Curcumin sensitizes response to Ara-C in xenograft model but not in AML cells

To address the synergistic effect between Curcumin and Ara-C, HL-60 cells were xenografted into male NSG-immunodeficient mice by tail vein as previously described. At week 3 after HL-60 cells injection, the symptoms of disease appeared. After the establishment of disease, mice were treated with either Curcumin + Ara-C (Cur/A) or PBS + Ara-C (PBS/A) for 5 days consecutively. BM samples were collected on day 8 and monoclonal cells (MNCs) were isolated (Fig. 1A). CCK-8 assay showed that MNCs isolated from Cur/A group displayed significantly reduced cell proliferation rate compared to PBS/A group (Fig. 1B). Apoptosis assay also revealed that Cur/A led to markedly increase of MNCs compared to the PBS/A group (Fig. 1C). To further confirm this observation, we performed in vitro assay that HL-60 and THP-1 cells were treated with serial concentrations of Curcumin (100, 50, 25, 12, 6, 3, 1 and 0 μM) and Ara-C (3 μM) (Cur/A or PBS/A treatment). Inconsistent with what observed in the animal model, we did not see significant alteration in Cur/Ara co-treated group compared to PBS/A-treated group as evidenced by CCK-8 and apoptosis assay (Fig S1A–D).

Fig. 1 Curcumin and Ara-C displayed synergistic effect in mice but not in cell lines. A Establishment of AML mouse model and experimental design. B, C MNCs isolated from Cur/A AML mice showed significantly increased apoptosis in concomitant with decreased cell proliferation. *p < 0.05 Full size image

Curcumin leads to altered intestinal microbiota

Due to the inconsistence between the in vivo and in vitro analysis, we asked whether this inconsistence is associated with intestinal microbiota which have been established relationship with AML [17]. To address this question, we first compared the microbiota profiling between Cur/A-treated mice and PBS/A-treated mice. α-diversity analysis revealed an unchanged enrichment of bacteria species as determined by Shannon index and Chao-1 (Fig. 2A). However, principal component analysis (PCA) based on species-level revealed a significant alteration of intestinal microbiota composition between Cur/A and PBS/A group (Fig. 2B). Heatmap showed that Lactobacillus Acidophilus (L. acidophilus), Bifidobacterium bifidum (B. bifidum) and Lactobacillus reuteri (L. reuteri) were significantly increased while pathologic bacteria, including Bacteroides fragilis (B. fragilis), Escherichia coli (E. Coli), Fusobacterium nucleatum (Fn) and Akkermansia muciniphila (A. muciniphila) were significantly reduced (Fig. 2C and Table S1). These observations indicate that Curcumin combined with Ara-C is associated with a different microbiota composition compared to Ara-C alone.

Fig. 2 Curcumin treatment led to intestinal microbiota alteration. A α-diversity showed no significant change of bacteria diversity alteration with curcumin treatment as evidenced by Chao-1 and Shannon index analysis. B PCA assay revealed that Curcumin treatment resulted in markedly altered microbiota composition. C Heatmap showed that Curcumin induced probiotics enrichment while pathogenic bacteria were reduced Full size image

Altered intestinal microbiota is involved in Curcumin-mediated increased response to Ara-C

To determine whether altered microbiota induced by Curcumin is associated with sensitized response to Ara-C, we used germ free mice (GF) to develop AML xenograft model, followed by Curcumin and Ara-C treatment as mentioned above. Then MNCs were isolated for viability and apoptosis assay (Fig. 3A). The time of the appearance of disease symptoms in GF mice was similar with the SPF mice. Interestingly, the synergistic effect between Curcumin and Ara-C disappeared in GF mice (Fig. 3B), suggesting Curcumin regulate microbiota to enhance response to Ara-C in AML. To further confirm this observation, GF mice were treated with stools of Cru/A-treated mice or stools of PBS/A mice (Fig. 3C). To confirm the colonization of bacteria, we checked the volume of altered bacteria using qPCR. It showed that B.fragilis, E. Coli and Fn were significantly reduced in Cur/A-treated-mice-stool-gavaged mice in concomitant with decreased B. fragilis, E. Coli, Fn and A.muciniphila, indicating a successful colonization (Fig. S2). Therefore, we treated the mice with Ara-C. Consistently, Cur/A-treated-mice-stool-gavaged mice displayed significantly reduced viability in concomitant with increased apoptosis in MNCs (Fig. 3D). These results indicate that Curcumin regulates intestinal microbiota to sensitize response to Ara-C.

