ENGINE: a Phase III randomized placebo controlled study of enzastaurin/R-CHOP as frontline therapy in high-risk diffuse large B-cell lymphoma patients with the genomic biomarker DGM1

Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous and aggressive tumor that affects mainly adults in their 60s–70s. This B-cell lymphoid malignancy is the most common histologic subtype of non-Hodgkin lymphoma [1–5] with an incidence of nearly 25–40% of all new cases of non-Hodgkin lymphoma diagnosed worldwide. Additionally, increases in the incidence of DLBCL have recently been reported in Central and South America as well as in Asia [3,4].

Despite the evaluation of a number of other agents in randomized trials, combination of rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP) administered every 21 days has remained the standard of care for all DLBCL subtypes for over 2 decades. While R-CHOP cures over 60% of DLBCL patients, 5-year survival rates vary from 90% for low-risk DLBCL patients to less than 40–50% for high-risk DLBCL patients as defined by the international prognostic index (IPI) score [4–9]. This heterogeneity is not well understood, and there is an urgent need to improve R-CHOP therapy, with one possible method being adding strategic agents to this backbone as an up-front option for high-risk patients [8]. Substantial effort has been devoted to developing clinical models and molecular and gene expression profiling as tools to identify DLBCL patients with comparatively inferior prognosis to R-CHOP therapy in order to determine which patients are the most suitable candidates for novel targeted therapies [4–8,10,11].

Analysis of samples from high-risk IPI patients that failed R-CHOP, identified PKC-β as a potential therapeutic target in this group [12–14]. PKC-β belongs to the PKC family known for their important regulatory functions within normal cells. The PKC family has also been a focus of research due to its role in the pathophysiology of cancer, such as promoting and/or inhibiting tumor growth, differentiation, tumor-associated angiogenesis and inhibition of apoptosis [11–14]. The contribution of the different PKC isozymes in cancer growth or suppression is variable [15,16]. In particular, PKC-β is a component of the B-cell receptor in both normal and clonal B-lymphocytes, and it is associated with tumor-stimulated angiogenesis upon its regulation of the VEGF pathway [16] and also with survival of these clones. High levels of PKC-β are associated with an inferior prognosis in some specific DLBCL subtypes making it a potential therapeutic target [5,13–18].

Enzastaurin

The role of PKC-β in the proliferation and survival of clonal B-lymphocytes led to the development of selective drugs targeting this regulatory pathway. Similar to other protein kinases, PKC consists of a catalytic and a regulatory region, and the enzyme structure is linked to their activation through molecules such as diacylglycerol (DAG) alone or DAG and calcium. In general, PKC inhibitors can be classified into three broad categories: PKC inhibitors that interfere with the binding site of second messengers and DAG; molecules that prevent the binding of receptor for activated C kinase and lastly; ATP competitive inhibitors that prevent activation of PKC enzymes [19].

Enzastaurin is an acyclic bisindolylmaleimide and an oral formulated selective serine-threonine ATP-competitive inhibitor of PKC-β kinase and the cell cycle regulatory signaling transduction pathway PI3K/AKT [19]. Enzastaurin inhibits the phosphorylation of downstream signal proteins associated with tumor suppression and induced angiogenesis [19]. The antitumor activity of enzastaurin has been demonstrated in vitro and in various cancer models including gastric and breast cancers, brain tumors and DLBCL [12,15–20].

Enzastaurin was developed by Eli Lilly and Company, IN, USA (Lilly) and evaluated in 51 clinical studies and 15 clinical pharmacology studies with 39 single-agent studies (enzastaurin alone or enzastaurin combined with non-neoplastic agents) and 27 combination studies (enzastaurin plus another anticancer agent). As of September 2013, among the patients exposed to enzastaurin, 3149 were cancer patients and 188 were healthy individuals. Lilly’s enzastaurin studies in DLBCL patients are summarized in Table 1 below.

Based on several Phase I studies Lilly conducted the recommended Phase II dose of enzastaurin in DLBCL was 500 mg. The safety and tolerability of enzastaurin 500 mg was confirmed in subsequent Phase II studies, with no serious adverse grade 3/4 effects or drug-associated deaths [15,17,21].

