Current challenges and novel treatment strategies in double hit lymphomas

Introduction: challenging lymphomas

Traditional pathological diagnosis of lymphoma has been based upon morphological appearance of involved tissue under light microscopy and definition of lineage by immunohistochemistry (IHC) or flow cytometry. Over recent years there has been increased recognition of the limitations of this approach in adequately differentiating disease entities in terms of clinical outcome, especially among aggressive B-cell lymphomas. Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma in Western countries [Friedberg and Fisher, 2008; Friedberg, 2012] and is defined by the World Health Organization (WHO) as a ‘neoplasm of large B lymphoid cells’ with a ‘diffuse growth pattern’ [Harris et al. 2008]. However, amongst patients with DLBCL, clinical behavior and responses to therapy can be heterogeneous [Barrans et al. 2010].

The mainstay of therapy for aggressive B-cell lymphoma is R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisolone) or R-CHOP-like regimens. Chemoimmuno-therapy in DLBCL has improved 2-year overall survival (OS) to greater than 70% [Coiffier et al. 2002; Sehn et al. 2005; Pfreundschuh et al. 2006]. However, an ongoing challenge remains how to deal with early relapsing or primary refractory disease. These patients continue to have poor outcomes even after salvage with high-dose therapy (HDT) and autologous stem-cell transplant (SCT) and are candidates for augmentation of first-line therapy, including addition of novel agents with a view to improving outcomes [Gisselbrecht et al. 2010]. Understanding the genetic drivers of poor prognosis aggressive B-cell lymphomas is key to identifying high-risk individuals, and the prospective identification of patients at risk is a necessary first step.

Patients with high-grade clinically aggressive lymphomas and transformed disease are an enriched population for so-called ‘double hit’ lymphoma (DHL) [Pedersen et al. 2012]. Thangavelu and others in 1990 [Thangavelu et al. 1990] provided an early description of lymphoid malignancies characterized by both t(14;18) and t(8;14) or t(8;22) with resulting concurrent overexpression of both BCL2 and MYC. These lymphomas displayed aggressive biological behavior. The definition of DHL has been expanded over time to refer to high-grade B-cell lymphomas with recurrent chromosomal breakpoints resulting in activation of two or more oncogenes [Aukema et al. 2011]. It is therefore an historical quirk that we retain the name DHL even in an era where we recognize that multiple genetic events occur in the development of most B-cell lymphomas [Cigudosa et al. 1999; Lohr et al. 2012]. Although DH chromosomal changes are recognized in a number of B cell malignancies including follicular lymphoma (FL), acute lymphoblastic leukemia and B cell lymphoma unclassifiable, by far the most common histological subgroup affected is DLBCL. For the purposes of this review unless otherwise specified below DHL will refer to DLBCL with chromosomal translocations that result in overexpression of both BCL2 and MYC proteins.

DHL is important because many authors now believe that standard chemo-immunotherapy alone is insufficient therapy [Friedberg, 2012; Lin et al. 2012] with reports of a median OS of approximately one year [Green et al. 2012b]. Nonetheless, barriers to improving current outcomes are multiple and include: (1) defining a homogenous group of lymphomas to be included in the definition of this disease; (2) optimizing accurate, widely available and timely detection of this high-risk entity in clinical practice; and (3) identification of augmented regimens with meaningfully enhanced antilymphoma efficacy.

Defining DHL: genetics and phenotype

Overexpression of MYC and BCL2 proteins in the same B cell classically arises as a result of chromosomal translocations whereby the relevant genes are juxtaposed to the promoter regions of genes that are constitutively active during B-cell development, most typically the immunoglobulin heavy chain (IGH) gene. The 14;18 translocation juxtaposes the IGH and the BCL2 genes, resulting in overexpression of the anti-apoptotic protein BCL2. t(14;18) is the hallmark of follicular lymphoma, occurring in 85% of patients of European background and 70% of patients with Asian background [Biagi and Seymour, 2002]. The vast majority of cases of follicular lymphoma have high level of expression of BCL2 protein [Pezzella et al. 1990]. In contrast t(14;18) occurs less frequently in DLBCL where it is seen in between five and 30% of cases [Barrans et al. 2003]. BCL2 promotes the survival of cells in the face of major stressors such as DNA-damage and growth factor deprivation. Not only is avoidance of apoptosis through overexpression of BCL2 a recognized hallmark of cancer driving its development [Hanahan and Weinberg, 2000] it also mediates resistance to chemotherapy in experimental models of lymphoma [Schmitt and Lowe, 2001]. This latter attribute likely explains the association of BCL2 overexpression with poorer prognosis in patients with DLBCL treated with CHOP in the pre-rituximab era [Mounier et al. 2003]. It may also underpin the inability of chemotherapy alone to cure follicular lymphoma.

