MYC-associated and double-hit lymphomas: A review of pathobiology, prognosis, and therapeutic approaches

Authors


Abstract

Aberrant expression of the v-myc avian myelocytomatosis viral oncogene homolog (MYC) proto-oncogene has known transformative potential in healthy human cells. Chromosomal MYC rearrangements and consequent MYC overexpression is the defining lesion in Burkitt lymphoma (BL), conferring a highly proliferative state. However, abnormalities of MYC are increasingly appreciated in non-BL histologies, including diffuse large B-cell lymphoma (DLBCL) and B-cell lymphomas intermediate between BL and DLBCL, with a particularly aggressive clinical phenotype. Although there are conflicting data regarding prognostic implications of isolated MYC aberrancy in these non-BLs, the co-occurrence of MYC rearrangements and either the antiapoptotic gene B-cell chronic lymphocytic leukemia/lymphoma 2 (BCL2) or the transcriptional repressor BCL6 leads to an entity termed double-hit B-cell lymphoma (DHL) (or triple-hit if all 3 abnormalities are observed) with a particularly poor prognosis and no established treatment paradigms. Notably, a distinct pattern of gene expression profiling has been noted when MYC is overexpressed in BL compared with other lymphomas, supporting the notion that, although MYC promotes target gene transcription, the target genes vary by disease subtype. The frequency of MYC activity depends on the method of detection and ranges from 5% to 10% using fluorescence in situ hybridization but up to 30% of DLBCL using immunohistochemistry. Standard therapies developed for DLBCL are less effective when the disease is driven by MYC, leading to lower response rates and response durations. An important clinical challenge is to pre-emptively identify MYC-associated lymphomas and to subsequently develop trials specifically for this group of patients. However, the design of such studies is complicated by variable definitions of MYC-associated lymphoid malignancies and the lack of effective therapies to date. The objective if the current review was to evaluate the implications of MYC aberrancy with respect to the B-cell lymphoma double-hit and triple-hit phenotypes and to consider the available data for clinical and practical management. Cancer 2014;120:3884–3895. © 2014 American Cancer Society.

INTRODUCTION

Lymphomas are a heterogeneous group of >60 mature malignancies of B-cells and T-cells.[1] In addition to the pathologic subtype, several biologic events contribute to clinical behavior and occasionally drive management options. Among these biologic events are complex genetic, epigenetic, and proteomic aberrations that can be detected by a variety of technologies. It is now known that abnormalities of the v-myc avian myelocytomatosis viral oncogene homolog (MYC) proto-oncogene, classically described in Burkitt lymphoma (BL), occur in other lymphoma subtypes more frequently than previously appreciated and confer an adverse prognosis. When MYC abnormalities coincide with other poor prognostic markers, a new genre of lymphomas, defined more by biologic features than by histologic features, emerge. These lymphomas, which classically include MYC along with chronic lymphocytic leukemia/lymphoma 2 (BCL2) rearrangements, have been dubbed double-hit lymphomas (DHLs) and are an emerging therapeutic challenge. It is noteworthy that DHLs are not restricted to any 1 histologic subtype and can be present in diffuse large B-cell lymphoma (DLBCL), B-cell lymphoma unclassifiable (BCLU), follicular lymphoma (particularly in transformed cases), and mantle cell lymphoma (MCL), among others. The current definition of DHL in the most recent World Health Organization classification is the co-occurrence of MYC and BCL2 rearrangement as detected by fluorescence in situ hybridization (FISH) or standard cytogenetics. However, MYC can also be amplified, mutated, or otherwise overexpressed as detected by FISH, cytogenetics, comparative genome hybridization, or immunohistochemistry (IHC). Similarly, BCL2 and BCL6 abnormalities can be detected with conventional cytogenetics, interphase FISH, or IHC. The clinical impact of the means of detection (ie, gene rearrangement vs protein overexpression) is a controversial topic, and the current data are addressed below. In this review, we discuss the role of MYC in normal and malignant lymphocytes, the various means of MYC detection, the most common partner rearrangements and mutations, and clinical considerations.