Pharmacology studies have proved that enzastaurin is primarily metabolized in the liver with minimal renal elimination. When enzastaurin was studied in cancer patients with hepatic impairment, drug exposures tended to be lower compared with patients with normal hepatic function and decreased with increasing severity of hepatic dysfunction in previous studies. Thus, it is not recommended to administer enzastaurin to patients with moderate or severe hepatic impairment.

In addition, enzastaurin is metabolized primarily by CYP4503A (CYP3A) as demonstrated in the H6Q-MC-JCAJ Phase II trial using enzastaurin alone in patients diagnosed with recurrent high-grade glioma [16]. Patients on enzyme-inducing anti-epileptic drugs such as phenytoin and carbamazepine had significantly decreased enzastaurin exposures. Thus, concomitant medications that are potent inducers or inhibitors of isozyme CYP3A affect enzastaurin exposure.

Despite the favorable and promising preclinical antitumor activity of enzastaurin in several cancer subtypes, it has been very difficult to replicate these results in clinical trials. Enzastaurin was evaluated as single agent as well as in combination with various chemotherapy regimens in several clinical trials, all of which have failed to support a statistically and clinically significant beneficial effect of enzastaurin, with the exception of in DLBCL. Specifically, in a Phase II study of enzastaurin administered as a single agent in patients with relapsed or refractory DLBCL (n = 55) [10], four patients (7%; 95% CI: 2–18%), including three complete responders and one patient with stable disease, experienced free from progression ranging 20–60 months after study entry. This led to the development of a Phase III clinical trial entitled: ‘Preventing Relapse in Lymphoma Using Daily Enzastaurin’ (PRELUDE) [17].

The PRELUDE study [17] compared disease-free survival (DFS) after maintenance therapy with enzastaurin versus placebo in patients with DLBCL in complete response (CR) and with a high risk of relapse after first-line therapy. This was a multinational, double blind, randomized, placebo-controlled study that enrolled 758 high-risk DLBCL (IPI ≥3) patients in CR or complete response unconfirmed and/or with a negative fluorodeoxyglucose-PET scan post completion of six to eight cycles of R-CHOP. Patients were randomly selected in a 2:1 ratio to receive either enzastaurin 500 mg daily or an identical placebo as maintenance for 3 years. The primary end point was DFS defined as lack of disease progression or death. The data collected was analyzed 36 months after the last enrolled patient initiated treatment [17]. The DFS hazard ratio (HR) for enzastaurin compared with placebo was 0.92 (95% CI: 0.69–1.22; 2-sided log-rank p = 0.54). The DFS at 24 and 48 months were 78 and 70% for the enzastaurin arm, and 75 and 71% for placebo, respectively [17]. The overall survival (OS) at 24 and 48 months was 87 and 81% for enzastaurin, and 89 and 82% for placebo [17]. No difference was found in 4-year DFS (70 vs 71%) or OS (81 vs 82%) between the enzastaurin and placebo arms, demonstrating the lack of benefit of enzastaurin when given only as maintenance in previously untreated DLBCL [17].

The benefit of enzastaurin concurrent with R-CHOP, versus R-CHOP alone, in previously untreated DLBCL was then evaluated in the randomized Phase II study (S028, n = 101) [20]. In this trial, 101 patients were randomized 3:2 to receive R-CHOP/enzastaurin (n = 58) or R-CHOP (n = 43) as first-line therapy for intermediate or high-risk DLBCL. Following the initial induction treatment period, patients in the enzastaurin group with CR or partial response (PR) also continued to receive enzastaurin for up to 3 years or until disease progression. The median progression-free survival (PFS) for patients in the enzastaurin arm was 36.2 months versus 22.6 months in the control arm (HR = 0.73; p = 0.151). After adjusting for imbalances in baseline IPI score (≤2 vs >2) and Eastern Cooperative Oncology Group performance status (0 vs 1 or 2), the HR was 0.59 (95% CI: 0.32–1.09; one sided p = 0.048). In a post-hoc subgroup analysis, patients with an IPI score ≥3 had a median PFS that was not evaluable in the enzastaurin arm compared with 8.9 months in the control arm (HR = 0.50; p = 0.037). The 2-year OS was 75 and 69% in the overall population, and 71 and 55% in IPI ≥3 patients (p = 0.186) [20].

