Article | Published:

Influence of BCL2L11 polymorphism on osteonecrosis during treatment of childhood acute lymphoblastic leukemia

The Pharmacogenomics Journalvolume 19pages3341 (2019) | Download Citation

Abstract

Osteonecrosis (ON) is corticosteroid-related complication, reported in children with acute lymphoblastic leukemia (ALL). We have previously found that polymorphisms in BCL2L11 gene coding for pro-apoptotic Bim protein influence reduction of overall survival (OS) in a corticosteroid (CS) dose-dependent manner in childhood ALL patients. The same set of SNPs was here investigated for an association with CS-related ON assessed retrospectively in 304 children with ALL from Quebec (QcALL cohort) who received Dana-Farber Cancer Institute (DFCI) ALL treatment protocols. Two-year cumulative incidence of symptomatic ON was 10.6%. Two BCL2L11 polymorphisms, the 891T>G (rs2241843) in all QcALL patients and 29201C>T (rs724710) in high-risk group were significantly associated with ON, P = 0.009 and P = 0.003, respectively. The association remained significant in multivariate model (HR891TT = 2.4, 95% CI 1.2–4.8, P = 0.01 and HR29201CC = 5.7, 95% CI 1.6–20.9, P = 0.008). Both polymorphisms influenced viability of dexamethasone treated lymphoblastoid cell lines (P ≤ 0.03). The 891T>G influenced Bim gamma isoform levels (0.03) and its association with ON was also confirmed in replication DFCI cohort (N = 168, P = 0.03). QcALL children had a high incidence of ON during therapy, which was highly associated with BCL2L11 polymorphisms.

Introduction

Osteonecrosis (ON) is one of the major complications of childhood acute lymphoblastic leukemia (ALL) treatment [1,2,3]. It can be symptomatic and cause severe pain, joint damage, articular collapse or it can remain asymptomatic and cause no disabilities. Although ON induced by chemotherapy appears to have multiple causes, the corticosteroids (CS), are considered one of the main culprits [4,5,6,7,8]. CS, like prednisone (PDN) and dexamethasone (DXM) are widely used in ALL protocols [1, 9,10,11] for their capacity to induce apoptosis of leukemia cells. However, it has been demonstrated that CS are also involved in osteoporosis and vertebral fractures due to change in the number or function of osteoclasts/osteoblasts, eventually leading to bone loss by increased bone resorption [12,13,14]. Cellular apoptotic effect of CS is essentially mediated through glucocorticoid receptors (GRs), a ligand-activated transcription factor, ubiquitously expressed in various types of cells [15], including osteoblasts and osteocytes [16]. GR-ligand complexes translocate to the nucleus and initiate a transcriptional up-regulation of pro-apoptotic genes in a cell- and gene- specific manner [17, 18]. At physiological concentrations, the CS are key regulators of bone development and cell differentiation. However, CS may have an unfavorable effect when used at higher concentrations and for a prolonged time. For example, the GR is reversibly down regulated in such cases, as a part of physiological feedback loop or up-regulated as a part of a positive autoregulation mechanism [19, 20].

Other drugs used in ALL therapy may also contribute to the development of osteonecrosis. Concurrent use of asparaginase and CS can potentiate each other’s effect by affecting dexamethasone clearance, asparaginase activity or inducing a hypercoagulable state [21]. The risk of ON may be influenced by clinical prognostic factors, like age at diagnosis higher than 10 years [1], female gender [2] and higher body mass index [22]. Recently, the genetic polymorphisms identified through candidate gene approaches and genome-wide association studies (GWAS) have been linked to the ON development [23,24,25,26,27].

Several genes may interact with the action of CS. Microarray profiling in CS-induced apoptosis models, has shown that CS exposure induces or represses the transcription of pro- or anti-apoptotic genes. [28]. Among these, BCL2L11 gene encoding Bim protein was found to be upregulated by CS. Bim protein is a member of the Bcl-2 family; it contains a BH3 homology domain required for apoptosis and a hydrophobic C-terminal domain required for its translocation to mitochondrial membrane where it may interact with other pro-apoptotic members [29]. An increase in the mRNA and Bim isoforms in ALL cells pre-exposed to DXM [30] has been shown suggesting that the level of Bim expression in leukemia cells may be an important determinant of the sensitivity of ALL cells to CS-induced apoptosis. BCL2L11 has been also described as a key regulator of osteoblast apoptosis; it is reported as induced by DXM in a dose- and time-dependent manner in murine osteoblasts, as well as in primary human bone marrow stromal cells [31].

Eight tag SNPs in BCL2L11 was previously described by our group [32] in the context of genetic association analysis with ALL event-free survival (EFS) and overall survival (OS). Our present study aims to establish whether polymorphisms in BCL2L11 gene may affect the risk of symptomatic ON in childhood ALL. The study was conducted in two independent patients’ cohorts. In addition, functional effect of associated polymorphisms was assessed through their impact on mRNA levels and cellular proliferation in cultured lymphoblastoid cell lines (LCLs).

