Cytogenetic Variation of B-Lymphoblastic Leukemia with Intrachromosomal Amplification of Chromosome 21 (Iamp21) a Multi-Institutional Series Review

AJCP / Original Article

Cytogenetic Variation of B-Lymphoblastic Leukemia With Intrachromosomal Amplification of Chromosome 21 (iAMP21) A Multi-Institutional Series Review

Ryan C. Johnson, MD,1 Olga K. Weinberg, MD,2 Michael J. Cascio, MD,3 Gary V. Dahl, MD,1 Bryan A. Mitton, MD,1 Lewis B. Silverman, MD,2 Athena M. Cherry, PhD,1 1 1

Daniel A. Arber, MD, and Robert S. Ohgami, MD, PhD Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

From the 1Stanford University Medical Center, Stanford, CA; 2Boston Children’s Hospital, Boston, MA; and 3Oregon Health and Science University, Portland.

Key Words: Hematopathology; Flow cytometry; Pediatric acute lymphoblastic leukemia; iAMP21; Cytogenetics and molecular genetics; Clinical and molecular epidemiology

Am J Clin Pathol July 2015;144:103-112

DOI: 10.1309/AJCPLUYF11HQBYRB

ABSTRACT The editors of the 2008 WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues sought to place Objectives: B-lymphoblastic leukemia (B-ALL) with a priority on the most recent cytogenetic and molecular intrachromosomal amplification of chromosome 21 findings to establish a more definitive characterization and (iAMP21) is a relatively uncommon manifestation of classification of both acute lymphoblastic and acute myeloid acute leukemia and limited predominantly to the pediatric leukemias, with some leukemias defined by their specific population. Case-specific information regarding flow translocation independent of the blast count observed. This cytometric, morphologic, and laboratory findings of this classification often provides a characteristic cytogenetic cor- subtype of leukemia is currently lacking. relation to the morphologic and flow cytometric expression Methods: We searched the databases of three large pattern observed. For example, although B-lymphoblastic institutions for lymphoblastic leukemia with iAMP21 from leukemia/lymphoma (B-ALL) with t(5;14)(q31;q32) IL3- 2005 through 2012 and analyzed the clinicopathologic IGH accounts for less than 1% of all B-ALL cases, the features. characteristic consistent finding of an associated nonclonal eosinophilia merits inclusion of this leukemia as one with a Results: We identified 17 cases with five or more RUNX1 recurrent cytogenetic abnormality.1 More important, how- signals on interphase nuclei, 14 of which were consistent ever, classification of leukemias with recurrent cytogenetic with the Children’s Oncology Group (COG) definition for abnormalities serves to stratify subtypes of leukemia based iAMP21—namely, the presence of three or more RUNX1 on their overall prognosis and facilitate appropriate chemo- signals on one marker chromosome. These cases showed therapeutic regimen selection.2 a statistically significant lower peripheral WBC count and Other less common recurrent cytogenetic abnormalities older age at diagnosis compared with all pediatric cases have also been recently reported but have not yet become of B-ALL. We also identified three cases with increased established entities. Such is the case with B-ALL with intra- RUNX1 signals scattered on multiple marker chromosomes chromosomal amplification of chromosome 21 (iAMP21). that did not meet the COG definition of iAMP21 but showed These cases are identified on routine karyotype by the similar 21q instability and older age at presentation. absence of a second normal copy of chromosome 21 and Conclusions: Our findings not only demonstrate that B-ALL concurrent fluorescence in situ hybridization (FISH) studies, with iAMP21 is truly a distinct clinicopathologic entity but which reveal an abnormal number and pattern of RUNX1 also suggest that a subset of cases of B-ALL with iAMP21 (AML1) signals when probes for t(12;21) ETV6-RUNX1 are can show variable cytogenetic features. used. The current International Standard for Human Cyto- genetic Nomenclature for this cytogenetic finding is nuc ish (RUNX1 × n), where n is the number of identified RUNX1 signals per interphase nucleus; for enumerable RUNX1

© American Society for Clinical Pathology Am J Clin Pathol 2015;144:103-112 103 DOI: 10.1309/AJCPLUYF11HQBYRB Johnson et al / B-Lymphoblastic Leukemia With iAMP21 signals, the nomenclature is nuc ish (RUNX1 amp).3 For sim- responsible for the breakpoint in a study by Sinclair et al21 plicity, however, this has been referred to in the literature as revealed that other genes local to the region around RUNX1, iAMP21. Although B-ALL with iAMP21 is not recognized such as PDE9A, COL6A, and DSCAM, may be involved. by the 2008 edition of the WHO Classification of Tumours Among cases of B-ALL with iAMP21, some consistent of Haematopoietic and Lymphoid Tissues, various groups clinicopathologic findings have been noted8-16: B-ALL with have defined iAMP21 in studies by the presence of three or iAMP21 tends to present in older children with pancytope- more RUNX1 signals on an abnormal chromosome 21 and/ nia or mildly elevated WBC counts (usually ≤10 × 109/L, or five or more total signals per interphase nucleus.4-7 The with most ≤50 × 109/L) at diagnosis.22 A low WBC count at Children’s Oncology Group (COG), however, currently diagnosis in B-ALL in general usually results in a reduced requires the presence of three or more RUNX1 signals on a risk stratification for treatment. However, the prognosis of single-marker chromosome to qualify as iAMP21.5 patients with B-ALL with iAMP21 is consistently poor,