Normal and Transforming Functions of MYC

MYC is a nuclear phosphoprotein whose name is derived from the avian malignancy myelocytomatosis, a condition caused by the viral gene (v-MYC), which was cloned and sequenced in 1983.[2] The homologous human counterpart of v-MYC is termed cellular MYC (c-MYC) (herein referred to as MYC), and localizes to chromosome 8, where its rearrangement is a hallmark of BL.[3] MYC operates as a transcription factor capable of either promoting or repressing the expression of a vast array of genes, together accounting for at least 10% of the human genome.[4] In general, genes targeted by MYC include mediators of metabolism, biosynthesis, and cell cycle progression,[5] such that aberrant MYC expression is associated with uncontrolled cell growth, division, and metastasis,[6] whereas loss or inhibition of MYC expression impedes growth, promotes differentiation, and sensitizes cells to DNA damage.[7] Specific gene targets of MYC generally have been identified by following messenger RNA (mRNA) levels after experimentally controlled activation of the c-MYC gene.[8] These targets have been extensively reviewed.[4] Some of the most biologically important targets are thought to be cyclins and cyclin-dependent kinases (CDKs), resulting in accelerated cell cycling[9]; down-regulation of phosphatase and tensin homolog (PTEN) with consequent up-regulation of the phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway[10]; and stabilization of the proapoptotic protein and tumor suppressor p53,[11] which can bypass the apoptotic BCL2 program.[12] Additional transcriptional targets include a large number of microRNAs (miRNAs), including both tumor suppressor and oncogenic miRNAs.[13, 14] Given the emerging pathogenetic role of miRNAs in the methylation status of cancer cells, this link between the regulatory function of MYC over miRNAs is particularly noteworthy. This transformative capacity of MYC is often in concert with other oncogenes and viruses, including the rat sarcoma (RAS) oncogene[15] and Epstein-Barr virus (EBV).[16] This latter correlation between EBV and MYC is complex, because EBV-associated proteins in turn potentiate MYC activity.[17]

Although the transcriptional role of MYC has been well described, there are also nontranscriptional functions that only recently have been appreciated, including regulation of mRNA translation and direct regulation of DNA replication (for a review, see Cole and Cowling[18]). For instance, MYC directly promotes methyl cap formation on the 5′ end of pre-mRNA for many genes, including cyclin D1 and CDK-9, and resulting in enhanced mRNA translation.[19] Recent preclinical models have further demonstrated that MYC binds directly to, and promotes the activity of, replication complexes responsible for DNA synthesis, such that MYC levels correspond with the degree of DNA replication.[20]

MYC Overexpression in Non-Burkitt B-Cell Lymphoma

The constellation of MYC effects on genes involved in proliferation has led to the concept of MYC-driven lymphomas, which have been described in both laboratory[21] and clinical[22, 23] studies. The classic MYC-driven lymphoma is BL, in which balanced rearrangements between chromosome 8, where MYC resides, and either chromosome 14 (immunoglobulin [Ig] heavy chain), chromosome 22 (IgG lambda light chain), or chromosome 2 (IgG kappa light chain), lead to a highly proliferative lymphoid malignancy with a propensity for extranodal involvement, particularly in immunocompromised patients.

The non-BL subtype of lymphoma most commonly associated with MYC is DLBCL, and, by extension, lymphomas that share morphologic features of DLBCL and BL, officially termed B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and BL (BCLU). When DLBCL cases harbor MYC aberrations, a distinct gene expression profile (GEP) is observed. Dave and colleagues reported that MYC abnormalities in BL tend to be simple and control several genes involved in proliferation.[24] In contrast, MYC abnormalities in DLBCL-targeted genes are associated with nuclear factor κB (NF-κB) signaling and antiapoptotic cascades. Those authors demonstrated that a GEP-based molecular profile was capable of accurately identifying 100% of BLs, as previously identified by a combination of IHC and FISH, in patients who had inferior outcomes when treated with DLBCL regimens such as rituximab with cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). In addition, they identified 9 of 53 DLBCLs (17%) that had a BL-like molecular profile, which otherwise would not have been identified as BL-like by standard measures.

The relative frequency of MYC abnormalities in non-BL lymphomas depends on the method of detection, and this has complicated both the definition and impact of the terms MYC-associated lymphomas and DHL. In clinical settings, MYC aberrations can be detected by standard cytogenetics, interphase FISH, comparative genomic hybridization, and, most recently IHC. Early reports evaluating MYC aberrancy using FISH or cytogenetics indicated an incidence between 2% and 17% in unselected, sequential patients with DLBCL.[25-27] A Canadian study constructed tissue microarrays and observed that 12 of 135 patients (8.8%) had detectable MYC using FISH; when comparing clinical characteristics between the 2 groups, there were no distinguishing clinical features, but patient samples that were positive for MYC had significantly higher proliferation rates.[26] In contrast, Japanese investigators observed a relatively similar incidence (11% MYC-positive using FISH) but also noted that MYC-positive patients had a worse performance status, higher baseline lactate dehydrogenase (LDH) levels, and increased bone marrow involvement.[25] It is worth noting that none of the patients in that series received rituximab, whereas all patients in the Canadian series received treatment with R-CHOP.

The relatively recent availability of a reliable commercial IHC stain for MYC protein offers a less costly and less laborious means of detecting MYC overexpression and represents a key step forward in studying MYC-associated lymphomas (Fig. 1A-C). With IHC, the frequency of MYC protein overexpression appears much higher than what is detected by FISH or standard cytogenetics. For example, a retrospective evaluation of 442 patients with de novo DLBCL enrolled in the German RICOVER study indicated that the frequency of MYC translocations versus IHC overexpression was 8.8% versus 31.8%, respectively.[28] Similarly, an international collaboration reported that only 11% of 167 patients with DLBCL had MYC translocations when using FISH, whereas IHC indicated that 29% of patients were positive for MYC protein expression.[29] Those authors also performed a validation study in an additional 140 patients and reported MYC translocations in 13% versus 37% of patients with low versus high MYC protein expression, respectively. It is noteworthy that the definition of MYC positivity by IHC is not universally standardized; however, the sum of the literature supports that having at least 40% of malignant lymphocytes with nuclear MYC expression is considered positive. Whether or not higher levels of positivity confer worsening prognosis is not clear.

