Systematic application of fluorescence in situ hybridization and immunophenotype profile for the identification of ZNF384 gene rearrangements in B cell acute lymphoblastic leukemia

Abstract

Introduction

ZNF384 gene fusions resulting from translocations with several partner genes have been described in B cell acute lymphoblastic leukemia (B-ALL) with a characteristic immunophenotype (aberrant CD13 and or CD33 with dim CD10). The prognosis of patients with this rearrangement appears to depend on the fusion partner. ZNF384 rearrangements have been identified by high through put technologies such as RNA sequencing in most of the studies published. We tested the feasibility of using the characteristic immunophenotype as a tool to screen for patients with ZNF384 translocations which can be subsequently confirmed by cytogenetic / molecular methodologies.

Methods

ZNF384 rearrangements in B-ALL patients at diagnosis with CD10 <80% and were negative for the BCR-ABL1 fusion (n = 109) were identified by fluorescence in situ hybridization followed by confirmation by reverse transcriptase-polymerase chain reaction and Sanger sequencing. The end of induction measurable residual disease evaluated by flow cytometry for these patients was obtained from patient records.

Results

ZNF384 translocations were identified in 14 patients and were cytogenetically cryptic in 13. EP300-ZNF384 was the most common fusion partner (n = 12), while TAF15-ZNF384 and TCF3-ZNF384 were identified in 1 patient each. End of induction MRD by flow cytometry was positive in 5 of 8 patients with the EP300-ZNF384 fusion treated at our center.

Conclusion

Our findings show a practical approach for the identification of ZNF384 gene rearrangements by widely available technologies and indicate that the response to therapy may be heterogeneous even in this subset, which has been reported as having a favorable prognosis.

1. Introduction

B cell acute lymphoblastic leukemia (B-ALL) is a heterogeneous disease characterized by the presence of distinct genetic subtypes which impact response to therapy.^1,2 Gene fusions involving the zinc-finger protein 384 (ZNF384) gene on chromosome 12p13.31 have been reported in up to 6% of children, 7.3% adults, and 15% of adolescents and young adults(AYA) with B-ALL.^3–5 The ZNF384 gene encodes a transcription factor that regulates promoters of the extracellular matrix genes⁶ and is known to be involved in B-ALL and in mixed phenotype acute leukemia (MPAL) through fusions with several partner genes.⁵,^7–10 Fusions involving the ZNF384 gene have been reported to be mutually exclusive of other genetic subtypes and thus constitute a distinct subgroup in B-ALL.⁶ ZNF384 rearranged (ZNF384r) leukemia shows a unique gene expression profile and displays a characteristic immunophenotype with weak CD10 and aberrant expression of the myeloid markers, CD13, and/ or CD33.^11–13

The genes most frequently involved in fusions with ZNF384 include EP300 (22q13.2), TCF3 (19p13.3), and TAF15 (17q12), and the prognosis appears to vary by fusion partner. The EP300-ZNF384 fusion appears to be favorable while the TCF3-ZNF384 fusion is frequently associated with late relapses and a poor prognosis.^11,14 Since the ZNF384 gene and several of the fusion partner gene loci are present adjacent to the telomeric ends of their respective chromosome arms, ZNF384 gene fusions are cytogenetically cryptic and hence cannot be identified by analysis of G-banded chromosomes.¹⁰

ZNF384r leukemia’s are characterized by overexpression of the FLT3 gene¹⁵ and sunitinib, a FLT3 inhibitor, has been shown to induce molecular response in a patient with EP300-ZNF384 rearrangement.¹⁶ Although the biological basis of FLT3 overexpression in such patients is not well understood, identification of such fusions may thus help in providing the option for targeted therapy in patients who remain refractory/relapse on conventional chemotherapy. There is a paucity of data in this unique subtype of patients from India. We describe a series of B-ALL patients with ZNF384 gene rearrangements diagnosed and treated at our center.

