Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Oct 10.
Published in final edited form as: J Clin Oncol. 2024 Aug 5;42(29):3478–3490. doi: 10.1200/JCO.24.00500

Impact of Age on Pharmacogenomics and Treatment Outcomes of B-cell Acute Lymphoblastic Leukemia

Satoshi Yoshimura 1,#, Zhenhua Li 1,#, Yoshihiro Gocho 1, Wenjian Yang 1, Kristine R Crews 1, Shawn H R Lee 1,2,3, Kathryn G Roberts 4, Charles G Mullighan 4, Mary V Relling 1, Jiyang Yu 5, Allen E J Yeoh 2,3, Mignon L Loh 6, Caner Saygin 7, Mark R Litzow 8, Sima Jeha 9, Seth E Karol 9, Hiroto Inaba 9, Ching-Hon Pui 9, Marina Konopleva 10, Nitin Jain 11, Wendy Stock 7, Elisabeth Paietta 12, Elias Jabbour 11, Steven M Kornblau 11, William E Evans 1, Jun J Yang 1,9
PMCID: PMC11458355  NIHMSID: NIHMS1996246  PMID: 39102629

Abstract

Purpose:

Acute lymphoblastic leukemia (ALL) can occur across all age groups, with a strikingly higher cure rate in children compared to adults. However, the pharmacological basis of age-related differences in ALL treatment response remains unclear.

Patients and Methods:

Studying 767 children and 309 adults with newly diagnosed B-cell ALL enrolled on frontline trials at St. Jude Children’s Research Hospital, MD Anderson Cancer Center, the Alliance for Clinical Trials in Oncology, and the ECOG-ACRIN Cancer Research Group, we determined the ex vivo sensitivity of leukemia cells to 21 drugs. 23 ALL molecular subtypes were identified using RNA-seq. We systematically characterized the associations between drug response and ALL genomics in children, adolescents and young adults, and elderly adults. We evaluated the impact of age-related gene expression signature on ALL treatment outcomes.

Results:

Seven ALL drugs (asparaginase, prednisolone, mercaptopurine, dasatinib, nelarabine, daunorubicin, and inotuzumab ozogamicin) showed differential activity between children and adults, of which six were explained by age-related differences in leukemia molecular subtypes. Adolescents and young adults showed similar patterns of drug resistance as older adults, relative to young children. Mercaptopurine exhibited subtype-independent greater sensitivity in children. Transcriptomic profiling uncovered subclusters within CRLF2-, DUX4-, and KMT2A-rearranged ALL that were linked to age and cytotoxic drug resistance. In particular, a subset of children had “adult-like” ALL based on leukemia gene expression patterns across subtypes, despite their chronological age. Resistant to cytotoxic drugs, pediatric patients with adult-like ALL exhibited poor prognosis in pediatric ALL trials, even after adjusting for age and minimal residual diseases.

Conclusion:

Our results provide pharmacogenomic insights into the age-related disparities in ALL cure rates and identify leukemia prognostic features for treatment individualization across age groups.

Introduction

Most cancers occur primarily in adults, although a few affect individuals across the entire lifespan.1 Such malignancies often exhibit significant age-related differences in disease biology,2 a phenomenon particularly prominent in the case of acute lymphoblastic leukemia (ALL).3 Children with ALL generally have outstanding treatment outcomes with overall cure rates exceeding 85%,4–6 whereas long-term survival for adult ALL remains suboptimal (ranging from 50% to 75%).7–9 With comprehensive elucidation of the molecular subtypes of ALL, especially B-cell ALL (B-ALL), this disparity has been partly explained by the distinct pattern of disease taxonomy. For instance, favorable subtypes such as ETV6::RUNX1 and high hyperdiploid are overrepresented in pediatric ALL, and high-risk subtypes including BCR::ABL1, BCR::ABL1-like, and KMT2A B-ALL are more prevalent in adults.3,10,11

Early studies of ALL sensitivity to cytotoxic drugs documented that ex vivo sensitivity is related to treatment outcome,12,13 and subsequently provided initial indications of relative resistance to cytarabine, asparaginase, daunorubicin and glucocorticoids in adults, although these studies suffered from small sample sizes and very limited molecular subtyping.14,15 However, the inherently inferior drug response in adult ALL can potentially be overcome with increasing dosages and duration of chemotherapy, as suggested by the recent success in applying pediatric-inspired treatment regimens for young adults.16,17 It is plausible that the ALL drug response phenotype spans a continuum across various age groups with variable contribution by leukemia genomics. Additionally, there is a particular paucity of data on age-related differences in ALL response to targeted therapeutics, including tyrosine kinase inhibitors and antibody-conjugated drugs.18–23 Without comprehensive characterization of both ALL drug sensitivity and leukemia genomics, the biological and pharmacological basis of age-related disparities in ALL remain poorly understood.

Pharmacotyping, involving the ex vivo profiling of drug sensitivity of primary tumor specimens, directly assesses cancer cell response to therapeutic agents, providing a phenotypic readout that complements genomics-based precision medicine.24,25 This is important because many tumor genomic aberrations are not known to be therapeutically targetable. Moreover, drug response is almost always multifactorial and predicting it from tumor genotypes alone is challenging, if not impossible.26 In fact, recent trials incorporating drug sensitivity profiles into personalized therapy for adults with aggressive hematological malignancies demonstrated both the feasibility of this approach and clinical benefits.24,27 For pediatric ALL, our group and others recently reported that leukemia pharmacotypes are associated with treatment response and survival outcomes,25,28,29 pointing to their potential prognostic utility in risk stratification. ALL drug sensitivity profiling can also provide new insights into the pharmacological heterogeneity of this disease and, in some cases, lead to the discovery of novel therapeutic vulnerabilities.28,30–32

In this study, we systematically evaluated ALL drug sensitivity in 767 children and 309 adults, demonstrating age-related differences in response to 21 therapeutic agents. In parallel, we comprehensively determined 23 B-ALL molecular subtypes in this cohort and examined their contribution to the association of age with drug resistance. Our results revealed age-related differences in drug response within ALL subtypes, in part explained by leukemia transcriptomic heterogeneity.

