Use of Whole-Genome Sequencingto Diagnose a Cryptic Fusion OncogeneJohn S. Welch, MD, PhDPeter Westervelt, MD, PhDLi Ding, PhDDavid E. Larson, PhDJeffery M. Klco, MD, PhDShashikant Kulkarni, PhDJohn Wallis, PhDKen Chen, PhDJacqueline E. Payton, MD, PhDRobert S. Fulton, MSJoelle Veizer, BSHeather Schmidt, BSTammi L. Vickery, BSSharon HeathMark A. Watson, MD, PhDMichael H. Tomasson, MDDaniel C. Link, MDTimothy A. Graubert, MDJohn F. DiPersio, MD, PhDElaine R. Mardis, PhDTimothy J. Ley, MDRichard K. Wilson, PhD
ACUTE PROMYELOCYTIC LEUKE-mia (APL) is commonly(�90%) associated withPML-RARA (NCBI Entrez
Gene 5371 and 5914) fusion tran-scripts resulting from pathogenic t(15;17) translocations.1,2 Unusual cytoge-netic rearrangements (eg, insertions and3, 4, or even 8-way translocations)2-4 canalso lead to PML-RARA formation. Al-ternative PML-RARA fusions and splice
variants exist but are not detected bystandard reverse transcription poly-merase chain reaction (RT-PCR)5-7; al-ternative X-RARA fusions also may ex-ist and may be responsive to all-transretinoic acid (ATRA) (eg, NuMA1-RARA, NPM1-RARA, STAT5B-RARA,PRKAR1A-RARA, FIP1L1-RARA,BCOR-RARA, and the non-RARA trans-location NUP98-RARG)1,8-12 or ATRA re-sistant (PLZF-RARA).1 Timely and ac-curate diagnosis of APL is essential,because the addition of ATRA to che-motherapy leads to substantially im-proved outcomes (5-year event-free sur-vival of 69%, compared with 29% in
patients receiving chemotherapyalone).13
CASE HISTORYA 39-year-old woman with acute my-eloid leukemia (AML) in first remis-
See also pp 1568 and 1596.
Author Affiliations: Departments of Medicine (DrsWelch, Westervelt, Tomasson, Link, Graubert, DiPer-sio, and Ley and Ms Heath), Pathology and Immu-nology (Drs Klco, Kulkarni, Payton, and Watson), Ge-netics (Drs Kulkarni, Mardis, Ley, and Wilson), andPediatrics (Dr Kulkarni), and Genome Institute (DrsDing, Larson, Wallis, Chen, Watson, Mardis, Ley, andWilson and Mr Fulton and Mss Veizer, Schmidt, andVickery), Washington University, St Louis, Missouri.Corresponding Author: Richard K. Wilson, PhD, Ge-nome Institute, Washington University School of Medi-cine, 4444 Forest Park Blvd, PO Box 8501, St Louis,MO 63108 ([email protected]).
Context Whole-genome sequencing is becoming increasingly available for researchpurposes, but it has not yet been routinely used for clinical diagnosis.
Objective To determine whether whole-genome sequencing can identify cryptic,actionable mutations in a clinically relevant time frame.
Design, Setting, and Patient We were referred a difficult diagnostic case of acutepromyelocytic leukemia with no pathogenic X-RARA fusion identified by routine meta-phase cytogenetics or interphase fluorescence in situ hybridization (FISH). The casepatient was enrolled in an institutional review board–approved protocol, with consentspecifically tailored to the implications of whole-genome sequencing. The protocol usesa “movable firewall” that maintains patient anonymity within the entire research teambut allows the research team to communicate medically relevant information to thetreating physician.
Main Outcome Measures Clinical relevance of whole-genome sequencing andtime to communicate validated results to the treating physician.
Results Massively parallel paired-end sequencing allowed identification of a cyto-genetically cryptic event: a 77-kilobase segment from chromosome 15 was inserteden bloc into the second intron of the RARA gene on chromosome 17, resulting in aclassic bcr3 PML-RARA fusion gene. Reverse transcription polymerase chain reactionsequencing subsequently validated the expression of the fusion transcript. Novel FISHprobes identified 2 additional cases of t(15;17)–negative acute promyelocytic leuke-mia that had cytogenetically invisible insertions. Whole-genome sequencing and vali-dation were completed in 7 weeks and changed the treatment plan for the patient.