Overexpression of the MYC proto-oncogene due to its juxtaposition with IGH in the t(8;14) or with IGL or IGK in the t(8;22) and t(2;8) light chain variants respectively, influences a multitude of cellular functions including proliferation [Anderson et al. 2015], metabolism and protein synthesis. When a MYC translocation occurs in isolation or in the context of a simple karyotype, as in the case of Burkitt lymphoma, the increase in proliferation is accompanied by increased apoptosis [Fanidi et al. 1992; Murphy et al. 2008]. Morphologically this is recognized by the classical ‘starry sky’ pattern of Burkitt lymphoma. Detection of a MYC rearrangement is thought to confer inferior outcome in DLBCL [Savage et al. 2009; Barrans et al. 2010; Valera et al. 2013]. However in DLBCL, MYC translocations are infrequent as a single hit and are usually found as part of a complex karyotype [Boerma et al. 2009; Smith et al. 2010].

The combination of increased proliferation and reduced apoptosis, conferred by co-expression of MYC and BCL2, results in the ‘perfect storm’ of oncogenesis resulting in aggressive difficult to treat malignancies [Fanidi et al. 1992; Johnson et al. 2009; Li et al. 2012; Pedersen et al. 2012; Cheah et al. 2015]. Although MYC overexpression in isolation results in aggressive lymphoma morphology [Mohamed et al. 2001], at least two studies have found that MYC protein overexpression is clinically more meaningful when associated with BCL2 [Green et al. 2012b; Johnson et al. 2012]. In a study by Barrans and others, MYC rearrangements (often t(8;22)) occurred in approximately 14% of untreated de novo DLBCL, with the incidence of concurrent MYC and BCL2 rearrangements being approximately 10% [Barrans et al. 2010]. The incidence of DH status increases to around 20% amongst patients with aggressive lymphoma transformed from more indolent diseases such as follicular lymphoma [Pedersen et al. 2012]. This may represent acquisition of MYC during the process of transformation. Paradoxically, other studies suggest that the incidence of DHL is approximately the same amongst cohorts of relapsed versus newly diagnosed disease [Cuccuini et al. 2012].

While cytogenetically defined BCL2⁺/MYC⁺ DHL accounts for 60–90% of all DHL [Aukema et al. 2011; Petrich et al. 2014], other combinations of translocations seen include BCL6⁺/MYC⁺ (approximately 5%), BCL2⁺/BCL6⁺/MYC⁺ (also known as triple hit lymphoma [THL], approximately 8%) [Petrich et al. 2014] as well as rare translocations involving other genes including CCND1, BCL3 and PAX5 [Cheah et al. 2015]. The importance of DHL has been increasingly recognized because of its profound impact on clinical outcome, particularly in DLBCL where early relapse following rituximab-containing chemo-immunotherapy portends a poor prognosis [Gisselbrecht et al. 2010]. One study by Niitsu and others, demonstrated 2-year survival of 23% in MYC⁺/BCL2⁺, compared with 65% in MYC⁺/BCL2^-, 81% in MYC^-/BCL2⁺ and 84% in MYC^-/BCL2^- lymphomas [Niitsu et al. 2009].