Methods

Patients and samples

The primary ALL pharmacotyping cohort in this study consisted of 767 pediatric (≤18 years old) and 309 adult (≥19 years old) newly diagnosed B-ALL patients (hereafter referred to as “pediatric/children” or “adult/adults”, respectively, unless otherwise specified). In general, pediatric samples were obtained from three consecutive ALL Total Therapy protocols led by St. Jude Children’s Research Hospital (SJCRH): Total XV (NCT00137111),33 XVI (NCT00549848),34 and XVII (NCT03117751); adult samples were collected from MD Anderson Cancer Center, the ECOG-ACRIN E1910 clinical trial (NCT02003222), and the Alliance A041501 trial (NCT03150693). Sample collection and testing were strictly based on the availability of specimens without other filtering. This study was approved by the respective institutional review boards and informed consent was obtained from parents, guardians and/or patients, as appropriate. Additional samples and data were obtained for validation of pharmacotyping and treatment outcome analyses, which are described in Supplemental Methods.

These leukemic samples were subjected to ex vivo drug sensitivity testing, RNA sequencing (RNA-seq), and DNA methylation profiling, the details of which can be found in Supplemental Methods.

Statistical analyses

Drug sensitivity was quantified as LC50, the concentration lethal to 50% of the cells, and each value was normalized to the range between 0 (most sensitive) and 1 (most resistant) (See Supplemental Methods). Unless otherwise stated, associations between two categorical variables were assessed using χ2 test. Associations involving a continuous variable (including drug LC50 values) and a categorical variable were assessed using Kruskal-Wallis test (for more than two categories) or two-sided Mann-Whitney U-test (for two categories). To address multiple comparisons, the Bonferroni correction was applied, with an adjusted P-value <0.05 to determine significance. To test if the LC50 differences between the age groups were independent of subtype, multivariate linear regressions were performed with normalized LC50 as dependent variable and age group, sex, ancestry, and ALL subtype as independent variables, and covariates were considered contributing toward sensitivity to the drug when their coefficients were negative. All analyses were performed with R (version 4.1.0; http://www.r-project.org). Further details of statistical analyses are provided in Supplemental Methods.

Results

B-ALL patient characteristics in children and adults

The B-ALL pharmacotyping cohort consisted of 1,076 patients with newly-diagnosed disease: 767 pediatric (≤18 years old) and 309 adult (≥19 years old) leukemia samples were evaluated for ex vivo sensitivity to a panel of 21 drugs, yielding a total of 7,975 unique LC50 measurements (Figure 1A). There was no difference in sex distribution between the two age groups (Table 1 and Figure 1B). There were more patients of African descent in children than in adults, and Admixed Americans were over-represented in adults, which was likely driven by population demographics within catchment areas of our pediatric vs adult institutions.35,36 However, there are complex interactions between ALL subtype, genetic ancestry, and age that may confound these analyses.37 In total, we identified 23 ALL molecular subtypes defined by fusion genes, gene expression profile, and/or sequence mutations.38 Consistent with prior reports,3,11 seven molecular subtypes occurred with different incidence by age. Specifically, ETV6::RUNX1 and high hyperdiploid were exclusive to children, while BCR::ABL1, KMT2A, CRLF2, BCR::ABL1-like, and low hypodiploid were more prevalent in adults than children.

Figure 1. Overview of pharmacotyping of pediatric and adult ALL.

Figure 1.

Open in a new tab

A, Drug sensitivity was tested for primary diagnostic leukemia samples of 767 pediatric (≤18 years old) and 309 adult (≥19 years old) B-ALL patients for 21 drugs. In the word cloud, the font size of each drug name is proportional to the number of samples tested for the drug. A total of 7,975 LC50 measurements were obtained. This panel was generated using BioRender.com. B, Circos plot of the age group, sex, ancestry group, subtype and availability of drug sensitivity data of the patients tested in this study.

Table 1.

B-ALL sample characteristics

All samples (N = 1,076) Pediatric (≤18 yrs, n = 767) Adult (≥19 yrs, n = 309) P-value
Age, median (IQR), years 8.0 (3.6–29.0) 4.8 (2.9–9.2) 47.0 (36.0–59.5)
Sex, No. (%)
Male 576 (53.5) 413 163 0.80
Female 500 (46.5) 354 146
Ancestry, No. (%)
European 691 (64.2) 500 (65.2) 191 (61.8) 1.1× 10−4
Admixed American 164 (15.2) 100 (13) 64 (20.7)
African 119 (11.1) 100 (13) 19 (6.1)
Asian 16 (1.5) 12 (1.6) 4 (1.3)
Other 59 (5.5) 42 (5.5) 17 (5.5)
Unknown 27 (2.5) 13 (1.7) 14 (4.5)
Subtype, No. (%)
ETV6::RUNX1 200 (18.6) 200 (26.1) 0 (0) 5.0 × 10−4
High hyperdiploid 200 (18.6) 199 (25.9) 1 (0.3)
BCR::ABL1 128 (11.9) 24 (3.1) 104 (33.7)
KMT2A 99 (9.2) 50 (6.5) 49 (15.9)
CRLF2 70 (6.5) 30 (3.9) 40 (12.9)
DUX4 57 (5.3) 41 (5.3) 16 (5.2)
PAX5 alt 56 (5.1) 49 (6.4) 7 (2.3)
BCR::ABL1-like 46 (4.3) 21 (2.7) 25 (8.1)
TCF3::PBX1 41 (3.8) 34 (4.4) 7 (2.3)
Low hypodiploid 22 (2) 6 (0.8) 16 (5.2)
ETV6::RUNX1-like 19 (1.8) 19 (2.5) 0 (0)
MEF2D 16 (1.5) 10 (1.3) 6 (1.9)
ZNF384 16 (1.5) 10 (1.3) 6 (1.9)
Near haploid 9 (0.8) 9 (1.2) 0 (0)
PAX5 P80R 9 (0.8) 5 (0.7) 4 (1.3)
iAMP21 8 (0.7) 7 (0.9) 1 (0.3)
BCL2/MYC 6 (0.6) 3 (0.4) 3 (1)
NUTM1 6 (0.6) 6 (0.8) 0 (0)
ZNF384-like 3 (0.3) 1 (0.1) 2 (0.6)
KMT2A-like 2 (0.2) 1 (0.1) 1 (0.3)
TCF3::HLF 2 (0.2) 2 (0.3) 0 (0)
IKZF1 N159Y 1 (0.1) 1 (0.1) 0 (0)
B-other 60 (5.6) 39 (5.1) 21 (6.8)
Open in a new tab

