size (figure 8C and online supplementary figure S17), suggesting FACT complex is a potential target for HCC therapy. We previously reported that sorafenib, an FDA-approved tyrosine kinase inhibitor for the first-line treatment of patients with advanced HCC, significantly increased ROS levels in HCC cells and the cancer cells activate antioxidant pathways to counteract.22 23 Given the critical role of FACT complex in oxidative stress response, we ask whether inhibition of FACT complex could improve the efficacy of sorafenib treatment. To this end, we showed that FACT complex depleted cells were more sensitive towards sorafenib treatment when compared with the wildtype controls (figure 8D). In addition, when compared with single treatments, Curaxin and sorafenib co-treatment exhibited a more potent effect in suppressing HCC proliferation and inducing cell apoptosis (figure 8E,F). In line with the in vitro data, we found that co-treatment of Curaxin and sorafenib achieved a better tumour inhibition effect in nude mice model (figure 8G). Moreover, Curaxin and sorafenib treatments are well tolerated in mice (online supplementary figure S18), highlighting the translational value of this new therapeutic target for patients with HCC. Together, our findings established the therapeutic potential of FACT complex inhibitor Curaxin in HCC treatment as single agent or as combined agent with sorafenib.

Figure 8

Curaxin and FACT complex knockout sensitised the effect of sorafenib treatment. (A) Curaxin effectively inhibited MHCC97L and Huh7 proliferation at low concentration. (B) Hepatocellular carcinoma cell lines MHCC97L and Huh7 were more sensitive towards Curaxin treatment compared with immortalised liver cell line, MIHA. (C) Curaxin treatment (5 days on, 2 days off, by oral gavage for 2 weeks) significantly reduced MHCC97L tumour growth in subcutaneous implantation. (D) SUPT16H knockout reduced cell survival of MHCC97L on sorafenib treatment. (E) Co-treatment of Curaxin and sorafenib induced cell apoptosis. (F) Co-treatment with Curaxin enhanced the antiproliferation effects of sorafenib in both MHCC97L and Huh7 cell lines. (G) Mice treated with both sorafenib and Curaxin for 2 weeks significantly reduced growth rate when compared with single treatments.

Discussion

Histone chaperones are a diverse group of proteins that bind to and escort histones during nuclear transport. They also regulate nucleosomal histone incorporation, exchange and displacement in chromatin.30 Histone chaperones work hand in hand with histone modification enzymes and chromatin remodellers to control the dynamic chromatin organisation in eukaryotic cells. Over the past decade, overwhelming evidence have indicated that mutation and deregulation of histone-modifying enzymes and chromatin remodellers are critically involved in genomic instability and aberrant transcriptional programmes in human cancers. However, the implications of histone chaperones in human carcinogenesis received very little attention. In this study, we determined the expression changes of various histone chaperones and identified a frequent deregulation of FACT complex in HCC. FACT complex possesses intrinsic histone chaperone function and is a critical component for gene transcription. It acts after the binding of transcriptional factors to overcome the nucleosome obstacles, which enables the rapid passage of RNA polymerase during transcriptional elongation. When a transcribing RNA polymerase approaches, SSRP1 subunit holds the H3–H4 tetramer and SUPT16H subunit removes one of the two H2A–H2B dimers from the nucleosome to form a transient hexasome structure that facilitates the passage of RNA polymerase. After the passage of RNA polymerase, FACT complex disposes the original H2A–H2B dimer back into place.31 This dual nucleosome disassembly and reassembly functions of FACT complex are important to enable a rapid transcriptional elongation and at the same time preserve the histone modification signature and maintain chromatin integrity of actively transcribing genes.

The expression of FACT complex is associated with cell differentiation stages. Several recent studies also reported that FACT complex is abundantly expressed in human cancers.32 In this study, we first unveiled that histone chaperone FACT complex, with its two subunits, SUPT16H and SSRP1, were frequently upregulated in HCC and associated with poor prognosis of patients with HCC. Using CRISPR-based gain-of-function and loss-of-function models, we demonstrated that FACT functioned as an oncogenic complex to promote HCC growth and metastasis. A small molecular inhibitor of FACT complex, Curaxin, is emerging as a promising anticancer agent in different cancer models.33 Curaxin was first identified from a chemical library screening as a compound that can simultaneously inhibit NF-κB (nuclear factor kappa B) and activate P53 (tumour protein 53). It is believed that Curaxin binds to DNA and ‘traps’ FACT complex in the chromatin, while the mechanism of action of Curaxin molecule remains to be further investigated.34 In this study, we also showed that Curaxin treatment effectively suppressed HCC tumour growth in both in vitro and in vivo models. Additionally, Curaxin also sensitised HCC cells to sorafenib. These findings demonstrated the therapeutic potential of Curaxin, as a single treatment or in combination with sorafenib, for the treatment of patients with HCC.