DHL are associated with a high prevalence of independent adverse risk factors in DLBCL including B symptoms, high lactate dehydrogenase, advanced age, extra nodal disease, bone marrow involvement, higher Ki-67 (median 80%), Ann Arbor stages III and IV as well as central nervous system (CNS) disease [Niitsu et al. 2009; Savage et al. 2009; Barrans et al. 2010; Snuderl et al. 2010; Friedberg, 2012; Cheah et al. 2015]. Even accounting for high international prognostic index (IPI) scores, DHL status confers a poor prognosis [Le Gouill et al. 2007]. Unlike many conventional DLBCL risk factors the adverse prognosis portended by DHL status is not negated by chemo-immunotherapy with rituximab [Savage et al. 2009]. DHL affects adults of all ages [Cheah et al. 2015] and there is a male predominance [Oki et al. 2014]. Among 129 DHL patients from MD Anderson the most predictive adverse clinical features at diagnosis were poor performance status and stage four disease [Oki et al. 2014].

Optimizing diagnosis of DHL: practical versus purist

Accurate detection of adverse prognostic markers such as MYC and BCL2 as well as other genes conferring DH or TH status including BCL6, CCND1, BCL3 and PAX5 is key to identifying high-risk patients [Petrich et al. 2014; Cheah et al. 2015]. DHL patients are ideal candidates for clinical trials of novel agents and optimization of standard chemotherapeutic approaches, so early and accurate detection is essential.

In routine clinical practice there are two key methods available for detecting dysregulated MYC and BCL2: fluorescence in situ hybridization (FISH) to identify translocations and IHC to identify protein overexpression. FISH is the gold standard for detection of high-risk DH or TH translocation status. It is a reproducible technique that correlates with clinical outcome [Niitsu et al. 2009] and is used to define DHL in the majority of clinical trials. Some authors advocate FISH testing in all new diagnoses of DLBCL [Landsburg et al. 2014], as it is a both sensitive and specific test for DHL status (Figure 1). Over time FISH has become cheaper and hence more widely available, with establishment of standardized reference laboratories. However it can be time-consuming meaning that results may not always be available to prospectively use in planning treatment decisions. In addition, optimal tissue samples are required for this technique.

OPEN IN VIEWER

Figure 1.

Detection of DHL by FISH. The G-banded karyotype was 49,XY,+8,t(8;14)(q24;q32),+12, t(14;18)(q32;q21),+r[11]/46,XY[18]. (A) FISH using the LSI MYC (8q24) dual color break apart rearrangement probe (Vysis) showed two intact copies of MYC and one rearranged (split) MYC signal. In metaphases the rearranged 5’MYC signal (red) was located on the derivative chromosome 8 [der(8)der(8;14)] and the 3’MYC signal (green) was located on a derivative chromosome 14 [the der(14)t(8;14)]. (B) FISH using the LSI IGH/BCL2 t(14;18) dual color dual fusion translocation probe (Vysis) identified two IGH/BCL2 fusion signals: one on a derivative chromosome 14 [der(14)t(14;18)] and one on the derivative chromosome 18 [der(18)t(14;18)]. Additional IGH signals (green) were seen on the der(8)t(8;14) and the der(14)t(8;14). A BCL2 signal (red) was present on the intact chromosome 18.

IHC to detect double protein expression is often used clinically to rapidly screen for DHL and when positive every effort should be made to confirm DHL status using FISH. IHC lacks standardization with no reliable cut off for positive used between clinical trials [De Jong et al. 2007]. Furthermore there is currently a paucity of evidence to support changing clinical practice based on the results of IHC alone. It also lacks sensitivity and may miss translocation-positive cases where the translocation is present in less than 40% of cells [Horn et al. 2013]. However, IHC to detect double protein expression is an inexpensive and widely available technique that is commonly utilized in clinical practice (Figure 2). More recently, adoption of standardized IHC methods has been associated with greater predictive value for this technique. With many current trials prospectively examining outcome in patients both by standardized IHC and FISH, the role of IHC in the detection of DHL is likely to become clearer.

OPEN IN VIEWER

Figure 2.

Detection of DHL by IHC. IHC of DH DLBCL is characterized by (A) diffuse growth of large B lymphoid cells on H+E with (B) high Ki-67, (C) high BCL2 and (D) high MYC expression.