The impact of age on ALL ex vivo sensitivity to chemotherapeutic agents

Using normalized LC50 as a measurement of drug sensitivity, we first compared results between children and adults. A total of 21 drugs were tested, representing six drug classes: antimetabolites (cytarabine, mercaptopurine, and nelarabine), non-antimetabolite cytotoxics (asparaginase, bortezomib, daunorubicin, prednisolone, and vincristine), kinase inhibitors (CHZ868, dasatinib, gilteritinib, ibrutinib, momelotinib, ruxolitinib, trametinib, palbociclib, and idelalisib), antibody-drug conjugates (inotuzumab ozogamicin, hereafter referred to as inotuzumab), histone deacetylase inhibitors (panobinostat and vorinostat), and BH3-mimetics (venetoclax) (Supplemental Table 1). Of these, seven drugs showed different activity between age groups (Figure 2A): children were more sensitive (i.e., lower LC50) to asparaginase (P=5.7×10−13), prednisolone (P=2.4×10−10), mercaptopurine (P=1.9×10−9), daunorubicin (P=0.0019), and inotuzumab (P=0.0019); and adults were more sensitive to dasatinib (P=3.7×10−8) and nelarabine (P=3.9×10−4) (Figure 2B-G).

Figure 2. ALL drug sensitivities differ between children and adults.

Figure 2.

Open in a new tab

A, The difference of median LC50 between age groups. The x-axis indicates the subtraction of median LC50 (pediatric minus adult) for each drug while the y-axis shows the significance of the difference using two-sided Mann–Whitney U-test. The red line denotes nominal P-value of 0.0024 (adjusted P-value equal to 0.05 after correction for multiple testing). The size of circles indicates the number of samples tested. B-H, Overall drug LC50 distributions are shown in violin plots comparing age groups for asparaginase (B), prednisolone (C), mercaptopurine (D), dasatinib (E), nelarabine (F), daunorubicin (G), and inotuzumab (H). The number of patients in each category is indicated in parenthesis and represents biologically independent samples. The median LC50 for each age group is shown as a bold black line. Nominal P-values comparing LC50 values are as shown and were determined by two-sided Mann–Whitney U-test.

We also explored drug sensitivity profiles of AYA ALL (i.e., age from 15 to 39 years).39 Compared to children 14 years old and younger, AYA ALLs showed a higher LC50 for asparaginase, mercaptopurine, and prednisolone (adjusted P<0.05); by contrast, AYAs showed no difference in drug LC50 when compared to adults older than 40 (Supplemental Figure 1 and Supplemental Table 2). These observations suggested that AYA B-ALL displayed a closer resemblance to adults than to children, at least in terms of drug sensitivity.

The interaction between age and subtype and its effects on ALL drug sensitivity

Because ALL drug sensitivity can be substantially influenced by leukemia molecular subtype,25 we next performed multivariate analysis incorporating subtypes, sex, and ancestry included as covariates for the seven drugs with significant age differences in the univariate analysis (Supplemental Figure 2). Six (asparaginase, prednisolone, dasatinib, nelarabine, daunorubicin, and inotuzumab) were no longer significant after adjusting for subtype, suggesting age-dependent subtype distribution was the main driver of their differential activities between children and adults (Supplemental Table 3). We also performed multivariate analysis treating age as a continuous variable and the results were largely the same (Supplemental Table 4). To identify subtypes which most significantly contributed to age-related differences in sensitivity to each ALL drug, we performed forward stepwise regression analysis, using LC50 and age as dependent and independent variables, respectively. In each step, we selected the subtype that most effectively attenuated the association between age and drug sensitivity, i.e., including the subtype in the regression model resulted in the greatest increase of the P-value for age. These iterative steps continued until age no longer retained significance in the multivariate model (adjusted P-value >0.05). In this analysis, we showed that age-related differences in sensitivity to dasatinib and inotuzumab were driven by a single subtype, namely BCR::ABL1 and high hyperdiploid, respectively. By contrast, pediatric ALL sensitivity to asparaginase and prednisolone was driven by two or more subtypes: ETV6::RUNX1 and high hyperdiploid for asparaginase, and ETV6::RUNX1, high hyperdiploid, and TCF3::PBX1 for prednisolone (Figure 3A, Supplemental Figure 3, and Supplemental Table 5).

Figure 3. The difference in drug response between pediatric and adult ALL can be attributed to distinct subtype compositions.

Figure 3.

Open in a new tab

A, Contribution to the significance (as measured by P-value in log scale) of the association between age and drug LC50 by subtype is determined using forward stepwise regressions. For each drug, the right end of the bar indicates the P-value (in log scale) for the association between age and LC50 in univariate analysis, while an arrowhead in the bar points to the P-value for the same association after a molecular subtype is added to the multivariate model. The red line indicates where adjusted P-values for age are 0.05. B, Comparison of the LC50 of mercaptopurine in CRLF2, DUX4, and KMT2A ALL subtypes. The median LC50 for each group is shown as a bold red line. Nominal P-values comparing LC50 values are as shown and were determined by two-sided Mann–Whitney U-test.