Patients and methods

Study population

Discovery (QcALL) cohort consisted of 304 Caucasian children from Quebec diagnosed with ALL at the University Health Center (UHC) Sainte Justine, Montreal, Canada, for whom both genotype and clinical data were available. Patients were treated in accordance with Dana-Farber Cancer Institute ALL Consortium protocols (DFCI 87-01, 91-01, 95-01, or 00-01) between January 1987 and July 2005. The replication (DFCI) cohort consisted of 168 Caucasian patients with available clinical and genotype data who underwent treatment with DFCI 00-01 protocol in the 9 remaining DFCI Consortium Institutions. Study protocol was approved by respective institutionnal ethics committee and informed consent was obtained from all subjects. Detailed description of these cohorts and DFCI treatment regimens is provided elsewhere [9,10,11, 32, 33]. Patient characteristics are presented in Table 1. Regarding CS treatment [32], all patients received PDN during the induction phase (40 mg/m2/day); CS were administered during the intensification and continuation phases as 5-day pulses every 3 weeks, until the completion of therapy. On protocols 87-01 and 95-01, PDN was used during these treatment phases, on Protocol 91-01 DXM was used instead of prednisone, and on Protocol 00-01, patients were randomized to receive either PDN or DXM. Standard-risk (SR) patients received DXM at a dose of 6 mg/m2/day or PDN at a dose of 40 mg/m2/day and high-risk (HR) patients received doses 3 times higher than those received by SR patients during both the intensification and continuation phases, except on protocol 00-01 when HR patients received the same dose as SR patients during the continuation phase.

Table 1 Characteristic of ALL patients in the discovery (QcALL) and validation (DFCI) cohort

Information on symptomatic osteonecrosis (NCI grades 2–4, Table 2) was collected from patients medical files and was defined as persistent bone pain or motor function limitations associated with ON lesion confirmed by radiological examination such as X-ray radiography, computed tomography (CT) scan or magnetic resonance imaging (MRI). ON was graded in accordance with the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events, version 3.0 [34].

Table 2 Characteristics of ALL patients with or without symptomatic osteonecrosis in QcALL cohort

Genotyping

The eight tag SNPs (rs2241842, rs2241843, rs73954926, rs72836346, 7582030, rs724710, rs72836345, rs6750142), previously described by our group [32], located in regulatory and coding regions of BCL2L11 have been investigated for the association with symptomatic ON in QcALL cohort. Positively associated SNP was analyzed in replication DFCI cohort. The genotypes were either already available or obtained as previously described using PCR—coupled Allele Specific Oligonucleotide (ASO) hybridization assay [32, 35].

Statistical analysis

Association of genotype with the presence of ON (at least one incident during the 2 years of therapy) was assessed by chi-square test (two-sided) in all patients and in patients assigned to the HR group due to higher CS doses administered in these patients. False discovery rate (FDR) correction was performed to adjust for multiple testing [36]; significant results at FDR< 0.05 were retained. Cumulative incidence during 2-years of ALL treatment was estimated in Kaplan–Meier survival analysis framework. Hazard ratio (HR) and a 95% confidence interval (95% CI) were estimated by univariable and multivariable Cox's regression analysis. The common prognostic risk factors, including age, sex, CS type and risk groups were analyzed beside genotype in the multivariable models. Statistical analysis was performed by SPSS statistical package (Chicago, IL), version 22.0.

Cellular proliferation assay and Bim expression studies

In vitro cellular viability following addition of DXM and PDN to the cultured lymphoblastic cell lines (LCLs from HapMap subjects of European origin obtained from Coriell biorepositories) was previously described by Gagné V. et al. [32]. The assay was here extended to include larger number of LCLs and it was performed in 73 and 92 LCLs exposed to DXM and PDN, respectively. Around 5x104cells/well were seeded and submitted either to seven different concentrations of PDN ranging from 0.75 µM to 750 µM or eight different concentrations of DXM ranging from 2.8X10−3 µM to 560 µM and tested in duplicate. Cell proliferation was measured by Cell Proliferation Reagent WST-1 after 48 h incubations time. The concentrations that inhibit the growth of 50% of the cells (IC50) were derived by sigmoid curve fitting. Three LCL groups were defined based on the observed IC50, with sensitive (0.0005–1.9 µM) intermediate (10.9–93 µM) and resistant phenotype (111.2–526.9 µM) (Supplementary Figure 1), and were compared to the genotypes significantly associated with ON, using chi-square analysis.

For analysis of total Bim and its isoforms, same dataset, as obtained in our previous study [32] was used for the analysis with 891T>G variation. Total RNA from LCLs treated with PDN (75 μM prednisolone) and DXM (0.28 µM) was extracted with Qiagen kit and was reverse transcribed. Quantitative PCR was carried out using the Syber Green detection system (Applied Biosystems). The expression was measured by relative quantification normalized to B2-microglobulin (primers: 5′-TACTCTCTCTTTCTGGCCTG-3′ and 5′-GGATGGATGAAACCCAGACA-3′) calculated using the comparative cycle threshold (C T) method. [37] Primers 5′-TGACCGAGAAGGTAGACAAT-3′ and 5′-GCCATACAAATCTAAGCCAGT-3′ targeting alternatively spliced exon 3 was used to amplify non BH3-containing gamma (γ) isoforms [38, 39]. For remaining Bim isoforms, primers 5′-GAGATATGGATCGCCCAAGA-3′ and 5′-CAATGCATTCTCCACACCAG-3′ were used to amplify last two exons common of all isoforms except γ. Semi-quantitative analysis for each of three major RNA isoform (Bim S, Bim L and Bim EL) was carried out by using primers described in [40] whereas γ1 and γ2 assessment was carried out by same PCR primers described above for total γ estimation. Expression of total Bim and its isoforms was compared between genotypes using ANOVA or non-parametric tests. Total Bim mRNA and total γ were also quantified in untreated LCLs. Within- and between-subject effect in treated and untreated LCLs in relation to the genotype was performed by repeated measures ANOVA.