The first reports of B-ALL with increased RUNX1 with patients demonstrating worse event-free survival (EFS) Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021 signals were made serendipitously when FISH probes for and overall survival (OS) when treated with standard-risk the t(12;21)(q22;q22) ETV6/RUNX1 translocation revealed chemotherapy regimens.7 Although these children typically increased signals of the RUNX1 probe.8-16 All of these cases demonstrate hematologic disease remission following induc- showed, in common, a loss of a normal chromosome 21 with tion chemotherapy, the relapse rate is very high and, subse- the presence of a derivative marker chromosome. The num- quently, relatively refractory to treatment.7 However, more ber of RUNX1 probe signals was consistent in each patient recent work has shown that treating B-ALL with iAMP21 but varied highly between patients, ranging from four to 15 with high-risk chemotherapeutic regimens ameliorates the copies. It was thus proposed in these initial studies that the poor EFS and OS findings.4,5 derivative chromosome was composed of variable fragments Given the relative heterogeneity in RUNX1 signals per of the long arm of chromosome 21, which contained the marker chromosome reported between leukemic clones RUNX1 gene. It has since been estimated that B-ALL with within and between patients, as well as the current two iAMP21 accounts for 1% to 2% of all pediatric B-ALLs.7 working definitions of B-ALL with iAMP21, we sought to The cytogenetic underpinnings of increased RUNX1 determine the heterogeneity of iAMP21 in a limited multi- signals are not completely understood; however, the institutional series. In addition, we wanted to compare our most widely cited mechanism involves multiple breakage- relative frequency of B-ALL with iAMP21, as well as mor- fusion-bridge (BFB) cycles and chromosome region 21q phologic and flow cytometric data with series and studies instability. By morphologic analysis and whole-chromo- reported in the literature to date. some painting probes for chromosome 21, Kuchinskaya et al17 showed that bone marrow lymphoblasts and cultured lymphoblasts from patients with iAMP21 showed Materials and Methods abnormal anaphase bridges of chromosome 21 material between mitotically active cells in a subset of relapsed cases. Multiple chromosome color banding using seven Patient Cohort and Cytogenetic Studies probes spanning the 21q region on metaphase cells showed All aspects of this study were approved by the institu- highly variable arrangements of probes, including cases tional review boards of all participating institutions where where probes proximal to RUNX1 were identified distal to applicable. We reviewed the cytogenetics reports of all repeated sequences of RUNX1; this arrangement suggested adult and pediatric cases of B-ALL within the database of a repeated cycle of chromosomal inversion with subse- the Stanford University Department of Pathology, Boston quent duplication. Independent of the exact banding pat- Children’s Hospital, and Oregon Health Sciences University tern, all color band plots showed a highly unstable region from 2005 through 2012 and identified cases where four or of chromosome 21q, with most cases also showing loss of more RUNX1 signals per nucleus were identified on FISH telomeric or subtelomeric sequences of 21q.18 analysis with the RUNX1/ETV6 probe set. Additional back- While one might presume that RUNX1 plays a central ground information regarding the general demographics of role in the leukemogenesis of B-ALL with iAMP21 cases, all acute B-ALLs, both pediatric and adult, were also con- RUNX1 itself has not been shown to be overexpressed com- currently obtained to determine the frequency of increased pared with other subtypes of B-ALL in expression profile RUNX1 signals and those that meet COG criteria for B-ALL analysis.19,20 Closer examination into the border between with iAMP21. the common region of amplification and normal comple- For cases with increased RUNX1 signals meeting cri- ment of chromosome 21 by array comparative genomic teria above, the following information was collected: CBC hybridization (aCGH) to find candidate genes or sequences count at diagnosis; patient demographics, including age and

104 Am J Clin Pathol 2015;144:103-112 © American Society for Clinical Pathology DOI: 10.1309/AJCPLUYF11HQBYRB AJCP / Original Article sex; patient presentation, including symptoms, signs, pres- cases of B-ALL diagnosed and treated at Stanford Univer- ence/absence of hepatosplenomegaly, or lymphadenopathy; sity from 2005 through 2012. central nervous system involvement at diagnosis; follow-up marrow and lumbar puncture studies; and chemotherapeutic regimen and patient outcome data (if available). Peripheral Results blood counts, bone marrow aspirate and trephine, and ancillary studies, including flow cytometry and cytogenetics, were also reviewed. Cytogenetic data that were collected Patient Cohort included karyotype at diagnosis, number of RUNX1 signals Demographic and presentation data are shown in identified by FISH using the ETV6/RUNX1 probe set, and ❚Table 1❚. In our combined series, 17 cases demonstrated concurrent FISH studies performed (eg, BCR/ABL1; trisomy five or more RUNX1 signals per interphase nucleus and thus

