CTP synthase 2 predicts inferior survival and mediates DNA damage response via interacting with BRCA1 in chronic lymphocytic leukemia

Blood samples of 96 de novo patients of CLL were collected at the Department of Hematology in Shandong Provincial Hospital after informed consent as approved by the ethics commission. The leukemic clone proportion in peripheral blood of the enrolled CLL patients was > 90%, which could be represented by peripheral blood mononuclear cells [19]. All the participants were aged between 29 and 85 years (mean age = 60.29 years, s. d. = 11.45). Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll density gradient centrifugation, which was described in previous study [20, 21]. CD19 + B cells from healthy donors were purified using CD19 + magnetic microbeads kit (Miltenyi Biotec, Bergisch Gladbach, Germany). Clinical characteristics of CLL patients, including age, stage, immunoglobulin mutational status, and cytogenetics (FISH) were recorded.

1030 clinically annotated CLL patients from multiple cohorts were enrolled in this study. Normalized microarray and RNA-seq data from the GSE50006, GSE22529, GSE55288, GSE22762, GSE39671 datasets were downloaded from the Gene Expression Omnibus (GEO) database. For the ICGC cohort, gene expression and clinical data were obtained from the International Cancer Genome Consortium database (https://​icgc.​org/​). The cut-off value was assessed based on median mRNA expression level. The Kaplan‒Meier method was used for survival analysis.

Human CLL cell lines were routinely cultured in IMDM (MEC-1) and RPMI-1640 (EHEB) medium with 10% fetal bovine serum (Gibco, MD, USA). The medium contained a 1% penicillin/streptomycin mixture. Cells were incubated at 37 °C in humidified air containing 5% CO2. Complete medium without Gln was purchased from M&C GENE (Beijing, China). Cytidine-triphosphate (HY-125818, MCE) were soluble in sterile saline solution to the storage concentration at 10 mM.

The sequences for CTPS2 shRNAs were as follows: shCTPS2#1, 5′-CCGAGGACCCTGTGAAATT-3′; shCTPS2#2,5′-GCAGTGATAGAGTTTGCAA -3′; shCTPS2#3, 5′-CCACAGAGTTTAGGCCAAA-3′. The shRNAs for CTPS2, lentivirus-BRCA1 and the negative control RNA (shControl) were synthesized and purified by OBIO (Shanghai, China). The shRNAs were cloned into lentiviral vectors and lentivirus infection was carried out according to the manufacturer’s instruction. Forty-eight hours after transfection, stably silenced clones were selected by 2 μg/ml puromycin. The medium was changed at twenty-four hours after transfection, and cells were harvested for subsequent analysis seventy-two hours after transfection. Plasmids delivery to overexpress CTPS2 was constructed by WZ Biosciences (Jinan, Shandong, China).

Total RNA was purified with Trizol reagent (Invitrogen, MA, USA) and reversed transcribed into cDNA using a reverse transcription kit (TaKaRa, Dalian, China). Real-time PCR was performed in a Light Cycler 480 II Real-Time PCR system (Roche Diagnostics, Basel, Switzerland) using TB Green (Takara). Real-time PCR of each sample was performed in triplicate. Results were obtained using the sequence detection software Light cycler 480 and analyzed using GraphPad Prism version 7.0 statistical software.

Whole cell lysates were extracted using radioimmunoprecipitation assay (RIPA) buffer together with a 1 × phosphatase inhibitor cocktail (PhosSTOP, Roche, Basel, Switzerland). The primary antibodies used were as follows: anti-CTPS2 (ab196016, Abcam, Cambridge, UK), anti-cyclin B1 (ab32053,Abcam), anti-CDK1 (ab18, Abcam), anti-phospho-H2AX (7631, Cell Signaling Technology, MA,USA), anti-phospho-ATM (13050, Cell Signaling Technology), anti-phospho-BRCA1 (14823, Cell Signaling Technology), anti-Bcl-2 (4223, Cell Signaling Technology), anti-Bax (14796, Cell Signaling Technology), anti-cleaved PARP (5625, Cell Signaling Technology), anti-p21 (2947, Cell Signaling Technology), β-actin (TA-08, Zhongshan Goldenbridge), and anti-GAPDH (TA-09, Zhongshan Goldenbridge, Beijing, China).

