Integrated Genomic Analysis of Chromosomal Alterations and Mutations in B-Cell Acute Lymphoblastic Leukemia Reveals Distinct Genetic Profiles at Relapse

The clonal basis of relapse in B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is complex and not fully understood. Next-generation sequencing (NGS), array comparative genomic hybridization (aCGH), and multiplex ligation-dependent probe amplification (MLPA) were carried out in matched diagnosis–relapse samples from 13 BCP-ALL patients to identify patterns of genetic evolution that could account for the phenotypic changes associated with disease relapse. The integrative genomic analysis of aCGH, MLPA and NGS revealed that 100% of the BCP-ALL patients showed at least one genetic alteration at diagnosis and relapse. In addition, there was a significant increase in the frequency of chromosomal lesions at the time of relapse (p = 0.019). MLPA and aCGH techniques showed that IKZF1 was the most frequently deleted gene. TP53 was the most frequently mutated gene at relapse. Two TP53 mutations were detected only at relapse, whereas the three others showed an increase in their mutational burden at relapse. Clonal evolution patterns were heterogeneous, involving the acquisition, loss and maintenance of lesions at relapse. Therefore, this study provides additional evidence that BCP-ALL is a genetically dynamic disease with distinct genetic profiles at diagnosis and relapse. Integrative NGS, aCGH and MLPA analysis enables better molecular characterization of the genetic profile in BCP-ALL patients during the evolution from diagnosis to relapse.


Introduction
Acute lymphoblastic leukemia (ALL) is a disease with specific genetic alterations associated with drug resistance, treatment failure and disease relapse [1,2]. Despite vast improvements in the treatment of childhood and adult ALL in recent years, the outlook for relapsed leukemia remains poor, highlighting the need for innovative treatment approaches [3]. It is well known that relapsed ALL is a heterogeneous disease and that distinct genetic alterations may be unique to small subgroups of patients [3]. Genomic studies of matched diagnosis-relapse samples from ALL patients have shed light on the clonal evolution that leads to relapse, the pathways associated with chemoresistance, and the potential targets for therapy [4][5][6][7][8][9]. However, the mechanisms that probably fuel an ALL relapse are not fully understood. A combined analysis of gene mutations and copy number alterations (CNAs) could provide valuable insight into the discovery of the patterns of clonal evolution and the biomarkers that predict a greater likelihood of relapse in ALL [3,10,11]. Here, we have performed an integrated and sequential genomic analysis combining next-generation sequencing (NGS), array comparative genomic hybridization (aCGH), and multiplex ligation-dependent probe amplification (MLPA) to identify the clonal shifts related to ALL progression.

Patients
Thirteen paired diagnosis and first relapse samples of B-cell precursor acute lymphoblastic leukemia (BCP-ALL) patients (4 children and 9 adults) were eligible for this study. The patients were treated in accordance with the risk-adapted protocols of PETHEMA (Programa Español de Tratamientos en Hematología) and SEHOP (Sociedad Española de Hematología y Oncología Pediátrica). The diagnosis of ALL was based on morphological, immunophenotypic and genetic features of leukemic blast cells, as described previously [12]. The patients' demographic information, clinical characteristics, risk classification, response to therapy and survival were recorded. The study was approved by the local ethical committee, the Comité Ético de Investigación Clínica, at the Hospital Universitario de Salamanca. Written informed consent was obtained from each patient or their legal guardian before entering the study. Table 1 shows the characteristics of the patients included in this study. The median age was 31 years (range 4-80 years). The median percentage of blast counts in their bone marrow was 82% (range 45-96%). Fifty-four per cent of the patients had none of the chromosomal abnormalities associated with poor risk (t(9;22), t(v;11q23) or hypodiploidy). Ninety-two per cent of patients died presenting a 5-year overall survival probability of 15% (median: 22 months, 95% CI: 3.2-40.8 (Table 1 and Table S1)).  1 Estimated by flow cytometry. 2 Includes the unfavorable abnormalities t(9;22), t(v;11q23) and hypodiploidy. 3 Risk group stratification was mainly designated according to the Programa Español de Tratamientos en Hematología (PETHEMA) protocols. 4 Time of relapse criteria: very early, earlier than 18 months after initial diagnosis and less than 6 months after the cessation of frontline treatment; early, more than 18 months after initial diagnosis, but less than 6 months after the cessation of frontline treatment; late, more than 6 months after the cessation of frontline treatment.