4, 10, and 17; and 11q23/MLL rearrangements). Metaphase would suggest intrachromosomal amplification of chromo- Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021 FISH data (G-banding of metaphase chromosomes followed some 21. However, three cases with metaphase FISH analy- by hybridization with the ETV6/RUNX1 probe set to localize sis demonstrated less than three RUNX1 signals per marker the RUNX1 signals) were recorded if available. chromosome, excluding them from the current formal COG Cases were considered to meet criteria for iAMP21 by definition of B-ALL with iAMP21. Of the 14 patients show- the presence of five or more RUNX1 probes per interphase ing cytogenetic findings consistent with COG criteria, 11 nucleus or by the presence of three or more RUNX1 probes were male and three female, with a male-to-female ratio per marker chromosome if sequential metaphase FISH was of 3.6:1. Although a statistical trend of male predominance performed. In all cases, the number of RUNX1 signals was among pediatric patients was observed, this was not statisti- determined using the Vysis LSI ETV6(TEL)/RUNX1(AML1) cally significant compared with all pediatric cases of B-ALL ES Dual Color Translocation Probe Set (Abbott, Abbott (P = .3919). Thirteen of the 14 patients were younger than Park, IL). All cases with a concurrent positive result for 18 years, and one patient sought treatment at age 20 years. ETV6-RUNX1, BCR-ABL1, hypodiploidy, hyperdiploidy/ The median age at presentation was 8 years, with a range trisomy 4 and 10, 11q23/MLL rearrangements, and trisomy/ from 5 to 20 years. Of the six patients diagnosed and treated polysomy of chromosome 21 if present were noted and at Stanford, these comprised 2.7% of 221 institutional cases excluded from assessment. of B-ALL. Comparing the 13 pediatric (aged <18 years) B-ALL with iAMP21 cases with all Stanford pediatric (aged Flow Cytometry <18 years) B-ALL cases without iAMP21, patients with Flow cytometry was performed using antibodies as iAMP21 sought treatment at an older mean age, which was listed in Supplemental Table 1 (which can be viewed at http:// statistically significant (P = .041). Hepatosplenomegaly was www.ascp.org/docs/default-source/pdf/press/johnsonjuly15. not noted in any of the patients, and only one patient had pdf). In brief, bone marrow aspirate specimens were lymphadenopathy in our total series. obtained in EDTA Vacutainers (BD Biosciences, Franklin Lakes, NJ); cells were washed in Dulbecco phosphate- B-ALL With iAMP21 Is Frequently Associated With buffered saline (DPBS) before labeling with antibodies Lower Peripheral Blood Blast Counts and Pancytopenia and subsequently treated with ammonium chloride RBC WBC data were available for all cases: the median lysis buffer (Pharmlyse; BD Biosciences) and then washed WBC count at presentation was 2.45 × 109/L, with a range again with DPBS. In the case of cytoplasmic or nuclear of 0.90 to 36.6 × 109/L. Twelve (85.7%) of 14 cases showed staining antibodies, samples were also permeabilized during a WBC count of 10 × 109/L or less, and all cases showed a antibody labeling. Samples were run on either CANTO II or WBC count of 40 × 109/L or less. Comparing all pediatric CALIBUR Flow Cytometers (BD Biosciences) and analyzed (aged <18 years) B-ALL with iAMP21 cases with all pedi- using FACSDiva or FCS Express (BD Biosciences). atric (aged <18 years) B-ALL cases diagnosed at Stanford University, patients with iAMP21 showed a lower mean Statistical Analysis peripheral WBC count at diagnosis, which was statistically Statistical analysis (t test, Fisher exact test) was per- significant (P = .0001). Complete CBC data were available formed using XLSTAT (Addinsoft, New York, NY); P for 11 patients: 10 of 11 showed anemia (hemoglobin range, values less than .05 were considered significant. Laboratory 3.4-13.3 g/dL; median, 7.0 g/dL), 10 of 11 showed throm- and demographic data for all pediatric (<18 years of age) bocytopenia (platelet count range, 13-153 × 109/L; median, cases that met COG criteria for B-ALL with iAMP21 from 49 × 109/L), and all cases showed neutropenia (absolute neu- all three institutional centers were compared with laboratory trophil count range, 0.0-0.9 × 109/L; median, 0.4 × 109/L). and demographic data from all pediatric (<18 years of age) Six (55%) of the 11 patients showed an absolute blast count

