Luks-pv Induces Apoptosis Through Methyltransferase Set8 in Human Acute Myeloid Leukaemia Cells


 Background We previously reported that LukS-PV induces apoptosis in human acute myeloid leukaemia (AML) cells. Furthermore, SET8 is a member of the SET domain-containing methyltransferase family and overexpressed in numerous tumours, including AML and indicated a poor prognosis. However, it is unclear whether LukS-PV induce apoptosis via SET8. In current study, we aimed to investigate the regulatory mechanisms of SET8 and effects on AML cells under the LukS-PV.Methods Flow cytometry was performed to detect apoptosis in AML primary blasts/cell lines treated with LukS-PV and invasion of AML cells in vivo. The expression of SET8 was quantified through reverse-transcription PCR and western blotting. Chromatin-immunoprecipitation (ChIP) sequencing and bioinformatic methods was performed to explore transcriptional target genes which regulated by SET8-H4K20me1 axis. All experiments were performed in triplicate, and all statistical analyses were conducted using SPSS 16.0.Results In this research, we found that LukS-PV induce cell apoptosis in vitro and inhibit cell invasion in vivo. Our results further confirmed SET8 and H4K20me1 were downregulated in LukS-PV-treated AML cells. Furthermore, we confirmed that LukS-PV induced apoptosis via downregulating SET8/H4K20me1. Finally, Genome-wide analysis identified PIK3CB is the target gene for cell apoptosis mediated by SET8/H4K20me1. In a nutshell, LukS-PV induces apoptosis via the PIK3CB/PI3K/AKT/FOXO1 signal pathway by targeting SET8.Conclusions The present results indicate that LukS-PV induces apoptosis in AML cells by downregulating SET8 and regulating downstream molecular targets, suggesting that SET8 is a potential target for AML therapy using LukS-PV for anti-leukaemia treatment.


Introduction
Acute myeloid leukaemia (AML) is a heterogeneous clonal disorder characterized by immature myeloid cell proliferation and bone marrow failure and a short course [24]. Treatments for AML primarily include radiotherapy, chemotherapy, and hematopoietic stem cell transplantation; however, the 5-year survival rate remains low [17,32]. Thus, it is important to identify more targeted therapies for AML.
In recent years, bacterial toxins have received increasing attention for the development of novel anticancer drugs owing to their speci city and cytotoxicity in target cells, and some bacterial toxincontaining anticancer drugs have entered clinical trials [19,31]. Panton-Valentine leucocidin (PVL) is an extracellular toxin produced by Staphylococcus aureus and comprises two subunits, LukS-PV and LukF-PV [1]. Our previous study reported that LukS-PV induces apoptosis in AML cells mediated by C5aR and inhibit AML cell proliferation in vitro and in vivo, and have no toxic side effects [2,25,26,37]. SET8 (also known as PR-Set7/9, SETD8, and KMT5A) is a member of the SET domain-containing methyltransferase family, and its activity is a speci c to H4K20me1 [18]. SET8 is overexpressed in numerous tumours including hepatocellular carcinoma, gastric cancer, small-cell lung cancer, breast cancer, ovarian cancer and AML [13,27,29,34,38]. High SET8 levels are also associated with poor survival in AML patients. Thus, SET8 is considered therapeutic targets and should be further studied in AML.
In this study, we investigated the effects of LukS-PV through downregulate SET8 on apoptosis of AML cells and further explored its molecular mechanism of action.

Cell culture and Reagent
Human acute leukaemia cell lines HL-60 and NB-4 were obtained from the Shanghai Institute for Biological Sciences (Shanghai, China). Cells were cultured in RPMI-1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10% foetal bovine serum (FBS, HyClone, Logan, UT, USA) and 1% penicillin/streptomycin in incubator at 37 °C with 5% CO 2 . The medium was changed every 2-3 d.
Total RNA extraction of peripheral blood from AML patient and normal people AML patients were diagnosed in accordance with clinical and laboratory criteria. Add ve times the volume of erythrocyte lysate to fresh peripheral blood, place it on a shaker for 15-20 minutes. Then, centrifuge for 5 minutes,1000 rpm/min, discard the supernatant, wash bottom cells twice with PBS, and re-lyse according to the lysis. Total RNA was extracted using TRIzol (Invitrogen, USA) in accordance with the manufacturer's instructions. This study was approved by the Ethics Committee and Institutional

Recombinant LukS-PV production and puri cation
The pET28a vector (Roche Diagnostics Corp, Switzerland) was used to generate six recombinant Histagged LukS-PV proteins. The LukS-PV sequence was ampli ed from PVL-positive S. aureus isolates.
PCR products were digested with XhoI and BamHI (Promega, Madison, WI, USA) and ligated into the pET28a vector. Recombinant LukS-PV was puri ed as previously described [21].