Conclusion Whole-genome sequencing can identify cytogenetically invisible onco-genes in a clinically relevant time frame.JAMA. 2011;305(15):1577-1584 www.jama.com
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
sion was referred to our institution forconsideration of allogeneic stem celltransplantation. She had initially pre-sented with hypofibrinogenemia, dis-seminated intravascular coagulopa-thy, and pancytopenia (white blood cellcount, 1300/µL; hemoglobin, 11.6 g/dL;platelets, 72 �103/µL). Her bone mar-row contained 61% atypical promyelo-cytes with invaginated nuclei (includ-ing bilobed forms) and dense primarygranules (FIGURE 1). She started in-duction chemotherapy with ATRA,cytarabine, and idarubicin. However,her metaphase cytogenetics (46, XX,del(9)(q12q32), del(12)(q12q21), −6,−16, add(16)(p13.2), �2 mar[13/20cells]) (FIGURE 2) revealed a complexpattern, which is associated with lessthan 15% long-term survival and istreated with allogeneic transplanta-tion during first remission wheneverpossible.14,15
Interphase fluorescence in situ hy-bridization (FISH) suggested a pos-sible fusion between chromosomes 15and 17 on der(17) but was most con-sistent with a RARA-PML fusion, not thepathogenic PML-RARA fusion charac-teristic of M3 AML (FIGURE 3). RT-PCRto detect a PML-RARA fusion tran-script was not performed at the refer-ring institution.
These findings led to a diagnostic co-nundrum, and ATRA was discontin-ued. Persistent AML was observed onday 14. The patient entered a com-plete remission following reinductionwith cytarabine, idarubicin, and eto-poside. She was then referred to our in-stitution for consideration of alloge-neic stem cell transplantation. At thattime, her bone marrow biopsy re-vealed no morphologic evidence ofAML and had normal metaphase cyto-genetics, normal interphase FISH re-sults, and no evidence of PML-RARA byRT-PCR. HLA typing identified 1matched sibling. This case posed a di-agnostic dilemma with prognostic andtherapeutic consequences: does the pa-tient have APL, or does she have AMLwith unfavorable-risk cytogenetics?
Because her leukemic cytomorphol-ogy was consistent with APL, we em-
pirically recommended 2 cycles of ar-senic trioxide consolidation, which shereceived.16
Little material from her original leu-kemia remained for subsequent evalu-ation, and no clinical samples wereavailable for FISH or RT-PCR. How-ever, 2 vials of bone marrow cells hadbeen cryopreserved under a researchprotocol at her referring institution.DNA and RNA were generated fromthese respective samples (the RNAsample was severely degraded). We ob-tained appropriate consent for whole-genome sequencing and completed thisanalysis using paired-end reads. Ourprimary goal was to determine if whole-genome sequencing could identify anactionable mutation (eg, a cryptic X-RARA rearrangement) in a clinically rel-evant time frame (6-8 weeks).
METHODSA “movable-firewall” within our re-search protocol allows for the commu-nication of clinically relevant findingsto the patient’s physician and to the pa-tient while strictly maintaining pa-tient anonymity among all research per-sonnel. Deidentified samples areentered into a tissue bank. Clinical in-formation (eg, age, sex, disease, treat-ment, outcome) is maintained in asso-ciation with deidentified codes only. Alist associating deidentified codes withpersonal patient information (name,date of birth, treating physician) ismaintained in a locked safe; a singleprotocol administrator has access to thislist, and the research team can com-municate medically relevant informa-tion to the administrator. The admin-istrator communicates this informationto the treating physician, who is re-sponsible for informing the patient ofthe whole-genome sequencing resultsand their clinical implications.
After obtaining explicit consent forwhole-genome sequencing with an in-stitutional review board–approved pro-tocol, DNA libraries were generatedfrom 1 cryovial of the original bonemarrow aspirate and from a skin punchbiopsy obtained during remission(matched normal cells). We gener-
ated 187.1 and 200.1 billion base pairsof DNA sequence from each of the re-spective samples, with an average readdepth of 43.7� and 46.8�, respec-tively. Library generation, sequenceproduction, and data analysis were per-formed as previously described.17-20 Ad-equate genome-wide coverage (�99.5%diploid coverage) was ensured by as-sessing the coverage of known single-nucleotide polymorphisms in the pa-tient’s genome, as defined with datacollected from the Affymetrix Genome-Wide Human SNP Array 6.0 (Af-fymetrix, Santa Clara, California).
All of the high-quality single-nucleotide variants (SNVs) found in tu-mor and skin samples from this pa-tient are available in the database ofgenotypes and phenotypes (dbGaP) ofthe National Center for BiotechnologyInformation (phs000159.v3.p2).
Reagents and methods for PCR vali-dation, RT-PCR, and FISH analyses, andthe National Center for BiotechnologyInformation Entrez Gene identifica-tion numbers of all genes relevant to thisarticle, are described in the eMethodssection available at http://www.jama.com.
RESULTSValidated whole-genome sequencing re-sults were completed and reported tothe patient’s physician 7 weeks after ob-taining the DNA samples.
The timeline of data production,analysis, and validation was as fol-lows: day 1, DNA samples logged in atWashington University Genome Insti-tute; day 5, libraries completed andsequencing begun; day 18, sequencecompleted; day 22, alignment to refer-ence sequence completed; day 24,prediction of SNVs completed; day25, structural variants predicted byBreakDancer20; day 52, insertionalfusion completely validated by PCRand results transmitted to the treatingphysician.