Results

Thirty-two patients were diagnosed with symptomatic ON in QcALL cohort. Symptoms consisted of pain, joint rigidity and limp with the involvement of one or several sites (mainly femoral head, followed by the ankle, knee, wrist or elbow). The diagnosis was based on X-rays (n = 23), CT scans (n = 4) or MRI (n = 5), Table 2. Most patients developed ON during consolidation phase (n = 23) and few patients had been diagnosed during induction (n = 4) or intensification (n = 5). The cumulative overall incidence of symptomatic ON during 2-years of therapy was 10.5%. Slightly higher (but not significantly different) frequency of ON was seen in patients that received DXM or were older than 10 years (Table 2). No difference in relation to sex was noted either in all patients or upon stratification by age.

Association analysis of eight tag SNPs with ON (Supplementary Table 1) revealed a significant association of two polymorphisms following multiple testing adjustment, 891T>G (rs2241843) in all patients (P = 0.009, Supplementary Table 1A) and 29201C>T (rs724710) in high-risk group (P = 0.003, Supplementary Table 1B). Patients with the 891TT genotype had more frequently ON compared to other genotypes, reflected also by higher cumulative incidence of ON (HR = 2.4, 95% CI = 1.2–4.8, P = 0.01, Fig. 1a). Similar increase in cumulative ON incidence was noted for the CC genotype of 29201C>T, which was particularly apparent in patients of high-risk group (P = 0.003, HR = 5.5, 95% CI = 1.5–19.8, Fig. 1b). The association of 891T>G polymorphism with ON remained significant in Cox regression multivariable model when adjusted for covariables such as age, sex, risk group or CS types (P = 0.01; HR = 2.4; 95 % CI, 1.2–4.8, Table 3A). In this model, the genotype was the only factor predictive of ON. The 29201C>T polymorphism remained significantly associated with ON in high-risk group multivariable model, in which the CC genotype, DXM and age above 10 years were all associated with higher cumulative incidence of ON (HR = 5.7, 95% CI 1.6–21.9 P = 0.008, HR = 5.9; 95 % CI, 1.8-19.9, P = 0.004, HR = 3.0; 95 % CI, 1.0-9.2, P = 0.05, respectively, Table 3B).

Fig. 1: Cumulative incidence of osteonecrosis in QcALL cohort in relationship to BCL2L11 polymorphisms 891T>G in a; and 29201C>T in b
Fig. 1

Total number of patients per genotype (number of patients with ON in brackets) is indicated next to the each curve. P value obtained by log-rank test and genotype-associated hazard ratio (HR) with 95% confidence interval (CI) are indicated

Table 3 Risk of osteonecrosis in QcALL patients associated with BCL2L11 891T>G (A) and 29201C>T (B) polymorphism in Cox regression multivariate mode

We next investigated whether 891T > G and 29201C > T may affect in vitro sensitivity to CS or Bim mRNA levels following exposure of LCLs to CS. DXM exposed cells with 891TT or 29201CC genotype were over-represented among sensitive LCLs (IC50 ≤ 1.9 µM, P = 0.01 and P = 0.03, respectively, Fig. 2) compared to those with intermediate or resistant phenotype. The percentage of viable cells after each DXM concentration relative to the genotypes is shown in Supplementary Table 2.

Fig. 2: Frequency of BCL2L11 genotypes in relation to cellular viability and mRNA expression
Fig. 2

In vitro sensitivity to DXM in relation to 891T>G a and 29201C>T b; relative mRNA levels of Bim γ isoforms in untreated and DXM treated LCLs in relation to 891T>G c. In a and b the genotype frequency is compared between sensitive (S) LCLs and LCLs with intermediate (I) and resistant (R) phenotype. p value estimated by chi-square is indicated on the plots. The number of individuals (and frequency) for each genotype (represented by black and gray part of the bar) is presented for each phenotype group. In c, difference in relative mRNA levels (normalized to B2-microglobulin) in untreated and DXM treated LCLs are compared to 891T>G genotype by repeated measures ANOVA. The horizontal arrow depicts the within-subject effect and vertical arrow depicts between-subject effect

We previously analyzed Bim mRNA and its isoforms with 29201C>T and have reported lower γ1 to γ2 ratio associated with the 29201CC genotype [32]. Analysis were here performed with 891T>G showing that DXM pre-treated LCLs with TT genotype have significantly lower total γ mRNA (Fig. 2c); When compared to γ levels in untreated LCLs significant within-and between-subject effect was seen (genotype by mRNA change, P = 0.05; between genotypes, P = 0.03), revealing lower γ isoforms induction in response to DXM in LCLs with the TT genotype. There was no significant association with total Bim mRNA or other Bim isoforms.