By contrast, mercaptopurine resistance in adults and sensitivity in children were independent of ALL subtype, and this was particularly notable within CRLF2, DUX4, and KMT2A ALL. Within each of these three subtypes, mercaptopurine LC50 was consistently lower in pediatric cases than in adults (CRLF2, P=0.0045; DUX4, P=0.020; KMT2A, P=0.032; Figure 3B and Supplemental Figure 4).

Age-dependent genomic heterogeneity within CRLF2, DUX4, and KMT2A subtypes associated with drug sensitivity

The associations between age and mercaptopurine response in cases harboring the same sentinel genomic abnormalities implied potential biological heterogeneity within each subtype. Using leukemia gene expression profiling, we performed unsupervised clustering within ALL subtypes and identified distinct clusters associated with both age and drug response within CRLF2, DUX4, and KMT2A subtypes. Surprisingly in all three subtypes, two clusters were identified, with one dominated by pediatric (hereafter termed as C1) and the other by adult (termed as C2) patients (Figure 4A, 4C, and 4E).

Figure 4. Age-associated heterogeneity within CRLF2, DUX4, and KMT2A ALL subtypes.

Figure 4.

Open in a new tab

A, UMAP of the gene expression profiles in CRLF2-rearranged B-ALL reveals two distinct subclusters that are associated with different age groups. B, Alluvial diagram of the sub-cluster identity, age group, presence of BCR::ABL1-like signature, and CRLF2 rearrangement partner in CRLF2 subtype. C, UMAP of the gene expression profiles of DUX4-rearranged B-ALL showing two clusters correlated with different age groups. D, Alluvial diagram showing the cluster identity, age group, CEBPA expression (high, above median; low, below median), and ERG deletion (ERGdel) status. E, UMAP of the gene expression profiles of KMT2A-rearranged B-ALL, two clusters with distinct age groups are observed. F, Alluvial diagram showing the associations among subcluster identity, age group, and KMT2A rearrangement partner within the KMT2A subtype. G, H, Comparison of LC50 of mercaptopurine (G) and prednisolone (H) among samples in different subclusters and age groups in CRLF2, DUX4, and KMT2A subtypes. C1 and C2 are combined for adult samples due to smaller number belonging to C1 (n=1 for both drugs). The median LC50 value for each group is indicated as a bold red horizontal line.

In CRLF2 rearranged ALL, C1 was dominated by pediatric patients (92%), compared to 24% in C2 (P=4.8×10−6; Figure 4A and 4B). Most cases in C2 displayed BCR::ABL1-like transcriptomic signature (90%, Figure 4B) with IGH::CRLF2 rearrangement (78%, Figure 4B), while C1 was mostly non BCR::ABL1-like (85%, P=2.7×10−8) with P2RY8::CRLF2 rearrangements (85%, P=2.5×10−6). Remarkably, 100% of Admixed American patients with CRLF2 ALL, regardless of their age, were found in C2 (P=0.0021; Supplemental Figure 5).

Similarly, in DUX4 ALL, all cases in C1 (n=22) were children (Figure 4C and 4D), compared to 45% in C2 (P=1.4×10−4). Meanwhile, all cases in C2 of DUX4 ALL showed high CEBPA expression (>median), compared to only 32% in C1 (P=1.8×10−8; Figure 4D), consistent with prior reports.38,40 ERG deletions, which were previously described as exclusively present in DUX4 cases,41 were more frequent in C1 than C2 (P=0.015; Figure 4D).

For KMT2A rearranged ALL, 96% of cases in C1 were pediatric, compared to 24% in C2 (P=7.8×10−10; Figure 4E and 4F). The two clusters were also distinct in terms of the partner gene of KMT2A rearrangements (P=1.3×10−8; Figure 4F): the majority (84%) of the cases in C2 had fusions with AFF1, followed by MLLT1 (9.5%); In C1, the rearrangement partners of KMT2A were more heterogeneous, with MLLT1 being the most frequent at 48% and only 12% with AFF1.

Even though the subclusters within CRLF2, DUX4, and KMT2A exhibited remarkable associations with age, there were notable exceptions. Across these subtypes, 38% to 50% of pediatric ALL cases clustered together with adults, pointing to “adult-like” leukemia despite their chronological age. Pharmacologically, these “adult-like” leukemias exhibited marked resistance to mercaptopurine and prednisolone, similar to adult ALL and much more so than cases in the pediatric clusters (C1; Figure 4G and 4H).

Identification of “adult-like” pediatric ALL and its prognostic impact

To further explore “adult-like” ALL in pediatric patients, we performed differential expression analysis comparing the two subclusters within each of the CRLF2, DUX4, and KMT2A subtypes and identified 228 genes that were commonly up-regulated in C2 clusters (false discovery rate <0.01; Figure 5A and Supplemental Table 6). Applying this to all B-ALL cases in our cohort, we sought to identify adult-like pediatric cases based on the degree of enrichment of this 228-gene expression signature,42 referred to as “adult-likeness index (ALI)” hereafter. ALI was strongly associated with ex vivo drug resistance, particularly to mercaptopurine and prednisolone, which remained significant even after adjusting for molecular subtypes (Supplemental Figure 6). To validate this finding, we first analyzed a previously published pharmacotyping cohort of 144 pediatric B-ALL cases from the Dutch Childhood Oncology Group and German Cooperative Study Group for Childhood ALL trials13 and observed a significant association between ALI and prednisolone resistance (P=0.0015). Additionally, we profiled an independent cohort of 37 B-ALL cases for mercaptopurine LC50 at SJCRH and again confirmed the association of ALI with drug resistance (P=0.037).

Figure 5. Adult-likeness index on differentially expressed genes predicts clinical outcome.

Figure 5.