Using massively parallel DNA se-quencing with paired-end reads, weidentified 2 sets of breakpoints be-tween chromosomes 15 and 17, whichoccur in the LOXL1/PML locus and
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
RARA locus, respectively (schemati-cally described in FIGURE 4). PCR am-plification across each predicted break-point validated the en bloc insertion ofa 77-kilobase (kb) segment of chromo-some 15 (containing parts of the LOXL1and PML genes) into intron 2 of RARA,the invariant site of RARA-associatedtranslocations (FIGURE 5A).
This insertion generates 3 novel fu-sion transcripts: bcr3 PML-RARA,LOXL1-PML, and RARA-LOXL1. Ex-pression of bcr3 PML-RARA was vali-dated by RT-PCR using 3 differentprimer pairs (using the degraded RNAfrom the original banked AML sample),including the Clinical Laboratory Im-provement Act (CLIA-CAP)–certifiedInVivoscribe PML/RARa Mix2b kit (In-vivoscribe Technologies, San Diego,California) (Figure 5B, eFigure 1A, anddata not shown). The RARA-LOXL1 fu-sion was out of frame and is predictedto encode a 67–amino acid protein(eFigure 1B and data not shown). TheLOXL1-PML fusion leads to alteredsplicing and a premature stop codonprior to the PML junction and is pre-dicted to encode a 573–amino acid pro-tein (eFigure 1C).
In addition, we identified and re-solved the breakpoints associated withall abnormalities observed with meta-phase cytogenetics, including del(9), del(12), and add(16)(p32.2); the latter wasin fact a translocation t(16;22)(p13.3;q13.31) (Figure 2, eFigure 2, and datanot shown). Two other large deletionsnot found by conventional cytogenet-ics were detected by whole-genome se-quencing: del(14) and del(19). The lat-ter was also identified in the skinsample, proving that it is an inheritedcopy number variant (Figure 5A). Thepredicted deletions of chromosomes 6and 16 were not detected by whole-genome sequencing. Instead, we iden-tified a 61-megabase inv(6)(p22.3;q14.1) and a translocation t(6;16)(q22.31;p13.3) (Figure 2). We furtheridentified and validated 12 SNVs withinprotein-coding sequences (FIGURE 6).SNV allele frequency was consistentwith the presence of 2 distinct leuke-mic clones in the bone marrow, reca-
pitulating the metaphase cytogenet-ics: 8 of 12 SNVs had a variant allelefrequency of 35% to 51%, and 4 of 12
SNVs had a variant allele frequency of13% to 21% (Figures 2 and 6). The sig-nificance of these somatic mutations for
Figure 1. Molecular Diagnostics of Case Patient
A B
Cytomorphology of initial bone marrow biopsy. A, Hematoxylin and eosin stain; original magnification �100.B, Hematoxylin and eosin stain; original magnification �600.
Figure 2. Metaphase Cytogenetics of Case Patient
Dominant cloneA Minor cloneB
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
Marker chromosomes
19 20 21 22 X Y
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
A, Dominant clone (46, XX, del(9)(q12q32), del(12)(q12q21), −6, −16, add(16)(p13.2), �2 mar [13 of 20cells]). B, Minor clone (46, XX, del(9)(q12q32), del(12)(q12q21)[6 of 20 cells]). Pink boxes indicate chromo-somal abnormalities.
Figure 3. FISH Analysis of Bone Marrow Aspirate Cells of Case Patient
A B
17
Fusion on der(17)FusionFusion
1515
Metaphase FISHInterphase FISH
Fluorescence in situ hybridization (FISH) performed using dual fusion, dual probes (Abbott/Vysis). A, Onefusion signal, 2 red signals (chromosomes 15), and 1 green signal (chromosome 17). B, Fusion on der(17).Probes labeled with Spectrum Orange (PML) and Spectrum Green (RARA).
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
disease pathogenesis is currently un-known.
We designed a new set of fosmid-based FISH probes (each 30-40 kilo-bases in size) for the detection ofinsertional fusions that target theminimal PML translocation region(the promoter/enhancer and exons1-3, which is roughly 30 kilobases)(FIGURE 7). We searched the Wash-ington University Department of
Pathology database for AML casesdiagnosed during the last 5 years. Weidentified 11 cases with features sug-gestive of APL (including any promy-elocytic morphology, characteristicCD33�CD34−HLA-DR− immunophe-notype, and variable-to-strongmyeloperoxidase staining by enzymecytochemistry) but that lacked nor-mal dual-fusion patterns by FISH.We found that 2 of these specimens
contained PML-RARA fusions result-ing from cryptic insertions: one wasassociated with an insertion of PMLinto the RARA locus (ins[17;15]) andthe other with insertion of RARA intot h e P M L l o c u s ( i n s [ 1 5 ; 1 7 ] )(FIGURE 8). Both cases (as well as theproband) had RT-PCR confirmationof a PML-RARA fusion (bcr1 andbcr3 isoforms), and all had featurestypical of APL (TABLE).