2. Methods

2.1. Patients

Consecutive B-ALL patients diagnosed in the Department of Haematology, Christian Medical College, Vellore between January 1, 2015, and September 30, 2020, with CD10 expression <80% and who were negative for the BCR-ABL1 fusion were selected for analysis.¹¹ Diagnosis and risk stratification of B cell ALL were based on morphology, immunophenotyping by multiparameter flow cytometry, cytogenetic analysis, and the RT-PCR detection of common fusion transcripts (BCR-ABL1, ETV6-RUNX1, TCF3-PBX1, and KMT2A-AFF1), as described previously.¹⁷ The clinical details including age, WBC count, bone marrow (BM) blast percentage, MRD status (positive >0.01%),^18,19 and treatment details were collected from the patients’ records.

2.2. Detection of ZNF384 rearrangement by FISH

FISH analysis was performed using a dual-color break apart probe, SPEC ZNF384 (ZytoLight dual-color break apart probe, Zytovision) on destained G-banded slides of bone marrow/ archived cell pellets. Slides were analyzed using the Axioimager M1 (Carl Zeiss) microscope and ISIS software (Metasystems).²⁰

2.3. Molecular analysis of ZNF384 fusion partners

Total RNA was extracted from bone marrow or peripheral blood samples at diagnosis. RNA (2 μg) was reverse transcribed to cDNA using random hexamers and Superscript RT Kit (Thermo Scientific). RT-PCR was performed to identify the ZNF384 fusion partner as described previously.^12,17 The PCR product was sequenced using the Big Dye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems) in an ABI Prism 3130 Genetic Analyzer (Applied Biosystems).

2.4. Treatment

Modified BFM 95 and GMALL protocols with risk stratification were used for treating children and adults, respectively.²¹ Postinduction, flow cytometry-based measurable residual disease (MRD) testing was performed.²² Patients with high-risk disease were offered allogeneic transplant.²¹

3. Results

3.1. Detection of ZNF384 rearrangements in BCP-ALL patients by FISH analysis

Of the 891 patients with B-ALL diagnosed during the study period (January 2015-June 2020), 158(68 children and 90 adults) had CD10 expression <80% after exclusion of patients with the BCR/ABL1 fusion (n = 16). Samples were not available for analysis in 49 patients. The algorithm of sample selection and evaluation is illustrated in Figure S1. ZNF384 gene rearrangements were identified in 14 of the 109 patients tested (12.8%). All 14 patients had weak to moderate expression of CD10 with aberrant CD13 and CD33 in 9 and either CD13/CD33 in 2 and 3 patients, respectively(Table 1).

Table 1. Clinical characteristics and laboratory findings in patients with ZNF384 gene rearrangements.

UPN.NO	Age/ Sex	Total WBC (/cu.mm)	Blast (%)	CD10 (%)	CD13 (%)	CD33 (%)	Karyotype	Fusion partner (RT-PCR)	End Induction MRD
FS 39/19	59/M	16 900	94	37.7	66.4	93.1	46,XY,t(12;17)(p13.3;q11.2) [16]/ 47,idem,+8[4]	TAF15	Positive (0.202%)
FS 42/19	33/M	6200	29	34.1	61.6	85	46,XY,add(7)(p11.1),del(7) (p11.1)[12]/46,XY[8]	TCF3 FISH only	aNot applicable
FS 52/19	23/M	34 000	88	47.6	60.4	15	46,XY[20]	EP300	Positive (0.411%)
FS 84/19	13/F	11 100	56	58	1.7	91.1	46,XX,del(7)(q22)[9]/ 46,XX[11]	EP300	Positive (1.735%)
FS 62/19	2/M	10 100	79	44.1	13.5	84.4	46,XY[20]	EP300-FISH only	aNot applicable
F1002/19	9/F	18 600	83	11.8	94.8	97.1	46,XX[20]	EP300	Positive (0.500%)
FS124/19	2/M	45 700	73	55.2	73.9	17.5	46,XY[22]	EP300	Negative
FS04/20	25/M	12 100	39	1.7	21.9	86.7	46 ~ 47,XY,del(1)(q32),?der (16), add(17)(q25), -20,+mar, +mar [cp7]/ 46,XY[19]	EP300	Positive (2.247%)
FS113/20	48/M	3200	96	17.9	90.4	80.4	46,XY,t(2;5)(p12;q35), del(13) (q13q32)[15]/ 46,XY[5]	EP300	Negative
FS124/20	22/M	5400	61	43.1	44.5	84.6	46,XY[20]	EP300	Negative
FS125/20	52/M	2000	90	7.5	42.3	83.9	46,XY,add(20)(q13.3) [2]/ 46,XY[18]	EP300-FISH only	Not available
FS127/20	20/F	5500	92	56.9	6.9	70.9	46,XX,ins(5;13) (q31;q13q33) [11]/ 46,XX[9]	EP300	aNot applicable
FS162/20	41/F	9100	82	0.3	46.8	86.4	46,XX[21]	EP300	Positive (0.463%)
FL 54/20	33/M	16 900	97	14	52.2	51	47,XY,+X[11]/48,idem, +12[4]/46,XY[6]	EP300-FISH only	aNot applicable