Open in a new tab

A, Venn diagram showing the up-regulated genes comparing C2 versus C1 within CRLF2, DUX4, and KMT2A subtypes. B,The correlation between adult-likeness index (ALI) and numerical age. The blue curve indicates the locally estimated scatterplot smoothing (LOESS) fitting of the data and the light blue area indicates the 95% confidence interval. C, The correlation between ALI and subdivided age groups. The red line indicates the cutoff (ALI=0.1) used in the subsequent analysis. D-G, Kaplan-Meier estimates of event-free or overall survival among pediatric B-ALL patients in SJCRH Total XV and XVI studies (D, E) and MaSpore 2003 and 2010 studies (F, G), stratified by ALI being ≥0.1 or <0.1. The numbers of patients at risk are shown beneath each panel.

ALI also varied by age: it was relatively high in infant ALL, but rapidly decreased in cases diagnosed at ages between 2 and 5; ALI gradually increased again with age and then plateaued at approximately age 30 and remained stable after that (Figure 5B). Overall, ALI displayed a bimodal distribution (≥ or <0.1, Figure 5C): there were few adult ALL cases with low ALI (<0.1; 4%), whereas a substantial number of pediatric patients had high ALI (≥0.1; 34%). Therefore, we defined “adult-like” cases as patients younger than 18 but with a leukemia ALI value of 0.1 or greater, and confirmed the presence of “adult-like” ALL across almost all molecular subtypes (Supplemental Figure 7).

We then examined if ALI is associated with clinical outcomes of pediatric ALL treated in frontline trials. RNA-seq and clinical data were analyzed for a total of 594 pediatric patients with B-ALL enrolled on SJCRH Total Therapy XV and XVI protocols.33,34 Compared to those with low ALI, patients with high ALI were significantly more likely to have positive (≥0.01%) minimal residual disease (MRD) at the end of induction (22% vs 12%; P=0.0056; Supplemental Figure 8) and were associated with worse outcomes (10-year event-free survivals, EFS, 80% vs 91%; P=7.2×10−4; overall survival, OS, 88% vs 95%; P=9.9×10−4; Figure 5D and 5E). ALI remained prognostic even after adjusting for NCI risk and MRD status (P=0.0064 and 0.013 for EFS and OS, respectively; Supplemental Table 7). In an independent validation cohort of 377 pediatric B-ALL patients treated on the Malaysia-Singapore (MaSpore) ALL frontline protocols,43,44 adult-like pediatric ALL patients (i.e., ALI ≥0.1) again exhibited significantly higher MRD positivity (69% vs 48%; P=5.8×10−4; Supplemental Figure 8) and worse EFS (77% vs 88%, P=0.020; Figure 5F) and OS (88% vs 94%, P=0.045; Figure 5G) than those with low ALI.

Discussion

While childhood ALL is largely curable with chemotherapy alone,4,5 the prognosis of adults with this cancer remains significantly worse.8,9 Because adult and pediatric ALL treatment regimens usually employ a similar set of chemotherapeutic agents,6,7 age-related differences in drug sensitivity have long been hypothesized yet not systematically investigated. Examining ALL ex vivo drug sensitivity and leukemia genomics, we showed significant differences in drug sensitivity (mostly cytotoxic agents) between children and adults, which were mainly explained by age-related differences in ALL subtypes. In particular, sensitivity to asparaginase and prednisolone in childhood ALL was driven by pediatric-specific subtypes, i.e., ETV6::RUNX1 and high hyperdiploid, in line with the previous reports.45–50

The exact mechanism by which age affects ALL biology is not clear. It is possible that pediatric and adult ALLs have differential cells of origin, i.e., arising in distinct hematopoietic compartments, even though the initiating oncogenic event is the same. Therefore, the transcriptomic program of the specific cell type might shape the drug response profile in patients from these age groups, in addition to onco-fusion genes such as KMT2A. Interestingly, we observed significant variability in ALI within infants with KMT2A-rearranged ALL: much higher in newborns than those older than 6 months, consistent with a previous report of age as a prognostic marker in infant ALL (Supplemental Figure 9).51

While our study primarily focused on leukemia genomics, the mechanism of drug sensitivity and resistance is multifactorial, many of which may differ by age. For example, epigenetic modifications (such as DNA methylation) are strongly influenced by age,52 which could also contribute to the difference in drug responses between pediatric and adult ALL. To explore this, we examined CEBPA, one of the top hits in the 228 genes that define ALI (Supplemental Table 6). As shown in Supplemental Figure 10, CEPBA methylation is highly dependent on age, partly explaining age-related differences in mercaptopurine and prednisolone LC50. Therefore, future studies should consider molecular profiling beyond genome sequencing to fully understand the biology of ALL drug sensitivity by age.

The prognostic impact of age in ALL has long been recognized, e.g., adults having worse outcomes than children and older children having worse outcomes than younger children. The exact age cutoff for ALL risk stratification remains a matter of debate. Within children, age older than 10 is empirically considered at higher risk for relapse, although leukemia or host biological features underlying this classification are unlikely to be categorical. Even worse, the designation of adult ALL is entirely based on the legal definition of majority, and there is little understanding of ALL heterogeneity within adults. Even though children are generally more drug-sensitive than adults, a significant proportion of them have ALL with genomic features commonly seen in adults. Children with this adult-like leukemia had a higher risk of relapse, and their leukemia cells were more resistant to cytotoxic agents. Therefore age of diagnosis is a crude proxy for high-risk leukemia, influencing both molecular subtypes and ALL drug sensitivity.

Our results raise several key questions that should be addressed in future clinical trials/studies. First, can we eliminate age entirely as a variable for ALL risk classification and replace it with molecular subtypes? This is unlikely because the impact of age on leukemia drug response cannot be completely explained by molecular subtypes, as seen with mercaptopurine. However, the precise prognostic value of age in the context of modern ALL taxonomy needs to be examined in large trials enrolling both pediatric and adult ALL patients. Second, do we need to intensify treatment for children with “adult-like” ALL (ALI ≥0.1), and if so, how? Similarly, are there any targeted therapeutics to be added for the treatment of adult ALLs since they already exhibit resistance to cytotoxic agents? For both adult-like pediatric ALL and adult ALL, more intensive chemotherapy is unlikely to be effective without causing excessive toxicity, and novel drugs should be prioritized in future trials.