Figure 4. Ins(17;15) and Resulting PML-RARA Fusion Identified by Whole-Genome Sequencing
Breakpoints in chromosomes 15 and 17 resulting in PML-RARA fusionA
Binding sites of primers for validation of 77-kb insertion from chromosome 15 to chromosome 17B
A and B, Schematic representation of ins(17;15) identified by whole-genome sequencing and resulting in PML-RARA fusion. Breakpoints (blue) are 72027045 and72104113 base pairs (bp) (chromosome 15) and 35742679 and 35742683 bp (chromosome 17), using NCBI36/hg18 build. Consistent with other translocations as-sociated with acute promyelocytic leukemia, the RARA breakpoint occurs before the DBD (DNA-binding domain) and preserves this entire domain. The fusion PML-RARA thus retains all the major protein domains of both proteins. ATG indicates coding sequence start; CC, coiled-coil domain; Kb, kilobase; ZF, zinc finger domain.Arrows indicate binding sites of primers used in Figure 5A.
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
COMMENTThe usefulness of massively parallelDNA sequencing has improved con-siderably with the introduction ofpaired-end reads, which allow for bet-ter mapping efficiency and more accu-rate identification of junctional break-points associated with structuralvariants (translocations, insertions, anddeletions).
In this report, we describe the use ofpaired-end–read whole-genome se-quencing for real-time oncologic diag-nosis and describe the genomic detailsof an oncogenic fusion gene created byan insertional event. Within 7 weeks,we completed the process of library gen-eration, massively parallel sequenc-ing, analysis, and validation of a novelinsertional fusion that created a clas-sic PML-RARA bcr3 variant. These find-ings altered the medical care of this pa-tient, who received ATRA consolidationinstead of an allogeneic stem cell trans-plant. The patient remains in first re-mission 15 months after her presenta-tion.
RT-PCR was not performed at the re-ferring institution, and results werenegative when evaluated at Washing-ton University when the patient was inremission. We did not initially per-form RT-PCR with the RNA generatedfrom the cryopreserved sample, be-
cause the RNA was severely degraded.Further, because this patient’s com-plex cytogenetics predicted an unfa-vorable prognosis, it was essential to de-termine whether the patient had arecognized PML-RARA fusion gene (be-cause t(15;17) supersedes other cyto-genetic findings and predicts a favor-able outcome in patients treated withATRA). Fortunately, the sequencing ofthe patient’s tumor genome resolved theconundrum, allowing the RT-PCR re-sults to serve as confirmatory proof ofthe fusion event (despite RNA degra-dation) and providing a novel mecha-nism for its formation, which assuredus that the diagnosis was correct.
Alternative laboratory approachescould have been used to detect poten-tial pathogenic RARA rearrangements(eg, nested PCR from a linker-ligatedlibrary, long-distance PCR,21 BAC clonescreening, targeted 3730 sequencing ofthe 48.5 kb RARA locus). However,such techniques are labor intensive andcan require large amounts of startingmaterial, personalized design, and it-erative troubleshooting. Moreover,many of these techniques have suc-cess rates not adequate for clinical prac-tice. In contrast, whole-genome se-quencing with paired-end libraries canbe accomplished with as little as 10 ngof starting DNA, is amenable to an au-
tomated “pipeline” strategy, and canconsistently deconvolute SNVs, smallinsertions and deletions, structural vari-ants, and clonality. Further, this ap-proach requires no custom reagents andno foreknowledge of genomic regionsthat must be assessed for diagnostic ac-curacy.
This study also confirms that recip-rocal RARA-PML fusions are not re-
Figure 5. PCR and Reverse Transcription (RT) PCR Validation of PML-RARA Expression From Case Patient
Tumor Control BlankDNAladder
bcr3PML-RARA
PCR validation of genomic DNAA RNA RT-PCR validation of PML-RARAexpression
B
Validation of t(17;15) insertion Validation of cryptic deletions
DNAladder
DNAladder
P1/P2(RARA-LOXL1)
P3/P4(PML-RARA) del(12) del(14)
P5/P6(LOXL1-PML) del(19)
N = Normal T = Tumor
bcr3PML-RARA
A, Polymerase chain reaction (PCR) of genomic DNA from the case patient’s skin (normal) and leukemia (tumor) using primers that span the junction of RARA-LOXL1(P1/P2), PML-RARA (P3/P4), del(12), del(14), LOXL1-PML (P5/P6), and del(19). Note amplification across fusion breakpoints in the leukemia sample but not in the skinsample for all but del(19). DNA ladder: 2176, 1766, 1230, 1033, 653, 517, 453, 394, 298, 234, and 154 base-pairs (bp). B, RNA prepared from cryopreserved leukemia cellsfrom case patient was amplified with a forward primer in PML exon 3 and a reverse primer in RARA exon 3. DNA ladder: 1500, 800, 500, 300, 200, 150, 100, and 50 bp.