Figure 1.

Images show representative immunohistochemical staining for v-myc avian myelocytomatosis viral oncogene homolog (MYC) and MYC abnormalities identified by fluorescence in situ hybridization (FISH). Double-hit lymphoma biopsy specimens are represented by (A) hematoxylin and eosin staining, (B) MYC immunohistochemical staining, and (C) CD20 immunohistochemical staining. (D) Results from conventional cytogenetics (karyotype) in a double-hit lymphoma specimen reveal an abnormal 8;14 translocation (arrows). FISH analyses demonstrate (E) an abnormal fusion probe (overlapping signals indicate abnormal translocation and approximation of MYC locus on chromosome 8 with immunoglobulin heavy-chain [IgH] locus on chromosome 14). LSI indicates locus-specific identifier, (F) an abnormal break-apart probe (arrows indicate the loss of a contiguous pattern of staining on opposite sides of the MYC breakpoint on chromosome 8), and (G) normal (nontranslocated) results.

MYC can also be amplified with increased copy numbers observed using break-apart probes (Fig. 1F). It has been demonstrated that MYC gain of copy number (GOC) correlates with increased levels of MYC RNA[30] and the corresponding MYC protein[31] and portends poor prognoses in patients with malignancies like melanoma,[31] prostate cancer,[32] and medulloblastoma,[33] among others. Despite the clear importance of MYC rearrangement in B-cell non-Hodgkin lymphoma (NHL), the role of MYC GOC in lymphoma is less well studied, and the studies evaluating its prognostic significance have yielded conflicting results. For instance, a study of 145 patients with DLBCL indicated that MYC GOC occurred with a frequency (7%) on par with that of MYC rearrangement (9%), and both types of MYC aberrations portended similarly poor outcomes when patients received R-CHOP and similar regimens.[34] Conversely, a study of 166 patients with DLBCL who received R-CHOP demonstrated that MYC GOC occurred with similar frequency (10%), was never a sole chromosomal aberration, and was not associated with inferior outcomes, except in patients who also carried deletion of the short arm of chromosome 8 (del 8p),[35] which has been described in other tumor types.[36] In addition, MYC GOC can be confused or misinterpreted if there is a global increase in DNA ploidy or if there is a complex karyotype with multiple numerical and/or structural chromosomal changes. The MYC complex karyotype is in contrast to the MYC simple karyotype observed in BL.

A more recent study by Valera and colleagues distinguished between patients who had GOC (3-4 copies) and amplification (>4 copies) with respect to c-MYC in DLBCL and observed that only the latter seemed to have unfavorable outcomes, although the numbers were small.[37] Likewise, an evaluation of 16 patients with B-cell NHL who had c-MYC amplification (>4 copies) observed a median survival of less than 6 months;[38] it is noteworthy that, of 6 patients who had c-MYC amplification, 4 were dead within 9 months, and 2 had achieved only a partial response as their best response. In any case, currently, it remains unclear whether those patients who carry MYC GOC along with rearrangements of BCL2 and/or BCL6 at the time of original diagnosis should be included in the definition of DHL.

Double-Hit and Triple-Hit Lymphomas

The term double-hit lymphoma (DHL) originally referred to patients who had concurrent MYC and BCL2 or BCL6 gene rearrangements detected by FISH or standard cytogenetics.[1] BCL2 is an antiapoptotic family of genes and proteins that lend a strong survival signal to malignant cells. The combination of a potent proliferation signal (MYC) and a potent survival signal (BCL2) clinically leads to a very aggressive and chemoresistant phenotype, with rare long-term survivors[39] (Table 1). Patients with classic, FISH-defined DHL have a very poor prognosis. For example, 1 study reported that 54 of 1260 patients (4%) who harbored concurrent 18q21 and 8q24 translocations were more likely to have clinical risk factors, such as a high International Prognostic Index (IPI) and bone marrow involvement, and had a median survival <1 year.[39] Other early reports on FISH-defined DHL support these findings; nearly all demonstrated that approximately 33% of patients with DLBCL have long-term survival, which is significantly inferior to those who have DLBCL without MYC or BCL2 rearrangements.[26, 27, 54] Paralleling the above discussion regarding MYC, the frequency of BCL2 protein overexpression in DLBCL is much more common than translocations; for example, Horn and colleagues observed that only 13.5% of patients with DLBCL harbored a BCL2 rearrangement compared with nearly 80% displaying strong BCL2 protein overexpression.[28] Less frequently, MYC rearrangements occur concurrently with BCL6 rearrangements, which also results in a poor overall prognosis, although data for this type of DHL are less robust. When all 3 FISH or cytogenetically identified lesions are present (MYC, MYC, and MYC), the term triple-hit lymphoma (THL) has been applied. Existing data suggest that this is a rare phenomenon and carries a poor prognosis, although whether such patients fare worse than those with DHL is not known.[55, 56]