Oligonucleotide Array Comparative Genomic Hybridizations (Array-CGH)
All samples were tested on an aCGH 12X135K array platform (Roche NimbleGen, Madison, WI, USA). Raw log 2 ratios were segmented using the copy number R package (version 1.20.0) [30]. We used the GISTIC algorithm-(version 2.0.23) to identify statistically significant minimal common altered regions (MCRs) and the broad CNAs present in the samples [31]. The Database of Genomic Variants from Toronto (DGV, http://dgv.tcag.ca/dgv/app/home) was used to exclude DNA variations located in regions with defined copy number variations. All CNAs with an overlap of more than 50% with respect to those reported in the DGV were excluded. MLPA reactions were performed using the SALSA MLPA P335-B1 ALL-IKZF1 probemix (MRC-Holland, Amsterdam, Netherlands) according to the manufacturer's instructions. DNA samples from three healthy donors were used as controls. The P335-B1 probemix contains probes for the following genes: IKZF1, CDKN2A/B, PAX5, EBF1, ETV6, BTG1, RB1, as well as genes from the X/Y PAR1 region (CRLF2, CSF2RA, IL3RA and P2RY8). MLPA amplification products were analyzed on an ABI 3130xl Genetic Analyzer (Applied Biosystems/Hitachi) with GeneMapper software V.3.7, using the Genescan 500LIZ internal size standard (Applied Biosystems). The copy number at each locus was estimated according to Schwab et al. [32].
According to the probemix contained in the P335-B1 MLPA kit, an integrative MLPA-aCGH analysis was performed to identify gene deletions in the IKZF1, CDKN2A/B, PAX5, EBF1, ETV6, BTG1 and RB1 genes, as well as genes from the X/Y PAR1 region (CRLF2, CSF2RA, IL3RA and P2RY8). The copy number at each locus was estimated according to the method of Schwab et al. [32], whereby values above 1.3, between 1.3 and 0.75, between 0.75 and 0.25, and below 0.25 were considered as gain, normal, hemizygous loss, and homozygous loss, respectively. It was possible to distinguish the gene deletions identified either by MLPA or by aCGH analysis, or by the both methods. The distributions of the probes in each platform are illustrated in Figure S1.

Statistical Methods
The differences between groups were compared by the chi-square, Fisher's exact, and Mann-Whitney tests, as appropriate. Values of p < 0.05 were considered to be statistically significant. Analyses were conducted using IBM SPSS version 21.0 (IBM Corp., Armonk, NY, USA).
The materials, procedures and statistical analyses are described in detail in the Supplementary File 1.

Recurring Genomic Alterations in Matched Diagnosis-Relapse BCP-ALL Samples
The aCGH detected 1451 somatic genetic lesions in 13 paired (diagnosis and relapse) BCP-ALL samples. The number of genetic lesions varied significantly between patients (1-287 lesions; median, 16 per sample). There were no significant differences in the number of CNAs between children and adults (p = 0.765), or between patients who had early relapses and those who did not (p = 0.731). Figure S2 show the main aCGH findings at diagnosis and relapse, with a significant increase in the number of lesions at relapse. There was a median of six alterations per sample at diagnosis and of 47 at relapse (p = 0.019). Figure S2 show the patterns and frequencies of DNA copy alterations observed in the 13 paired diagnosis/relapse samples. The most recurrent broad and focal copy number changes observed at diagnosis and/or relapse were: dup(1q) at 23%, dup(X) at 31%, dup(21) at 15%, del(7 or 7p) at 77%, dup(7q) at 31%, del(9p) at 62%, del(12p) at 15%, del(13q) at 23% and del(17p) at 15%. Of these, the deletions located on 7p and 9p were the most frequently focal and broad chromosomal alterations detected at diagnosis and/or relapse (77% and 62%, respectively (Table S1)).
The complementary integrative genomic analysis using aCGH, MLPA and NGS revealed that 100% of ALL patients showed at least one genetic alteration at diagnosis and relapse (mutation, loss and/or gain, or chromosomal rearrangement identified by Fluorescence in situ hybridization (FISH) ). The frontline risk-adapted protocols, outcome, clinical status, karyotype, FISH, NGS, aCGH and MLPA analysis of each patient are shown in Table S1.