Figure 6. Single-Nucleotide Variants in theTotal Bone Marrow Population of CasePatient
Ann
otat
ed g
enes
with
sing
le-n
ucle
otid
e va
riant
s
Percentage of each allele withleukemic somatic single-nucleotide variant
0 20 40 60
PITPNM1
SLC35A4
DYTN
ZNF687
PCSK2
PTK2
SH3D19
GPRC6A
C3orf54
CDC45L
DEGS2
ZFHX4
Single-nucleotide variants (SNVs) that occur in pro-tein coding sequences of the case patient in total bonemarrow cells. Two clones are identified based on the2 clusters of SNV frequency (clone 1: variant allele fre-quency, 35% to 51%; clone 2: variant allele fre-quency, 13% to 21% [see text]). This pattern is con-sistent with the cytogenetic findings of 2 geneticallydistinct clones (Figure 2).
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
quired for the development of APL.RARA-PML has been proposed to par-ticipate in APL pathogenesis, al-though it is identified in only approxi-mately 67% of APL cases by RT-PCR.22
Complex rearrangements and dele-tions can lead to unusual RARA-PMLtranscripts of uncertain signifi-cance.22-25 Furthermore, bcr3 RARA-PML does not independently lead toleukemia in a murine leukemiamodel.26,27 In this patient, the recipro-cal fusion RARA-PML was absent, andthe alternative RARA-LOXL1 was fusedout of frame.
Whole-genome sequencing re-solved not only the PML-RARA inser-tion event but also all other abnormali-ties observed during routine cytogenetictesting. Loss of chromosome 6 and16 were not detected with whole-genome sequencing; rather, we identi-fied an additional inversion and trans-location, inv(6)(p22.3;q14.1) and t(6;16)(q22.31;p13.3). This suggests thatgenetic information on chromosomes6 and 16 was actually present within the2 marker chromosomes but that chro-mosomal banding patterns were dis-rupted by a 3-way translocation.
FISH has been used to suggest thatinsertional translocations may occur int(15;17)–negative promyelocytic leu-kemia.2,22,28-30 The commercially avail-able dual-fusion dual-probe strategy
Figure 7. Fosmid Selection and Position of Homology to Chromosomes 15 and 17
Schematic representation of fosmid homology to the STROML1/PML locus and to the CDC6/RARA locus. ATG indicates coding sequence start; CC, coiled-coil do-main; Kb, kilobase; ZF, zinc finger domain. bcr1, bcr2, and bcr3 indicate breakpoints in PML associated with different-sized PML-RARA fusion transcripts.
Figure 8. Interphase and Metaphase Fluorescence In Situ Hybridization (FISH) of BoneMarrow Aspirate Cells of Additional Patients
A B
17
17
Fusion on der(17)
Fusion
Interphase FISH Metaphase FISH
15
15
C D
17
17
17
17
Fusion
Interphase FISH Metaphase FISH
Fusion on der(15)
1515
15
15
Fusion events in additional patients 1 (A and B) and 2 (C and D). A, One fusion signal, 2 red signals (chromo-somes 15), and 1 green signal (chromosome 17). B, Fusion event on der(17), consistent with ins(17;15). C,One fusion signal, 2 green signals (chromosomes 17), and 1 red signal (chromosome 15). D, Fusion event onder(15). Probes labeled with Spectrum Orange (STROML1/PML) and Spectrum Green (CDC6/RARA).
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
(Abbott/Vysis, Abbott Park, Illinois)uses large probes (between 239 and 417kb, schematically described in eFig-ure 3A and B). However, the minimalrequired PML insertion region (the pro-moter/enhancer and exons 1-3) isnearly one-tenth the size of theseprobes. The large probes (239� kb) im-prove sensitivity for conventional t(15;17) detection, but they make accuratediagnosis of small insertional events dif-ficult or even impossible.2,31 Alterna-tive cosmid-based strategies improvethe ability to detect small insertions,2
but the reagents are not widely avail-able. Because of these issues, we de-signed a new set of fosmid-basedprobes, which are publicly available.Using these probes, we identified fur-ther cases of PML-RARA insertional fu-sions missed by conventional cytoge-netics and FISH.
Insertional translocations are likelyto be underdiagnosed because of thetechnical difficulties associated withFISH using conventional probes (ashighlighted by this case). This may beespecially true of tumor types interro-gated with FISH break-apart strategiesand tumors that lack an alternative mo-lecular diagnostic assay (eg, Burkittlymphoma).32
Diagnostic whole-genome sequenc-ing remains cost-prohibitive for uni-versal application in patients with can-cer (approximately $40 000 for eachtumor/normal pair at the current time).This price has been decreasing rapidlyover the last several years, while the ex-pansion of AML-associated genes andmutations (eg, PML-RARA, ETO-AML1, CBFB-MYH11, BCR-ABL, FLT3,NPM1, KRAS, DNMT3A, TET2, IDH1/2,RUNX1, CEBPA) is increasing the cu-mulative cost of molecular tests thatmay be relevant for diagnosis and risk-prediction.