In this study, we found that the interaction between SUPT16H and SSRP1 is essential for the stabilisation of FACT complex. In our knockout models, we found that knockout of SUPT16H led to the degradation of the untargeted SSRP1 protein and vice versa. This phenomenon was very prominent in knockout HCC cells but could have been missed in RNAi knockdown models as the residual SUPT16H and SSRP1 proteins would be sufficient to stabilise the FACT complex. SUPT16H and SSRP1 subunits are mutually dependent, and consistent with this notion, the expression of SUPT16H and SSRP1 was found to be strikingly correlated with each other in primary HCC and NT livers. The degradation of the untargeted SUPT16H or SSRP1 subunit appeared to be an event independent of their mRNA expression. Interestingly, a recent study showed that SUPT16H protein in yeast was regulated by an E3 ligase san1 and followed by proteasomal degradation.35 San1 homolog does not exist in the human genome, and we unprecedentedly showed that the degradation of SSRP1 in human HCC was regulated by another E3 ubiquitin ligase KEAP1 and plays a critical role in KEAP1-mediated oxidative stress response pathway. Our findings suggested that the post-translational regulation of FACT complex might have important biological implications.

It has previously been reported that FACT complex is important for NF-κB signalling. Knockdown of either SUPT16H or SSRP1 decreased the expression of NF-κB downstream genes.34 However, the underlying molecular mechanisms by which FACT complex promoted cancer progression remained unknown. In this study, we demonstrated that FACT complex played a central role in oxidative stress response. Oxidative stress elicited by the accumulation of ROS due to altered cellular metabolism is a common feature of cancer cells. Low level of ROS may cause oxidative DNA damage that eventually increases the likelihood of DNA mutation and cancer development. However, excessively high level of ROS could be detrimental to the cancer cells and induce apoptosis. Therefore, a higher antioxidant-producing capacity is necessary to enable cancer cells to survive oxidative stress.28 The KEAP1/NRF2 pathway is the key pathway providing defence against oxidative stress. Under normal conditions, KEAP1 negatively regulates NRF2 protein stability via ubiquitin–proteasome degradation pathway. Under oxidative stress, KEAP1 is inactivated and thus relieving NRF2 from degradation. Stabilised NRF2 translocates into the nucleus, where it binds to ARE and displaces BACH1 from ARE to activate an array of antioxidant genes to balance the intercellular ROS level. KEAP1/NRF2 is one of the most frequently mutated pathways in human HCC and associated with drug resistance.36 A recent study in a hepatocarcinogen-induced HCC mouse model showed that Nrf2 mutations were present in all stages of HCC.37 Here, we demonstrated that the expression of SUPT16H and SSRP1 was controlled by NRF2 in response to oxidative stress. Moreover, as mentioned above, the protein stability of FACT complex is regulated by KEAP1 through a mechanism similar to NRF2. Thus, elevated ROS level, on one hand, increased FACT complex expression at mRNA level and on the other hand stabilised the FACT complex at protein level. Upregulation of FACT complex, in turn, accelerates the transcription elongation of NRF2 and its downstream antioxidant genes. This positive feedback loop between NRF2 and FACT complex further increases expression of antioxidant genes to counteract excessive ROS. In addition, analysis of TCGA data revealed that NRF2 and KEAP1 were recurrently mutated in human HCC. Therefore, future studies should address whether HCC with NRF2 and KEAP1 mutations is more sensitive towards FACT complex inhibition treatment in preclinical and clinical settings.

How cancer cells respond to environmental cues remains a central question of cancer biology. Signal transduction and transcriptional control are believed to be the two critical underlying mechanisms. Here, we disclosed that oxidative stress activates the expression of FACT complex through NRF2/KEAP1 pathway. Upregulation of FACT complex facilitates transcriptional elongation of antioxidant genes via its histone chaperone functions. This mechanism enables cancer cells to respond to oxidative stress rapidly and efficiently, therefore maintaining cancer cell survival during tumour development and under sorafenib treatment (figure 9). Our study exemplifies the critical role of histone chaperone FACT complex in human HCC and provides a molecular connexion between metabolic and transcriptional regulation in carcinogenesis.

Figure 9

Summary of pathological and underlying regulatory mechanisms of FACT complex in hepatocellular carcinoma.

Methods

HCC cell line

Human hepatoma cell lines HepG2 and Huh-7 as well as human embryonic kidney cell line 293FT were obtained from ATCC. HCC cell line MHCC97L is a gift from Dr Z Y Tang (Fudan University, Shanghai, China). Immortalised hepatocyte cell line MIHA is a gift from Shanghai Institute of Cell Biology, Chinese Academy of Sciences. All cell lines except HepG2 were cultured in DMEM supplemented with 10% fetal bovine serum (FBS). HepG2 was cultured in MEM containing 10% FBS.

Establishment of SUPT16H and SSRP1 knockout cells

Knockout of SUPT16H and SSRP1 was established by CRISPR gene editing system obtained from Professor Zhang Feng (Board Institute, USA). MHCC97L and Huh-7 cells were first engineered to express Cas9. Two guide RNAs were then delivered to target SUPT16H or SSRP1, respectively. Specific single-guided RNAs were cloned and packaged into 293FT. Luciferase-labelled Cas9-expressing MHCC97L and Huh-7 were then infected with lentivirus and selected by puromycin. Knockout efficiency was confirmed at protein level by Western blot.