Heterogeneous Patterns of Genetic Evolution in Paired Diagnosis and Relapse Samples
Tables S2-S4 show the regions of statistically significant recurrent amplification and deletion that were retained, lost, or acquired as new lesions at relapse, respectively (q-value < 0.05).
Diagnostics 2020, 10, x FOR PEER REVIEW  8 of 14   Patient ID  ID1  ID2 ID3 ID4 ID5  ID6  ID7 ID8  ID9 ID10 ID11 ID12    The MLPA-aCGH analysis revealed that most of the patients (5/7, 71%) with deletions of IKZF1 (7p) at diagnosis retained this genetic alteration at relapse (children: ID1 and ID4, and adults: ID6, ID9 and ID12). By contrast, only three (ID2, ID9, ID12) of seven patients (42.9%) with CDKN2A/B and/or PAX5 deletions (9p) retained these deletions at relapse. It should be noted that the adult patient ID13, who presented an elevated number of CNAs at relapse, acquired new lesions in the chromosomal regions that harbored the EBF1, CDKN2A/B, IKZF1, BTG1 and RB1 genes. Finally, the losses on 17p were also identified in two patients (child ID2 and adult ID9) at diagnosis, being retained at relapse in child ID2. An insertion/deletion mutation in TP53 was also identified in this pediatric patient at diagnosis and relapse (Tables S1 and S2, Figures 1, 3 and 4). Table 2 details the mutations observed at both times and describes their mutational burden. Interestingly, two TP53 mutations were acquired at relapse (ID3, ID7), whereas all three TP53 mutations, which were detected from diagnosis, had an increase in their mutational burden at relapse (ID2, ID4, ID7 ( Figure 5)). Thus, in the pediatric patient ID2, an increase in the TP53 mutant clone burden was observed at relapse (53% to 71%), as well as in the pediatric patient ID4 (11% to 21%) and the adult patient ID7 (3.5% to 26%), respectively.  Table 2 details the mutations observed at both times and describes their mutational burden. Interestingly, two TP53 mutations were acquired at relapse (ID3, ID7), whereas all three TP53 mutations, which were detected from diagnosis, had an increase in their mutational burden at relapse (ID2, ID4, ID7 (Figure 5)). Thus, in the pediatric patient ID2, an increase in the TP53 mutant clone burden was observed at relapse (53% to 71%), as well as in the pediatric patient ID4 (11% to 21%) and the adult patient ID7 (3.5% to 26%), respectively. It is worth mentioning that the pediatric patient ID3, who was stratified as at low risk of disease at diagnosis, had acquired a mutation in TP53 by relapse. Likewise, it should be noted that one adult patient (ID7) showed a co-occurrence of mutations in two different genes, TP53 and PAX5. In the TP53 gene, the splice mutation (c.-8_4del12) was detected only at relapse (mutational burden: 15%) whereas the missense mutation (c.818G > C) was present at both times (mutational burden: 3.5% vs. 26%). In the PAX5 gene, only one missense mutation (c.399T > A) was observed at diagnosis (mutational burden: 20%). This patient had also acquired a deletion in the IKZF1 gene by relapse (Table S1 and Figures 3 and S3). The integrative NGS, aCGH, and MLPA analysis enabled a better molecular characterization of the genetic profile in ALL patients during the evolution of their disease from diagnosis to relapse.

Discussion
The present study, carried out in sequential ALL patients at diagnosis and relapse, showed that all 13 ALL patients exhibited heterogeneous clonal changes in terms of CNAs and mutations between It is worth mentioning that the pediatric patient ID3, who was stratified as at low risk of disease at diagnosis, had acquired a mutation in TP53 by relapse. Likewise, it should be noted that one adult patient (ID7) showed a co-occurrence of mutations in two different genes, TP53 and PAX5. In the TP53 gene, the splice mutation (c.-8_4del12) was detected only at relapse (mutational burden: 15%) whereas the missense mutation (c.818G > C) was present at both times (mutational burden: 3.5% vs. 26%). In the PAX5 gene, only one missense mutation (c.399T > A) was observed at diagnosis (mutational burden: 20%). This patient had also acquired a deletion in the IKZF1 gene by relapse (Table S1 and Figure 3 and Figure S3). The integrative NGS, aCGH, and MLPA analysis enabled a better molecular characterization of the genetic profile in ALL patients during the evolution of their disease from diagnosis to relapse.