Two polymorphisms were further analyzed for an association with ON in the replication DFCI cohort. There was a higher frequency of 891TT genotype among patients who developed ON compared to G allele carriers (57 vs. 43 %, P = 0.03, Fig. 3), whereas an association with 29201C > T did not reveal significant results.

Fig. 3: The frequency of 891 T>G BCL2L11 genotype in relation to osteonecrosis in replication cohort
Fig. 3

The bars represent the frequency of TT genotypes versus other genotype group in patients of DFCI replication group with and without osteonecrosis (+ and −, respectively). The number of patients represented by each bar, p value for the difference between genotype groups and odds ratio (OR) with 95% confidence interval (CI) are indicated

Discussion

Overall cumulative incidences of symptomatic ON related to ALL chemotherapy vary widely among the studies [2, 3, 22] ranging from 1% [3] to 11% [16]. It seems dependent on other clinical risk predictors and was reported to be higher in girls, patients older than 10 years [1, 8, 41] older children that received DXM [33] and in those that received higher cumulative doses of corticosteroids [3, 22]. We noticed an increase in frequency, although non-significant, in patients older than 10 years and those that received DXM (Table 2), whereas in high-risk patients, both factors were significant predictors of symptomatic ON in multivariate model (Table 3). Majority of symptomatic ON cases reported in our study occurred in the continuation phase, few patients nevertheless developed ON during induction therapy (Table 2), which is in accordance with a previous hypothesis [2, 3] that some subjects may manifest a precondition and develop early symptoms of bone morbidity.

In our attempt to determine the genetic component of ON, we investigated implications of the eight tag SNPs that we previously described in BCL2L11 gene [32]. BCL2L11 is involved in the default mechanism of apoptosis in normal and malignant cells through GR-CS mediated up-regulation. Functional defects in Bim protein isoforms may have consequences at cellular level, including increased osteoblast and osteocyte apoptosis, as previously described by Espina et al., 2008 [31]. Indeed, they have shown that Bim regulates osteoblast apoptosis and that gene silencing with short interference RNA reduced apoptosis induced by DXM. Apoptosis of osteoblasts, osteocytes and osteoclasts is the cytological basis for osteonecrosis [42, 43]. Using the rabbit model of femoral head necrosis, Kabata et al demonstrated the direct link between apoptosis of osteoblasts and osteoclasts with osteoporosis, damage of the mechanical bearing structure of the bone and development of femoral head necrosis [44]. Among SNPs investigated in the BCL2L11 gene, two polymorphisms, 29201C > T and 891T > G, seem to influence the risk of ON development (Fig. 1, Supplementary Table 1). The 891T > G is a potential regulatory SNP, given its location in the 5′UTR of BCL2L11 gene; The 5′UTR regions contain important cis-regulatory elements and their disruption or introduction could affect gene expression and protein abundance [45]. Indeed, 891T>G is predicted to affect two transcription factor binding sites. Particularly interesting is the binding site for retinoblastoma binding protein 2 (RBP2) that has repressive activity in osteogenic differentiation of human adipose-derived stromal cells [46] RBP2 is physically and functionally associated with transcription factor RUNX2. The recent study analyzing the molecular mechanisms underlying Bim up-regulation found that DXM treatment induces RUNX2 in parallel with BIM induction [47] The 891TT genotype was associated with increased risk of symptomatic ON (Fig. 1); it remained significant ON predictor in multivariate model (Table 3) and was successfully replicated in the DFCI cohort (Fig. 3). Furthermore, the same genotype conferred higher sensitivity to DXM in LCLs (Fig. 2) that could explain higher incidence of ON seen in patients with this genotype. Several studies have shown relationship between cellular sensitivity to DXM and Bim expression [48, 49]. We did not notice the significant change in the total Bim levels in relation to 891T>G polymorphism, but, among several Bim isoforms, γ isoform without apoptotic BH3 domain was less responsive to induction in LCLs with TT genotypes (Fig. 2). The DXM concentration used in our mRNA analysis (0.28 μM) was based on prior IC50 experiments (Supplementary Table 2), and correspond to those (1 nM—1 μM) associated with Bim induction, which is usually detectable 48 h after exposure. [31, 50]. Nevertheless, other study designs, including different CS concentrations or cell types might be needed to provide additional insight into the relationship between studied polymorphism and Bim mRNA levels.