Using a cut off for positive of ⩾40% MYC positive cells, concurrent MYC and BCL2 were detected by IHC in 21% of DLBCL [Johnson et al. 2012]. Using a cut off for positive of ⩾50%, it was shown that among 77 cases of DLBCL, all patients with a MYC translocation were detected by IHC [Kluk et al. 2012]. Furthermore, in this study 4 out of 77 patients who were positive for MYC by IHC but negative by FISH demonstrated increased MYC transcription by gene expression testing and all IHC MYC positive cases demonstrated a poor prognosis [Kluk et al. 2012]. Others suggest that IHC MYC ⩾ 70% is the optimal cut off giving a sensitivity of 100% and a specificity of 93% for MYC expression as measured by quantitative real-time polymerase chain reaction (QRT-PCR) [Green et al. 2012a]. Among 193 patients with B-cell lymphoma treated with R-CHOP, 29% demonstrated increased MYC and BCL2 expression by IHC: all of whom had a poor outcome [Green et al. 2012b]. In this study only 11% had a detectable MYC rearrangement by FISH and only 6% were DHL by FISH [Green et al. 2012b]. The apparent incidence of DH status appears to increase from around 10% of DLBCL when defined by FISH to between 20% and 30% with IHC [Green et al. 2012b]. The higher incidence of DH status with IHC over FISH likely pertains to detection of protein overexpression driven by mechanisms other than translocations e.g. microRNA overexpression, amplification or mutation [Cheah et al. 2015] (Table 1). Similar issues arise in detection of BCL2 abnormalities by IHC.

OPEN IN VIEWER

Table 1.

Detection of MYC and BCL2. Comparison of FISH and IHC for detecting MYC and BCL2.

Sensitivity	Specificity	References
167 DLBCL patients; IHC positive for MYC at 40%; BCL2 and MYC detected in 21%. MYC IHC and MYC FISH strongly correlated (p < 0.01)		Johnson et al. [2012]
193 DLBCL patients treated with R-CHOP; IHC MYC positive at >40%; BCL2 and MYC detected in 29%; patients with IHC BCL2 and MYC had worse outcome	Only 6% were DHL by FISH.	Green et al. [2012]
77 DLBCL patients; MYC IHC positive >50%; all MYC translocation positive cases detected by IHC	4/77 cases had positive IHC but were negative by FISH	Kluk et al. [2012]
219 DLBCL + Burkitt lymphoma patients; MYC IHC graded weak to strong; 93% of subjects with MYC translocations had ⩾80% MYC positive cells on IHC; ROC analysis ⩾ 70% optimal cut off with sensitivity of 100%	Only 3% of non MYC rearranged cases had ⩾80% MYC staining on IHC (p < 0.0001); specificity of a MYC cut off of ⩾70% is 93%	Green et al. [2012]

DLBCL, diffuse large B-cell lymphoma; IHC, immunohistochemistry; FISH, fluorescence in situ hybridization; DHL, double hit lymphoma; R-CHOP, rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone; ROC, receiver operating characteristic

All new diagnoses of DLBCL should have IHC for MYC and BCL2 at a minimum [Friedberg, 2012] and when one or both is positive FISH should be performed for MYC, BCL2 and BCL6 to confirm DHL status. However, the cut off for positive IHC remains controversial in the literature, and in cases that are borderline positive, confirmatory testing with FISH is necessary. A standardized approach to the diagnosis of this entity will facilitate assessment of incidence and optimized management approaches.

Improving prognosis in DHL: incremental versus paradigm shifts

Given the poor outcome seen in DLBCL that relapses early post-chemo-immunotherapy [Gisselbrecht et al. 2010] optimizing therapy remains the key clinical challenge in DHL [Klapper et al. 2008; Oki et al. 2014; Pedersen et al. 2014]. The challenges here are twofold: (1) optimizing chemo-immunotherapy regimens to minimize the incidence of primary refractory and early relapsing disease (collectively, early treatment failure) and (2) salvaging individuals with poor response to initial therapy.