It should also be noted that our drug sensitivity assay relies on cultured ALL cells and does not reflect host immune surveillance,53,54 or systemic drug metabolism and pharmacodynamics,55 all of which can affect drug response in vivo. More sophisticated assays or algorithms that incorporate host effects are needed in the future to address these limitations and enable better prediction of treatment outcomes in patients.

In summary, we systematically compared the in vitro drug response of pediatric and adult patients with ALL, uncovering the genomic basis of differing drug sensitivity between children and adults and their association with treatment response. These findings have potential implications for the development of the next generation of ALL therapy across age groups.

Supplementary Material

PV Data Sharing Statement
PV Data Supplement

Context Summary.

Key objective

To characterize the pharmacogenomic basis of age-related disparity in ALL treatment outcomes

Knowledge generated

Evaluating ex vivo sensitivity to 21 chemotherapeutics and leukemia genomics of 1,076 children and adults with B-ALL, we observed marked resistance to cytotoxic agents in adults, which was strongly driven by differences in ALL subtype by age. In addition, transcriptomic profiling analysis revealed intra-subtype heterogeneity that is highly associated with age. We identified a significant proportion of pediatric ALL with “adult-like” molecular features, characterized by unique genomic aberrations, resistance to cytotoxic agents, and poor outcomes in frontline trials.

Relevance (written by Dr. Smita Bhatia)

These results provide pharmacogenomic insights into age-related disparities in ALL outcomes and could be harnessed to develop risk-based therapeutic approaches.

Acknowledgments

We thank the patients and families who participated in the clinical trials included in this study for donating ALL specimens for research, and the clinicians and research staff for their assistance in sample collection, processing, and curation. We also thank A. John, L. Rowland, B. Smart, H. Williams, D. Maxwell, and J. Hunt for their technical assistance in drug-sensitivity profiling. We appreciate the thoughtful discussions with S. Brady. This work was supported by the American Lebanese Syrian Associated Charities; and the following National Institutes of Health (NIH) grants: R35 CA197695 (to C.G.M.), R35 GM141947 (to J.J.Y.), P30 CA021765 (SJCRH Cancer Center Support Grant), P50 GM115279 (Center for Precision Medicine of Leukemia Grant, to M.V.R., C.G.M., W.E.E., and J.J.Y.), U10CA180820 (to M.R.L. and E.P.), UG1CA189859 (to E.P.), and UG1CA232760 (to M.R.L.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

This study was partly presented at 65th ASH Annual Meeting and Exposition, San Diego, CA, USA, in December 2023.

Footnotes

Conflict of interest

W.E.E. reports serving as a member of the scientific advisory board for Princess Maxima Centre for Childhood Cancer and serving as a board member for BioSkryb Genomics, neither of which pertains to the submitted work. C.G.M. reports receiving grants from Pfizer and AbbVie and personal fees from Amgen and Illumina outside the submitted work. J.J.Y. reports receiving research funding from Takeda Pharmaceutical Company unrelated to this submitted work and having a patent pending for Methods for Determining Benefit of Chemotherapy. M.V.R. reports receiving investigator-initiated research funding from Servier. M. R. L. reports receiving research grants or contracts from AbbVie, Amgen, Astellas Pharma, Actinium Pharmaceuticals, Pluristem, and Sanofi, serving as a member of the advisory board for Jazz Pharmaceuticals, being involved in speakers bureau for Amgen and Beigene, and serving on a Data Safety Monitoring Committee for BioSight outside the submitted work.

Data availability

RNA-seq data can be accessed through the European Genome-phenome Archive (https://ega-archive.org/) under accession numbers EGAS00001001952, EGAS00001001923, EGAS00001000447, EGAS00001000654, EGAS00001003266, EGAS00001004739, EGAS00001005084, EGAS00001006336, EGAS00001007473, EGAS00001004810, and EGAS00001005180, or through St. Jude Cloud Genomics Platform (https://platform.stjude.cloud/data/cohorts) under the following datasets: i) Pan-Acute Lymphoblastic Leukemia dataset (SJC-DS-1009), ii) Real-time Clinical Genomics dataset (SJC-DS-1007), and iii) Genomes 4 Kids (SJC-DS-1004).