The 29201C>T is a synonymous polymorphism, located in exon 8 of BCL2L11 gene, which was predicted to affect transcriptional exonic splicing [32]. Similarly, to other studies in which synonymous SNPs are reported to affect mRNA splicing, stability, and protein folding [51], we previously reported that in DXM pretreated LCLs, the T allele affected γ isoforms formation resulting in higher γ1 to γ2 ratio [32], which could have explained lower sensitivity of TT genotype to CS and reduced OS in ALL patients [32]. In line with this observation is a recent study, which reported T allele–related decrease in basal expression of the Bim mRNA and higher T allele frequency in non responsive chronic myeloid leukaemia patients treated with imatinib. [52]. The CC genotype is in contrast expected to confer higher sensitivity to treatment, as indeed shown in our study through an association with increased ON risk mainly observed in HR QcALL group (Fig. 1) who received higher CS doses. This finding was further supported by higher in vitro sensitivity to DXM shown by cellular proliferation assay in LCLs (Fig. 2). The association with ON was not, however, replicated in DFCI cohort, which might be due to lower sample size and lower incidence of ON in this cohort. DFCI cohort was composed only of patients treated with 00-01 protocol in which high-risk patients received lower CS doses during maintenance compared to earlier DFCI protocols. It is worth noting that the 891TT genotype is always accompanied by the 29201CC genotype, so it is possible that diplotypes with these genotype combinations are contributing to observed functional effect and higher symptomatic ON incidence, whereas the 29201CC alone is not sufficient to modulate ON risk. Several limitations of our study should be nevertheless mentioned. The incidence of symptomatic ON data was collected retrospectively, based on patient medical records, and a reporting bias (under- or over-estimation of cases) is possible. Stratified analyses were underpowered due low number of patients per group, which was more evident in replication cohort.

Among other BCL2L11 polymorphisms reported in the literature, the most studied is an intronic deletion polymorphism [39, 53,54,55] that affect splicing resulting in expression of Bim isoforms lacking the pro-apoptotic BH3 domain. This polymorphism is mostly present in Asian population and was reported associated with therapy resistance or early relapse in chronic myeloid leukemia or lung cancer patients treated with tyrosine kinase inhibitors [52,53,54,55]. Two intronic BCL2L11 SNPs (rs6746608 and rs1261324) were also reported as associated with decreased risk of non-Hodgkin lymphoma [39].

Candidate genes and genome-wide association studies (GWAS) identified several polymorphisms underlying ON development in ALL patients [24,25,26,27], including the plasminogen activator inhibitor-1 (PAI-1) rs6092 SNP [27]; the rs10989692 in the glutamate receptor subunit 3A (GRIN3A) gene [25] and rs12714403 at the SH3YL1-acid phosphatase-1 (SH3YL1-ACP1) locus [26]. Glutamate receptor variants were associated with vascular phenotypes, like cerebral ischemia, arterial embolism and thrombosis [25], whereas ACP1 gene plays a role in lipid levels regulation and osteoblast differentiation [24]. Two additional SNPs were identified though GWAS conducted in children younger than 10 years, rs75161997 that is polymorphic in non-Caucasians and located near the bone morphogenesis protein 7 (BMP7) gene, and rs1891059 at PROX1-antisense RNA1 (PROX1-AS1) locus. [24].

In conclusion, our study identified additional pharmacogenetics markers contributing to ON development in childhood ALL patients, located in BCL2L11 gene. The TT891 on the background of CC29201 genotype increased the risk of ON during the treatment. The finding was further supported by increased in vitro cellular sensitivity to CS associated with these genotypes coupled with lower expression of γ Bim mRNA isoforms, suggesting a possible shift toward other BH3-containing isoforms which could induce osteoblast apoptosis, driving ON.

References

  1. 1.

    Patel B, Richards SM, Rowe JM, Goldstone AH, Fielding AK. High incidence of avascular necrosis in adolescents with acute lymphoblastic leukaemia: a UKALL XII analysis. Leukemia 2008;22:308–12.

  2. 2.

    Mattano LA Jr., Sather HN, Trigg ME, Nachman JB. Osteonecrosis as a complication of treating acute lymphoblastic leukemia in children: a report from the Children's Cancer Group. J Clin Oncol 2000;18:3262–72.

  3. 3.

    Arico M, Boccalatte MF, Silvestri D, Barisone E, Messina C, Chiesa R, et al Osteonecrosis: An emerging complication of intensive chemotherapy for childhood acute lymphoblastic leukemia. Haematologica 2003;88:747–53.

  4. 4.

    te Winkel ML, Appel IM, Pieters R, van den Heuvel-Eibrink MM. Impaired dexamethasone-related increase of anticoagulants is associated with the development of osteonecrosis in childhood acute lymphoblastic leukemia. Haematologica 2008;93:1570–4.

  5. 5.

    Belgaumi AF, Al-Bakrah M, Al-Mahr M, Al-Jefri A, Al-Musa A, Saleh M, et al Dexamethasone-associated toxicity during induction chemotherapy for childhood acute lymphoblastic leukemia is augmented by concurrent use of daunomycin. Cancer 2003;97:2898–903.

  6. 6.

    Chan KL, Mok CC. Glucocorticoid-induced avascular bone necrosis: diagnosis and management. Open Othoped J 2012;6:449–57.

  7. 7.

    Hyakuna N, Shimomura Y, Watanabe A, Taga T, Kikuta A, Matsushita T, et al Assessment of corticosteroid-induced osteonecrosis in children undergoing chemotherapy for acute lymphoblastic leukemia: a report from the Japanese Childhood Cancer andLeukemia Study Group. J Pediatr Hematol Oncol 2014;36:22–29.

  8. 8.

    Strauss AJ, Su JT, Dalton VM, Gelber RD, Sallan SE, Silverman LB. Bony morbidity in children treated for acute lymphoblastic leukemia. J Clin Oncol 2001;19:3066–72.

  9. 9.