Lentiviral transduction
The lentiviral vector used for SET8 silencing and overexpression (Hanbio, Shanghai, China) was transduced into HL-60 and NB-4 cells. As controls, lentiviral vectors containing short hairpin RNA sequences targeting a non-mammalian gene were used. After 48 h of transduction, the cells were screened using puromycin and cultured.

Flow cytometric analysis
To assess apoptosis, cells were harvested through centrifugation at 1000 rpm for 5 min, washed twice in cold PBS, resuspended in 500 µl staining buffer, and co-stained with Annexin V-PE and 7AAD (eBioscience, CA, USA) at room temperature for 15 min in the dark, and then analysed using a FACS Calibur ow cytometer (BD Biosciences, NJ, USA) in 1 h. For proliferation assay, cells were pre-treated with the kFlour647 Click-iT Flow detection kit (KeyGEN BioTECH, Jiangsu, China) in accordance with the manufacturer's instructions. The data were analysed by the FCS Express software.

Western blotting
Cells were lysed by RIPA with 1% PMSF (Beyotime, Shanghai, China) on ice for 30-60 min and then centrifuged at 12000 rpm for 5 min and the pellet was discarded. The protein samples and SDS were boiled for 15 min and the proteins were separated through SDS-PAGE and electro-transferred onto a 0.45µm nitrocellulose membrane (Millipore, Bedford, MA, USA). The membranes were blocked with Protein Free Rapid Blocking Buffer (EpiZyme, Jiangsu, China) and subsequently probed with primary antibodies. The anti-human SET8(Rabbit), PIK3CB(Rabbit), FOXO1(Rabbit), AKT(Rabbit), p-AKT(Mouse), BAK(Rabbit), Bcl2(Rabbit), H4K20me1(Rabbit), GAPDH(Mouse) and H4(Rabbit) antibodies were purchased from Cell Signaling Technology( Danvers, MA). Thereafter, the membranes were incubated with a horseradish peroxidase-conjugated secondary antibody for 1.5 h at room temperature. Immunoreactive bands were visualized using an enhanced chemiluminescence detection system.

Xenograft mice assay
Male BALB/c nude mice(4 weeks old) were obtained from nanjing China GemPharmatech. Ltd and maintained in a speci c pathogen-free facility at the Lab Animal Center of Anhui Medical University. Mice were injected intraperitoneally (i.p.) with cyclophosphamide (CTX, 100 mg/kg body weight) on each of 3 successive days and then randomized into 3 groups, including the normal control(5 mice), HL-60 group(10 mice) and NB4 group(10 mice). Mice in the normal control group received only PBS. The HL-60 group and NB4 group were injected 5*10^6 HL-60 or NB4 cells per mice via tail vein and then randomized into PBS group(5 mice) and LukS-PV group(5 mice) respectively. Then LukS-PV group mice were treat with LukS-PV (300 ug/kg body weight per mice) via tail vein for 3 successive days. After 30 days, mice were sacri ced and their spleens and peripheral blood samples were collected for next experiments. CD33, a myeloid lineage-speci c antigen, is a sialoadhesin family member that is normally expressed on precursor myeloid cells and it can be used as a speci c marker to observe leukemic cell proliferation and in ltration in a mouse leukemia model [7]. Hence, we used CD33-PE(BD, cat:555450) to assess invasion of AML cell in vivo via FCM. This study was approved by the Ethics Committee and Institutional Review

Statistical analysis
All data are expressed as mean ± standard deviation (SD) values, and all experiments were performed in triplicate. Statistical analyses were performed using t-tests or one-way ANOVA. All statistical analyses were conducted using SPSS 16.0 (SPSS Inc., Chicago, IL, USA). A value of p < 0.05 was considered statistically signi cant.

LukS-PV induce cell apoptosis in vitro and inhibit cell invasion in vivo
We randomly isolated bone marrow samples from 4 AML patients for in vitro culture and treated them with LukS-PV at different concentrations to detect apoptosis via ow cytometry, these results showed LukS-PV induces apoptosis in a dose-dependent manner in primary AML blasts (Fig. 1A). In order to further study the anti-leukemia activity of Luks-PV in vivo, we injected AML cell lines (HL-60 and NB4) into the tail vein of nude mice and treat with Luks-PV. The results showed that the spleen index of treatment group was lower than untreated group (Fig. 1B). Besides, FCM analysis showed the percentage of AML cells(CD33 + cells) in peripheral blood and spleen was lower in treatment group compare to untreated group ( Fig. 1C-D). These results indicated that Luks-PV not only induce AML cell apoptosis in vitro, but also inhibit cells invasion in vivo.