While these studies failed to meet their primary end point, the S028 trial data established the tolerability of concurrent R-CHOP/enzastaurin and the improved PFS in high-risk IPI patients after treatment with this combination. Collectively, these data stimulated further efforts to identify a pretreatment biomarker capable of predicting patients who benefit most from this combination.

Discovery & validation of DGM1

Even though enzastaurin has failed to demonstrate efficacy in overall population in PRELUDE and S028 trials [22,23], there were signs that there might be subsets of patients who might be benefited from the enzastaurin treatment. Another challenge is that there are no biopsies, which are typically required to study biomarkers for oncology drug, available for most of the patients enrolled in these studies. However, there are evidences suggesting that germline polymorphism could be biomarkers for cancer drug, such BRCA germline variation has been proved to be predicator for olaparib. Thus, DNA samples extracted from blood of patients (high risk IPI>= 3) enrolled in the PRELUDE study were genotyped using whole genome single nucleotide polymorphism (SNP) arrays and approximately 5 million SNPs were analyzed. It was found that a specific configuration of two SNPs (rs309605 and rs309604) located on chromosome 8 was strongly associated with survival in the enzastaurin arm. The p-value for this association was 5.8 × 10⁻⁹. To put this in context, a p-value less than 5 × 10⁻⁸ is generally suggestive of genome-wide significance. This biomarker is now referred to as DGM1. The closest gene to DGM1 is TRPS1, which encodes transcriptional repressor GATA binding 1 or TRPS1 protein. TRPS1 represses GATA-regulated genes and plays a central role in cell cycle control and tumor development, which may explain partially why it may be related to enzastaurin activity [24,25].

To confirm these findings from the PRELUDE trial, an analysis of the DGM1 biomarker was performed using archived DNA samples from patients enrolled in the S028 study. This investigation demonstrated significant improvement in OS (HR: 0.28; p = 0.018) for high-risk (IPI ≥3) DGM1+ patients receiving R-CHOP plus enzastaurin compared with high-risk DLBCL DGM1+ patients receiving R-CHOP alone (Figure 1). These results suggest that addition of enzastaurin to R-CHOP may significantly improve outcome in frontline high-risk DGM1+ DLCBL patients [18].

DGM1 was additionally evaluated in control arm patients who were treated with R-CHOP alone in S028 study. Results indicated that DGM1 status was not predictive of efficacy in the control (R-CHOP only) arm arguing against DGM1 as a prognostic biomarker. Similar result was observed in control arm patients who received placebo in Phase III PRELUDE trial.

ENGINE study

Study rationale

The Phase III ENGINE study was designed to prospectively validate the DGM1 biomarker analysis results from the Phase II study (S028) in patients with newly diagnosed high-risk DLBCL. The primary objective is the effect of adding enzastaurin to R-CHOP therapy in high-risk IPI ≥3 DLBCL patients who possess the DGM1 biomarker on OS. The study aims to determine whether there is markedly improved survival in this enriched population compared with the current standard of care, R-CHOP alone. This study is being conducted in the USA and China. The prevalence of biomarker-positive patients ranges between approximately 68% for patients of African ancestry living in the Southwest USA, approximately 90% for patients of European ancestry and approximately 95% of patients living in Beijing, China (International Genome Sample Resource -1000 Genomes Project).

Study design

ENGINE is a Phase III, randomized double blind, placebo-controlled, multicenter study in patients with treatment naive high-risk DLBCL. Approximately 235 patients will be randomized in a 1:1 ratio in the USA and China, stratified by IPI score (3 vs 4, 5) and region (USA vs China; Figure 2). DGM1 biomarker testing is performed in batches retrospectively after randomization.

All patients will receive up to six cycles (one cycle = 21 days) of treatment during the combination phase. PET-computed tomography will be used for staging and to assess radiographic response at the end of combination phase treatment. Each patient’s treatment assignment will be unblinded after combination phase tumor response assessment. Patients randomized to the enzastaurin arm who achieve a complete response (CR) or PR (at investigator’s discretion) by the Lugano Classification (Cheson 2014) will have the opportunity to continue in the single-agent phase of the study and receive single-agent enzastaurin for up to 2 additional years (Figure 2).