References

  • 1.White MC, Holman DM, Boehm JE, et al. : Age and cancer risk: a potentially modifiable relationship. Am J Prev Med 46:S7–15, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Chatsirisupachai K, Lesluyes T, Paraoan L, et al. : An integrative analysis of the age-associated multi-omic landscape across cancers. Nat Commun 12:2345, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Gu Z, Churchman ML, Roberts KG, et al. : PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia. Nat Genet 51:296–307, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Hunger SP, Mullighan CG: Acute Lymphoblastic Leukemia in Children. N Engl J Med 373:1541–52, 2015 [DOI] [PubMed] [Google Scholar]
  • 5.Pui CH, Yang JJ, Hunger SP, et al. : Childhood Acute Lymphoblastic Leukemia: Progress Through Collaboration. J Clin Oncol 33:2938–48, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Inaba H, Mullighan CG: Pediatric acute lymphoblastic leukemia. Haematologica 105:2524–2539, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Terwilliger T, Abdul-Hay M: Acute lymphoblastic leukemia: a comprehensive review and 2017 update. Blood Cancer J 7:e577, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Jabbour E, Pui CH, Kantarjian H: Progress and Innovations in the Management of Adult Acute Lymphoblastic Leukemia. JAMA Oncol 4:1413–1420, 2018 [DOI] [PubMed] [Google Scholar]
  • 9.Sasaki K, Jabbour E, Short NJ, et al. : Acute lymphoblastic leukemia: A population-based study of outcome in the United States based on the surveillance, epidemiology, and end results (SEER) database, 1980–2017. Am J Hematol 96:650–658, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Paietta E, Roberts KG, Wang V, et al. : Molecular classification improves risk assessment in adult BCR-ABL1-negative B-ALL. Blood 138:948–958, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Roberts KG, Gu Z, Payne-Turner D, et al. : High Frequency and Poor Outcome of Philadelphia Chromosome-Like Acute Lymphoblastic Leukemia in Adults. J Clin Oncol 35:394–401, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Pieters R, Huismans DR, Loonen AH, et al. : Relation of cellular drug resistance to long-term clinical outcome in childhood acute lymphoblastic leukaemia. Lancet 338:399–403, 1991 [DOI] [PubMed] [Google Scholar]
  • 13.Holleman A, Cheok MH, den Boer ML, et al. : Gene-expression patterns in drug-resistant acute lymphoblastic leukemia cells and response to treatment. N Engl J Med 351:533–42, 2004 [DOI] [PubMed] [Google Scholar]
  • 14.Styczynski J, Pieters R, Huismans DR, et al. : In vitro drug resistance profiles of adult versus childhood acute lymphoblastic leukaemia. Br J Haematol 110:813–8, 2000 [DOI] [PubMed] [Google Scholar]
  • 15.Maung ZT, Reid MM, Matheson E, et al. : Corticosteroid resistance is increased in lymphoblasts from adults compared with children: preliminary results of in vitro drug sensitivity study in adults with acute lymphoblastic leukaemia. Br J Haematol 91:93–100, 1995 [DOI] [PubMed] [Google Scholar]
  • 16.Stock W, Luger SM, Advani AS, et al. : A pediatric regimen for older adolescents and young adults with acute lymphoblastic leukemia: results of CALGB 10403. Blood 133:1548–1559, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Huguet F, Chevret S, Leguay T, et al. : Intensified Therapy of Acute Lymphoblastic Leukemia in Adults: Report of the Randomized GRAALL-2005 Clinical Trial. J Clin Oncol 36:2514–2523, 2018 [DOI] [PubMed] [Google Scholar]
  • 18.Shen S, Chen X, Cai J, et al. : Effect of Dasatinib vs Imatinib in the Treatment of Pediatric Philadelphia Chromosome-Positive Acute Lymphoblastic Leukemia: A Randomized Clinical Trial. JAMA Oncol 6:358–366, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Kantarjian H, Short NJ, Jain N, et al. : Frontline combination of ponatinib and hyper-CVAD in Philadelphia chromosome-positive acute lymphoblastic leukemia: 80-months follow-up results. Am J Hematol 98:493–501, 2023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Jeha S, Coustan-Smith E, Pei D, et al. : Impact of tyrosine kinase inhibitors on minimal residual disease and outcome in childhood Philadelphia chromosome-positive acute lymphoblastic leukemia. Cancer 120:1514–9, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Rousselot P, Coude MM, Gokbuget N, et al. : Dasatinib and low-intensity chemotherapy in elderly patients with Philadelphia chromosome-positive ALL. Blood 128:774–82, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Pennesi E, Michels N, Brivio E, et al. : Inotuzumab ozogamicin as single agent in pediatric patients with relapsed and refractory acute lymphoblastic leukemia: results from a phase II trial. Leukemia 36:1516–1524, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kantarjian HM, DeAngelo DJ, Stelljes M, et al. : Inotuzumab Ozogamicin versus Standard Therapy for Acute Lymphoblastic Leukemia. N Engl J Med 375:740–53, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kornauth C, Pemovska T, Vladimer GI, et al. : Functional Precision Medicine Provides Clinical Benefit in Advanced Aggressive Hematologic Cancers and Identifies Exceptional Responders. Cancer Discov 12:372–387, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Lee SHR, Yang W, Gocho Y, et al. : Pharmacotypes across the genomic landscape of pediatric acute lymphoblastic leukemia and impact on treatment response. Nat Med 29:170–179, 2023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Malone ER, Oliva M, Sabatini PJB, et al. : Molecular profiling for precision cancer therapies. Genome Med 12:8, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Malani D, Kumar A, Bruck O, et al. : Implementing a Functional Precision Medicine Tumor Board for Acute Myeloid Leukemia. Cancer Discov 12:388–401, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Frismantas V, Dobay MP, Rinaldi A, et al. : Ex vivo drug response profiling detects recurrent sensitivity patterns in drug-resistant acute lymphoblastic leukemia. Blood 129:e26–e37, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Eaton A, Wong V, Schiff D, et al. : Tumor Genomic Profiling and Ex Vivo Drug Sensitivity Testing for Pediatric Leukemia and Lymphoma Patients. J Pediatr Pharmacol Ther 27:123–131, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Laukkanen S, Gronroos T, Polonen P, et al. : In silico and preclinical drug screening identifies dasatinib as a targeted therapy for T-ALL. Blood Cancer J 7:e604, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Gocho