SET8 is upregulated in AML cells
We rst analyze AML patient peripheral blood RNA-seq data from TCGA database and normal people peripheral blood RNA-seq data from GTEx database [33] to explore the expression of SET8 between AML patient and normal people. The analysis results showed that the expression of SET8 mRNA in AML patients was signi cantly higher than normal people, and indicated a poor prognosis ( Fig. 2A-B). Then, we sampled peripheral blood from 20 AML patients and 20 healthy controls and quanti ed SET8 expression in peripheral blood leukocytes of the two groups to veri ed results of database. Consequently, RT-PCR and western blotting revealed that SET8 was signi cantly upregulated in AML patients than in the healthy controls ( Fig. 2C-D), especially in M3 AML (Fig. 2E). These results indicated that SET8 involved in the leukaemia pathogenesis and maybe a potential therapeutic target in AML.

LukS-PV can downregulate SET8 and H4K20me1
Since SET8 was upregulated in leukaemia cells and is associated with leukaemia pathogenesis, concurrent with a previous study reporting that LukS-PV induces apoptosis in AML cells. To investigate whether LukS-PV exerts anti-leukaemia effects by regulating SET8, we rst treated HL-60 and NB-4 cells with different concentrations of LukS-PV and quanti ed SET8 expression via RT-PCR and western blotting. Consequently, SET8 gene and protein were signi cantly downregulated after exposure with LukS-PV ( Fig. 3A-B). Further, we assessed SET8 gene and protein expression levels in LukS-PV-treated HL-60 and NB-4 cells at different timepoints, the results showed SET8 gene and protein were signi cantly downregulated after exposure with LukS-PV ( Fig. 3C-D). These results indicate that LukS-PV downregulates SET8 in a dose-and time-dependent manner.
As SET8 is a member of the SET domain-containing methyltransferase family that speci cally catalyses monomethylation of histone H4 Lys-20 (H4K20me1), so we also treated HL-60 and NB-4 cells with different concentrations of LukS-PV and quanti ed H4K20me1 expression via western blotting. Consequently, H4K20me1 levels were signi cantly lower in than those in control cells (Fig. 3E). Similarly, we determined H4K20me1 levels in LukS-PV-treated HL-60 and NB-4 cells at different timepoints, and our results were consistent with SET8 downregulation, and H4K20me1 levels were reduced (Fig. 3F). These results indicate that LukS-PV can downregulate H4K20me1 through SET8 in a dose-and time-dependent manner.