Acute promyelocytic leukemia islikely to be an indicator of the varietyof mutations that can present with simi-lar morphology and clinical course. APLmay be associated with diverse PML-RARA splice variants, unusual translo-cations/insertions, multiple RARA fu-sion partners, and even a RARG fusion
partner.1,8-12 The diagnostic strength ofwhole-genome sequencing is that it is ageneric and stable platform for detec-tion of mutations, and special ap-proaches are not required for specific di-agnostic settings. All classes of mutationsare detected in a totally unbiased fash-ion, allowing for confirmation of a sus-pected diagnosis, even if caused by a rareor unusual mutation; these data can beobtained and interpreted in a clinicallyrelevant time frame.
The time required to completewhole-genome sequencing is rapidlydecreasing. The sequencing timeline forthis patient involved 25 days to gener-ate and analyze whole-genome sequenc-ing and an additional 27 days for or-thogonal validation; this makesvalidation a significant but necessarybottleneck in the overall time to com-plete clinical-grade sequencing (eFig-ure 4). As costs continue to decline, in-creasingly deep coverage will becomepossible (eg, 60� coverage instead of30�), allowing for improved variant de-tection that may reduce the need forvalidation. With these improvements,clinical-grade whole-genome sequenc-ing should soon be possible within 4weeks of sample collection.
Meaningful diagnostic time framesare dependent on the cancer type beingassessed. Some cancers (eg, follicularlymphoma and myelodysplastic syn-drome) have long diagnostic windowsbefore patients need therapy; others (eg,breast, lung, colon) have moderatelylong diagnostic windows (those inwhich definitive chemoradiotherapyoften is not considered until after sur-gical resection and appropriate heal-ing), and a few (eg, AML, acute lym-phoblastic leukemia [ALL], Burkittlymphoma, blast-phase chronic my-elogenous leukemia, small-cell lungcancer) require urgent chemotherapy.The critical decision in the treatmentin AML and ALL is not which induc-tion therapy to use (because uniformapproaches to remission-induction arecurrently used for both diseases) butwhether patients should receive con-solidation therapy with chemothera-peutic approaches or should receive al-logeneic transplantation. This decisionis generally made within 6 to 8 weeksof initial presentation; for AML andALL, a 6-week time frame for whole-genome sequencing is therefore clini-cally relevant. However, to fully use thispotentially transformative technology
Table. Clinical and Pathologic Features of Acute Promyelocytic Leukemia in 3 Patients WithInsertional PML-RARA
FeatureCase
Patient
Additional Patients
1 2
Insertion ins(17;15) ins(17;15) ins(15;17)
Clinical featuresAge, y 39 33 30
Sex Female Female Male
WBC count, cells/µL 1300 400 700
Platelets, �103/µL 72 30 17
Hemoglobin, g/dL 11.6 9.8 8.0
Peripheral blasts, % 4 14 37
Bone marrowblasts/promyelocytes, %
61 100 100
Flow cytometryimmunophenotype
CD34 − − −
HLA-DR ND − −
CD117 ND � �
CD33 � � �
CD13 � � −
CD56 ND − �
Abbreviations: ND, not determined; WBC, white blood cell.
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
Downloaded From: http://jama.jamanetwork.com/ on 02/25/2013
to make informed clinical decisions,standards will have to be developed thatallow for CLIA-CAP certification ofwhole-genome sequencing and for di-rect reporting of relevant results to treat-ing physicians.Author Contributions: Drs Ding and Larson had fullaccess to all of the data in the study and take respon-sibility for the integrity of the data and the accuracyof the data analysis.Study concept and design: Welch, Westervelt, Ding,Larson, Klco, Watson, Tomasson, Link, Graubert,DiPersio, Mardis, Ley, Wilson.Acquisition of data: Klco, Kulkarni, Fulton, Veizer,Vickery, Heath, Watson, Wilson.Analysis and interpretation of data: Welch, Westervelt,Ding, Larson, Klco, Kulkarni, Wallis, Chen, Payton,Fulton, Veizer, Schmidt, Vickery, Mardis, Ley,Wilson.Drafting of the manuscript: Welch, Westervelt, Ley.
Critical revision of the manuscript for importantintellectual content: Welch, Westervelt, Ding,Larson, Klco, Kulkarni, Wallis, Chen, Payton,Fulton, Veizer, Schmidt, Vickery, Heath, Watson,Tomasson, Link, Graubert, DiPersio, Mardis, Ley,Wilson.Statistical analysis: Ding, Larson, Wallis, Chen, Schmidt,Vickery.Obtained funding: Ley, Wilson.Administrative, technical, or material support: Welch,Klco, Fulton, Vickery, Heath, Watson, Graubert, Ley,Wilson.Sequence analysis: Ding, Larson, Wallis, Chen, Schmidt,Vickery.Drs Welch and Westervelt contributed equally to thiswork.Conflict of Interest Disclosures: All authors have com-pleted and submitted the ICMJE Form for Disclosureof Potential Conflicts of Interest. Dr Westervelt re-ported receiving lecture fees from Celgene and Novar-tis; Dr DiPersio reported receiving consulting and lec-ture fees from Genzyme. No other authors reporteddisclosures.