Table 1. Incidence of MYC Aberrancy and Double-Hit Lymphoma in Reported Series of Patients With Diffuse Large B-Cell Leukemia
ReferenceTotal No.No. MYC+ (%)No. With DHL (%DHL)aOutcome/Comments
FISH    
van Imhoff 2006[40]6610 (15)6 (9)The study demonstrated that GC had better outcomes than NGC in transplanted patients
Yoon 2008[34]15614 (9)5 (3.4)GOC was present in 7%; MYC translocation was more common in GC vs NGC and was associated with worse OS
Copie-Bergman 2009[41]682 (3)NRPatients with a positive FISH index defined as positivity for at least 2 of 3 NGC markers (MUM1/IRF4, forkhead box protein P1, BCL6 rearrangement) had worse OS (P=.01).
Tibiletti 2009[42]7412 (16)NRBCL2 present in 21%, BCL6 present in 45%
Obermann 2009[43]2209 (4)NRMedian OS was 9 mo for MYC+ vs 80 mo for those without MYC
Stasik 2010[30]521 (2)NRGOC present in 38%
Barrans 2010[2][7]24534 (14)24 (10)THL present in 3%; OS worse in MYC-rearranged patients (2-y OS, 35% vs 61%)
Tapia 2011[44]459 (20)NRMYC by IHC correlated with MYC translocation
Kluk 2012[22]565 (9)NR+IHC >50% for MYC in 19%; high MYC protein overexpression correlated with inferior OS when R-CHOP was used
Johnson 2012[2][9]29035 (12)15 (5)14% MYC by FISH in GC vs 9% in NGC
Green 2012[4][5]18921 (11)11 (6)High Bcl-2 and MYC by IHC was present in 29% and predicted lower RR, PFS, and OS
Cuccuini 2012[4][6]16127 (17)21 (13)4-y PFS (18% vs 42%) and OS (29% vs 62%) were lower in MYC+ relapsed patients undergoing HSCT
Valera 2013[3][7]21915 (7)9 (4)ICN in 2%; high MYC expression correlated with worse OS regardless of IPI
Horn 2013[2][8]44240 (9)35 (8)IHC or FISH/IHC scores predict outcome in DLBCL patients independent of the IPI
Caponetti 2013[47]17814 (7)NRMYC patients without BCL2 or BCL6 rearrangements have similar survival to MYC patients
Wang 2013[48]11541 (36)24 (21)The majority of MYC also had BCL2 rearrangements
Howlett 2013[4][9]280NR34 (12)Those able to undergo allo-SCT had significantly longer PFS than those not undergoing SCT
Tsai 2013[50]560NR34 (6)Escalated induction not associated with improved outcomes
Landsburg 2013[51]5326 (49)17 (32)Clinical features fail to predict presence of MYC rearrangements
Totals3469 (2629b/2776c)315 (12)235 (8) 
    IHC Cutoff, %
ReferenceTotal No.No. MYC+ (%)No. With DHL (%DHL)aMYC+BCL2+
  1. Abbreviations: −, negative; +, positive; allo-SCT, allogeneic stem cell transplantation; Bcl-2, B-cell chronic lymphocytic leukemia/lymphoma 2; Bcl-6, B-cell chronic lymphocytic leukemia/lymphoma 6; DHL, double-hit lymphoma; FISH, fluorescence in situ hybridization (by translocation unless otherwise specified); GC, germinal center; GOC, gain of copy number; HSCT, hematopoietic stem cell transplantation; ICN, increased copy number; IHC, immunohistochemistry; IPI, International Prognostic Index; IRF4, interferon regulatory factor 4; MUM1, melanoma-associated antigen (mutated) 1; MYC, V-myc myelocytomatosis viral oncogene homolog; NGC, nongerminal center; NR, not reported; OS, overall survival; PFS, progression-free survival; R-CHOP, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone; RR, response rate; SCT, stem cell transplantation; THL, triple-hit lymphoma.

  2. a

    The percentage DHL (%DHL) is given as a proportion of the total number of patients with diffuse large B-cell lymphoma (DLBCL) evaluated, not as a proportion of patients with MYC-aberrant DLBCL.

  3. b

    Values are the total number of patients included in the studies that reported rates of MYC rearrangement.

  4. c

    Values are the total number of patients included in the studies that reported rates of DHL.