    Silverman LB, Stevenson KE, O'Brien JE, Asselin BL, Barr RD, Clavell L, et al Long-term results of Dana-Farber Cancer Institute ALL Consortium protocols for children with newly diagnosed acute lymphoblastic leukemia (1985-2000). Leukemia 2010;24:320–34.

  10. 10.

    Silverman LB, Gelber RD, Dalton VK, Asselin BL, Barr RD, Clavell LA, et al Improved outcome for children with acute lymphoblastic leukemia: results of Dana-Farber Consortium Protocol 91-01. Blood 2001;97:1211–8.

  11. 11.

    Moghrabi A, Levy DE, Asselin B, Barr R, Clavell L, Hurwitz C, et al Results of the Dana-Farber Cancer Institute ALL Consortium Protocol 95-01 for children with acute lymphoblastic leukemia. Blood 2007;109:896–904.

  12. 12.

    Mazziotti G, Giustina A, Canalis E, Bilezikian JP. Glucocorticoid-induced osteoporosis: clinical and therapeutic aspects. Arq Bras Endocrinol Metabol 2007;51:1404–12.

  13. 13.

    Alos N, Grant RM, Ramsay T, Halton J, Cummings EA, Miettunen PM, et al High incidence of vertebral fractures in children with acute lymphoblastic leukemia 12 months after the initiation of therapy. J Clin Oncol 2012;30:2760–7.

  14. 14.

    Weinstein RS, Jilka RL, Parfitt AM, Manolagas SC. Inhibition of osteoblastogenesis and promotion of apoptosis of osteoblasts and osteocytes by glucocorticoids. Potential mechanisms of their deleterious effects on bone. J Clin Invest 1998;102:274–82.

  15. 15.

    Bookout AL, Jeong Y, Downes M, Yu RT, Evans RM, Mangelsdorf DJ. Anatomical profiling of nuclear receptor expression reveals a hierarchical transcriptional network. Cell 2006;126:789–99.

  16. 16.

    Abu EO, Horner A, Kusec V, Triffitt JT, Compston JE. The localization of the functional glucocorticoid receptor alpha in human bone. J Clin Endocrinol Metab 2000;85:883–9.

  17. 17.

    Biddie SC, Conway-Campbell BL, Lightman SL. Dynamic regulation of glucocorticoid signalling in health and disease. Rheumatology (Oxford) 2012;51:403–12.

  18. 18.

    Huang DC, Strasser A. BH3-Only proteins-essential initiators of apoptotic cell death. Cell 2000;103:839–42.

  19. 19.

    Webster JC, Cidlowski JA. Downregulation of the glucocorticoid receptor. A mechanism for physiological adaptation to hormones. Ann N Y Acad Sci 1994;746:216–20.

  20. 20.

    Tonko M, Ausserlechner MJ, Bernhard D, Helmberg A, Kofler R. Gene expression profiles of proliferating vs. G1/G0 arrested human leukemia cells suggest a mechanism for glucocorticoid-induced apoptosis. FASEB J 2001;15:693–9.

  21. 21.

    Liu C, Janke LJ, Kawedia JD, Ramsey LB, Cai X, Mattano LA Jr, et al Asparaginase potentiates glucocorticoid-induced osteonecrosis in a mouse model. PLoS ONE 2016;11:e0151433.

  22. 22.

    Niinimaki RA, Harila-Saari AH, Jartti AE, Seuri RM, Riikonen PV, Paakko EL, et al High body mass index increases the risk for osteonecrosis in children with acute lymphoblastic leukemia. J Clin Oncol 2007;25:1498–504.

  23. 23.

    Bernbeck B, Mauz-Korholz C, Zotz RB, Gobel U. Methylenetetrahydrofolate reductase gene polymorphism and glucocorticoid intake in children with ALL and aseptic osteonecrosis. Klin Padiatr 2003;215:327–31.

  24. 24.

    Karol SE, Mattano LA Jr., Yang W, Maloney KW, Smith C, Liu C, et al Genetic risk factors for the development of osteonecrosis in children under age 10 treated for acute lymphoblastic leukemia. Blood 2016;127:558–64.

  25. 25.

    Karol SE, Yang W, Van Driest SL, Chang TY, Kaste S, Bowton E, et al Genetics of glucocorticoid-associated osteonecrosis in children with acute lymphoblastic leukemia. Blood 2015;126:1770–6.

  26. 26.

    Kawedia JD, Kaste SC, Pei D, Panetta JC, Cai X, Cheng C, et al Pharmacokinetic, pharmacodynamic, and pharmacogenetic determinants of osteonecrosis in children with acute lymphoblastic leukemia. Blood 2011;117:2340–7. quiz 2556

  27. 27.

    French D, Hamilton LH, Mattano LA Jr, Sather HN, Devidas M, Nachman JB, et al A PAI-1 (SERPINE1) polymorphism predicts osteonecrosis in children with acute lymphoblastic leukemia: a report from the Children's Oncology Group. Blood 2008;111:4496–9.

  28. 28.

    Medh RD, Webb MS, Miller AL, Johnson BH, Fofanov Y, Li T, et al Gene expression profile of human lymphoid CEM cells sensitive and resistant to glucocorticoid-evoked apoptosis. Genomics 2003;81:543–55.

