Diallyl disulfide down‐regulates calreticulin and promotes C/EBPα expression in differentiation of human leukaemia cells

Abstract Diallyl disulfide (DADS), the main active component of the cancer fighting allyl sulfides found in garlic, has shown potential as a therapeutic agent in various cancers. Previous studies showed DADS induction of HL‐60 cell differentiation involves down‐regulation of calreticulin (CRT). Here, we investigated the mechanism of DADS‐induced differentiation of human leukaemia cells and the potential involvement of CRT and CCAAT enhancer binding protein‐α (C/EBPα). We explored the expression of CRT and C/EBPα in clinical samples (20 healthy people and 19 acute myeloid leukaemia patients) and found that CRT and C/EBPα expressions were inversely correlated. DADS induction of differentiation of HL‐60 cells resulted in down‐regulated CRT expression and elevated C/EBPα expression. In severe combined immunodeficiency mice injected with HL‐60 cells, DADS inhibited the growth of tumour tissue and decreased CRT levels and increased C/EBPα in vivo. We also found that DADS‐mediated down‐regulation of CRT and up‐regulation of C/EBPα involved enhancement of reactive oxidative species. RNA immunoprecipitation revealed that CRT bound C/EBPα mRNA, indicating its regulation of C/EBPα mRNA degradation by binding the UG‐rich element in the 3′ untranslated region of C/EBPα. In conclusion, the present study demonstrates the C/EBPα expression was correlated with CRT expression in vitro and in vivo and the molecular mechanism of DADS‐induced leukaemic cell differentiation.

cells through the induction of differentiation. 7 Therefore, the development of novel differentiation inducing drugs for blocking AML is of clinical significance.
Diallyl disulfide (DADS), the main active component of the cancer fighting allyl sulfides found in garlic, was reported to reduce the initiation of carcinogen induced cancers and inhibit the proliferation of various types of cancer cells. 8 Previous studies confirmed that DADS can inhibit the proliferation of human leukaemia cells in vivo in a dose dependent manner. Several studies showed that a small dose of DADS (<1.25 mg/L) can induce human leukaemia cell differentiation, and proteomic analysis suggested that calreticulin (CRT) was involved in DADS mediated induction of differentiation in HL 60 cells. 1,9 CRT is a multiprocess calcium buffering chaperone of the endoplasmic reticulum that is essential for numerous cellular functions. [10][11][12] Some studies have identified increased expression of CRT in many tumour tissues or cells. 13 CCAAT enhancer binding protein-α (C/EBPα) is a basic leucine zipper transcription factor that is crucial for normal neutrophil differentiation. 14 Deregulation of C/EBPα function by genomic mutations, transcriptional, and posttranscriptional suppression, or phosphorylation-dependent inactivation is a common event in subgroups of AML patients. 15,16 Previous studies have shown that in some leukaemic cells, the inhibition of C/EBPα transcription and translation level is closely related to CRT. 17 The downregulation of CRT by siRNA can increase the expression of C/EBPα. These findings suggest that CRT may be an important factor that regulates C/EBPα in leukaemia differentiation. 18 In the present study, we hypothesised that DADS down-regulates CRT and promotes C/EBPα expression in inducing differentiation of human leukaemia cells. These findings may lead to a better understanding of the molecular mechanisms of the down-regulation of CRT in DADS-induced differentiation of leukaemia cells and will provide essential knowledge for the development of differentiation inducers to treat leukaemia.

| Cell culture and treatments
HL-60 cells were purchased from Central South University (Hengyang, China) and cultured in RPMI 1640 (Gibco, California, CA, USA) supplemented with 10% heat-inactivated foetal calf serum.
Cultured cells were maintained in 5% CO 2 and humidified air at 37°C. Cell cultures were replaced with fresh medium every 2-3 days.
Cells in logarithmic growing phase were used for experiments. DADS was diluted to 1.25 mg/L in culture medium. N-acetyl-cysteine (NAC) was diluted to 10 mM in culture medium.