Discussion
The present study, carried out in sequential ALL patients at diagnosis and relapse, showed that all 13 ALL patients exhibited heterogeneous clonal changes in terms of CNAs and mutations between diagnosis and relapse, involving the acquisition, loss and maintenance of lesions at relapse. The shared lesions between the relapse clone and the predominant clone at diagnosis suggest a common pre-leukemic origin [22], while the acquired lesions provide unequivocal evidence of a second clone that was present as a minor population at diagnosis, but acquired different genetic alterations before emerging as the relapse clone [5].
An integrated NGS, aCGH, and MLPA analysis allowed for the identification of alterations on the IKZF1 (7p) and TP53 (17p) genes in paired diagnostic and relapse samples; these were more frequent at relapse than at diagnosis. Both genetic events could have strongly influenced disease relapse and the short survival of these patients. In this study, TP53 is the most frequently mutated gene at relapse (31%). TP53 abnormalities (deletion and/or mutation) have been associated with a resistance to treatment and worse prognosis in childhood and adult ALL [8,33]. TP53 gene abnormalities have a key role in ALL relapse, as they independently predict a high risk of treatment failure in ALL patients [8,24,34]. The presence of TP53 alterations has been associated with a reduced response rate to induction therapy and correlated with a shortened duration of survival, even after successful reinduction therapy [35,36]. Different therapeutic strategies to target mutant p53 have been developed for the high risk TP53-mutant ALL, such as the use of the small molecule APR-246 which exhibits antileukemia activity in TP53mut BCP-ALL, targeting non-functional mutant p53 and restoring its tumor suppressive function [37].
IKZF1 was the most frequently deleted gene, the incidence of deletions being greater at relapse than at diagnosis. Similarly, as seen in previous studies, the frequency of IKZF1 deletions was higher in adults than in children [38]. IKZF1 deletions have been associated with a higher risk of relapse in ALL and have been shown to be a hallmark of BCR-ABL1-positive ALL, although they have also been identified in a fraction of BCR-ABL1-negative ALL patients [10,[38][39][40][41][42][43], as noted in our study. In recent years, IKZF1 deletions and TP53 alterations are being recognized as important markers of poor prognosis in ALL after a first relapse, mainly in children [9,39,44]. Therefore, these alterations could contribute to the re-stratification of risk for ALL patients and to proposing timely therapeutic strategies such as treatment intensification and identifying candidates for transplantation or for inclusion in clinical trials due to their high risk of suffering a second relapse [34,45,46].
In the present study, heterogeneous patterns of genetic evolution in paired diagnostic and relapse samples were observed that are consistent with those reported in previous studies [47]. In particular, two TP53 mutations were only detected at relapse (patients ID3 and ID7), whereas all three TP53 mutations increased their mutational burden between diagnosis and relapse (patients ID2, ID4 and ID7). It should be noted that two of the six TP53 mutations identified had not been reported in the genomic databases.
ALL is clonally heterogeneous and genetic lesions in minor clones may confer resistance to therapy and promote disease relapse (e.g., TP53, IKZF1, CREBBP) [22]. The low proportion of the minor relapse subclone at diagnosis suggests that the leukemia at diagnosis contains genetically diverse subclones. Therapy would aim to select the eventual dominant relapse clone whose alterations confer resistance to treatment [48]. Thus, TP53 mutations could be considered as driver mutations that probably confer a selective growth advantage on ALL tumor cells at relapse [24]. Preclinical studies and clinical experience have shown that leukemic blasts are more resistant at relapse than at diagnosis. Mechanisms of resistance may include the selection of a pre-existing resistant subclone or the acquisition of additional genomic lesions under the selective pressure of chemotherapy, as observed in our study [4].

Conclusions
In summary, the present study provides additional evidence that the clonality of ALL is genetically dynamic from diagnosis to relapse. The integrative NGS, aCGH, and MLPA analysis enabled a better molecular characterization of the genetic profile in ALL patients during the evolution of their disease, showing distinct genetic profiles at diagnosis and relapse. With this study, the utility of simultaneously identifying CNAs and mutations at the time of diagnosis and relapse was evidenced, which is clinically important to predict the evolution of the patients. New genomic strategies to identify various genetic lesions from a single sample and in a single experiment are currently being solved by designing specific panels for each type of hematological disease, which contribute to the improvement of the risk stratification, promoting the use of personalized treatment in ALL.