  29. 29.

    O'Connor L, Strasser A, O'Reilly LA, Hausmann G, Adams JM, Cory S, et al Bim: a novel member of the Bcl-2 family that promotes apoptosis. EMBO J 1998;17:384–95.

  30. 30.

    Iglesias-Serret D, de Frias M, Santidrian AF, Coll-Mulet L, Cosialls AM, Barragan M, et al Regulation of the proapoptotic BH3-only protein BIM by glucocorticoids, survival signals and proteasome in chronic lymphocytic leukemia cells. Leukemia 2007;21:281–7.

  31. 31.

    Espina B, Liang M, Russell RG, Hulley PA. Regulation of bim in glucocorticoid-mediated osteoblast apoptosis. J Cell Physiol 2008;215:488–96.

  32. 32.

    Gagne V, Rousseau J, Labuda M, Sharif-Askari B, Brukner I, Laverdiere C, et al Bim polymorphisms: influence on function and response to treatment in children with acute lymphoblastic leukemia. Clin Cancer Res 2013;19:5240–9.

  33. 33.

    Vrooman LM, Stevenson KE, Supko JG, O'Brien J, Dahlberg SE, Asselin BL, et al Postinduction dexamethasone and individualized dosing of Escherichia Coli L-asparaginase each improve outcome of children and adolescents with newly diagnosed acute lymphoblastic leukemia: results from a randomized study--Dana-Farber Cancer Institute ALL Consortium Protocol 00-01. J Clin Oncol 2013;31:1202–10.

  34. 34.

    DCTD NCI (2006). National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0.

  35. 35.

    Labuda D, Krajinovic M, Richer C, Skoll A, Sinnett H, Yotova V, et al Rapid detection of CYP1A1, CYP2D6, and NAT variants by multiplex polymerase chain reaction and allele-specific oligonucleotide assay. Anal Biochem 1999;275:84–92.

  36. 36.

    Storey JDTJ, Siegmund D. . Strong control, conservative point estimation and simultaneous conservative consistency of false discovery rates: a unified approach. J R Stat Soc 2004;66:187–205.

  37. 37.

    Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (San Diego, Calif) 2001;25:402–8.

  38. 38.

    Anczukow O, Rosenberg AZ, Akerman M, Das S, Zhan L, Karni R, et al The splicing factor SRSF1 regulates apoptosis and proliferation to promote mammary epithelial cell transformation. Nat Struct Mol Biol 2012;19:220–8.

  39. 39.

    Kelly JL, Novak AJ, Fredericksen ZS, Liebow M, Ansell SM, Dogan A, et al Germline variation in apoptosis pathway genes and risk of non-Hodgkin's lymphoma. Cancer Epidemiol Biomark Prevent 2010;19:2847–58.

  40. 40.

    Miao J, Chen GG, Yun JP, Chun SY, Zheng ZZ, Ho RL, et al Identification and characterization of BH3 domain protein Bim and its isoforms in human hepatocellular carcinomas. Apoptosis 2007;12:1691–701.

  41. 41.

    Mattano LA Jr, Devidas M, Nachman JB, Sather HN, Hunger SP, Steinherz PG, et al Effect of alternate-week versus continuous dexamethasone scheduling on the risk of osteonecrosis in paediatric patients with acute lymphoblastic leukaemia: results from the CCG-1961 randomised cohort trial. Lancet Oncol 2012;13:906–15.

  42. 42.

    Sato M, Sugano N, Ohzono K, Nomura S, Kitamura Y, Tsukamoto Y, et al Apoptosis and expression of stress protein (ORP150, HO1) during development of ischaemic osteonecrosis in the rat. J Bone Joint Surg Br 2001;83:751–9.

  43. 43.

    Zalavras C, Shah S, Birnbaum MJ, Frenkel B. Role of apoptosis in glucocorticoid-induced osteoporosis and osteonecrosis. Crit Rev Eukaryot Gene Expr 2003;13:221–35.

  44. 44.

    Kabata T, Kubo T, Matsumoto T, Nishino M, Tomita K, Katsuda S, et al Apoptotic cell death in steroid induced osteonecrosis: an experimental study in rabbits. J Rheumatol 2000;27:2166–71.

  45. 45.

    Nott A, Meislin SH, Moore MJ. A quantitative analysis of intron effects on mammalian gene expression. RNA (New York, NY) 2003;9:607–17.

  46. 46.

    Ge W, Shi L, Zhou Y, Liu Y, Ma GE, Jiang Y, et al Inhibition of osteogenic differentiation of human adipose-derived stromal cells by retinoblastoma binding protein 2 repression of RUNX2-activated transcription. Stem Cells 2011;29:1112–25.

  47. 47.

    Heidari N, Miller AV, Hicks MA, Marking CB, Harada H. Glucocorticoid-mediated BIM induction and apoptosis are regulated by Runx2 and c-Jun in leukemia cells. Cell Death Dis 2012;3:e349.

  48. 48.

    Bachmann PS, Gorman R, Papa RA, Bardell JE, Ford J, Kees UR, et al Divergent mechanisms of glucocorticoid resistance in experimental models of pediatric acute lymphoblastic leukemia. Cancer Res 2007;67:4482–90.