❚Table 1❚ Summary of Cases of B-Lymphoblastic Leukemia With iAMP21 RUNX1 RUNX1 Copies/ WBC Predominant Aberrant Other Patient Copies/ Marker Age, y/ Count, × Blast Antigen Cytogenetic No. Cell Chromosome Sex 109/L Morphology Expression Abnormalitiesa Outcome 1 5 4 on mar1 6/M 6.1 L1 dpCD33 del(11q), –21 Induction failure, subsequent SCT, died of GVHD, respiratory compromise, and multiorgan failure 19 mo from dx 2 >5 >5 on mar1 8/M 2.4 L1 dpCD4 del(8p), –21 CR1 × 12 mo 3 5 4 on mar1 6/M 1.5 L1 dpCD117 NK CR1 × 13 mo 4 6 5 on mar1 5/M 36.6 L1 dpCD33 +X, –13, –21 CR1 × 14 mo 5 8 NP 11/M 0.9 L2 dpCD7 NK CR1 × 7.5 y 6 >5 NP 8/M 3.7 L1 None NK CR1 × 8 y 7 >5 NP 17/M 2.4 L1 CD13, NK Relapse 58 mo from dx, Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021 CD33 following SCT, CR2 × 3 y 8 >5 NP 7/F 2.5 L1 None NK CR1 × 7 y 9 3~7 NP 8/M 23.7 L1 None NK CNS relapse 17 mo from dx, reinduction, CR2 × 3 y 10 >10 >10 on mar1 8/F 1.6 L1 None +X, –20, –21 First relapse 14 mo from dx; died after second relapse 11 mo later 11 6-7 6-7 on mar1 20/M 2.2 L1 None –13, –14, +21 CR1 × 30 mo 12 3-8 NP 9/F 6.5 L1 None add(11)(p15), Lost to follow-up r(21) 13 5-6 NP 5/M 5.5 L2 None +X, –21 Lost to follow-up 14 3-5 NP 8/M 1.6 L1 None –11, add(16) Lost to follow-up (p13.3), –21 Ab 5-6 2 on mar1, 2 on 15/M 3.1 L1 pCD13, –8, der(16) CR1 × 9 mo mar2 pCD15 t(1;16), –21 Bb 3-5 2 on mar1, 1 on 12/F 20.3 L2 None del(16)(p13.1), Induction failure, subsequent mar2 –21 SCT, CR1 × 3 y Cb 4-6 1 on mar1-3, 2 on 7/M 1.1 L1 None +X, inv(9) CR1 × 8 y mar4 (p11q13)c, +14, –15, –21

CNS, central nervous system; CR1, first complete remission; CR2, second complete remission; dp, dim partial; dx, diagnosis; GVHD, graft vs host disease; mar, marker; NK, normal karyotype; NP, not performed; p, partial; SCT, stem cell transplant. a One case (case 11) showed a nonconstitutional gain of chromosome 21. b Does not meet Children’s Oncology Group criteria for iAMP21. of less than 0.5 × 109/L, and four (46%) of the 11 cases had CD56 was noted. No aberrant coexpression of myeloper- 1% or fewer blasts in the peripheral blood. Although blast oxidase was present; however, other myeloid marker aber- counts in the peripheral blood varied, all cases demonstrated rancy expression was seen in four (29%) cases. Three cases an effaced bone marrow completely replaced by blasts with showed expression of one marker (CD117, CD13, or CD33), minimal residual background trilineage hematopoiesis. In while one case showed expression of both CD13 and CD33. both the peripheral blood and bone marrow aspirate mate- CD7 and CD4 were each aberrantly expressed in two cases. rial, no associated monocytosis, eosinophilia, or basophilia was identified ❚Image 1❚. Cytogenetic Abnormalities Associated With B-ALL With iAMP21 B-ALL With iAMP21 Shows a Predominantly L1 Blast All cases had a routine karyotype available for review. Morphology and a Precursor/Common B-Cell Phenotype Seven (50%) of 14 cases classically showed a replacement Twelve of 14 cases showed an L1-predominant blast of a normal chromosome 21 (–21) by an overtly abnormal morphology, and two of 14 cases showed an L2-predomi- marker chromosome—an effect of multiple additions of nant blast morphology at diagnosis. Flow cytometric data RUNX1 and other genes into one of the chromosome 21 cop- were available for all 14 cases of B-ALL with iAMP21. All ies. One of these cases showed an abnormal ring chromo- cases demonstrated a precursor B-cell or common B-cell some formation of chromosome 21. However, six (43%) of phenotype, and all showed expression of CD10, CD19, 14 cases did not show an obviously abnormal chromosome CD22, TdT, CD38, and HLA-DR, with most cases show- 21 but showed an apparent normal karyotype. One case ing expression of CD34 (13 of 14), CD45 (typically dim or (case 11) showed a nonconstitutional gain of an extra copy dim partial, 13 of 14), and CD20 (typically dim, 10 of 14). of chromosome 21. The most common additional cytoge- No aberrant expression of CD2, CD3, CD5, CD8, CD16, or netic abnormality was a gain in the sex chromosome (+X),

106 Am J Clin Pathol 2015;144:103-112 © American Society for Clinical Pathology DOI: 10.1309/AJCPLUYF11HQBYRB AJCP / Original Article