Note: Abbreviations: CD, Cluster of differentiation; MRD, Measurable residual disease; RT-PCR, Reverse transcriptase-polymerase chain reaction; WBC, White blood cell.

^aPatients who discontinued treatment at the study center.

3.2. EP300 was the most frequent fusion partner

Archived cDNA/RNA samples were available for 10 of the 14 patients with ZNF384 gene rearrangements. The EP300-ZNF384 fusion was identified in 9 and the TAF15-ZNF384 in one patient, respectively.

The breakpoints were identified by sequencing to characterize the fusion genes. In all the 9 patients with the EP300-ZNF384 fusion, the resulting sequence revealed that the exon 6 of EP300 was fused to exon 3 of ZNF384 (Figure 1A). In the patient with the t(12;17), exon 9 of the TAF15 gene was fused to exon 3 of ZNF384 (Figure 1B).

Figure 1. Identification of ZNF384 gene fusions.

Sequence of the (A) EP300-ZNF384 fusion involving exon 6 of EP300 and exon 3 of ZNF384 (UPN. No FS 84/19 ) (B) TAF15-ZNF384 fusion involving exon 9 of TAF15 and exon 3 of ZNF384 (UPN. No FS 39/19) (C-D) Inverted DAPI view of metaphase FISH image (1Fusion 1Red 1Green) showing the (C) t(12;22) with the 3’ZNF384 signal(spectrum green) on the derivative(der) chromosome 22 (UPN. No FS 62/19) (D) t(12;19) with the 3’ZNF384 signal (spectrum green) on the derivative chromosome 19 (UPN. No FS 42/19) (E) Representative image of G-banded karyotype showing the reciprocal t(12;17) (F) Inverted DAPI view of metaphase FISH image showing the t(12;17) with the 3ZNF384 signal(spectrum green) on the derivative chromosome 17 (UPN. No FS 39/19)

Matched metaphase FISH analysis in 4 patients for whom samples were not available for RT-PCR analysis showed the t(12;22)- EP300-ZNF384 in 3(Figure 1C), and the t(12;19)- TCF3-ZNF384 in 1(Figure 1D).

Among the 14 patients who showed ZNF384 gene rearrangement by FISH, 9 had karyotypic abnormalities of which only 1 showed an abnormality involving the ZNF384 locus on 12p13.31, namely the t(12;17) (Figure 1E,F). Normal karyotypes were reported in 7 patients while the rest had structural abnormalities that did not involve the ZNF384 gene locus (Table 1).