Cell viability was assayed using the Cell Counting Kit-8 (CCK-8) (Dojindo, Japan) as previously described. Briefly, the cell density was adjusted to 1 × 10⁴ cells/100 μl/well and inoculated in 96-well plates. Thereafter, CCK-8 reagents were added to each well and incubated for 3 h. Finally, optical density values were read using the SpectraMax M2 Microplate Reader (Molecular Devices, CA, USA) at 450 nm to ascertain cell proliferation capability.

The analyses of the cell cycle distributions and the measurements of the percentage of apoptotic cells were performed by flow cytometry using Navios flow cytometer (Beckman Coulter, CA, USA). For cell cycle analysis, cells were washed with PBS, fixed with cold 70% alcohol at-20 °C overnight before stained with PI (BD Biosciences, MA, USA). The apoptosis detection was performed by using the Annexin V-PE/7AAD apoptosis detection kit and according to manufacturer’s instructions (BD Biosciences). The percentage of cells in the indicated cell cycle phase was calculated with ModFit LT software.

RNA was exacted using the Trizol reagent (Invitrogen, MA, USA). RNA-seq experiments were performed by Novogene (Beijing, China). After reverse transcription and cDNA library enrichment, six RNA-sequencing (RNA-seq) libraries (three for shCTPS2 group and three for shControl group) were generated following the manufacturer’s recommendations. Purified library DNA was captured on an Illumina flowcell for cluster generation and sequenced on an Illumina HiSeq platform, and the fragments per kilobase of transcript per million fragments mapped (FPKM) of each gene were calculated. Differentially expressed genes were determined by the DESeq2 Bioconductor/R package (https://​doi.​org/​10.​18129/​B9.​bioc.​DESeq2). The visualization of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were implemented using the R package ClusterProfiler (v3.0.0). Gene counts for heatmap generation were converted into z-scores and input into the ComplexHeatmap Bioconductor/R package [22].

Cells were fixed with 4% paraformaldehyde/PBS for 15 min and permeabilized with 0.5% Triton X-100 (in 1 × PBS) for 20 min at room temperature. Primary antibodies for staining CTPS2 (ab32087, Abcam), BRCA1 (Santa Cruz, TX, USA) and p-H2AX (ab81299, Abcam) were used at 1:200 for 1 h at room temperature. After three washes with PBS, the sections were incubated with the secondary antibody in PBS with 2% normal serum for 1 h at room temperature, followed by three washes in PBST. Subsequently, cells were stained with 4, 6-diamidino-2-phenylindole dihydrochloride (DAPI; a DNA-specific fluorescent dye) for 5–10 min. Stained cells were covered with coverslips and visualized using confocal microscope.

Alkaline single-cell gel electrophoresis assay was performed according to the protocol from Trevigen (4250–050-K). Briefly, cells were plated on CometSlides after dilution in low melting point agarose. Lysis solution overnight allowed for electrophoretic separation of DNA fragments and DAPI was used to stain the DNA fragments. At least 50 cells per treatment group were imaged and images were analyzed with Comet Assay IV Software Version 4.3 to calculate the tail moment.

The data were expressed as mean values ± standard error of mean (SEM) from at least three independent experiments. The significance of differences between groups was analyzed using Students t-tests and one-way analysis of variance (ANOVA). The overall survival (OS) time was calculated from the date of diagnosis to the date of death or the last follow-up date. The Kaplan–Meier method and log-rank tests were used for survival analysis. Cox proportional hazards regression models were used for univariate and multivariable analysis. All statistical analyses were performed with SPSS Statistics version 20.0 and GraphPad Prism version 7.0 statistical software. p < 0.05 was considered statistically significant.

To study the clinical relevance, the mRNA expression level of CTPS2 was examined in 96 CLL primary samples from Shandong Provincial Hospital CLL (SPHCLL) database. Quantitative real-time PCR confirmed the significant upregulation of CTPS2 in CLL patients compared to purified normal CD19 + B cells derived from healthy donors, with an over tenfold increase (Fig. 1A, p <0.01). CTPS2 expression was also confirmed to be higher in MEC-1 (1.68-fold change, p = 0.04) and EHEB cells (2.5- fold change, p < 0.001) than in normal B cells (Additional file 1: Fig. S1A). What’s more, the protein expression levels of CTPS2 in the CLL patients were increased significantly compared with those of the control group (Fig. 1B). Equivalent results could be obtained through bioinformatic analysis in another three independent cohorts (GSE5006, GSE22529 and GSE55288, Fig. 1C–E). Based on median expression level of mRNA, patients were divided into two groups: CTPS2^high and CTPS2^low. Notably, the CTPS2^high group was significantly involved in undesirable prognostic indicators, including unfavorable therapeutic effect (Fig. 1F), high-risk Binet stage (Fig. 1G), IGHV un-mutation (Fig. 1H) and 11q23- (Fg. 1I) through analyzing ICGC database.

To explore the clinical and pathological role of CTPS2 in CLL, we analyzed the correlation of CTPS2 expression levels with clinical parameters. As Table 1 shown, high levels of CTPS2 expression were associated with the lack of IGHV mutation (p = 0.035) and unfavorable cytogenetics (p = 0.038), suggesting that upregulation of CTPS2 expression was associated with poor prognosis of CLL disease. Consistent with these data, Kaplan–Meier curves suggested a significant association between increased expression of CTPS2 and undesirable overall survival in CLL patients (Fig. 2A, p = 0.03). In the ROC curve analyses, the area under the ROC curve was 0.867, indicating a relatively accurate discrimination (Fig. 2B). In addition, it was validated that the CLL patients with CTPS2 overexpression had a significant decrease in overall survival both in GSE22762 (HR = 4.488, p = 0.001, Fig. 2C) and ICGC database (HR = 1.614, p = 0.049, Fig. 2D). Consistently, CTPS2 upregulation was also involved in unfavorable treatment-free survival in GSE22762 (HR = 2.715, p = 0.003, Fig. 2E) and GSE39671 cohorts (HR = 1.909, p = 0.008, Fig. 2F).

Furthermore, univariate Cox analysis implicated that the high expression of CTPS2 predicted adverse outcome in CLL patients (HR = 2.745, 95% CI 2.018–3.734, p < 0.001). Subsequently, the only two significant parameters, CTPS2 expression and β2-MG level, were further enrolled into multivariate analysis. The results of the multivariate Cox regression analysis suggested that CTPS2 was the independent prognostic indicator of overall survival (Table 2). Moreover, the same analysis was conducted in public databases, showing the independent prognostic value of CTPS2 in CLL (Fig. 2G).

Motivated by above finding, we sought to investigate the physiological phenotypes associated with CTPS2 in CLL. Three lentivirus-mediated RNA interference vectors targeting CTPS2 were used and two of them exhibited remarkable silence of CTPS2 at the mRNA and protein levels in MEC-1 (Fig. 3A, C) and EHEB cells (Fig. 3B, D). shCTPS2#1 demonstrated highest efficacy with 54% and 64% in MEC-1 and EHEB cells, respectively (p < 0.001). The CCK-8 assay was applied to assess the effect of CTPS2-loss on the proliferation of CLL cells. According to the results, stable transfection of shCTPS2 significantly suppressed CLL cell proliferation (Fig. 3E, F).

Subsequently, flow cytometry was performed to evaluate the apoptotic rates. CLL cells with CTPS2 knockdown showed obvious apoptotic rates relative to control transfected cells (Fig. 3G). Decreased expression levels of Bcl-2 and elevated levels of Bax as well as cleavage of the apoptotic marker PARP, were observed in CTPS2 knockdown cells (Fig. 3H). The cell cycle distribution was evaluated by flow cytometry, which demonstrated that silence of CTPS2 induced a blockade in cell cycle progression at the G2/M phase (Fig. 3I–K). As shown in Fig. 3L, the decreased levels of CDK1 and cyclin B1, checkpoint proteins of G2/M cell cycle, were observed. Concurrently, the protein p21 was increased. Therefore, the defection of CTPS2 restrained the growth of CLL cells predominantly through inhibition of cell apoptosis and induction of cell cycle arrest.

To investigate whether CTPS2 promoted CLL progression through affecting CTP synthesis, exogenous CTP was addicted to shCTPS2 and shControl of EHEB cells. As Fig. 4A showed, the addition of CTP significantly rescued the impeded proliferation caused by CTPS2 knockdown. Subsequently, plasmid with CTPS2 overexpression was transfected into EHEB and CLL primary cells, both with effective efficiency (Fig. 4B, C, Additional file 1: Figure S1B). Subsequently, CTPS2 group showed a positive impact on CLL cell proliferation through CCK-8 cytotoxicity assay (Fig. 4D, E). In the Annexin V-PE/7AAD apoptosis assay, CTPS2 enhancement showed a significant attenuation on the percentage of Annexin V + cells (Fig. 4F, Additional file 1: Figure S1C). After successfully establishment of CTPS2-overexpressing cell models, we further detected the effects of Gln and CTP deficiency. EHEB cells with or without CTPS2 overexpression was cultured under Gln-deficient or Gln-rich conditions. Gln-deficient in the media substantially decreased proliferation of EHEB cells with partially recovered by CTPS2 enhancement (Fig. 4G). Meanwhile, CTPS2 upregulated could also alleviated the aberrant apoptosis caused by Gln defection (Fig. 4H, Additional file 1: Figure S1D). Notably, comparable rescued effects were achieved by exogenous CTP addition under Gln defection (Fig. 4I, Additional file 1: Fig. S1E). These data suggested that CTPS2 induced Gln utilization and CTP synthesis to promoted CLL proliferation and survival.

To further understand the molecular mechanisms by which CTPS2 enhanced the development and progression of CLL, we further inspected our RNA-seq data. It had been found that 8115 mRNAs were differentially expressed (DE) in shCTPS2 cells, of which 4580 were downregulated and 3535 were upregulated. (Additional file 2: Figure S2A). A heatmap analysis showed that genes involved in apoptosis and DDR signaling were enriched in shCTPS2 (Fig. 5A). Notably, an inverse trend between the RNA-seq results and protein level was observed in Bcl-2 as well as Bax. To validate the RNA-seq data, we conducted additional quantitative PCR experiments. Unlike protein level, no significant alteration was observed in Bax and Bcl-2 mRNA level, suggesting a potential role of CTPS2 in post-transcriptional modification (Additional file 2: Figure S2B). In addition, subsequent pathway enrichment analyses revealed that the DE molecules were mainly related to ribonucleotide metabolic processes, regulation of cell cycle and DDR pathway (Fig. 5B; Additional file 2: Figure S2C). In the DDR-related genes, BRCA2 and BRCA1 showed the greatest change with 0.73-fold and 0.85-fold variation, respectively. Correlation network indicated the interaction between CTPS2 and DDR-related genes, like ATM, RAD51C, Chk1 et al. (Fig. 5C). A further set of correlation analyses was performed in CLL patients from three independent cohorts which showed strong association between CTPS2 and DDR-related regulators (Fig. 5D).

To explore whether CTPS2 can regulate the DDR pathway in CLL cells, the single cell gel electrophoresis or comet assay was performed to estimate overall DNA damage. As Fig. 5E showed, shCTPS2 group displayed longer tail moment, indicating enhanced DNA damage. Moreover, CTPS2 silence induced elevated expression of DNA damage marker (p-H2AX) in CLL cells (Fig. 5F). Subsequently, we examined the phosphorylation of ATM, BRCA1, and H2AX proteins by western blot. As Fig. 5G showed, the shCTPS2 groups were noted to have increased levels of phosphorylated ATM and H2AX, implicating increased levels of DNA damage. Whereas the phosphorylation of BRCA1 (mediate chromosome damage repair by homologous recombination) was decreased, suggesting the alleviated DNA repair potential in shCTPS2 groups. These results indicated that knockdown of CTPS2 was involved in the activation of DDR signaling, thereby beneficial to the DNA damage and reducing DNA repair.

To clarify the specific association between CTPS2 and the DDR pathway, BRCA1 attracted our particular attention as the significant correlation with CTPS2 among the key components of the DDR pathway. To determine whether CTPS2 and BRCA1 co-localized in CLL cells, we performed immunofluorescence staining with anti-CTPS2 and anti-BRCA1 antibodies in MEC-1 and EHEB cells. The confocal immunofluorescent images illustrated the colocalization of CTPS2 and BRCA1 in CLL cells (Fig. 6A). Meanwhile, a decrease in BRCA1 protein levels in shCTPS2 groups was also corroborated by confocal microscopy. It was observed that endogenous CTPS2 was localized in both the cytoplasm and nucleus, consistent with BRCA1 in MEC-1 and EHEB cells. Subsequently, a co-immunoprecipitation (Co-IP) experiment was further carried out, suggesting the potential interaction between CTPS2 and BRCA1 in CLL cells (Fig. 6B). Analyzing the BRCA1 mRNA expression by qPCR, it was found that BRCA1 was frequently upregulated in CLL patients (Fig. 6C). The expression was also identified to be elevated in the other three cohorts (GSE5006, GSE22529 and GSE55288, Additional file 3: Figure S3A–C). However, the probabilities of overall survival did not differ significantly between BRCA1^high and BRCA1^low groups in SPHCLL database (Additional file 3: Figure S3D). We further analyzed the prognostic effect of BRCA1 in public data. CLL patients with BRCA1^high showed adverse overall survival only in GSE22762 (Additional file 3: Figure S3E). No significance was observed in the larger ICGC data, which was coincided with TTFT (Additional file 3: Figure S3F–H). On the other hand, the expression of BRCA1 showed a close relationship with CTPS2 through correlation analysis (r = 0.51, p = 0.006, SPHCLL, Fig. 6D). Subsequently, a lentiviral vector encoding BRCA1 overexpressing (LV-BRCA1) was further constructed and transduced into MEC-1 cells to investigate the role of BRCA1 in CLL (Additional file 3: Figure S3I, J). Overexpression of BRCA1 could accelerate CLL cell growth at 72 h but mildly effect on 24 and 48 h, which may result from the spillover effect (Additional file 3: Figure S3K).

To understand the regulatory relationship between the two proteins, CTPS2 expression in BRCA1-overexpressed cells was detected by qPCR. BRCA1 enhancement showed no significant effect on CTPS2 expression (Fig. 6E, Additional file 3: Figure S3J), which suggested that BRCA1 functioned as the downstream of CTPS2. To further validate the assumption, we tested whether enhancement of BRCA1 rescued the phenotype associated with CTPS2 knockdown. Stably BRCA1-upregulated lentivirus was transfected into shCTPS2 MEC-1 cells (Fig. 6F, G). Short hairpin (sh) scrambled control and lentivirus vector control were transduced into cells as experimental controls. As expect, the expression of CTPS2 was not affected after additional BRCA1 upregulation (Additional file 3: Figure S3L). It was observed that elevated BRCA1 expression was capable of rescuing the decrease in CLL cell growth (Fig. 6H) and elevation in cell apoptosis causing by CTPS2 knockdown Fig. 6I. Moreover, immunofluorescence staining of p-H2AX (Fig. 6J) and comet assay (Fig. 6K, L) showed that overexpression of BRCA1 rescued CTPS2 shRNA-induced DNA damage elevation. Furthermore, overexpression of BRCA1 overcame the upregulation of p-ATM and p-H2AX after CTPS2 knockdown (Fig. 6M). Taken together, our findings provided supportive evidence that CTPS2 could interact with BRCA1, thereby influencing the biological progress of CLL cells through DDR signaling (Fig. 7).