A B Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

C ❚Image 1❚ Prototypical case of B-lymphoblastic leukemia (B-ALL) with intrachromosomal amplification of chromosome 21 (iAMP21). A, Peripheral blood smear showing typical features of B-ALL with iAMP21: anemia, neutropenia, and thrombocytopenia. The absolute blast count is usually 10 × 109/L or less, with some cases showing pancytopenia with rare circulating blasts (Wright-Giemsa, ×400). B, Bone marrow aspirate smear showing L1 blasts that comprise the predominant morphologic type in most cases (Wright- Giemsa, ×400). C, Bone marrow core biopsy specimen showing a hypercellular marrow with sheets of blasts, as is typical of B-ALL (H&E, ×200).

which was seen in three cases, and two cases demonstrated ❚Image 2❚. However, all cases showed instability in the distal loss of 11q or the entire chromosome 11. In these two cases arm of chromosome 21 and similar presentation as the cases with –11/–11q abnormalities, loss of the MLL gene locus that met COG criteria. was confirmed by one signal present on the MLL break- The first case (patient A) showed a 47,XY,-8,der(16) apart probe study. Other abnormalities are summarized in t(1;16)(q21;q12~13),-21,+2mar karyotype. Interphase FISH Table 1. Although other clonal cytogenetic anomalies were showed five or more RUNX1 signals per cell, consistent present in most cases, no other World Health Organization with inclusion criteria for most studies assessing iAMP21.5-7 (WHO)–defined recurrent genetic abnormalities were pres- However, when metaphase FISH using the ETV6/RUNX1 ent in our cases to warrant consideration for another B-ALL probe set was performed to localize the extra RUNX1 prognostic entity. Constitutional trisomy 21 was not present signals, two RUNX1 signals localized to each of the in any of the cases. two marker chromosomes. The second patient (patient B) showed a 47,XX,del(16)(p13.1),-21,+mar1,+mar2 karyo- Presentation and Outcome Data for B-ALL With type. Metaphase FISH revealed that two RUNX1 signals iAMP21 and Variant RUNX1 Amplification Patterns localized to one marker chromosome and at least one In addition, three cases showed an abnormal RUNX1 RUNX1 signal localized to the other marker chromosome. signal pattern on the ETV6/RUNX1 probe set but did not The third patent (patient C) showed a 50,XY,+X,inv(9) meet the current COG definition for iAMP21 requiring (p11q13)c,+14,-15,-21,+mar1,+mar2,+mar3,+mar4 karyo- three or more RUNX1 signals on one marker chromosome type. Metaphase FISH analysis revealed that two RUNX1

D E Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

F ❚Image 1❚ (cont) D, The routine karyotype of B-ALL with iAMP21 typically shows loss of a normal chromosome 21 and the presence of a marker chromosome containing extra material from chromosome 21. Other cytogenetic abnormalities may be present but are variable. E, Composite image showing multiple clustered RUNX1 probes (red) relative to ETV6 probe (green) using the RUNX1/ETV6 fluorescence in situ hybridization (FISH) probe set (×1,000). F, Metaphase FISH analysis using the RUNX1/ETV6 probe set confirms that all of the clustered red probes on interphase FISH are localized to one chromosome (arrows) (×1,000). The solitary RUNX1 signal corresponds to the normal chromosome 21 (arrowhead).

signals localized to one marker chromosome and one signal compromise, and multiorgan failure and died 15 months to each of the three remaining marker chromosomes. These after transplant. Five of the 6 patients (patients 2, 3, 4, 5, and three patients had similar low to moderate WBC counts at 11) treated with high-risk therapy have maintained complete diagnosis (3.1, 20.3, and 1.1 × 109/L, respectively), low to remission to date. moderate absolute blast count (0.4, 15.6, and <0.1 × 109/L, Five of 11 patients were placed on standard-risk chemo- respectively), and older age at diagnosis (15, 12, and 7 therapy, consisting of a three-drug standard induction with years, respectively). In these three cases, all showed flow vincristine, dexamethasone, and pegylated-asparaginase. One cytometry findings typical of B-ALL with iAMP21. All patient (patient 7) showed concurrent extramedullary central three expressed CD10, CD19, CD22, TdT, CD38, CD34, nervous system and medullary relapse of leukemia 58 months and HLA-DR, with dim CD45 expression. One case showed after initial diagnosis, underwent reinduction chemotherapy partial expression of CD15 and partial expression of CD13. followed by stem cell transplant, and is currently in second Treatment data were available for 11 of the 14 patients clinical remission 3 years to date. Another patient (patient 10) with B-ALL with iAMP21 (Table 1). Six of 11 patients were relapsed 14 months from original diagnosis and underwent initiated on a high-risk or very high-risk chemotherapeutic reinduction therapy, including intravenous methotrexate, eto- protocol. One patient (patient 1) failed induction therapy poside, and cyclophosphamide, but unfortunately relapsed but, following extended induction, underwent successful again 11 months later and died of leukemia. Three of the 7/8 allele-matched umbilical cord stem cell transplant; how- 5 patients (patients 6, 8, and 9) treated with standard-risk ever, the patient developed graft vs host disease, respiratory therapy have maintained complete remission to date.