IHC     
Kluk 2012[22]5610 (18)NR≥50NA
Johnson 2012[29]30789 (29)64 (21)≥4050
Green 2012[45]18593 (50)54 (29)≥4070
Valera 2013[37]21928 (13)NR≥40NA
Horn 2013[28]442141 (32)NR≥40%NA
Hu 2013[52]466300 (64)158 (34)≥4070
Molina 2013[53]577170 (30)121 (21)≥4070
Pooled analysis of MYC+ by IHC2252 (1535)c831 (37)397 (26)  

The observation that IHC routinely detects a greater proportion of both MYC and BCL2 overexpression compared with FISH techniques raises an important question: is expression of both MYC and BCL2 protein (as detected by IHC) prognostically relevant independent of genetic rearrangements? In fact, there is now a growing body of data supporting the notion that patients with this so-called double protein-expressing or, simply, double-expressing lymphoma have a poor prognosis on par with FISH-defined DHL; however, there is significant discordance on the degree of immunohistochemical positivity associated with adverse outcomes.[22, 28, 45, 52],[57, 58] Green and colleagues defined double-expressing lymphoma using the median percentage of MYC-positive cells (40%) and BCL2-positive cells (70%) cells in approximately 300 patients.[58] Twenty-nine percent of those patients were positive for both MYC and BCL2 protein; and, although there were no important clinical differences between FISH-defined DHL and double-expressing lymphoma, the outcomes were worse for patients with double-expressing lymphoma who received R-CHOP, with 3-year progression-free survival (PFS) and overall survival (OS) rates of 39% and 43%, respectively, versus 75% and 86%, respectively. Johnson and colleagues used a similar MYC IHC cutoff point (≥40%) but a lower BCL2 cutoff point (≥50%) but still reported 5-year PFS and OS rates of 32% and 36%, respectively. Perry and colleagues used an even lower cutoff point for BCL2 (30%) but a higher cutoff point for MYC (50%); however, they similarly reported that patients with double-expressing lymphoma had a 9-fold increased risk of death.[57] The largest series to date evaluating the outcomes of patients with double-expressing lymphoma who received treatment with R-CHOP evaluated 700 patients with de novo DLBCL and reported 5-year PFS and OS rates of 27% and 30%, respectively[52]; patients without double-expressing lymphoma in that study had superior 5-year PFS and OS rates of 73% and 75%, respectively (P < .0001). It is noteworthy that this large study confirmed earlier reports that neither BCL2 nor MYC protein overexpression independently confers an adverse prognosis; rather, it is the simultaneous presence of both risk factors that is central to imparting unfavorable risk. Further supporting the adverse significance of IHC-detected MYC is the up-regulation of MYC target genes, similar to patterns observed with genetic rearrangements.[22] These reports highlight the reality that there is currently no universally accepted cutoff point for either BCL2 or MYC positivity using IHC. Nevertheless, these reports consistently and firmly establish that double-expressing lymphoma is a clinically relevant entity with a poor prognosis.

Testing for MYC and DHL

An important practical question is whether or not all DLBCL specimens should be tested for MYC and BCL2 or whether there are clinical or other biologic clues predictive of DHL phenotype. Savage and colleagues evaluated clinical features and reported no statistically significant differences in median age, sex, performance status, LDH level, IPI score, BCL2 protein expression, or cell of origin (COO) among 8.8% of patients with DLBCL who harbored an MYC rearrangement and those lacking the MYC aberrancy. However, patients with MYC-aberrant DLBCL were significantly more likely to have a high Ki67 index: 58% of patients had a Ki67 index >80%.[26] In contrast, a Japanese study of 252 patients with DLBCL indicated that the 28 patients (11%) who had 8q24/c-MYC aberrations were more likely to have a poor performance status, an increased LDH level, and increased bone marrow involvement at presentation.[25] There also are some data suggesting that the partner gene in MYC rearrangement may be an important predictor of outcome, with patients who have disease involving Ig partners faring significantly worse than those with non-Ig partners[54]; thus, detection of this MYC rearrangement subtype may be warranted. Any MYC aberrancy, including double-expressing lymphoma and the DHL phenotype, correlate with high IPI scores.[26, 29, 37, 39] Despite these somewhat conflicting reports, the profound negative prognostic implications of harboring dual MYC and BCL2 aberrations suggests that most patients should be tested for DHL, at least using IHC, which is less costly and easier to perform. In the future, these findings may have management implications as well.

Outside of a brisk proliferation rate, there are few other biologic clues to underlying DHL. A seminal observation, now more than a decade old, is that the COO in DLBCL had an impact on clinical outcome in both the prerituximab and postrituximab eras independent of clinical features and IPI.[59-63] Patients with a germinal center (GCB) phenotype have superior outcomes compared with those with non-GCB (also termed activated B-cell like [ABC]) phenotypes. Several initial reports indicated that MYC rearrangement in DLBCL is strongly associated with CD10 positivity and GCB derivation.[28, 37] In 1 multicenter evaluation, 98% of FISH-defined DHLs were of GCB derivation.[64] However, the apparent strong association between GCB phenotype and MYC may be abrogated when IHC is used instead of FISH to define MYC-associated lymphomas. The largest report to date on DHL, which includes nearly 900 patients, indicated that dual MYC/BCL2 protein expression was actually higher in the non-GCB subset of patients. Some investigators have observed that outcomes for patients with double-expressing lymphoma are uniquely and relatively worse for those with GCB subtypes but not those with non-GCB DLBCL.[57] Another group similarly observed that patients with double-expressing lymphoma had an equally poor outcome independent of the COO; in the latter study, MYC/BCL2 coexpression actually trumped the COO in terms of prognostic utility.[52] Thus, it appears that the decision on whether or not to test for DHL or double-expressing lymphoma should be independent of the COO, because the presence of double hits and/or double-expressing lymphoma appears to be adversely prognostic in both subsets.