  49. 49.

    Bachmann PS, Gorman R, Mackenzie KL, Lutze-Mann L, Lock RB. Dexamethasone resistance in B-cell precursor childhood acute lymphoblastic leukemia occurs downstream of ligand-induced nuclear translocation of the glucocorticoid receptor. Blood 2005;105:2519–26.

  50. 50.

    Wang Z, Malone MH, He H, McColl KS, Distelhorst CW. Microarray analysis uncovers the induction of the proapoptotic BH3-only protein Bim in multiple models of glucocorticoid-induced apoptosis. J Biol Chem 2003;278:23861–7.

  51. 51.

    Hunt R, Sauna ZE, Ambudkar SV, Gottesman MM, Kimchi-Sarfaty C. Silent (synonymous) SNPs: should we care about them? Methods Mol Biol 2009;578:23–39.

  52. 52.

    Augis V, Airiau K, Josselin M, Turcq B, Mahon FX, Belloc F. A single nucleotide polymorphism in cBIM is associated with a slower achievement of major molecular response in chronic myeloid leukaemia treated with imatinib. PLoS ONE 2013;8:e78582.

  53. 53.

    Ng KP, Hillmer AM, Chuah CT, Juan WC, Ko TK, Teo AS, et al A common BIM deletion polymorphism mediates intrinsic resistance and inferior responses to tyrosine kinase inhibitors in cancer. Nat Med 2012;18:521–8.

  54. 54.

    Katagiri STT, Tadokoro K, Ohyashiki JH, Ohyashiki K. BCL2L11 (BIM) deletion polymorphism, rather than BIM SNP, is an effective predictor of early molecular relapse after ABL tyrosine kinase inhibitor discontinuation in patients with chronic myeloid leukemia. Integr Cancer Sci Therap 2015;2:242–4.

  55. 55.

    Lee JH, Lin YL, Hsu WH, Chen HY, Chang YC, Yu CJ, et al Bcl-2-like protein 11 deletion polymorphism predicts survival in advanced non-small-cell lung cancer. J Thorac Oncol 2014;9:1385–92.

Download references

Acknowledgements

We thank all patients and their parents who consented to participate in genetics studies related to leukemia. Leukemia Lymphoma Society of Canada, Canadian Institutes of Health Research and Charles Bruneau Foundation supported this study. Dana-Farber

Cancer Institute ALL treatment protocols are supported by the National Cancer Institute/NIH grant 5 P01CA06848

Author contributions

M.K. designed the study; M.P., V.G., M.Y. and B. S-A. performed experiments; M.P. and S.J-G performed medical chart reviews; C.L. JM.L., N.A, S.E.S., D.N. J.K, L.B.S and D.S. contributed to sample and clinical data collection and interpretation; M.P., V.G. and M.K. performed the data analysis; M.P. and M.K. drafted the article; All authors contributed to the interpretation of data and revised the manuscript.

Author information

Affiliations

  1. Charles-Bruneau Cancer Center, CHU Sainte-Justine Research Center, Montreal, QC, Canada

    • Maria Plesa
    • , Vincent Gagné
    • , Sanja Glisovic
    • , Melissa Younan
    • , Bahram Sharif-Askari
    • , Caroline Laverdière
    • , Nathalie Alos
    • , Jean-Marie Leclerc
    • , Daniel Sinnett
    •  & Maja Krajinovic
  2. Department of Pediatrics, Faculty of Medicine, University of Montreal, Montreal, QC, Canada

    • Caroline Laverdière
    • , Nathalie Alos
    • , Jean-Marie Leclerc
    • , Daniel Sinnett
    •  & Maja Krajinovic
  3. Department of Pharmacology, Faculty of Medicine, University of Montreal, Montreal, QC, Canada

    • Maria Plesa
    •  & Maja Krajinovic
  4. Department of Pediatric Oncology, Dana-Farber Cancer Institute, Boston, MA, USA

    • Stephen E Sallan
    •  & Lewis B Silverman
  5. Division of Hematology/Oncology, Children’s Hospital, Boston, MA, USA

    • Stephen E Sallan
    •  & Lewis B Silverman
  6. Department of Biostatistics & Computational Biology, Dana-Farber Cancer Institute, Boston, MA, USA

    • Donna Neuberg
  7. Department of Pathology, Brigham and Women’s Hospital, Boston, MA, USA

    • Jeffery L Kutok

Authors

  1. Search for Maria Plesa in:

  2. Search for Vincent Gagné in:

  3. Search for Sanja Glisovic in:

  4. Search for Melissa Younan in:

  5. Search for Bahram Sharif-Askari in:

  6. Search for Caroline Laverdière in:

  7. Search for Nathalie Alos in:

  8. Search for Jean-Marie Leclerc in:

  9. Search for Stephen E Sallan in:

  10. Search for Donna Neuberg in:

  11. Search for Jeffery L Kutok in:

  12. Search for Lewis B Silverman in:

  13. Search for Daniel Sinnett in:

  14. Search for Maja Krajinovic in:

Conflict of interest

The authors declare that they have no competing interests

Corresponding author

Correspondence to Maja Krajinovic.

Electronic supplementary material

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/s41397-017-0002-4