108 Am J Clin Pathol 2015;144:103-112 © American Society for Clinical Pathology DOI: 10.1309/AJCPLUYF11HQBYRB AJCP / Original Article

A B Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

C ❚Image 2❚ Three cases with variant amplification of chromosome 21 (increased RUNX1 signals) but not meeting Children’s Oncology Group (COG) criteria for intrachromosomal amplification of chromosome 21 (iAMP21). A-C, Case A: B-lymphoblastic leukemia (B-ALL) showed five RUNX1 signals per nucleus with at least two clustered per nuclear region suggestive of amplification (A; ×1,000); karyotype analysis shows a 47,XY male karyotype with loss of chromosome 21 and presence of two additional marker chromosomes (B); metaphase fluorescence in situ hybridization (FISH) reveals that two RUNX1 signals are clustered on the two marker chromosomes (C; ×1,000).

Treatment data were available for all three patients who predominantly L1 morphology, appear in older children, and did not meet COG criteria for B-ALL with iAMP21. Given have lower peripheral WBC and blast counts. In addition, we their age at diagnosis, two of the three patients (patients A and have identified a subset of patient cases with variant amplifi- B) were placed on a high-risk chemotherapeutic protocol. One cation of chromosome 21 who show similar pathologic fea- of the high-risk patients (patient B) failed induction therapy tures as those patients with B-ALL with iAMP21. Although and demonstrated positive minimal residual disease studies a statistical difference was noted in the WBC count and age at the end of induction. This patient was subsequently treated when comparing iAMP21 with all cases of B-ALL, treating with clofarabine and cytarabine, followed by an umbilical cord the remainder of cases of ALL without iAMP21 as a mono- stem cell transplant, and is currently 3 years in clinical remis- lithic entity has its limitations since subgroups of B-ALL may sion. The other two patients (patients A and C) showed no also present similarly (eg, B-ALL with t(9;22) BCR-ABL1 evidence of clinical, morphologic, or cytogenetic relapse and also tends to occur in older children). Thus, the presentation are in complete remission 9 months and 8 years, respectively. of B-ALL in an older child with a lower WBC count is characteristic but not specific for B-ALL with iAMP21. Intrachromosomal amplification of chromosome region 21q is thought to arise via multiple cycles of chromosomal Discussion BFB formation.17,18 Although the end result of this amplifica- Our findings demonstrate that patients with B-ALL with tion may not be appreciated on routine karyotype depending iAMP21 have blasts with a common/precursor phenotype and on the number of amplification cycles that occurred prior to

D E Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

F ❚Image 2❚ (cont) D-F, Case B: Composite image of B-ALL showed a variable pattern of RUNX1 signals, with most showing between four and five RUNX1 signals (arrow) (D; ×1,000); karyotype analysis shows a 47,XX female karyotype with loss of chromosome 21 and presence of two marker chromosomes (E); metaphase FISH analysis shows that two RUNX1 signals were localized to one marker chromosome and one RUNX1 signal localized to the second marker chromosome (F; ×1,000).

leukemogenesis, all cases are currently defined based on the In addition, since the karyotypes were derived from abnormal RUNX1 pattern via the ETV6/RUNX1 FISH probe cultured cells to capture metaphase events, it is uncertain set. It is important to note, however, that although a working whether relative instability of the marker chromosome in definition has been established by international groups and the iAMP21 in vitro accounts for this karyotypic manifestation. COG,5 the degree of amplification of 21q22 varies between However, given the current leading proposed mechanism for patients and may even vary within the same patient. iAMP21 as chromosome 21 instability through BFB cycles, In our characterization of B-ALL with iAMP21, we it is conceivable that multiple marker chromosomes could uncovered three cases with four to six RUNX1 signals per be formed in this clonal process. Since no other recurrent cell where RUNX1 signals were not localized to a single- cytogenetic abnormality was present in any of these cases marker chromosome but located on two or several distinct yet all shared a loss of a normal chromosome 21, we feel that marker chromosomes. In all three instances, criteria used these cases may represent variant cytogenetic manifestations in retrospective and prospective studies (≥5 RUNX1 signals of iAMP21. Additional studies employing metaphase FISH per cell) were met2,4,6,7; however, these cases did not meet analysis, aCGH, or next-generation sequencing as well as the stricter COG criteria (≥3 RUNX1 signals on one marker clinical follow-up data on a larger cohort of cases are needed chromosome). These cases similarly had a relatively low to help validate our findings. WBC count in older pediatric patients. There are too few Metaphase FISH studies are commonly performed at cases to draw any definitive conclusions about this group of most academic centers and can help provide further informa- cases and our standard cases of iAMP21. tion as to the precise location of the extra RUNX1 signals.