Management Considerations in Non-BL MYC-Associated Lymphoma and DHL

Less than 33% of patients with either DHL or double-expressing lymphoma achieve long-term survival with current therapy. In light of such poor outcomes, both intensified chemoimmunotherapy regimens and consolidative stem cell transplantation (SCT) have been evaluated, albeit all in retrospective settings and primarily reflecting FISH-defined DHL. In 1 of the first of such reports, Johnson and colleagues noted that, among 54 patients with double-expressing lymphoma, the 6 patients who received intensified induction, with or without autologous SCT (ASCT), fared at least as poorly as those who received R-CHOP.[39] In 2012, Abramson and colleagues reported their experience evaluating 34 patients with DHL according to FISH criteria[65] and noted seemingly favorable outcomes among patients who received an infusional dose-adjusted (DA) regimen of etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (EPOCH) (median OS, 34 months vs 8 months for those who received R-CHOP). Other retrospective studies have suggested that intensified induction, such as DA-EPOCH; hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (Hyper-CVAD); and cyclophosphamide, vincristine, doxorubicin, and cytarabine with methotrexate/etoposide, ifosfamide, and cytarabine (CODOX-M/IVAC), do not improve outcomes, at least in the relatively small patient numbers studied.[50, 66]

Two large retrospective series evaluating outcomes of patients with FISH-defined DHL were recently presented. Gandhi and colleagues reported on a series of 106 patients with DHL according to FISH criteria[56] with the aim of evaluating the roles of 1) intensification of frontline therapy, and 2) consolidative high-dose chemotherapy (HDT) plus ASCT (HDT-ASCT). No significant difference in survival was noted based on the type of induction (see Table 2), although the use of intensified backbones (DA-EPOCH, Hyper-CVAD, CODOX-M/IVAC) was associated with significantly higher rates of complete response (CR) compared with the use of a CHOP backbone. Among the patients in that study who achieved complete remission, there was a trend toward improved OS with HDT-ASCT, but this was not statistically significant (5-year OS rate, 92% vs 58%; P = .22). The authors concluded that intensification of therapy may be considered for the sake of improving the odds of achieving a CR. Similarly, Oki and colleagues noted in their single-center series of 56 patients[67] that complete remission rates in response to frontline DA-EPOCH (83%) and Hyper-CVAD (68%) were higher than those observed among patients who received R-CHOP (50%). However, among those who achieved complete remission, there was no significant difference in 3-year OS among those who underwent hematopoietic SCT in first remission versus those who were observed after induction (66% vs 43%, respectively; P=.8).

Table 2. Reported Outcomes From Retrospective Studies With Various Regimens in Patients With Double-Hit Lymphoma
ReferenceNo. With DHLDetection MethodTherapyOutcomesCharacteristics
  1. Abbreviations: CHOP-R-14, 14-day cycle of rituximab with standard cyclophosphamide, doxorubicin, vincristine, and prednisone; CHOP-R-21, 21-day cycle of rituximab with standard cyclophosphamide, doxorubicin, vincristine, and prednisone; CNS, central nervous system involvement; CODOX-M/IVAC, cyclophosphamide, vincristine, doxorubicin, and cytarabine with methotrexate/etoposide, ifosfamide, and cytarabine; DHL, double-hit lymphoma; DLBCL, diffuse large B-cell lymphoma; EN, extranodal; FISH, fluorescence in situ hybridization; FL, follicular lymphoma; GC, germinal center; GCB, germinal center B-cell like diffuse large B-cell lymphoma; IHC, immunohistochemistry; IPI, International Prognostic Index; LDH, lactate dehydrogenase; Leuk phase, leukemic phase; OS, overall survival; p53, tumor protein 53; PFS, progression-free survival; PS, performance status; R-CHOP, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone; R-EPOCH, rituximab with etoposide, doxorubicin, vincristine (by 96-h infusion), cyclophosphamide, and prednisone; R-Hyper-CVAD, rituximab with hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone.

  2. a

    This was a comparison with patients who had DLBCL without the DHL phenotype.

  3. b

    The cutoff was 70% for B-cell chronic lymphocytic leukemia/lymphoma 2 (Bcl-2) and 40% for V-myc myelocytomatosis viral oncogene homolog; (MYC) by IHC.

  4. c

    The results from this study suggested that MYC/Bcl coexpression is more critical than cell of origin in conferring a worse prognosis for patients with DLBCL who receive R-CHOP.

  5. d

    Twelve percent of patients had triple-hit lymphoma.