LukS-PV induces apoptosis in AML cells by downregulating SET8/H4K20me1
To determine whether LukS-PV exerts anti-leukaemia effects by downregulating SET8, we rst transfected HL-60 and NB4 cells with siRNAs to silence or overexpress SET8, and then quanti ed SET8 and H4K20me1 expression by RT-PCR and western blotting (Fig. 4A-B). Then, we assessed apoptosis via ow cytometry, the results showed both early and late apoptosis were signi cantly increased in set8 knockdown AML cell lines (Fig. 4C-D). Furthermore, we overexpressed or knocked down SET8 in AML cells and then treated the cells with 2.0 µM LukS-PV, followed by the assessment of apoptosis via ow cytometry. Consequently, the LukS-PV-treated cells displayed a signi cantly higher extent of apoptosis than the PBS control group, and this effect was further enhanced in SET8 knockdown cells but markedly alleviated in SET8-overexpressing cells. In other words, the effect of LukS-PV on apoptosis can be reversed by SET8 overexpression (Fig. 4E-F). In accordance, LukS-PV was able to decrease the levels of SET8, H4K20me1, Bcl-2 and increase protein levels of Bax, and this effect was further enhanced in SET8 knockdown cells but markedly alleviated in SET8-overexpressing cells (Fig. 4G-H). These results indicated that LukS-PV induced cell apoptosis in AML cells by downregulating SET8 expression.
PIK3CB is the target gene for cell apoptosis mediated by SET8 /H4K20me1 To further explore the molecular mechanism that LukS-PV induces apoptosis by downregulating SET8 in AML cells, we speculated whether it functions by regulating downstream target genes through H4K20me1. To verify this speculation, target genes regulated by LukS-PV-mediated H4K20me1 were determined through analysis of genome-wide transcriptional targets of LukS-PV through ChIP-sEq. ChIP experiments were rst performed with HL-60 cells with antibodies against H4K20me1 after LukS-PV treatment. Thereafter, H4K20me1-associated DNA sequences in LukS-PV-treated cells were ampli ed under nonbiased conditions, labelled, and sequenced. Through HiSeq2000 with a p-value cut-off of 10 − 5 , we identi ed 2450 H4K20me1-speci c binding peaks, of which 731 were upregulated and 1719 were downregulated (Supplement Table S1).
Since SET8 promotes the transcription of target genes through H4K20me1, and LukS-PV downregulates SET8, we focused on the downregulated peaks related genes binding by LukS-PV-mediated H4K20me1. GO-based biological process analysis indicated these genes to be signi cantly enriched in transcription coactivator activity, magnesium ion binding, phosphatidylinositol bisphosphate binding, potassium ion transmembrane transporter activity and so on (Fig. 5A). Furthermore, these genes were then classi ed into various cellular signalling pathways, using the Molecule Annotation System software with a p-value cut-off of 10 − 3 . These signalling pathways included protein export, axonal guidance, leukocyte transendothelial migration, and cAMP and Wnt pathways, which are critical for cell growth, survival, and apoptosis (Fig. 5B). As our previous study illuminated that LukS-PV induces apoptosis in AML cell through C5aR, and considering Yoshihiro Maeda et al. reported that C5aR plays a crucial role in invasion of metastatic renal cell carcinoma via PI3K pathways [35], we identi ed a potential gene PIK3CB as a target for LukS-PV induces cell apoptosis mediated by SET8 /H4K20me1 (Fig. 5C). Further, we veri ed this nding through ChIP-PCR in HL-60 and NB4-cells. Accordingly, upon LukS-PV treatment, the binding of H4K20me1 at the PIK3CB promoter was signi cantly reduced (Fig. 5D). Finally, we treated HL-60 and NB4 with LukS-PV at different concentrations, and PIK3CB gene and protein expression levels were further reduced in a dose-dependent manner on qRT-PCR and western blotting (Fig. 5E-F).
PIK3CB was identi ed as a downstream target of LukS-PV induces apoptosis by downregulating SET8/H4K20me1, hence, we speculated whether it plays a role in cell apoptosis. To investigate the biological function of PIK3CB, we tested its effect on apoptosis through inhibition with GSK2636771 (a PIK3CB inhibitor) in SET8-overexpressing cells. The ow cytometry results showed GSK2636771 induced apoptosis in control SET8-expressing cells; however, this effect was alleviated in SET8-overexpressing cells in both HL-60 and NB-4 cells (Fig. 5G-H). Together, our results indicate that LukS-PV induces cell apoptosis through downregulate expression of the target gene PIK3CB mediated by SET8 /H4K20me1 in AML cells.

LukS-PV induces apoptosis via the PIK3CB/PI3K/AKT/FOXO1 signal pathway by targeting SET8
It was reported that PIK3CB inhibits transcription factor FOXO1 by regulating AKT phosphorylation, and then inhibits apoptosis by regulating apoptotic proteins Bax and Bcl2. Therefore, we further veri ed this molecular mechanism by western blotting, and the results were in line with our expectations. We found that 2.0 µM LukS-PV treated cells had lower levels of PIK3CB, pAKT and anti-apoptotic Bcl2 but higher levels of FOXO1 and pro-apoptotic Bax than PBS-treated control cells in both HL-60 and NB4 cells. These effects were further enhanced in SET8 knockdown cells but markedly alleviated in overexpressed SET8 cells (Fig. 6A).
Similarly, we determined the levels of associated proteins in primary AML blasts via western blotting. In accordance, the result showed that 2.0 µM LukS-PV signi cantly decreased the levels of SET8, H4K20me1, PIK3CB, pAKT and anti-apoptotic Bcl2 but increased the levels of FOXO1 and pro-apoptotic Bax in comparison with the PBS-treated control group (Fig. 6B). Together, our results indicate that LukS-PV induces apoptosis via the PIK3CB/PI3K/AKT/FOXO1 signal pathway by targeting SET8 in primary AML blasts.
In conclusion, this study demonstrated that LukS-PV decreased SET8 expression and regulated the PI3K/AKT/FOXO1 signal pathway midated by SET8, thereby inducing apoptosis in AML cells (Fig. 6C).