Approaches to optimizing traditional chemotherapeutic regimens, to minimize the incidence of early treatment failure, is critical, as outcomes in these patients remain poor even with HDT followed by autologous SCT [Gisselbrecht et al. 2010]. Variations on R-CHOP, including R-CHOP administered every 14 instead of every 21 days or the addition of etoposide to the regimen, have resulted in incremental benefits in survival of questionable significance, often at the expense of increased toxicity [Pfreundschuh et al. 2004]. Given the poor results with traditional chemo-immunotherapeutic approaches in the early treatment failure setting, much of the current research effort is directed towards evaluation of the safety and efficacy of a multitude of novel agents in the treatment of high-risk lymphoma such as DH DLBCL.

The key challenges in the field of novel agents for DHL and DLBCL in general are (1) defining which combinations of agents optimize clinical outcome while minimizing toxicity and (2) ordering the utilization of new agents and standard chemotherapeutics for example sequential therapy versus maintenance therapy [Anderson et al. 2015]. These challenges are compounded by the fact that DHL, like many cancers, is a disease of aging with a median age of diagnosis in the seventh decade of life [Green et al. 2012b; Johnson et al. 2012]. While the optimal therapeutic approach for this difficult group of patients remains unknown, ongoing clinical trials particularly with novel agents offer the hope of better outcomes for patients in the future [Friedberg, 2012]. Given that DHL patients are often poorly recognized and represent only a subset of poor prognosis DLBCL, much of the available data regarding optimizing treatment in DHL must be inferred from studies in a wider group of relapsed and refractory DLBCL patients.

Novel targeted approaches to the treatment of DHL

Given the poor prognosis associated with DHL status, it is an area of active preclinical and early phase clinical research for exploring novel approaches to the treatment of difficult lymphomas. As DHL is characterized by overexpression of both BCL2 and MYC, approaches targeting these specific pathways are particularly exciting areas of exploration (Table 2). Due to the relative rarity of DHL, there are few studies looking at this population in isolation. However many lessons pertaining to the optimization of outcomes in DHL may be drawn from studies in the broader group of relapsed or refractory DLBCL.

OPEN IN VIEWER

Table 2.

Targeted therapy for DHL. The two approaches involve (a) inhibition of BCL2 or (b) inhibition of MYC. Ideally the combination of both approaches would be tested in clinical trials along with chemo-immunotherapy.

Targeted therapy for DHL	Proposed mechanism	Preliminary clinical results
BCL2 inhibitors
Navitoclax	Inhibition of BCL2, BCLxL and BCLw	4/4 evaluable patients with R/R DLBCL progressed [Wilson et al. 2010]
Venetoclax	Selective inhibition of BCL2	28% ORR in R/R DLBCL [Davids et al. 2014]
MYC inhibition
Alisertib	Inhibitor of aurora A kinase, a downstream target of myc, which when blocked results in apoptosis [den Hollander et al. 2010]	3/21 patients with R/R DLBCL responded [Friedberg et al. 2014]
JQ1	BET bromodomain (BRD4) inhibitor resulting in down regulation of MYC transcription [Delmore et al. 2011]	Not in clinical trials
GSK525762	BET bromodomain inhibitor	Results awaited [ClinicalTrials.gov identifier: NCT01943851]
CPI0610	BET bromodomain inhibitor	Results awaited [ClinicalTrials.gov identifier: NCT01949883]

DLBCL, diffuse large B-cell lymphoma; DHL, double hit lymphoma; R/R, relapsed/refractory; ORR, overall response rate.

Conclusions: towards a brighter future

DHL remains a challenging disease on a number of levels. To ensure appropriate diagnosis, there needs to be widespread availability of fast, reliable and affordable detection methods. Once diagnosed, specific treatment strategies are needed, but as yet there is no universally accepted therapeutic approach. While it is widely acknowledged that standard R-CHOP chemotherapy is insufficient to cure many of these cases, alternatives are not clear. Likely ways forward for this disease will come in two guises: (1) intensified regimens using standard agents such as DA-R-EPOCH and (2) inclusion of novel agents that either target the malignant B cell generally (e.g. immunotherapy) or the roots of the DHL problem, overexpression of MYC and BCL2 proteins. The greatest challenge going forward in this disease will be finding new agents that work to ameliorate the poor prognosis conferred by DH status. Wherever possible every effort should be made to accurately diagnose DHL using FISH and confirmed cases should be enrolled in clinical trials both in the upfront and relapsed/refractory setting.