Fig. 3 Curcumin-induced microbiota alteration sensitized response to Ara-c. A Establishment of AML mouse model in SPF and GF mice. B Apoptosis and proliferation assay demonstrated that Curcumin had no effect on sensitizing response to Ara-c under GF condition. C Experimental design for stool transplantation. D Apoptosis and proliferation assay revealed that stool of Cur/A-treated mice sensitized response to Ara-C. *p < 0.05 Full size image

Curcumin enhances intestinal intact to sensitize response to AML

Intestinal intact is important in separating intestinal pathogen from epithelial cells to avoid pathogen invasion [18]. Mucus and tight junction proteins (TJPs) are key component to maintain intestinal intact [19]. As we observed the reduction of bacteria with the function of degrading intestinal mucus in Curcumin-treated mice, we asked whether Curcumin could strengthen intestinal intact. To address this question, FITC-dextran was administrated to mice 4 h before killing. Then, blood was collected for FITC determination. Consistent with our hypothesis, the concentration of FITC-dextran was significantly reduced in serum of Cur/A-treated mice (Fig. 4A). Tight junction proteins ZO-1, occludin and claudin-1 were also increased in Cur/A-treated mice (Fig. 4B). Next, we investigated whether enhanced intestinal intact could sensitize response to Ara-C in AML. Probiotics (VSL#3) was previously reported to strengthen intestinal intact. AML mice were treated with VSL#3 (15 mg in 200ul PBS per mouse and treated to mice one week before Ara-c treatment) plus Ara-C (VSL/A) or PBS/A (Fig. 4C). It turned out that MNCs of VSL/A-treated AML mice showed significantly decreased proliferation rate and increased apoptosis compared to PBS/A-treated AML mice (Fig. 4D). These results indicate that Curcumin enhances response to Ara-C by strengthening intestinal intact.

Fig. 4 Curcumin treatment enhanced intestinal intact. A Fitc-Dextran assay showed that Curcumin treatment led to decreased intestinal permeability. B TJPs, ZO-1, Occludin and Claudin-1, were up-regulated in Cur/A-treated mice. C Probiotics treatment reduced intestinal permeability compared to control group. D Apoptosis and proliferation assay showed significantly increased apoptosis and decreased proliferation rate in MNCs of probiotics-treated AML mice. *p < 0.05 Full size image

Curcumin-induced microbiota alteration affects intracellular cholesterol synthesis

Cellular metabolites were proved to be associated with chemoresistance [20]. To disclose the mechanism of Curcumin sensitizing response to Ara-C, we performed metabolomics for BM MNCs collected from Cur/A- and PBS/A-treated mice. The composition of metabolites in MNCs of the two groups of AML mice was significantly different as determined by principal composition analysis (PCA) (Fig. 5A). Heatmap showed that in Cur/A-treated MNCs, some metabolites were decreased while others were increased (Fig. 5B and Table S2). Of note, cholesterol, which was enriched in PBS/A-treated MNCs, was located at the outlier of volcano plot, whereas squalene, which was enriched in Cur/A-treated MNCs, was located at the outlier of volcano plot (Fig. 5C). Next, we further examined the cholesterol level directly in the MNCs. Consistently, cholesterol level was significantly reduced in Cur/A-treated MNCs compared to PBS/A-treated MNCs (Fig. 5D).Correlation analysis was performed to determine the potential association between altered gut microbes and cholesterol of PBS/A-treated mice. It showed that A. muciniphila and E.Coli were positively correlated with cholesterol, whereas L. acidophilus and B. bifidum were negatively correlated with cholesterol (Fig. 5E). These observations indicate that Curcumin-induced microbiota alteration reduced intracellular cholesterol level of MNCs.

Fig. 5 Curcumin-induced microbiota alteration led to change of metabolites in MNCs of AML mice. A PCA analysis saw an altered metabolomics of MNCs between Cur/A- and PBS/A-treated AML mice. B Heatmap showed markedly altered metabolites. C Volcano plot revealed that cholesterol and squalene were located at the outlier of significantly changed metabolites. D Curcumin treatment led to significantly down-regulated cholesterol in MNCs of AML mice. **p < 0.01, ****p < 0.0001 Full size image

Curcumin-induced microbiota sensitizes response to Ara-C by suppressing SQLE

It was reported that cholesterol accumulation was associated with chemoresistance [21]. On the other hand, we noticed that Squalene acts as the substrate of Squalene Epoxidase (SQLE) for cholesterol biosynthesis. So, we hypothesized that SQLE is involved in the Curcumin-mediated sensitized response to Ara-C in AML. WB showed that SQLE expression was inhibited in MNCs of Cur/A-treated AML mice compared to MNCs of PBS/A-treated mice (Fig. 6A) To further confirm the function of SQLE, THP-1 cells were transfected with SQLE plasmids. As a result, SQLE over-expression led to a significant reduction of cell apoptosis as well as increased cell viability (Fig. 6B). To further assess the function of SQLE we used Terbinafine, a selective SQLE inhibitor, to treat AML mice. Similar with what observed in cells, SQLE inhibition resulted in sensitized response to Ara-C in MNCs of AML mice (Fig. 6C). These results suggest that Curcumin sensitizes response to Ara-C in AML by suppressing SQLE.