Funding/Support: This study was supported by theWashington University Cancer Genome Initiative, byNational Institutes of Health (NIH) grant PO1CA101937, by Barnes-Jewish Hospital Foundationgrant 00335-0505-02 (Dr Ley), and by NIH grant U54HG003079 (Dr Wilson). Dr Welch is a fellow of theLeukemia Lymphoma Society.Role of the Sponsor: The funding organizations hadno role in design and conduct of the study; the col-lection, management, analysis, and interpretation ofthe data; or the preparation, review, or approval ofthe manuscript.Online-Only Material: The eMethods and eFigures1-4 are available at http://www.jama.com.Additional Contributions: We thank the Washing-ton University Cancer Genomics Initiative for its sup-port and Charles W. Caldwell, MD, PhD, and CarlFreter, MD (both at University of Missouri-ColumbiaSchool of Medicine), for referring the patient and forcollecting, storing, and contributing the cryopre-served bone marrow cells used for this study. Neitherof these individuals received compensation for theircontributions.
REFERENCES
1. Jurcic JG, Soignet SL, Maslak AP. Diagnosis andtreatment of acute promyelocytic leukemia. Curr On-col Rep. 2007;9(5):337-344.2. Grimwade D, Biondi A, Mozziconacci MJ, et al;Groupe Francais de Cytogenetique Hematologique,Groupe de Francais d’Hematologie Cellulaire, UK Can-cer Cytogenetics Group and BIOMED 1 EuropeanCommunity-Concerted Action “Molecular Cytoge-netic Diagnosis in Haematological Malignancies.”Characterization of acute promyelocytic leukemia caseslacking the classic t(15;17): results of the EuropeanWorking Party. Blood. 2000;96(4):1297-1308.3. Abe S, Ishikawa I, Harigae H, Sugawara T. A newcomplex translocation t(5;17;15)(q11;q21;q22) in acutepromyelocytic leukemia. Cancer Genet Cytogenet.2008;184(1):44-47.4. Hiorns LR, Swansbury GJ, Catovsky D. An eight-way variant t(15;17) in acute promyelocytic leuke-mia elucidated using fluorescence in situ hybridization.Cancer Genet Cytogenet. 1995;83(2):136-139.5. Pandolfi PP, Alcalay M, Fagioli M, et al. Genomicvariability and alternative splicing generate multiplePML/RAR alpha transcripts that encode aberrant PMLproteins and PML/RAR alpha isoforms in acute pro-myelocytic leukaemia. EMBO J. 1992;11(4):1397-1407.6. Park TS, Kim JS, Song J, et al. Acute promyelo-cytic leukemia with insertion of PML exon 7a and par-tial deletion of exon 3 of RARA: a novel variant tran-script related to aggressive course and not detectedwith real-time polymerase chain reaction analysis. Can-cer Genet Cytogenet. 2009;188(2):103-107.7. Reiter A, Saussele S, Grimwade D, et al. Genomicanatomy of the specific reciprocal translocation t(15;17) in acute promyelocytic leukemia. Genes Chromo-somes Cancer. 2003;36(2):175-188.8. Dong S, Tweardy DJ. Interactions of STAT5b-RARalpha, a novel acute promyelocytic leukemia fu-sion protein, with retinoic acid receptor andSTAT3 signaling pathways. Blood. 2002;99(8):2637-2646.9. Yamamoto Y, Tsuzuki S, Tsuzuki M, Handa K,Inaguma Y, Emi N. BCOR as a novel fusion partnerof retinoic acid receptor alpha in a t(X;17)(p11;q12)variant of acute promyelocytic leukemia. Blood. 2010;116(20):4274-4283.10. Catalano A, Dawson MA, Somana K, et al. ThePRKAR1A gene is fused to RARA in a new variant acutepromyelocytic leukemia. Blood. 2007;110(12):4073-4076.