Discussion
Bacterial toxins reportedly have speci c cytotoxic effects on target cells including tumour cells, and they have received increasing attention in the development of anti-rumour drugs. LukS-PV is the S component of PVL secreted by S. aureus. We previously reported that LukS-PV induces apoptosis in AML cells mediated by C5aR and inhibit AML cell proliferation in vitro and in vivo and have no obvious side effects on mice [2,25,26,37]. However, the precise mechanisms underlying histone modi cation regulation by LukS-PV exerts its anti-leukaemia effects in AML cells have remained poorly understood.
AML is complex disease with a diverse genetic landscape. Recent studies have revealed that epigenetic dysregulation plays an important role in leukaemia pathogenesis [6,11]. Further histone-modifying enzymes reportedly contribute to leukaemia pathogenesis. For example, MLL1 is a member of the SET family of methyltransferases and has been reported affect stem cell differentiation in leukaemia patients [10]. SETD2 is a histone methyltransferase that regulates leukaemia pathogenesis [30,39]. This phenomenon is precisely based on the reversibility of epigenetic modi cations may facilitate targeted leukaemia therapy [12]. For example, azacitidine and decitabine are DNA methyltransferase inhibitors approved for clinically treating AML [8,15]. EZH2 and LSD1 are reportedly involved in leukaemia pathogenesis, and these targeted drugs have entered clinical trials [14,22]. In summary, histone modi cations are potentially promising for targeted therapy for leukaemia.
SET8 is a member of the SET domain-containing methyltransferase family speci cally targeting H4K20me1 [18]. SET8 is involved in vital cellular processes, including transcriptional regulation [5,20], Sphase progression [28], genome replication and stability [4,16], and DNA repair [36]. Aberrant SET8 expression has been reported in numerous tumours. High SET8 levels are also associated with poor survival in cancer patients. Zhou et al reported that SET8 and H4K20me1 levels were higher in AML cells than in control cells, and that SET8 can regulate the Wnt signalling pathway and is associated with leukaemia pathogenesis [38]. Our observation showed that SET8 is downregluate targeted by LukS-PV revealed a mechanism underlying SET8 regulation, thus underscoring the importance of SET8 regulation in AML.
Previous studies have reported that SET8 is a speci c monomethylation catalytic enzyme for H4K20me1 [18]. Therefore, we further found that LukS-PV reduced H4K20me1 levels by downregulating SET8 in a dose-and time-dependent manner. To further assess the downstream mechanism of LukS-PV in exerting anti-leukaemia effects through the SET8-H4K20me1 axis, we performed ChIP-seq to identify H4K20me1 downstream target genes. Finally, a potential target gene PIK3CB was identi ed by classifying target genes, and we tested its effect on apoptosis through PIK3CB inhibitor GSK2636771. Together, our study indicate that LukS-PV induces cell apoptosis through downregulate expression of the target gene PIK3CB mediated by SET8 in AML cells.
The phosphatidylinositol 3-kinase (PI3K) pathway plays pivotal roles in cell growth, proliferation, and survival by integrating extracellular growth signals [3]. PI3K proteins fall into three classes, of which class IA are heterodimeric lipid kinases, consisting of a p85 regulatory subunit and a p110 catalytic subunit. The p110 catalytic subunits are p110α, β, and δ. PIK3CB (PI3-kinase p110β) is a member of the PI3K pathway, and hyperactivation of this pathway contributes to cancer progression in humans [9]. Akt, a serine/threonine protein kinase, is one of the best-characterized targets of the PI3K pathway. Recent studies have reported numerous Akt substrates including BAD, CREB, GSK-3β, mTOR, FOXO1, and PTEN [23]. Initiation of these signalling cascades regulates cellular proliferation, motility, and survival in cancer cells. The present ChIP-seq experiments and bioinformatics analysis revealed that PIK3CB is the transcriptional target gene for LukS-PV-mediated H4K20me1. Furthermore, our results show that LukS-PV decreases AKT phosphorylation and increases FOXO1 levels, thus inducing apoptosis by downregulating anti-apoptotic protein Bcl2 and upregulating pro-apoptotic protein Bax in HL-60 and NB-4 cells.
In conclusion, these results show that LukS-PV induces apoptosis in AML cells via the PI3K/AKT/FOXO1 signal transduction pathway by targeting SET8, implying that SET8 is a potential target for LukS-PVmediated treatment of AML.

Consent for publication
Not applicable Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information les.

Competing interests
The authors declare no con ict of Interest.