110 Am J Clin Pathol 2015;144:103-112 © American Society for Clinical Pathology DOI: 10.1309/AJCPLUYF11HQBYRB AJCP / Original Article

G H Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021

❚Image 2❚ (cont) G-I, Case C: B-ALL showed greater than I four RUNX1 signals per interphase nucleus (arrows) with two signals appearing clustered (G; ×1,000); karyotype analysis showed a 50XY,+X karyotype with four marker chromosomes (H); metaphase FISH studies confirmed that two RUNX1 signals were on one marker chromosome and one signal each on the remaining three marker chromosomes (I; ×1,000). All three cases (Table 1) showed a low WBC count and absolute blast count, as well as presentation in older children. Although the interphase FISH findings are compatible with iAMP21, these cases do not meet the strict COG definition of three or more RUNX1 signals on a single- marker chromosome to qualify as iAMP21. For metaphase FISH images C, F, and I, marker chromosomes are indicated with arrows; the solitary RUNX1 signal corresponds to normal chromosome 21 (arrowheads).

Although the probe set used is the same one employed to the literature with modest frequency (~10% of cases).4 It is detect the t(12;21) translocation, metaphase FISH requires unclear whether these more common concurrent cytogenetic culturable leukemia cells. Cases with aberrant RUNX1 signal manifestations play any role in the leukemogenesis of B-ALL patterns may include constitutional or acquired trisomy 21, with iAMP21. Next-generation sequencing and/or aCGH may acquired tetrasomy 21, translocations involving the RUNX1 shed more light on the pathophysiology underlying chromo- gene, or multiple marker chromosomes derived from chro- some 21 instability in B-ALL with iAMP21. mosome 21 in addition to iAMP21. Metaphase FISH or None of our cases that met criteria for iAMP21 showed a aCGH should be able to resolve many of these possibilities. concurrent translocation corresponding to WHO-recognized Regarding other cytogenetic abnormalities that occurred B-ALL with recurrent genetic abnormalities. Although infre- concurrently with iAMP21 in our series, three patients showed quent, iAMP21 with concurrent t(9;22)(q34;q11.2) BCR- gains in the X sex chromosome, and two patients showed ABL14 as well as t(12;21)(p13;q22) ETV6-RUNX14,16,23,24 loss of one 11q23 locus, resulting in an abnormal pattern on has been reported in the literature. It would be difficult in FISH for MLL rearrangements. Chromosome 16 was also these circumstances to predict prognosis based on these commonly involved in translocations as well as deletions. Of concurrent findings, and further studies are required to concurrent abnormalities reported in the literature, gain of the determine the prognostic or treatment impact that such com- X sex chromosome was the most common occurrence and is binatorial cytogenetic abnormalities might have on patients. estimated to occur in approximately 10% to 20% of cases.4,22 B-ALL with iAMP21 tends to have a low WBC count Abnormalities involving 11q have also been reported in in older children, and our series of patients supports much