11. Kondo T, Mori A, Darmanin S, Hashino S, TanakaJ, Asaka M. The seventh pathogenic fusion geneFIP1L1-RARA was isolated from a t(4;17)-positive acutepromyelocytic leukemia. Haematologica. 2008;93(9):1414-1416.12. Such E, Cervera J, Valencia A, et al. A novel NUP98/RARG gene fusion in acute myeloid leukemia resem-bling acute promyelocytic leukemia. Blood. 2011;117(1):242-245.13. Tallman MS, Andersen JW, Schiffer CA, et al. All-trans retinoic acid in acute promyelocytic leukemia:long-term outcome and prognostic factor analysis fromthe North American Intergroup protocol. Blood. 2002;100(13):4298-4302.14. Hill BT, Copelan EA. Acute myeloid leukemia: whento transplant in first complete remission. Curr Hema-tol Malig Rep. 2010;5(2):101-108.15. Burnett AK, Hills RK, Green C, et al. The impacton outcome of the addition of all-trans retinoic acidto intensive chemotherapy in younger patients withnonacute promyelocytic acute myeloid leukemia: over-all results and results in genotypic subgroups definedby mutations in NPM1, FLT3, and CEBPA. Blood. 2010;115(5):948-956.16. Powell BL, Moser B, Stock W, et al. Arsenic tri-oxide improves event-free and overall survival for adultswith acute promyelocytic leukemia: North AmericanLeukemia Intergroup Study C9710. Blood. 2010;116(19):3751-3757.17. Ley TJ, Mardis ER, Ding L, et al. DNA sequenc-ing of a cytogenetically normal acute myeloid leukae-mia genome. Nature. 2008;456(7218):66-72.18. Mardis ER, Ding L, Dooling DJ, et al. Recurringmutations found by sequencing an acute myeloid leu-kemia genome. N Engl J Med. 2009;361(11):1058-1066.19. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mu-tations in acute myeloid leukemia. N Engl J Med. 2010;363(25):2424-2433.20. Chen K, Wallis JW, McLellan MD, et al. Break-Dancer: an algorithm for high-resolution mapping ofgenomic structural variation. Nat Methods. 2009;6(9):677-681.21. Kim MJ, Cho SY, Kim MH, et al. FISH-negativecryptic PML-RARA rearrangement detected by long-distance polymerase chain reaction and sequencinganalyses: a case study and review of the literature. Can-cer Genet Cytogenet. 2010;203(2):278-283.22. Lafage-Pochitaloff M, Alcalay M, Brunel V, et al.Acute promyelocytic leukemia cases with nonrecip-
rocal PML/RARa or RARa/PML fusion genes. Blood.1995;85(5):1169-1174.23. Borrow J, Goddard AD, Sheer D, Solomon E. Mo-lecular analysis of acute promyelocytic leukemia break-point cluster region on chromosome 17. Science. 1990;249(4976):1577-1580.24. Walz C, Grimwade D, Saussele S, et al. AtypicalmRNA fusions in PML-RARA positive, RARA-PMLnegative acute promyelocytic leukemia. Genes Chro-mosomes Cancer. 2010;49(5):471-479.25. Alcalay M, Zangrilli D, Fagioli M, et al. Expres-sion pattern of the RAR alpha-PML fusion gene in acutepromyelocytic leukemia. Proc Natl Acad Sci U S A.1992;89(11):4840-4844.26. Pollock JL, Westervelt P, Kurichety AK, Pelicci PG,Grisolano JL, Ley TJ. A bcr-3 isoform of RARalpha-PML potentiates the development of PML-RARalpha-driven acute promyelocytic leukemia. Proc Natl AcadSci U S A. 1999;96(26):15103-15108.27. Walter MJ, Ries RE, Armstrong JR, Park JS, MardisER, Ley TJ. Expression of a bcr-1 isoform of RARalpha-PML does not affect the penetrance of acute promy-elocytic leukemia or the acquisition of an interstitialdeletion on mouse chromosome 2. Blood. 2007;109(3):1237-1240.28. Goldschmidt N, Yehuda-Gafni O, Abeliovich D,Slyusarevsky E, Rund D. Interstitial insertion of RAR�gene into PML gene in a patient with acute promy-elocytic leukemia (APL) lacking the classic t(15;17).Hematology. 2010;15(5):332-337.29. Hiorns LR, Min T, Swansbury GJ, Zelent A, DyerMJ, Catovsky D. Interstitial insertion of retinoic acidreceptor-alpha gene in acute promyelocytic leuke-mia with normal chromosomes 15 and 17. Blood. 1994;83(10):2946-2951.30. Grimwade D, Gorman P, Duprez E, et al. Char-acterization of cryptic rearrangements and varianttranslocations in acute promyelocytic leukemia. Blood.1997;90(12):4876-4885.31. Sanz MA, Grimwade D, Tallman MS, et al. Man-agement of acute promyelocytic leukemia: recom-mendations from an expert panel on behalf of theEuropean LeukemiaNet. Blood. 2009;113(9):1875-1891.32. May PC, Foot N, Dunn R, Geoghegan H, NeatMJ. Detection of cryptic and variant IGH-MYC rear-rangements in high-grade non-Hodgkin’s lym-phoma by fluorescence in situ hybridization: implica-tions for cytogenetic testing. Cancer Genet Cytogenet.2010;198(1):71-75.
WHOLE-GENOME SEQUENCING TO DIAGNOSE CRYPTIC ONCOGENE