Johnson 2012[29]14FISHR-CHOP-14/R-CHOP-215-y PFS, 27%; 5-y OS, 18%Higher LDH,a worse PS,a higher IPI scoresa
23IHCR-CHOP-14/R-CHOP-215-y PFS, 36%; 5-y OS, 32%
Green 2012[45, 58]11FISHR-CHOP3-y PFS, 46%; 3-y OS, 46%Higher LDH,a more advanced stage,a higher IPI scoresa; ≥2 EN, 27%; GC, 92%
54IHCR-CHOP3-y PFS, 39%; 3-y OS, 43%
Abramson 2012[65]28FISHR-CHOP, 44%PFS, 8 mo; OS, 11 moPrior FL, 29%; advanced stage, 79%; EN, 77%; CNS, 17%; Leuk phase, 26%; high-risk IPI, 50%; high LDH, 88%
R-EPOCH, 35%PFS, 21 mo; OS, 34 mo
CODOX-M/IVAC, 18%NR
Hu 2013[52]10FISHR-CHOPMedian PFS, 12 mo; median OS,: 20 moAdvanced stage, 67%; B-symptoms, 38%; ≥2 EN, 28%; high LDH, 64%; high-risk IPI, 49%; p53 mutation, 25%c
157IHCbR-CHOP5-y PFS, 27%; 5-y OS, 30%
Gandhi 2013[56]106FISHdR-CHOP, n=33PFS, 7.7 mo; OS, 34 moMost common histology, DLBCL; GCB phenotype, 87%; prior FL, 29%
R-EPOCH, n=28PFS, 21.2 mo; OS, 25.5 mo
R-Hyper-CVAD, n=15PFS, 8.1 mo; OS, 29.8 mo
R-CODOX-M/IVAC, n=6PFS, 4 mo; OS, 6.4 mo

In another single-center experience, Howlett and colleagues presented data on 37 patients with FISH-defined DHL,[49] observing that intensified regimens followed by SCT (most were allogeneic [allo]-SCTs; hazard ratio, 0.079; P=.016), but not early intensification without transplantation (hazard ratio, 0.53; P=.237), were associated with improved PFS compared with R-CHOP alone. However, no direct comparison of patients who received intensified regimens and subsequent SCT, compared with those who received intensified induction regimens without SCT, was offered in that analysis. Another recent retrospective study of patients who had DHL according to FISH criteria sought to evaluate different SCT strategies by comparing 9 patients who underwent allo-SCT with 9 patients who did not. Those investigators reported that the estimated 3-year PFS rate (76% vs 22%; P=.015) and 3-year OS rate (78% vs 14%; P=.014) favored those who underwent allo-SCT.[68] Given the retrospective nature of these reports, it is very likely that the patients who received escalated induction therapy, HDT-ASCT, and/or allo-SCT were chosen based on such features as younger age, better performance status, and (in the case of SCT) based on response to induction therapy, each of which correlates with more favorable disease characteristics and with improved clinical outcomes. Moreover, the lack of a survival benefit associated with escalated induction regimens observed in the studies by Gandhi and colleagues and Oki et al,[56, 67] despite favorable CR rates, suggests that relapse of DHL offsets most, if not all, early benefit of intensification. The role of escalated induction therapy for those patients who have double-expressing lymphoma or DHL, as defined by MYC GOC criteria, remains poorly explored.

In the relapsed setting, traditional approaches to relapsed DLBCL do not appear to be as effective, particularly with HDT-ASCT. In a subset analysis of the prospective CORAL study, which evaluated chemoimmunotherapy for patients with relapsed/refractory DLBCL, Cuccuini and colleagues noted that MYC aberrations imparted a similarly poor prognosis whether discovered as rearrangement or GOC.[46] Among 161 available tissue specimens, 28 (17%) had MYC rearrangements; these patients were more likely to be male, to have more advanced disease, and to have an elevated LDH level; but, notably, they did not have a higher frequency of early relapse from induction R-CHOP chemotherapy. However, the response to salvage therapy in that study was inferior for MYC-positive patients (25% vs 45%; P=.0497) and translated into fewer patients proceeding to transplantation. Moreover, among the MYC-positive patients who did undergo HDT-ASCT, the 4-year PFS rate was only 14%. Several small reports suggest that allo-SCT may be beneficial in a subset of patients with DHL, but most patients with relapsed DHL have very limited options and constitute a population in need of better treatments.

Novel Approaches

The increased awareness of this disease, coupled with better recognition of its aggressiveness and appreciation of its biology, hopefully will lead to more targeted trials. The lack of a clear benefit from intensification of therapy (with or without SCT) suggests that incorporating novel and targeted agents should be pursued, and several candidate agents are available. For instance, ABT-263 and ABT-199, BH3-targeted inhibitors of BCL2, both exhibit enhanced cell kill in resistant lymphoma cell lines that overexpress both MYC and BCL2.[69, 70] Furthermore, when combined with cytotoxic agents, ABT-263 exerted additive and/or synergistic apoptotic effects. Mechanism-associated thrombocytopenia caused by collateral effects on MCL-1 and B-cell lymphoma-extra large (BCL-XL) has halted the development of ABT-263, but ABT-199 has exhibited dramatic activity in patients with NHL[71] and is currently being investigated in clinical trials combining it with rituximab and chemotherapy. Because MYC regulates the expression of Aurora-A and B kinases and the transcripts of both kinases are elevated in MYC-driven NHL, investigating Aurora kinase inhibitors in the setting of DHL is also appealing.[72] Alisertib, a selective Aurora-A kinase inhibitor, has exhibited preliminary activity in patients with relapsed/refractory NHL, including transformed follicular lymphoma and BL.[73] A study evaluating this drug in combination with cytotoxic chemotherapy in patients with aggressive NHL, including DHL, is planned.