© American Society for Clinical Pathology Am J Clin Pathol 2015;144:103-112 111 DOI: 10.1309/AJCPLUYF11HQBYRB Johnson et al / B-Lymphoblastic Leukemia With iAMP21 of the data noted in current literature. All cases showed a 8. Alvarez Y, Coll MD, Bastida P, et al. AML1 amplification precursor or common B-lymphoblast phenotype, with a in a child with acute lymphoblastic leukemia. Cancer Genet Cytogenet. 2003;140:58-61. subset showing aberrant myeloid-associated antigen expres- 9. Busson-Le Coniat M, Nguyen Khac F, Daniel MT, et al. sion. Since cases are currently defined by the COG as three Chromosome 21 abnormalities with AML1 amplification in or more RUNX1 signals per single-marker chromosome acute lymphoblastic leukemia. Genes Chromosomes Cancer. and by most studies as five or more signals per interphase 2001;32:244-249. nucleus, consideration for other cytogenetic variants of 10. Niini T, Kanerva J, Vettenranta K, et al. AML1 gene amplification: a novel finding in childhood acute iAMP21 should be made when formalizing a final defini- lymphoblastic leukemia. Haematologica. 2000;85:362-366. tion of this entity. Ancillary information, such as metaphase 11. Dal Cin P, Atkins L, Ford C, et al. Amplification of FISH or aCGH analysis, may lead to more inclusive criteria AML1 in childhood acute lymphoblastic leukemias. Genes for iAMP21 and allow patients with this high-risk B-ALL Chromosomes Cancer. 2001;30:407-409. appropriate high-risk stratification chemotherapy. 12. Mathew S, Rao PH, Dalton J, et al. Multicolor spectral Downloaded from https://academic.oup.com/ajcp/article/144/1/103/1761609 by guest on 30 September 2021 karyotyping identifies novel translocations in childhood acute lymphoblastic leukemia. Leukemia. 2001;15:468-472. Corresponding author: Ryan C. Johnson, MD, Dept of Pathology, 13. Morel F, Herry A, Le Bris M-J, et al. AML1 amplification in Stanford Hospital and Clinics, 300 Pasteur Dr, L235 MC 5324, a case of childhood acute lymphoblastic leukemia. Cancer Stanford, CA 94035; [email protected]. Genet Cytogenet. 2002;137:142-145. Acknowledgments: The authors thank Dana Bangs for 14. Penther D, Preudhomme C, Talmant P, et al. Amplification providing the cytogenetic images, Norm Cyr for his assistance in of AML1 gene is present in childhood acute lymphoblastic production of the figures for this article, and Mark Fleming, MD, leukemia but not in adult and is not associated with AML1 DPhil, at Boston Children’s Hospital for his continued support gene mutation. Leukemia. 2002;16:1131-1134. and collaboration. 15. Jabber Al-Obaidi MS, Martineau M, Bennett CF, et al. ETV6/AML1 fusion by FISH in adult acute lymphoblastic leukemia. Leukemia. 2002;16:669-674. 16. Mikhail FM, Serry KA, Hatem N, et al. AML1 gene over- References expression in childhood acute lymphoblastic leukemia. Leukemia. 2002;16:658-668. 1. Borowitz MJ, Chan JKC. B lymphoblastic leukemia/ lymphoma with recurrent cytogenetic abnormalities. In: 17. Kuchinskaya E, Nordgren A, Heyman M, et al. Tiling- Swerdlow SHCE, Harris NL, Jaffe ES, et al, eds. WHO resolution array-CGH reveals the pattern of DNA copy Classification of Tumours of Hematopoietic and Lymphoid number alterations in acute lymphoblastic leukemia with 21q Tissues. Lyon, France: IARC; 2008:110-112. amplification: the result of telomere dysfunction and breakage/ fusion/breakage cycles? Leukemia. 2007;21:1327-1330. 2. Moorman AV, Ensor HM, Richards SM, et al. Prognostic effect of chromosomal abnormalities in childhood B-cell 18. Robinson HM, Harrison CJ, Moorman AV, et al. precursor acute lymphoblastic leukaemia: results from the Intrachromosomal amplification of chromosome 21 UK Medical Research Council ALL97/99 randomised trial. (iAMP21) may arise from a breakage-fusion-bridge cycle. Lancet Oncol. 2010;11:429-438. Genes Chromosomes Cancer. 2007;46:318-326. 3. International Standing Committee on Human Cytogenetic 19. Strefford JC, Van Delft FW, Robinson HM, et al. Complex Nomenclature. ISCN 2013: An International System for Human genomic alterations and gene expression in acute lymphoblastic Cytogenetic Nomenclature. Basel, Switzerland: Karger; 2013. leukemia with intrachromosomal amplification of chromosome 21. Proc Natl Acad Sci USA. 2006;103:8167-8172. 4. Harrison CJ, Moorman AV, Schwab C, et al. An international study of intrachromosomal amplification of 20. Van Delft FW, Bellotti T, Luo Z, et al. Prospective gene chromosome 21 (iAMP21): cytogenetic characterization and expression analysis accurately subtypes acute leukaemia outcome. Leukemia. 2014;28:1015-1021. in children and establishes a commonality between hyperdiploidy and t(12;21) in acute lymphoblastic leukaemia. 5. Heerema NA, Carroll AJ, Devidas M, et al. Intrachromosomal Br J Haematol. 2005;130:26-35. amplification of chromosome 21 is associated with inferior outcomes in children with acute lymphoblastic leukemia 21. Sinclair PB, Parker H, An Q, et al. Analysis of a treated in contemporary standard-risk Children’s Oncology breakpoint cluster reveals insight into the mechanism of Group studies: a report from the Children’s Oncology Group. J intrachromosomal amplification in a lymphoid malignancy. Clin Oncol. 2013;31:3397-3402. Hum Mol Genet. 2011;20:2591-2602. 6. Harrison CJ, Haas O, Harbott J, et al; Biology and 22. Soulier J, Trakhtenbrot L, Najfeld V, et al. Amplification of Diagnosis Committee of International Berlin-Frankfürt- band q22 of chromosome 21 including AML1 in older children Münster Study Group. Detection of prognostically relevant with acute lymphoblastic leukemia: an emerging molecular genetic abnormalities in childhood B-cell precursor acute cytogenetic subgroup. Leukemia. 2003;17:1679-1682. lymphoblastic leukaemia: recommendations from the 23. Haltrich I, Csóka M, Kovács G, et al. Six cases of rare Biology and Diagnosis Committee of the International gene amplifications and multiple copy of fusion gene in Berlin-Frankfürt-Münster Study Group. Br J Haematol. childhood acute lymphoblastic leukemia. Pathol Oncol Res. 2010;151:132-142. 2013;19:123-128. 7. Moorman AV, Richards SM, Robinson HM, et al. Prognosis 24. Ma SK, Wan TS, Cheuk AT, et al. Characterization of of children with acute lymphoblastic leukemia (ALL) additional genetic events in childhood acute lymphoblastic and intrachromosomal amplification of chromosome 21 leukemia with TEL/AML1 gene fusion: a molecular (iAMP21). Blood. 2007;109:2327-2330. cytogenetics study. Leukemia. 2001;15:1442-1447.