| Patient samples and clinical data
Thirty-nine samples were included in our study, with a mean age of 42.6 ± 14.8 years (median, 46 years; range, 13-87 years). We numbered all the samples from 1 to 39. All patients were diagnosed with leukaemia between 2015 and 2016 in the First Affiliated Hospital of University of South China. We tracked the patients' treatment. We detected the patients before and after treatment, respectively. Patient characteristics are presented in Table 3. The study was approved by the local Human Research Ethics and Animal Care Committees.

| Bone marrow biopsy
Bone marrow aspirate smears from all patients were stained with a Wright-Giemsa preparation. Bone marrow core biopsies were fixed in B5 fixative, rinsed with 10% neutral buffered formalin, and subjected to limited decalcification in RDO Gold Working Solution (Apex Engineering, Aurora, IL, USA) prior to routine processing.

| Small interference RNA
Small interfering RNA (siRNA) transfection reagent, siRNA transfection medium, were purchased from Santa Cruz Biotech-nology (Dallas, TX, USA). Cells were collected through centrifugation and the concentration was adjusted to 1-2 × 10 6 cells/mL. The cells were then washed once with 2 mL siRNA transfection medium. Two various preparations were used: Solution A, consisting of 6 μL CRT siRNA duplex in 100 μL siRNA transfection medium, and solution B, including 6 μL siRNA transfection reagent in 100 μL siRNA transfection medium. Solution A was then directly added to solution B using a pipette, and the solutions were mixed gently and incubated for 30 minutes at room temperature. siRNA transfection medium (0.8 mL) was added to each tube containing the siRNA transfection reagent mixture, and the mixture was overlaid onto cells. Cells were incubated for 5-7 hours at 37°C in a CO 2 incubator, and then 1 mL of normal growth medium containing two times the normal serum amount (8% calf serum) was directly added without removing the transfection mixture. Cells were incubated for an additional 18-24 hours, after which the medium was removed and replaced with fresh normal growth medium. At 24-72 hours, cells were harvested for analysis.

| Wright and Giemsa staining
The smears were incubated with Wright Stain (1 mL) for 2 minutes and then rinsed with 2.0 mL distilled water or phosphate buffer pH SUN ET AL. | 195 6.5 for another 2 minutes. The stained smears were rinsed with water or phosphate buffer pH 6.5 until the edges were faintly pinkish-red. For Giemsa stain, smears were incubated for 10 minutes in one volume Wright Stain and four volumes phosphate buffer, pH 6.5. Smears were then dried by careful blotting.

| Reverse-transcription polymerase chain reaction analysis
Expression levels of the CRT and C/EBPα gene in HL60 cells were measured by semiquantitative reverse-transcription polymerase chain reaction (RT-PCR). β-actin was used as an internal control. All primer sequences were designed by Premier 5.0 software (Premier Biosoft International, Palo Alto, CA, USA) and synthesised by Takara Bio Inc.
(Otsu, Japan). The primer sequences were as Table 1. 10 000-fold diluted SYBR Green, was used. All PCR reactions were performed at 95°C for 5 minutes, followed by 40 cycles of 95°C for 10 seconds, 59°C for 15 seconds, and 72°C for 20 seconds. To establish the melting curve, the samples were then heated between 72 and 99°C, increasing 1°C every 5 seconds. The results were quantified by dissociation and amplification curves.

| Western blot analysis
Total protein was extracted from HL-60 cells and mice tumours, and protein concentration was quantified by the Pierce BCA Protein Assay kit (Thermo Fisher Scientific, Inc.). The protein samples were separated by 7.5%-12% SDS-PAGE gels (Applygen, Beijing, China) and transferred onto polyvinylidene fluoride membranes (120 minutes at 100 V) using standard procedures. The membranes were blocked in TBST (PBS containing 0.1% Tween) containing 5% non-fat dry milk for 120 minutes at room temperature and then incubated with Mouse body weight was measured twice a week. After 21 days of treatment, all of the mice were killed 12 hours after the final injection. Tissues were flash-frozen in liquid nitrogen or fixed in formalin and embedded in paraffin. During the treatment period, the mice were weighed twice weekly and monitored for any overt signs of toxicity.

| Hematoxylin-eosin staining
Multiple organs, including brain, heart, lungs, liver, kidneys, and spleen, were removed from killed mice at end-point. Tumour tissue sections were fixed in either alcohol or an aldehyde-based fixative and then processed to obtain 5 μm sections and placed on slides.
The slides were rinsed for 1 minute in H 2 O and then stained with 1% eosin Y solution with agitation for 10-30 seconds. Xylene was 750 bp used to extract the alcohol. One or two drops of mounting medium were added, and the sections were covered with a cover slip.