3.3. Treatment outcome

Of the 10 patients for whom molecular characterization was done, 3 children were treated with a modified BFM protocol while 7 adult patients received the modified GMALL protocol. MRD status at the end of induction was available for 9 patients of whom 6 (66.7%) were positive (Table 1). This included one patient with the TAF15-ZNF384 fusion and 5 with the ZNF384-EP300 fusion. The median follow-up since treatment initiation for the 10 patients who were treated at our center was 13.4 months (interquartile range: 9.0-39.1 months). Hence, although, we noted high rates of detectable MRD at the end of induction therapy, we did not see any morphologic relapses. This was despite only 2 patients being able to undergo stem cell transplantation (Supplementary results).

4. Discussion

Advanced genomic approaches have led to the classification of BALL, specifically B-other ALL into novel genetic subtypes, some with a clear potential for targeted therapeutic approaches.^2,15 In our cohort, we used a strategy wherein Ph-negative patients with low CD10 expression were tested for ZNF384 rearrangements. In one of the largest studies on ZNF384 rearranged B-ALL, the CD10 expression was reported to range from 0.39% to 67.34%.¹⁴ Since the phenotypic association has been well described previously, we tested for this abnormality in a selected group of 109 patients. Using this strategy, among the tested samples, we identified ZNF384 rearrangements in 6% of children and 11% of adults. ZNF384 fusions constituted 2% (4/193) of children and 5% (10/192) of adults with B-other ALL in the study group. There is a wide range of frequencies of ZNF384 gene fusion in literature depending on the specific subtype of B-ALL that is analyzed. Studies on B-ALL show the ZNF384 in 1%-5% of children and 7.3% in adults, whereas it is seen in 5% and 6% of children with B-other ALL and adults with Philadelphia-negative ALL, respectively. Although the frequencies reported by us are comparable to published reports, a limitation in this study is that the efficiency of this selection strategy based on immunophenotype has not been compared with a control group with CD10 >80%.

Similar to previous reports, EP300 was the most frequent fusion partner in our cohort of patients with ZNF384 gene rearrangements.¹¹ ZNF384r ALLs do not exhibit other unique presenting features except for the distinct immunophenotype.

The EP300-ZNF384 and the TCF3-ZNF384 fusions identified in this study were the result of cytogenetically cryptic translocations as has been reported previously,¹⁰ while only the TAF15-ZNF384 fusion could be identified by karyotyping. A striking difference is the presence of chromosomal abnormalities in 9 patients in contrast to the majority of previous reports which have shown cryptic ZNF384 rearrangements in patients with normal karyotypes.^10,14 The clinical significance of this observation has to be explored further.

The prognosis of patients with ZNF384r may be dependent on the fusion partner. Previous studies have shown that the EP300-ZNF384 group to be associated with good prognosis and the TCF3-ZNF384 to be associated with late relapse.¹² However, 6 of the 9 patients treated in our center were MRD positive at the end of induction including 5 with EP300-ZNF384 suggesting that this subgroup may require a longer follow-up to assess their prognostic impact. Due to a short follow-up period, this does not appear to translate into an adverse clinical outcome in our cohort.

The limitations of this study include the fact that the initial screening was done by FISH analysis followed by RT-PCR for only the patients who were positive by FISH. Cryptic insertions might not have been identified by FISH. In addition, we identified only one patient each with the TCF3 and TAF15- ZNF384 gene fusions, and hence, the clinical outcome of this subgroup needs to be validated in a larger cohort with long-term follow-up data.

Screening for ZNF384 rearrangements in patients with B-other ALL who show an aberrant expression of CD13 and /or CD33 with dim to moderate CD10 expression may help in identifying this subgroup of patients. Aberrant expression of CD13 and CD33 has also been reported in Philadelphia chromosome-positive B-ALL.²³ Hence, in B-ALL patients with low CD10 and aberrant CD13/CD33, a possible reflex test could be either FISH using a break part probe for ZNF384 or RT-PCR for the commonly reported ZNF384 gene fusions after excluding BCR-ABL1 fusion.

Supplementary Material

Additional supporting information may be found online in the Supporting Information section.

Data Availability Statement

Data available on request from the authors.