Subcutaneous injection and orthotopic implantation models

For subcutaneous injection model, 2×106 SUPT16H activation, SSRP1 activation, FACT complex activation cells and their corresponding control cells was injected subcutaneously into different sides of 4-week-old BABL/c nude mice. After 4 weeks, mice were sacrificed. In vivo tumour growth was measured every 3–5 days. For orthotopic implantation, 1×106 of luciferase-labelled SUPT16H knockdown, SSRP1 knockdown, SUPT16H knockout MHCC97L cells was injected into the left liver lobe of male BALB/c nude mice 6–7 weeks old. After 5 weeks, mice were sacrificed. In vivo tumour growth and lung metastasis were detected with IVIS 100 Imaging System (Xenogen). For drug treatment, 2 weeks after the tumour implantation, mice were administered with (1) vehicle with DMSO, (2) 30 mg/kg sorafenib every day for 2 weeks and/or (3) 20 mg/kg Curaxin 5 days on, 2 days off for 2 weeks, through oral gavage. After 4 weeks, mice were sacrificed. Tumour sizes were measured with a calliper and calculated with the formula 0.5×length×width×width. Animal experiments were all carried out according to the Control of Hong Kong Animals Experiment Ordinance and the Institute’s principles on animal works.

4sU-DRB transcription elongation assay

5,6-Dichlorobenzimidazole 1-β-d-ribofuranoside (DRB, D1916; Sigma-Aldrich) was used at a final concentration of 100 µM to stop transcription. Cells were incubated in DRB for 3 hours. 4-Thiouridine (4sU, T4509; Sigma-Aldrich) was used at a final concentration of 1 mM to label newly synthesised RNA. Total RNA were extracted 0 min, 4 min and 8 min after DRB release. Total RNA was extracted with the miRNeasy kit (217004; Qiagen). Biotinylation and purification of 4sU-labelled RNA was done as described previously.38 4sU-labelled RNA was biotinylated using EZ-Link Biotin-HPDP (21341; ThermoFisher Scientific) dissolved in dimethylformamide (20673; ThermoFisher Scientific). Biotin-labelled RNA was then captured using Dynabeads MyOne Streptavidin T1 beads (65601; Life Technologies). Beads were magnetically fixed and cleaned up. 4sU-labelled RNA was eluted with 100 µL of freshly prepared 100 mM dithiothreitol (20291; ThermoFisher Scientific) and cleaned on RNeasy MinElute Kit (Qiagen). Eluted RNA samples were used for qRT-PCR with specific primers listed in online supplementary table S3.

Chemicals and drug dilution

Curaxin (19110; Cayman) was dissolved in DMSO for in vitro and in vivo use. Drugs were further diluted in physiological saline for in vivo studies. The volume was adjusted so that a 20 g mouse would receive a 20 mg/kg dose of Curaxin in 200 µL. Sorafenib (LC laboratory) was dissolved in DMSO and further diluted in physiological saline to reach a concentration that a 20 g mouse would receive a 30 mg/kg dose of sorafenib in 200 µL. For co-treatment group, Curaxin and sorafenib were diluted together so that a 20 g mouse would receive a 30 mg/kg dose of Curaxin and 30 mg/kg dose of sorafenib in 200 µL. SFN (S6317; Sigma) and MG132 (M7449; Sigma) were diluted in DMSO. tBHP (458139; Sigma) and H2O2 (386790; CalBioChem) were diluted in phosphate-buffered saline for in vitro use.

Statistical analysis

Statistical analysis was performed using GraphPad Prism V.7 software. Data are presented as mean±SD. Student’s t-test was used to compare parametric continuous variables. Correlation between genes was analysed by linear regression. Kaplan-Meier curve was used to compare HCC patient survival. P values less than 0.05 were considered statistically significant.

Detailed methodology can be found in online supplementary materials and methods.

Acknowledgments

We thank the Core Facility and Centre for Genomic Sciences of LKS Faculty of Medicine for their technical support. We also thank the Laboratory Animal Unit for animal housing.

References

View Abstract

Footnotes

  • Contributors Conceptualisation: JS, CC-LW, CMW. Methodology: JS, C-TL, CC-LW, CMW. Formal analysis: JS, C-TL, CMW. Investigation: JS, MC, DL, LW, FH-CT, DC-WC, CL-HC, JL, CC-LW, CMW. Resource: IOLN, CC-LW, CMW. Funding acquisition: IOLN, CC-LW, CMW. Writing original draft: JS, CC-LW, CMW. Writing—review and editing: JS, CC-LW, CMW. Supervision: CC-LW, CMW.

  • Funding This study was funded by RGC-TBRS 2016 (T12-704/16-R), Health and Medical Research Fund 2018 (05161786).

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article or uploaded as online supplementary information.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles