Skip to main content

Genome-wide association analysis identified splicing single nucleotide polymorphism in CFLAR predictive of triptolide chemo-sensitivity



Triptolide is a therapeutic diterpenoid derived from the Chinese herb Tripterygium wilfordii Hook f. Triptolide has been shown to induce apoptosis by activation of pro-apoptotic proteins, inhibiting NFkB and c-KIT pathways, suppressing the Jak2 transcription, activating MAPK8/JNK signaling and modulating the heat shock responses.


In the present study, we used lymphoblast cell lines (LCLs) derived from 55 unrelated Caucasian subjects to identify genetic markers predictive of cellular sensitivity to triptolide using genome wide association study. Our results identified SNPs on chromosome 2 associated with triptolide IC50 (p < 0.0001). This region included biologically interesting genes as CFLAR, PPIl3, Caspase 8/10, NFkB and STAT6. Identification of a splicing-SNP rs10190751, which regulates CFLAR alternatively spliced isoforms predictive of the triptolide cytotoxicity suggests its role in triptolides action. Our results from functional studies in Panc-1 cell lines further demonstrate potential role of CFLAR in triptolide toxicity. Analysis of gene-expression with cytotoxicity identified JAK1 expression to be a significant predictor of triptolide sensitivity.


Overall out results identified genetic factors associated with triptolide chemo-sensitivity thereby opening up opportunities to better understand its mechanism of action as well as utilize these biomarkers to predict therapeutic response in patients.


Triptolide is a biological diterpenoid derived from the Chinese herb Tripterygium wilfordii HOOK f. Triptolide has been shown to have anti-inflammatory and immunosuppressive activities and has been used in traditional Chinese medicine to treat several diseases, such as, rheumatoid arthritis, immune complex diseases, and systemic lupus erythematosus [1, 2]. It has been shown to have influence on several anti-tumor target genes and inhibit tumors by altering multiple signaling pathways, such as, inhibition of NFκB and c-KIT pathway [3], inhibition of Jak2 transcription [4], inducing apoptotic signals by activation of pro-apoptotic proteins [5, 6], activation of MAPK8/JNK [5, 6], and inhibition of heat shock response [7, 8]. Triptolide has also been shown to influence epigenetic modulation of genes by interaction with histone methyltransferase and demethylase [9]. In spite of the wide therapeutic properties of triptolide, poor water solubility has limited its clinical use in the past. However, recently a water-soluble analog of triptolide–Minnelide has shown promising results in pancreatic cancer cell lines, human xenograft models, as well as in mouse models of pancreatic cancer [10]. Minnelide has been shown to reduce tumor burden in preclinical models of osteosarcoma [11]. Taken together with the anti-tumor properties of triptolide and the recent development of triptolide analogs to overcome its water solubility, triptolide has emerged as a promising anti-tumor agent.

In the present study, we evaluated the impact of genetic variations and gene expression profiles predictive of triptolide cytotoxicity using Epstein-Barr-virus transformed lymphoblastoid cell lines (LCLs) that are part of International HapMap project ( [12]. HapMap LCLs has been used as model to identify genetic markers associated with in vitro chemo-sensitivity to several drugs [1316]. Genotype data is publically available allowing for genome-wide association analyses for biomarker identification. Our results validated some of the known genes/pathways as well as identified novel candidate genes/pathways of relevance to triptolide. We further validated the functional significance of CFLAR in pancreatic cell lines.


In vitro cytotoxicity assays

HapMap LCLs from subjects with European ancestry (CEU; n = 55 unrelated) were obtained from the Coriell Institute for Medical Research and were maintained as recommended. In vitro cytotoxicity was determined by treating LCLs with varying concentrations of triptolide (500, 50, 20, 10, 6.67,1.3 and 0 nM) for 48 hr followed by cell viability measurements using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays (Life Technologies, USA) and Synergy 5 multi-plate reader. Cytotoxicity assays were performed in duplicates and cell lines were randomly chosen to repeat on different dates to rule out any experimental variation. Panc-1, a pancreatic cell line (ATCC, USA) was used for functional validation of the top gene identified in the GWAS. Panc-1 was cultured in DMEM medium supplemented with 2 mM glutamine and 10 % fetal bovine serum.

Real-time quantitative PCR analysis

mRNA expression levels of CFLAR spliced isoforms were quantitated in LCLs and Panc-1 cell lines using CFLAR isoform specific oligonucleotide primers and 2^-ΔΔCT method as described in Additional file 1 and Fig. 2.

Genotyping of panc-1 cell lines

Genomic DNA from Panc-1 cell line was genotyped for CFLAR SNP rs10190751 (A/G) using TaqMan SNP genotyping assay (Applied Biosystems, Foster City, CA, USA). Genomic DNAs from HapMap cell lines with known genotype (AA, AG and GG) were used as controls.

Western blotting

Western plotting was performed using whole-cell lysates and CFLAR (Enzo lifesciences) or β-actin (Abcam) primary and mouse IgG secondary antibodies.

siRNA mediated knockdown of CFLAR in cancer cell lines

Mission esiRNA were procured from Sigma Aldrich. esiRNA were synthesized through in vitro transcription of a 300–600 bp gene specific dsRNA, which is further digested in to complex pool of siRNA using RNAses. This digested pools of esiRNA are verified by DNA sequensing and gel electrophoresis to ensure identity and high specificity (Sigma Aldrich). To ensure high specificity and efficacy of the esiRNA, the algorithm DEQOR is utilized (Design and Quality Control of RNAi, available to the public via CFLAR specific 424 base pair sequence used in this study to create esiRNA pool for CFLAR knockdown is “TCCATCAGGTTGAAGAAGCACTTG ATACAGATGAGAAGGAGATGCTGCTCTTTTTGTGCCGGGATGTTGCTATAGATGTGGTTCCACCTAATGTCAGGGACCTTCTGGATATTTTACGGGAAAGAGGTAAGCTGTCTGTCGGGGACTTGGCTGAACTGCTCTACAGAGTGAGGCGATTTGACCTGCTCAAACGTATCTTGAAGATGGACAGAAAAGCTGTGGAGACCCACCTGCTCAGGAACCCTCACCTTGTTTCGGACTATAGAGTGCTGATGGCAGAGATTGGTGAGGATTTGGATAAATCTGATGTGTCCTCATTAATTTTCCTCATGAAGGATTACATGGGCCGAGGCAAGATAAGCAAGGAGAAGAGTTTCTTGGACCTTGTGGTTGAGTTGGAGAAACTAAATCTGGTTGCCCCAGA”.

Panc-1 cells were transfected with 200 nM CFLAR-esiRNA and negative siRNA using Lipofectamine® 2000 (Life technologies) as per manufacturer’s protocol. Twenty-four hours post-transfections cell were treated with varying concentrations of triptolide and cell viability was determined 48 hr post-treatment using MTT assays. mRNA levels of all three isoforms of CFLAR siRNA were quantitated 24 hr post-transfection to check for the knockdown.

Transfection of CFLAR-Short and CFLAR-Long plasmid in Panc-1 cell line

Panc-1 cells were transfected with control and CFLAR expression plasmids (pEF6-V5, pEF-Flag A, pEF6-V5-CFLAR-S and pEF-Flag A-CFLAR-L) using Fugene HD reagent (Promega) followed by triptolide treatment and MTT assay as described above. Cell pellets were also collected for protein analysis.

Statistical analysis

As these cell lines are part of several publically available genotyping databases, genotype data was retrieved on all cell lines from the HapMap project (release 23). For 29 samples, data was also retrieved from the 1000 genomes project (20101123 version). mRNA expression was retrieved for all of the individuals from a publically available source ( IC50, concentration that kills 50 % of the cells, was calculated from a 4-parameter logistic model using the package drc v2.2.1 in R v2.14.0 [17].

Association analysis of triptolide cytotoxicity with genetic variation

SNP genotype (n = 4098136) data was retrieved from the HapMap (release 23) for all 55 samples (29 female and 26 male). SNPs were filtered using various quality control criteria as, build changes, call rate, compliance with Hardy Weinberg and minimum allele frequency (MAF) as described in Additional file 1. In total 1978803 SNPs in 55 individuals passed quality control measures. For genotype data for individuals in the 1000 genomes only a MAF filter was used, dropping SNPs with MAF < 0.05. Several SNPs overlap between the 1000 genomes and HapMap data, data obtained from the 1000 genomes project was used preferentially over data obtained from the HapMap. Since not all of the samples used have been sequenced as part of the 1000 genomes project as of current, SNPs in the 1000 genomes project, but not in HapMap (release 23) were imputed for samples not included in the 1000 genomes project. BEAGLE v3.3.1 [18] was used to impute SNPs with the reference of the 1000 genomes as described in Additional file 1.

Association analysis of triptolide cytotoxicity with gene expression variation

mRNA expression from the Illumina Sentrix Human-6 Expression BeadChip version 1 was normalized as per Stranger et al., using 47293 probes [19]. For this analysis, only the Caucasian samples were used and data was normalized using quantile normalization across replicates and median normalization across individuals. Original Illumina annotation was retrieved from ReMOAT [20]. Expression was not adjusted for gender (did not appear to affect association with phenotypes).

Integrative analyses

With the wealth of data produced by current genomic technologies, collection of multiple types of genomic data on a set of samples is becoming commonplace. New methods explore a multifactor approach that combine different kinds of genomic data, sometimes referred to as “integrative genomics” or “genomic convergence”, in which a multistep procedure is used to identify potential key drivers of complex traits that integrate DNA variation and gene expression data. To integrate the genotype, expression and drug cytotoxicity data, we first identified markers associated with triptolide IC50 using a liberal significance threshold of 0.001. Next, we determined which expression probe sets were associated with these IC50 associated SNPs (p-values ≤ 10−5) (i.e., eQTLs). Finally, to determine whether the expression probe sets associated with these SNPs were also associated with triptolide IC50 values, we identified which expression probe sets were associated with IC50 with a p-values ≤ 0.0001). A similar integrative analysis approach has been used successfully to detect novel candidate genes [21]. The association between IC50 and SNP genotypes modeled as count of rare alleles (additive genetic model), or log2 normalized mRNA expression and cytotoxicity phenotypes, or SNPs was quantified using a spearman correlation coefficient, and p-value calculated for the null hypothesis of no association using an F-test.


Genetic associations with triptolide IC50 in LCLs

We evaluated 55 LCLs derived from unrelated subjects with Caucasian ancestry for cellular sensitivity to triptolide. Triptolide IC50 values ranged from 4 to 34 nM indicating wide inter-individual variation in chemo-sensitivity. GWAS analysis identified 140 SNPs in 11 genes that were associated with triptolide IC50 at p < 10−5 (488 SNPs at p < 10−4; Fig. 1a). Significant proportion of SNPs (110 of 140: 78.6 %) clustered in ~293 kb region on chromosome 2 (Table 1), which maps to multiple biologically interesting genes (Genes important for cell division and Cancer development) including CFLAR, CLK1, FAM126B, NDUFB3, NIF3L1, ORC2 and PPIL 3 (Table 2; Fig. 1b). Other genes with significant SNPs included TP53BP2 (chr 1), MTSU2 (chr 13), ZNF532 (chr18) and FNDC3B (chr 3). Ingenuity pathway analysis of these genes mapped them to 4 networks (Table 3). These networks are involved in Cellular Movement, Inflammatory Response, Cell-To-Cell Signaling and Interaction; Cell Death and Survival, Cellular Function and Maintenance, Molecular Transport; Cancer, Dermatological Diseases and Conditions, Developmental Disorder; Cellular Movement, Hematological System Development and Function, Immune Cell Trafficking and carbohydrate metabolism.

Fig. 1

Genome-wide association study of SNPs with Triptolide cytotoxicity. a Manhattan plot showing association of SNPs with Triptolide IC50 (only SNPs with p <10−4 are included). b Genomic region on Chr 2 with strongest association with triptolide cytotoxicity. Y-axis represents -Log 10 (P value) and X-axis presents chromosomal location

Table 1 List of top 140 SNPs (p <0.00001) from GWAS analysis that were predictive of triptolide cytotoxicity in HapMap LCLs
Table 2 Summary of genes with SNPs significantly associated (p < 0.00001) with triptolide cytotoxicty in HapMAP LCLs
Table 3 Ingenuity pathway analysis tool mapped the genes identified in gene expression vs. cytotoxicity analysis to 5 networks

CFLAR (Caspase 8 and FADD like apoptosis regulator) with maximum number of most significant SNPs (n = 41; due to high LD as shown in Additional file 2: Figure S1) codes for protein c-FLIP. c-FLIP regulates apoptosis and is structurally similar to caspase-8, however, lacks the caspase activity. It has been implicated as a crucial link between cell survival and cell death pathways in mammalian cells.

Gene expression associations with triptolide IC50 in LCLs

Genome-wide gene expression analysis identified 14 probes that were associated with triptolide IC50 (p < 0.0001, Additional file 3: Figure S2 and Table 4). Some of the biologically interesting genes with expression levels associated with triptolide sensitivity included: JAK1 (Janus kinase 1), DTX1 (Deltex Homolog 1; positive regulator of Notch signaling pathway), AGL (Amylo-Alpha-1, 6-Glucosidase, 4-Alpha-Glucanotransferase involved in glycogen degradation), and MUC15 (Mucin 15, Cell Surface Associated). Pathway analysis using Ingenunity pathway analysis tool mapped these genes to JAK/STAT, IL, iNOS, EGFsignalining pathways (Additional file 4: Figure S3).

Table 4 Top probe and genes with expression levels in LCLs associated with Triptolide LC50

Integrated SNP-mRNA association analysis with IC50

We performed integrated analysis between SNPs-mRNA expression-IC50; we selected top SNPs that were significantly (p < 0.001) associated with cytotoxicity, and top gene expression signatures significantly (p < 0.001) associated with cytotoxicity for association with each other i.e. SNP vs. gene expression. This analysis basically identified eQTLs associated with triptolide IC50 and at p <0.0001 we identified 648 unique SNP-mRNA pairs that were associated with triptolide cytotoxicity however these SNP-mRNA pairs mapped to only 28 genes (Table 5), indicating association of multiple SNPs with one gene (which might be due to LD between SNPs). Some of the biologically interesting SNP-mRNA pairs included: association of multiple SNPs in chromosome 16 spanning genes CHST5, TMEM231, GABARAPL2 and ADAT1 with expression levels of AGL; SNPs in ASXL3 with ASCL4 expression; CAMTA1 SNPs and CRYGS expression; TIAM1 SNPs with DTX1 expression; SNPs in GPATCH1, and multiple SNPs on chr 5 and 14 with JAK1 expression. Multiple SNPs within CFLAR were associated with expression levels of MTVR1, PIP5K1B and C9orf19/GLRIP2.

Table 5 Summary of 3 way integrated analysis of Gene-Expression-SNP and triptolide cytotoxicity (p < 0.0001)

Functional significance of CFLAR splicing SNP

We selected CFLAR for functional validation in our study on the basis of following observations; i) higher number of SNPs in CFLAR were associated with the triptolide cytotoxicity (Tables 1 and 2); ii) CFLAR. SNP-mRNA pairs associated with triptolide cytotoxicity included three CFLAR mRNA probes additionally multiple SNPs within CFLAR were associated with each 3 mRNA probes. Majority of the significantly associated SNPs within CFLAR gene were intronic and occurred in high LD (Additional file 2: Figure S1). However, one CFLAR SNP, rs10190751 G > A, was present at the splice junction of exon 7 (Fig. 2a) and was significantly associated with triptolide cytotoxicity (Fig. 2b). We screened HapMap cell lines with AA, AG and GG genotype (3 cell lines in each genotype group) for long (CFLAR-L), short (CFLAR-S) and raji (CFLAR-R) forms of CFLAR splice variants. CFLAR-L form was present in all cell lines irrespective of the genotype whereas CFLAR-S form was only expressed in cell lines with at least one G allele (Fig. 3a). Real-time quantification of CFLAR splice variants showed significant association of AA genotype with low levels of CFLAR-L; complete absence of CFLAR-S and higher levels of CFLAR-R form (Fig. 3). Western blot analysis confirmed the association of C-FLIP protein isoform levels corresponding to rs10190751 genotype and CFLAR mRNA isoforms (Fig. 3c).

Fig. 2

Schematic representation of CFLAR gene and its splice variants. a The schematic of CFLAR gene shows all the exons and the location of rs10190751 SNP at the 3’splice site of intron 6. Presence or absence of this SNP regulates production of CFLAR-short form (CFLAR-S). Difference splice variants of CFLAR are also shown along with the isoform specific primer pair’s used in this study. b Box plot showing association of rs10190751 Splicing SNP with Triptolide LC50 in HapMap LCLs. Y-axis is Log 2 Triptolide IC50 and X-axis represents rs10190751 genotype. Box Plots show medians as a line between boxes representing the first and third quartiles; the whiskers represent the range after excluding outliers. The outliers are defined as data points that fall outside of the first and third quartiles by more than 1.5-times the interquartile range. Circles falling outside the whiskers represent outliers

Fig. 3

Correlation of rs10190751 genotype with CFLAR splice variants. a Isoform specific amplification of CFLAR splice-variants (CFLAR-L, CFLAR-S and CFLAR-R) in LCLs with AA, GG and AG genotype for rs10190751. b Real-time-mRNA quantification showing relative levels of CFLAR-L, CFLAR-S and CFLAR-R isoforms in different genotype groups. Box plots details are same as in figure 2B. c Western blotting showing protein levels of CFLAR-L and CFLAR-S/R forms in HapMap LCLs with AA, AG or GG genotype for rs10190751

SiRNA mediated functional studies on CFLAR

We selected Panc-1 for further investigation for impact of siRNA mediated transient knockdown of CFLAR on cellular sensitivity to triptolide. We selected Panc-1 for further investigation for impact of siRNA mediated transient knockdown of CFLAR on cellular sensitivity to triptolide. Panc-1 was selected based on the literature evidence of efficient use of triptolide in pancreatic cancer at pre-clinical and clinical level [Ref] as well as CFLAR being reported as a therapeutic target for triptolide in Pancreatic cancer [ref]. Genotype of Panc-1 for rs10190751, was identified as GG, therefore all isoform of CFLAR expressed in this cell line and it makes this cell line a perfect model to do functional validation for different isoform. In a pancreatic cancer cell line, Panc-1, siRNA mediated knockdown resulted in significant reduction in the CFLAR-L, CFLAR-S and CFLAR-R isoforms and significant increase in sensitivity to, triptolide (Fig. 4a and b).

Fig. 4

Impact of siRNA mediated knockdown or overexpression of CFLAR isoforms on Triptolide sensitivity in Panc-1 cancer cell lines. a Impact of siRNA medicated knockdown of CFLAR on mRNA expression levels of CFLAR-L, CFLAR-S and CFLAR-R isoforms b Impact of siRNA medicated knockdown of CFLAR on cellular cytotoxicity to Triptolide. c Western blot showing over expression of CFLAR-L and CLFAR-S isoforms as compared to cells transfected to control plasmids. d Impact of overexpression of CFLAR-L and CFLAR-S isoforms in cellular sensitivity to Triptolide

Over expression of CFLAR-long and short isoform

Since Panc-1 demonstrated change in chemo-sensitivity in siRNA mediated knockdown of CFLAR, we further overexpressed CFLAR-L and CFLAR-S forms in Panc-1 cell lines. Transient transfection of most abundant isoforms of CFLAR was done in Panc-1 cell line using pEFA-CFLAR-L and pEF6-V5-CFLAR-L plasmid for Long and Short form of CFLAR respectively. Compared to cells transfected with control plasmids the level of CFLAR protein isoforms were significantly increased (Fig. 4c). Over-expression CFLAR-L or CFLAR-S isoforms resulted in significant decrease in sensitivity for triptolide, (Fig. 4d).


Triptolide is a diterpenoid triepoxide and has been used in traditional Chinese medicine for years. Poor water solubility and toxicity has limited its use in clinics, however recent advances focused on developing triptolide derivatives with better solubility such as MC002 [22], omtritolide [23], minnelide [10] etc. are showing promising advancements especially in pancreatic cancer. The anticancer activity of triptolide has been associated with its ability to inhibit various pro-proliferative or anti-apoptotic factors thereby inducing apoptosis [16]. Triptolide has been implicated in activation of both intrinsic and extrinsic apoptotic pathways by inducing caspase-8, −9 and 3 as well as by inducing cleavage of PARP [24, 25]. Given the fact that triptolide has a promising potential as a therapeutic agent, we designed this study to identify the genomic markers associated with triptolide cytotoxicity using LCLs from International HapMap project.

Our results identified a QTL on chromosome 2 consisting of several SNPs with significant association with triptolide IC50. This region on chromosome 2 included biologically interesting genes such as CLK1, PPL3, NIF3L1, CFLAR, NDUFB3, CASP10, CASP8 etc. (Tables 1 and 2) with important roles in apoptosis and cell cycle regulation pathways. Of particular interest was significant over-representation the top most significant SNPs in CFLAR gene (Caspase 8 and FADD like apoptosis regulator). CFLAR gene encodes for protein c-FLIP, best-known for its anti-apoptotic regulatory role by inhibiting TNF-alpha, FAS-L and TRAIL induced apoptosis [26]. Although most of the SNPs in CFLAR were intronic, one splicing SNP, rs10190751 (3’ splice site of intron 6) was of particular interest [27]. CFLAR-protein, c-FILP exists in several isoforms due to alternate splicing, the most studied forms include long (C-FLIP-L) and short (C-FLIP-S) isoforms of 55 kD and 26 kD, respectively. CFLAR gene has 14 exons and inclusion or exclusion of intron 6 or exon 7 regulates the expression of long, or short or raji forms. CFLAR long form (CFLAR-L) skips exon 7 and is expressed as a full-length protein of 480 amino acids. CFLAR short form (CFLAR-S) includes exon 7 thereby changing the reading frame, creating an early stop codon, and hence a shorter isoform with 221 amino acids. C-FLIP-L is composed of two death effector domains (DEDs) at the amino terminus and a caspase homologous domain, structurally similar to caspase 8 and caspase 10 at carboxy terminus. In contrast C-FLIP-S has two DEDs but lacks caspase homology domain. Presence of rs10190751 regulates the splicing event with rs10190751-A allele resulting in lack of expression of the short form (Fig. 4). In addition to these isoforms recently cFLIP-R forms has been identified in the Raji cells [27]. Due to intronic insertion; CFLAR-R isoform has a premature stop codon resulting in a protein with 212 amino acids and like the CFLAR-S isoform lacks caspase like domain.

Although the characterization of the functional differences of these isoforms is still ongoing, cell type specific pro-apoptotic role of CFLAR-L has been reported. CFLAR-L expression levels are considered critical factor in determining the balance between apoptotic and pro-survival signaling. The CFLAR-L has also been shown to play critical role in autophagy, necroptosis and apoptosis in T-lymphocytes with CFLAR-L deficiency triggering severe cell death upon stimulation [28]. In spite of its major role in regulating death receptor signaling, it has been shown to be involved in regulation of apoptosis by several other mechanism including; modulating the activity of ripoptosome [29] regulation of nectroptosis by preventing caspase 8 activation [3032], inhibiting autophagosome formation by interfering with conjugation of LC3 and in NFkB signaling with its ectopic expression resulting in NFkB activation [3335].

Given the important role of CFLAR (CFLIP) as a key inhibitor of processing and activation of caspase 8; its prognostic and therapeutic relevance in AML [36] as well as in development of drug resistance [37] we designed this study to further explore the clinical significance of the CFLAR and its genetic variation especially the splicing SNP (regulating CFLAR-L and CFLAR-S forms) as biomarker of risk of disease as well as with development drug resistance. Our results of siRNA mediated knock down and overexpression of CFLAR in pancreatic cancer cell lines further provides evidence of its involvement in chemo-sensitivity to triptolide.

Gene expression levels of JAK1, AGL, and DTX1 genes, all involved in cell-to cell signaling (Additional file 4: Figure S3) has been associated with triptolide cytotoxicity analysis. JAK1, Janus Kinase 1 is involved in interferon-alpha/beta and -gamma signal transduction pathways and is a critical component of JAK/STAT pathway; AGL is member of 4 alpha-glucanotransferase and is involved in glycogen degradation; DTX1, deltex homolog 1 is involved in NOTCH signaling pathway which is a critical for cell fate determination and has been implicated in several diseases as well as tumorogenesis [38]. In our integrative exploratory analysis we identified several biologically interesting gene-SNP-gene-expression pairs as TIAM1-DTX1, ASXL3: ASCL4, GPATCH2: JAK1, CAMPTA1-CRYGS, ERBB4-NADSYN1 etc.

In recent years there has been significant evidence suggesting triptolide mediated inhibition of ATPase activity of XPB, thereby by influencing transcription as well as Nucleotide excision repair [39]. XPB, also known as ERCC3 is a subunit of transcription factor TFIIH. Triptolide has been shown to influence gene expression by globally reducing gene expression although to not to same extent for all genes by blocking transcription initiation [40, 41]. Antiproliferative effects of triptolide due to inhibition of XPB/TFIIH has also been shown to phenocopy JNK-dependent apoptosis phenotype in Dp53 deficient wing disc cells in Drosophila [42]. This global reduction of transcription caused by triptolide, correlates well with the phenotypes observed in tumour cells and in inflammation. If we take in account these evidences, and if the treatment with triptolide, reduce global transcription, cells with reduction of the CFLAR mRNA isoforms by the splicing SNP will be even more sensitive, since this gene may negatively modulates apoptosis. The KD and overexpression results using Panc-1 cells incubated with triptolide may also be explained in part by taking in account a reduction in global transcription caused by triptolide.

In conclusion, our results identified CFLAR as an important predictor of triptolide cytotoxicity. Splicing SNP-rs10190751 regulates production of CFLAR- long and short isoforms, which are associated with triptolide cytotoxicity. The central role of anti-apoptotic protein c-FLIP (CFLAR product) in regulating death receptor signaling points to the fact that this splicing SNP might of importance to other chemotherapeutic agents. Up-regulation of c-FLIP has been associated with poor clinical outcome and thus could be reliable prognostic factor for several types of cancer, however the significance of CFLAR genetic variation as predictor of therapeutic efficacy has not been explored so far, thus opening up opportunities for future studies.


Triptolide being an emerging drug, provides us a reason to do a genome wide association study to identify specific genetic polymorphism which may affect triptolide induced cytotoxicity. We observed significant association of triptolide IC50 with SNPs located in biological important genes from apoptotic pathway, such as CFLAR, PPIL3, caspase 8/10, NfKb and STAT6. CFLAR is an upstream regulator of apoptotic pathway. Due to its important position as a regulator of apoptosis, we validated its functional role in triptolide induced cytotoxicity in pancreatic cancer cell line. Our finding shows that CFLAR polymorphism plays important role in cancer cell death induced by triptolide. Further studies are needed to predict the therapeutic response in patients.

Availability of supporting data

Gene expression data is publically avialble from Gene Expression Omnibus under submission number series: GSE6536. Genotype data is avaiable at 23 (


  1. 1.

    Liu Z, Ma L, Zhou GB. The main anticancer bullets of the Chinese medicinal herb, thunder god vine. Molecules. 2011;16(6):5283–97.

    CAS  Article  PubMed  Google Scholar 

  2. 2.

    Pan J. RNA polymerase - an important molecular target of triptolide in cancer cells. Cancer Lett. 2010;292(2):149–52.

    CAS  Article  PubMed  Google Scholar 

  3. 3.

    Zhou GS, Hu Z, Fang HT, Zhang FX, Pan XF, Chen XQ, et al. Biologic activity of triptolide in t(8;21) acute myeloid leukemia cells. Leuk Res. 2011;35(2):214–8.

    CAS  Article  PubMed  Google Scholar 

  4. 4.

    Chen Q, Lu Z, Jin Y, Wu Y, Pan J. Triptolide inhibits Jak2 transcription and induces apoptosis in human myeloproliferative disorder cells bearing Jak2V617F through caspase-3-mediated cleavage of Mcl-1. Cancer Lett. 2010;291(2):246–55.

    CAS  Article  PubMed  Google Scholar 

  5. 5.

    Miyata Y, Sato T, Ito A. Triptolide, a diterpenoid triepoxide, induces antitumor proliferation via activation of c-Jun NH2-terminal kinase 1 by decreasing phosphatidylinositol 3-kinase activity in human tumor cells. Biochem Biophys Res Commun. 2005;336(4):1081–6.

    CAS  Article  PubMed  Google Scholar 

  6. 6.

    Liu Q, Chen T, Chen H, Zhang M, Li N, Lu Z, et al. Triptolide (PG-490) induces apoptosis of dendritic cells through sequential p38 MAP kinase phosphorylation and caspase 3 activation. Biochem Biophys Res Commun. 2004;319(3):980–6.

    CAS  Article  PubMed  Google Scholar 

  7. 7.

    Antonoff MB, Chugh R, Skube SJ, Dudeja V, Borja-Cacho D, Clawson KA, et al. Role of Hsp-70 in triptolide-mediated cell death of neuroblastoma. J Surg Res. 2010;163(1):72–8.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  8. 8.

    Huang M, Zhang H, Liu T, Tian D, Gu L, Zhou M. Triptolide inhibits MDM2 and induces apoptosis in acute lymphoblastic leukemia cells through a p53-independent pathway. Mol Cancer Ther. 2013;12(2):184–94.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  9. 9.

    Zhao F, Chen Y, Zeng L, Li R, Zeng R, Wen L, et al. Role of triptolide in cell proliferation, cell cycle arrest, apoptosis and histone methylation in multiple myeloma U266 cells. Eur J Pharmacol. 2010;646(1–3):1–11.

    CAS  Article  PubMed  Google Scholar 

  10. 10.

    Chugh R, Sangwan V, Patil SP, Dudeja V, Dawra RK, Banerjee S, et al. A preclinical evaluation of Minnelide as a therapeutic agent against pancreatic cancer. Sci Transl Med. 2012;4(156):156ra139.

    PubMed Central  Article  PubMed  Google Scholar 

  11. 11.

    Banerjee S, Thayanithy V, Sangwan V, Mackenzie TN, Saluja AK, Subramanian S. Minnelide reduces tumor burden in preclinical models of osteosarcoma. Cancer Lett. 2013;335(2):412–20.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  12. 12.

    International HapMap Project.

  13. 13.

    Wheeler HE, Dolan ME. Lymphoblastoid cell lines in pharmacogenomic discovery and clinical translation. Pharmacogenomics. 2012;13(1):55–70.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  14. 14.

    Shukla SJ, Dolan ME. Use of CEPH and non-CEPH lymphoblast cell lines in pharmacogenetic studies. Pharmacogenomics. 2005;6(3):303–10.

    CAS  Article  PubMed  Google Scholar 

  15. 15.

    Stark AL, Dolan ME. Lymphoblastoid cell lines in pharmacogenomics: how applicable are they to clinical outcomes? Pharmacogenomics. 2013;14(5):447–50.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  16. 16.

    Jiang J, Fridley BL, Feng Q, Abo RP, Brisbin A, Batzler A, et al. Genome-wide association study for biomarker identification of Rapamycin and Everolimus using a lymphoblastoid cell line system. Front Genet. 2013;4:166.

    PubMed Central  Article  PubMed  Google Scholar 

  17. 17.

    Ritz C. SJC: Bioassy Analysis using R. Journal of Statistical Software 2005, 12(5):1-22.

  18. 18.

    Browning BL, Browning SR. A unified approach to genotype imputation and haplotype-phase inference for large data sets of trios and unrelated individuals. Am J Hum Genet. 2009;84(2):210–23.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  19. 19.

    Stranger BE, Forrest MS, Dunning M, Ingle CE, Beazley C, Thorne N, et al. Relative impact of nucleotide and copy number variation on gene expression phenotypes. Science. 2007;315(5813):848–53.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  20. 20.

    Barbosa-Morais NL, Dunning MJ, Samarajiwa SA, Darot JF, Ritchie ME, Lynch AG, et al. A re-annotation pipeline for Illumina BeadArrays: improving the interpretation of gene expression data. Nucleic Acids Res. 2010;38(3):e17.

    PubMed Central  Article  PubMed  Google Scholar 

  21. 21.

    Huang RS, Duan S, Bleibel WK, Kistner EO, Zhang W, Clark TA, et al. A genome-wide approach to identify genetic variants that contribute to etoposide-induced cytotoxicity. Proc Natl Acad Sci U S A. 2007;104(23):9758–63.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  22. 22.

    Zhuang XM, Liu PX, Zhang YJ, Li CK, Li Y, Wang J, Zhou L, Zhang ZQ. Simultaneous determination of triptolide and its prodrug MC002 in dog blood by LC-MS/MS and its application in pharmacokinetic studies. Journal of ethnopharmacology 2013;150(1):131–7.

  23. 23.

    Kitzen JJ, de Jonge MJ, Lamers CH, Eskens FA, van der Biessen D, van Doorn L, et al. Phase I dose-escalation study of F60008, a novel apoptosis inducer, in patients with advanced solid tumours. Eur J Cancer. 2009;45(10):1764–72.

    CAS  Article  PubMed  Google Scholar 

  24. 24.

    Yang M, Huang J, Pan HZ, Jin J. Triptolide overcomes dexamethasone resistance and enhanced PS-341-induced apoptosis via PI3k/Akt/NF-kappaB pathways in human multiple myeloma cells. Int J Mol Med. 2008;22(4):489–96.

    CAS  PubMed  Google Scholar 

  25. 25.

    Carter BZ, Mak DH, Schober WD, McQueen T, Harris D, Estrov Z, et al. Triptolide induces caspase-dependent cell death mediated via the mitochondrial pathway in leukemic cells. Blood. 2006;108(2):630–7.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  26. 26.

    Miura K, Fujibuchi W, Unno M. Splice variants in apoptotic pathway. Exp Oncol. 2012;34(3):212–7.

    CAS  PubMed  Google Scholar 

  27. 27.

    Ueffing N, Singh KK, Christians A, Thorns C, Feller AC, Nagl F, et al. A single nucleotide polymorphism determines protein isoform production of the human c-FLIP protein. Blood. 2009;114(3):572–9.

    CAS  Article  PubMed  Google Scholar 

  28. 28.

    He MX, He YW. CFLAR/c-FLIPL: a star in the autophagy, apoptosis and necroptosis alliance. Autophagy. 2013;9(5):791–3.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  29. 29.

    Feoktistova M, Geserick P, Kellert B, Dimitrova DP, Langlais C, Hupe M, et al. cIAPs block Ripoptosome formation, a RIP1/caspase-8 containing intracellular cell death complex differentially regulated by cFLIP isoforms. Mol Cell. 2011;43(3):449–63.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  30. 30.

    Oberst A, Dillon CP, Weinlich R, McCormick LL, Fitzgerald P, Pop C, et al. Catalytic activity of the caspase-8-FLIP(L) complex inhibits RIPK3-dependent necrosis. Nature. 2011;471(7338):363–7.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  31. 31.

    Zhang H, Zhou X, McQuade T, Li J, Chan FK, Zhang J. Functional complementation between FADD and RIP1 in embryos and lymphocytes. Nature. 2011;471(7338):373–6.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  32. 32.

    Kaiser WJ, Upton JW, Long AB, Livingston-Rosanoff D, Daley-Bauer LP, Hakem R, et al. RIP3 mediates the embryonic lethality of caspase-8-deficient mice. Nature. 2011;471(7338):368–72.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  33. 33.

    Chaudhary PM, Eby MT, Jasmin A, Kumar A, Liu L, Hood L. Activation of the NF-kappaB pathway by caspase 8 and its homologs. Oncogene. 2000;19(39):4451–60.

    CAS  Article  PubMed  Google Scholar 

  34. 34.

    Kataoka T, Tschopp J. N-terminal fragment of c-FLIP(L) processed by caspase 8 specifically interacts with TRAF2 and induces activation of the NF-kappaB signaling pathway. Mol Cell Biol. 2004;24(7):2627–36.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  35. 35.

    Kataoka T, Budd RC, Holler N, Thome M, Martinon F, Irmler M, et al. The caspase-8 inhibitor FLIP promotes activation of NF-kappaB and Erk signaling pathways. Current biology : CB. 2000;10(11):640–8.

    CAS  Article  PubMed  Google Scholar 

  36. 36.

    McLornan D, Hay J, McLaughlin K, Holohan C, Burnett AK, Hills RK, et al. Prognostic and therapeutic relevance of c-FLIP in acute myeloid leukaemia. Br J Haematol. 2013;160(2):188–98.

    CAS  Article  PubMed  Google Scholar 

  37. 37.

    Lee S, Yoon CY, Byun SS, Lee E, Lee SE. The role of c-FLIP in cisplatin resistance of human bladder cancer cells. J Urol. 2013;189(6):2327–34.

    CAS  Article  PubMed  Google Scholar 

  38. 38.

    Ranganathan P, Weaver KL, Capobianco AJ. Notch signalling in solid tumours: a little bit of everything but not all the time. Nat Rev Cancer. 2011;11(5):338–51.

    CAS  Article  PubMed  Google Scholar 

  39. 39.

    Chen F, Gao X, Shilatifard A. Stably paused genes revealed through inhibition of transcription initiation by the TFIIH inhibitor triptolide. Genes Dev. 2015;29(1):39–47.

    PubMed Central  Article  PubMed  Google Scholar 

  40. 40.

    Tee WW, Shen SS, Oksuz O, Narendra V, Reinberg D. Erk1/2 activity promotes chromatin features and RNAPII phosphorylation at developmental promoters in mouse ESCs. Cell. 2014;156(4):678–90.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

  41. 41.

    Villicaña C, Cruz G, Zurita M. The genetic depletion or the triptolide inhibition of TFIIH in p53-deficient cells induces a JNK-dependent cell death in Drosophila. J Cell Sci. 2013;126(Pt 11):2502–15.

    Article  PubMed  Google Scholar 

  42. 42.

    Titov DV, Gilman B, He QL, Bhat S, Low WK, Dang Y, et al. XPB, a subunit of TFIIH, is a target of the natural product triptolide. Nat Chem Biol. 2011;7(3):182–8.

    PubMed Central  CAS  Article  PubMed  Google Scholar 

Download references


We are thankful to Professor Ingo Schmitz, (Helmholtz Center for Infection Research Inhoffenstr, Braunschweig, Germany) for providing CFLAR expression plasmids (pEF6-V5, pEF-Flag A, pEF6-V5-CFLAR-S and pEF-Flag A-CFLAR-L) and Dr. Gunda Georg for providing triptolide. This work was supported by Minnesota State partnership funds; the Mayo Foundation and the US National Institute of Health (R21-GM86689; R21-CA182715; P30-CA168524; P20-GM103418; R01-CA132946; R21-CA155524).

Author information



Corresponding author

Correspondence to Jatinder K. Lamba.

Additional information

Competing interests

The authors declare they have no competing interests.

Authors contributions

JKL, conceived the study design; LC, NB, TF and TMG performed the experiments; BLF and GDJ performed statistical analysis; All authors contributed to manuscript writing. All authors read and approved the final manuscript.

Additional files

Additional file 1:

Supplementary Notes.

Additional file 2: Figure S1.

LD plot generated using Haploview for chromosome 2 region flanking CFLAR gene in CEU population.

Additional file 3: Figure S2.

Genome-wide association study of mRNA expression with Triptolide cytotoxicity. Y-axis represents -Log 10 (P value) and X-axis presents chromosomal location.

Additional file 4: Figure S3.

Overlapping networks identified by pathway analysis of genes identified in gene expression analysis with Triptolide IC50 in HapMap LCLs using Ingenuity pathway analysis (IPA) tool.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chauhan, L., Jenkins, G.D., Bhise, N. et al. Genome-wide association analysis identified splicing single nucleotide polymorphism in CFLAR predictive of triptolide chemo-sensitivity. BMC Genomics 16, 483 (2015).

Download citation


  • Triptolide
  • Genome-wide association studies
  • Hap-map
  • Single nucleotide polymorphisms