Y, Liu J, Hu J, et al. : Network-based systems pharmacology reveals heterogeneity in LCK and BCL2 signaling and therapeutic sensitivity of T-cell acute lymphoblastic leukemia. Nat Cancer 2:284–299, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Diaz-Flores E, Comeaux EQ, Kim KL, et al. : Bcl-2 Is a Therapeutic Target for Hypodiploid B-Lineage Acute Lymphoblastic Leukemia. Cancer Res 79:2339–2351, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Pui CH, Campana D, Pei D, et al. : Treating childhood acute lymphoblastic leukemia without cranial irradiation. N Engl J Med 360:2730–41, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Jeha S, Pei D, Choi J, et al. : Improved CNS Control of Childhood Acute Lymphoblastic Leukemia Without Cranial Irradiation: St Jude Total Therapy Study 16. J Clin Oncol 37:3377–3391, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Pui CH, Sandlund JT, Pei D, et al. : Results of therapy for acute lymphoblastic leukemia in black and white children. JAMA 290:2001–7, 2003 [DOI] [PubMed] [Google Scholar]
  • 36.Jabbour E, Short NJ, Jain N, et al. : Ponatinib and blinatumomab for Philadelphia chromosome-positive acute lymphoblastic leukaemia: a US, single-centre, single-arm, phase 2 trial. Lancet Haematol 10:e24–e34, 2023 [DOI] [PubMed] [Google Scholar]
  • 37.Lee SHR, Antillon-Klussmann F, Pei D, et al. : Association of Genetic Ancestry With the Molecular Subtypes and Prognosis of Childhood Acute Lymphoblastic Leukemia. JAMA Oncol 8:354–363, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Brady SW, Roberts KG, Gu Z, et al. : The genomic landscape of pediatric acute lymphoblastic leukemia. Nat Genet 54:1376–1389, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Siegel SE, Stock W, Johnson RH, et al. : Pediatric-Inspired Treatment Regimens for Adolescents and Young Adults With Philadelphia Chromosome-Negative Acute Lymphoblastic Leukemia: A Review. JAMA Oncol 4:725–734, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Novakova M, Zaliova M, Fiser K, et al. : DUX4r, ZNF384r and PAX5-P80R mutated B-cell precursor acute lymphoblastic leukemia frequently undergo monocytic switch. Haematologica 106:2066–2075, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Zhang J, McCastlain K, Yoshihara H, et al. : Deregulation of DUX4 and ERG in acute lymphoblastic leukemia. Nat Genet 48:1481–1489, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Barbie DA, Tamayo P, Boehm JS, et al. : Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature 462:108–12, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Yeoh AE, Ariffin H, Chai EL, et al. : Minimal residual disease-guided treatment deintensification for children with acute lymphoblastic leukemia: results from the Malaysia-Singapore acute lymphoblastic leukemia 2003 study. J Clin Oncol 30:2384–92, 2012 [DOI] [PubMed] [Google Scholar]
  • 44.Yeoh AEJ, Lu Y, Chin WHN, et al. : Intensifying Treatment of Childhood B-Lymphoblastic Leukemia With IKZF1 Deletion Reduces Relapse and Improves Overall Survival: Results of Malaysia-Singapore ALL 2010 Study. J Clin Oncol 36:2726–2735, 2018 [DOI] [PubMed] [Google Scholar]
  • 45.Kaspers GJ, Smets LA, Pieters R, et al. : Favorable prognosis of hyperdiploid common acute lymphoblastic leukemia may be explained by sensitivity to antimetabolites and other drugs: results of an in vitro study. Blood 85:751–6, 1995 [PubMed] [Google Scholar]
  • 46.Krishna Narla R, Navara C, Sarquis M, et al. : Chemosensitivity of TEL-AML1 fusion transcript positive acute lymphoblastic leukemia cells. Leuk Lymphoma 41:615–23, 2001 [DOI] [PubMed] [Google Scholar]
  • 47.Ramakers-van Woerden NL, Pieters R, Loonen AH, et al. : TEL/AML1 gene fusion is related to in vitro drug sensitivity for L-asparaginase in childhood acute lymphoblastic leukemia. Blood 96:1094–9, 2000 [PubMed] [Google Scholar]
  • 48.Frost BM, Forestier E, Gustafsson G, et al. : Translocation t(12;21) is related to in vitro cellular drug sensitivity to doxorubicin and etoposide in childhood acute lymphoblastic leukemia. Blood 104:2452–7, 2004 [DOI] [PubMed] [Google Scholar]
  • 49.Kaspers GJ, Veerman AJ, Pieters R, et al. : In vitro cellular drug resistance and prognosis in newly diagnosed childhood acute lymphoblastic leukemia. Blood 90:2723–9, 1997 [PubMed] [Google Scholar]
  • 50.Usami I, Imamura T, Takahashi Y, et al. : Discontinuation of L-asparaginase and poor response to prednisolone are associated with poor outcome of ETV6-RUNX1-positive pediatric B-cell precursor acute lymphoblastic leukemia. Int J Hematol 109:477–482, 2019 [DOI] [PubMed] [Google Scholar]
  • 51.Pieters R, De Lorenzo P, Ancliffe P, et al. : Outcome of Infants Younger Than 1 Year With Acute Lymphoblastic Leukemia Treated With the Interfant-06 Protocol: Results From an International Phase III Randomized Study. J Clin Oncol 37:2246–2256, 2019 [DOI] [PubMed] [Google Scholar]
  • 52.Horvath S: DNA methylation age of human tissues and cell types. Genome Biol 14:R115, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Duault C, Kumar A, Taghi Khani A, et al. : Activated natural killer cells predict poor clinical prognosis in high-risk B- and T-cell acute lymphoblastic leukemia. Blood 138:1465–1480, 2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Li Y, Yang X, Sun Y, et al. : Impact of T-cell immunity on chemotherapy response in childhood acute lymphoblastic leukemia. Blood 140:1507–1521, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Pui CH, Relling MV, Evans WE: Role of pharmacogenomics and pharmacodynamics in the treatment of acute lymphoblastic leukaemia. Best Pract Res Clin Haematol 15:741–56, 2002 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

PV Data Sharing Statement
NIHMS1996246-supplement-PV_Data_Sharing_Statement.pdf (131.5KB, pdf)
PV Data Supplement
NIHMS1996246-supplement-PV_Data_Supplement.pdf (1.9MB, pdf)

Data Availability Statement

RNA-seq data can be accessed through the European Genome-phenome Archive (https://ega-archive.org/) under accession numbers EGAS00001001952, EGAS00001001923, EGAS00001000447, EGAS00001000654, EGAS00001003266, EGAS00001004739, EGAS00001005084, EGAS00001006336, EGAS00001007473, EGAS00001004810, and EGAS00001005180, or through St. Jude Cloud Genomics Platform (https://platform.stjude.cloud/data/cohorts) under the following datasets: i) Pan-Acute Lymphoblastic Leukemia dataset (SJC-DS-1009), ii) Real-time Clinical Genomics dataset (SJC-DS-1007), and iii) Genomes 4 Kids (SJC-DS-1004).

RESOURCES