Other targets include pathways that directly or indirectly interface with MYC function. For example, mTOR inhibitors seem to reduce malignant transformation induced by deregulated expression of c-MYC in B lymphocytes,[74] can have dramatic activity in in vivo models of MYC-positive lymphoma,[75] and form the rationale for completed and ongoing studies of these agents in aggressive NHL.[76, 77] In fact, upstream targeting of the same pathway at the level of PI3K with the oral drug idelalisib already has demonstrated significant activity in NHL.[78, 79] This agent will likely be approved in the very near future for patients with relapsed/refractory chronic lymphocytic leukemia,[80] making it an attractive option for future testing in patients with DHL. Another pathway that seems to be linked to MYC regulatory functions is NF-κB signaling. Klapproth et al investigated the role of NF-κB in mouse and human c-MYC–transformed lymphomas and observed that c-MYC overexpression sensitizes cells to NF-κB–induced apoptosis.[81] Those investigators further demonstrated that persistent inactivity of NF-κB signaling is a prerequisite for MYC-mediated tumorigenesis. Similarly, Evens and colleagues demonstrated that the second-generation proteasome inhibitor MLN9708 potently degraded MYC and effectively induced cell death at nanomolar concentrations in preclinical lymphoma models[82] and might support investigating NF-κB pathway inhibitors in DHL.

An intriguing potential target in MYC-associated malignancies is the relatively recent identification of super-enhancers at the promoter regions of oncogenes, including MYC. These bromodomain and extraterminal (BET) proteins are transcriptional activators associated with oncogenes, are amenable to small-molecule inhibition, and, for unexplained reasons, exhibit selective sensitivity toward MYC inhibition.[83] Recently, Emadali and colleagues discovered a gene-regulatory circuit involving the nuclear factor CYCLON, which characterizes aggressive disease and resistance to rituximab, in high-risk B-cell lymphoma. Those investigators observed that CYCLON knockdown inhibited MYC-overexpressing tumors in mice and modulated gene expression programs of biologic relevance to lymphoma.[84] Furthermore, CYCLON knockdown increased the sensitivity of human lymphoma B cells to rituximab in vitro and in vivo. This newly identified CYCLON appears to autonomously drive aggressive tumor growth and rituximab resistance. Targeting this pathway with a small-molecule inhibitor of BET proteins mimicked CYCLON knockdown, and may provide a rational approach to DHL. Currently, several inhibitors are entering early phase clinical trials. Collectively, available data support the finding that R-CHOP is an unsatisfactory therapy for DHL. We strongly encourage the enrollment of newly diagnosed patients onto prospective studies investigating rationally designed molecular therapeutics.

Summary

In summary, MYC-associated lymphomas are no longer relegated to the balanced rearrangements classically described in BL. Rather, there appears to be a spectrum of MYC-associated lymphomas with rearrangements, other chromosomal abnormalities, mutations, increased copy numbers, and increased protein expression; when coassociated with BCL2 or BCL6 rearrangements, the resultant phenotype is clinically aggressive DHL or THL. Although approximately 5% to 10% of DLBCLs and BCLUs are DHL according to FISH, the incidence increases to nearly 30% if protein overexpression is detected by immunohistochemistry. Unfortunately, there are no reliable clinical, morphologic, or COO data predictive of DHL, suggesting that more universal testing for MYC/BCL2 gene or protein should be implemented, particularly given the availability of a robust immunohistochemical stain. Given the clinical dilemma presented by MYC-associated lymphomas, DHLs, and THLs, there is clearly an unmet need to design trials specifically for these patients, and the first challenge will be to uniformly define MYC-associated lymphomas and DHLs. Although R-CHOP seems to be insufficient, the best augmented regimen, if any, remains unknown, and the current generation of trials should focus on the incorporation of biologic agents for a population badly in need of improved outcomes.

FUNDING SUPPORT

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

Dr. Petrich reports research funding, consultancy fees, and honoraria from Genentech. Dr. Nabhan reports grants and personal fees for advisory board meetings and speaking engagements from Genentech and Celgene. Dr. Smith reports fees for consulting and advisory board meetings from Celgene; fees for advisory board meetings and Data Safety and Monitoring Board Committee meetings from Genentech; fees for advisory board meetings from Pharmacyclics, TG Therapeutics, Seattle Genetics, Gilead, and Onyx; and she previously served on the speakers' bureaus of Celgene, Janssen, and Seattle Genetics.

Ancillary