| RNA immunoprecipitation
To determine CRT and C/EBPα mRNA interaction in HL-60 cells, the

| Statistical analysis
Results were presented as mean ± SD. All statistical analysis was performed using SPSS 18.0 software (Statistical Product and Service Solutions, Chicago, USA). The differences between groups were analysed by using the two-sample t test and paired t test. The paired t test was used in the pre-treatment and post-treatment groups, the two-sample t test was used in the pre-treatment group and the control group, as well as in the post-treatment group and the control group. The correlation analysis of two variables using linear regression analysis, twosided P < 0.05 values were considered statistically significant.

| Expression of CRT and C/EBPα in peripheral blood leukocytes of leukaemia patients
We performed biopsy and staining for all clinical samples from 20 healthy people and 19 AML patients. We confirmed the type of the 19 leukaemia patients, as shown in Figure 1A, sample #21 and #23 were diagnosed as AML without maturation (M1) and AML with maturation (M2), respectively. Compared to the control (sample #9), the bone marrow of leukaemia samples showed extremely active granular hyperplasia.
We next examined CRT and C/EBPα expression in leucocyte of clinical samples before and after clinical treatment by FCM immunophenotyping ( Figure 1B). CRT level was significantly higher and C/EBPα level was significantly lower in the pre-treatment samples compared to controls (P < 0.01; Table 3). We also compared the levels of CRT and C/EBPα between pre-treatment and posttreatment groups and detected significantly lower CRT and signifi- We also examined the relationships between CRT and C/EBPα in different patient groups, and correlation analysis revealed a positive correlation between CRT and C/EBPα. In the control group, CRT was positively correlated with C/EBPα (r = 0.633, P = 0.002), and there were correlations between CRT and C/EBPα in the pre-treatment group (r = 0.739, P = 0.000) and post-treatment group (r = 0.806, P = 0.000). In a conclusion, CRT was highly expressed and C/EBPα was expressed in low levels in the pre-treatment AML patient group compared to controls. And C/EBPα expression was negatively correlated with CRT.

DADS-induced differentiation of HL-60 cells
Our data indicate that CRT expression shows a relationship with C/ EBPα expression in AML. We next investigated the relationship between CRT and C/EBPα in HL-60 promyelocytic leukaemia cells during differentiation. We first treated HL-60 cells with 1.25 mg/L DADS for 48 hours and confirmed induction of cells to differentiate into granulocyte-like cells (Figure 2A). The morphology of the DADStreated cells was clearly changed compared to controls.
CD33 is transmembrane receptor expressed on cells of myeloid lineage and is usually considered to be myeloid specific. 19 We performed similar analyses for CD11b expression (Figure 3C). CD11b, also known as cluster of differentiation 11b molecule, is expressed on the surface of numerous leucocytes involved in the innate immune system, including monocytes, granulocytes, macrophages, and natural killer cells. [22][23][24] We found that while con-

| DADS impacted the tumour growth and CRT and C/EBPα expression of SCID mice with HL-60 cells engraftment
We next examined the effects of DADS on tumour growth in vivo. To monitor any possible toxicity arising from the treatment, the mice were weighed twice weekly during the 21 days treatment period, and no significant differences in body weight, food, or water consumption were observed between control and groups during the intervention.

| DADS induced CRT down-regulation and translocation involves the ROS pathway
We previously showed significantly reduced C/EBPα protein level in HL-60 cells with CRT overexpression. Our current data show that DADS reduced CRT expression levels and up-regulated C/EBPα protein levels ( Figure 2B). ROS are important second messengers in many cellular processes, including differentiation. Some antitumour drugs promote the expression of C/EBPα by increasing the production of ROS in tumour cells and releasing the binding of CRT and C/EBPα mRNA. 26 We thus speculated whether DADS-mediated regulation of CRT and C/EBPα involves ROS and whether CRT directly regulates C/EBPα mRNA.
We first examined the distribution of CRT in HL-60 cells. In control HL-60 cells, CRT was mainly localised in the cytoplasm, but its expression was reduced and CRT translocated to the plasma membrane surface upon DADS treatment ( Figure 4A). ROS generation is a crucial modulator in a variety of signaling pathways that are associated with autophagy and apoptosis. 27 To measure ROS, we next used DCFH-DA, which is rapidly oxidised to highly fluorescent dichlorofluorescein (DCF) in the presence of ROS. As shown in By trying to identify ROS pathway, we found that the significant ROS production triggered by DADS was largely inhibited by the antioxidant NAC. The data were not shown. HL-60 cells were pretreated with NAC for 1 hour, followed by DADS treatment for various time points. FCM analysis indicated that DADS-induced ROS production was reduced by NAC pre-treatment. Based on these results, we proposed that ROS levels were up-regulated by DADS, resulting in CRT translocation and a possible "eat-me" signal.
Some antitumour drugs release CRT by increasing the production of ROS in tumour cells, phosphorylating PERK/PKR and eIF2α, activating PI3K/AKT pathway, and displacing CRT/Annexin A1 to the cell membrane C/EBPα mRNA binding, promote C/EBPα expression. Our current data suggest that DADS induced CRT translocation, and PERK and PI3K/AKT pathway may be involved in this process. Next we tried to dissect this signal pathway. In addition, we found that the changes in CRT expression and C/EBPα expression triggered by DADS were largely inhibited by NAC treatment (Figure 4D and E).
We also observed that DADS treatment induced PI3K/AKT activation in HL-60 cells ( Figure 4F and G).

| DADS-induced differentiation of HL-60 cells down-regulates CRT and promotes translation of C/ EBPα mRNA
To further examine the relationship between CRT and C/EBPα mRNA, we performed RIP analysis. Our data showed that DADS treatment for 48 hours results in a marked reduction of CRT protein levels ( Figure 5A and B). We performed RIP assay with anti-CRT antibody and cell lysates from the HL-60 cells and found that C/EBPα mRNA was immunoprecipitated by anti-CRT antibody in untreated cells and that IP was reduced in the presence of DADS ( Figure 5C), which suggests CRT binds and targets C/EBPα mRNA and that DADS treatment reduces this interaction. As shown in Figure 5D and E, we detected the control and treatment group cells, Our previous studies demonstrated the down-regulation of CRT during DADS-induced differentiation in HL-60 cells and indicated that CRT was involved in cell proliferation, invasion, and differentiation in DADS-treated HL-60 cells. 1,10,35 Numerous studies have demonstrated that CRT could play an essential role in AML cell proliferation and invasion, and therefore may be an important target for AML. [36][37][38] At the same time, CRT is an important factor that regulates the differentiation of C/EBPα from leukaemia. 39 In this study, CRT was highly expressed and C/EBPα was expressed in low levels in the pre-treatment AML patient group compared to controls. We also found that C/EBPα expression was negatively correlated with CRT. These data suggest that CRT is closely associated with C/EBPα Several studies have shown that DADS (<1.25 mg/L) generates ROS and that this effect is important for the induction of differentiation. 6 DADS can induce ROS production, which is often associated with cancer progression. [43][44][45] Indeed, we found that intracellular levels of ROS were increased after treatment of HL-60 cells with DADS. We further found that the significant ROS production triggered by DADS was largely inhibited by NAC, and that NAC affected CRT and C/EBPα expression. These results indicated that DADS induced CRT down-regulation and translocation through the ROS pathway.
RNA immunoprecipitation revealed that CRT binds C/EBPα mRNA and likely mediates its degradation by binding the UG-rich element in the 3′ untranslated region of C/EBPα. 46 CRT was mainly localised in the cytoplasm of untreated HL-60 cells, but it translocated to the plasma membrane surface upon DADS treatment. Furthermore, C/EBPα mRNA was immunoprecipitated by anti-CRT antibody, suggesting that C/EBPα is a CRT target gene. Together these data indicate that CRT inhibits C/EBPα mRNA expression and that DADS-mediated differentiation releases CRT inhibition of C/EBPα mRNA expression.
In conclusion, the present study clearly demonstrates the correla-

ACKNOWLEDG EMENTS
The present study was sponsored by the National Natural Scientific