Study objectives

• Primary objective is to compare the effect of R-CHOP plus enzastaurin versus R-CHOP plus placebo on OS in treatment-naive patients with high-risk DLBCL who possess the DGM1 biomarker. Note: Both DGM1+ and DGM1 biomarker negative (DGM1-) patients will be enrolled but primary analysis will include only DGM1+ patients. Sites and sponsor will remain blinded to biomarker status of each patient;

• Secondary objectives are to compare combination phase CR rate and objective response rate (ORR) in DGM1+ patients; evaluate event-free survival (EFS) in all patients and in DGM1+ patients; determine OS of enzastaurin + R-CHOP in DGM1- patients and evaluate the safety profile of enzastaurin + R-CHOP.

Study key eligibility

Key inclusion criteria

• ≥18 years;

• Histologically confirmed CD20 positive DLBCL (MYC & BCL2 and/or BCL6 rearrangements eligible);

• Eastern Cooperative Oncology Group performance status 0, 1 or 2;

• IPI score ≥3;

• DGM1+ or DGM1 -;

• Left ventricular ejection (LVEF) ≥50% by echo or multiple gated acquisition (MUGA);

• Adequate organ function:

  	–	Total bilirubin ≤1.5× upper limit of normal (ULN; ≤5× ULN in the case of Gilberts syndrome, liver or pancreatic involvement by lymphoma);
  	–	ALT & AST ≤2.5× ULN (≤5× ULN if liver involvement);
  	–	Creatinine clearance ≥40 ml/min by Cockcroft-Gault equation;
  	–	Platelet ≥75 × 109/l (≥50 × 109 if BM involvement);
  	–	Hgb ≥8 g/dl (≥7 g/dl if BM involvement);
  	–	Absolute neutrophil count (ANC) ≥1.5 × 109 /l (≥1.0 × 109 if bone marrow [BM] involvement).

Key exclusion criteria

• History of indolent lymphoma or follicular grade 3b lymphoma;

• Primary mediastinal (thymic) large B-cell lymphoma; B-cell lymphoma unclassifiable;

• Known CNS involvement or Second primary malignancy (SPM);

• Use of a strong inducer or moderate/strong inhibitor of CYP3A4;

• History of long QT syndrome, QT corrected formula (QTcF) >450 ms (males) or >470 ms (females);

• Use of medication that can prolong QT/QTc;

• Ongoing > G2 peripheral neuropathy;

• Evidence of chronic hepatitis C by antibody to hepatitis C virus (HCV) with HCV-RNA (+);

• Evidence of chronic hepatitis B either HBsAg (+) or HBcAb (+) with hepatitis B virus (HBV)-DNA (+).

Study end points

• Primary end points:

  –

  OS in patients who are DGM1+. OS is defined as the elapsed time from the date of randomization to the date of death from any cause;

• Secondary end points:

  	–	CR in patients who are DGM1+;
  	–	ORR (CR and PR) in patients who are DGM1+;
  	–	OS in patients who are DGM1-;
  	–	EFS including EFS12 and EFS24, for all the patients, and for the patients who are DGM1+.

ENGINE trial versus other Phase III studies in frontline DLBCL

• Unique genomic biomarker;

• High risk patient population IPI = 3, 4, 5;

• Simplified screening procedures to allow quick treatment initiation, feasible to enroll high-risk patients;

• Eligibility based on local pathology diagnosis;

• Local safety labs;

• Primary end point OS with less frequent imaging schedule;

• 2 years of single agent phase after induction phase.

Statistical considerations

Efficacy analysis will be conducted on an intent-to-treat basis. The intent-to-treat population is defined as all patients who are randomized into the trial, regardless of whether they received study treatment. Safety analyses will include data from all patients who have received at least one dose of study treatment. The primary end point is OS in the DGM1+ patients. Study patients will be randomized 1:1 to R-CHOP plus enzastaurin or R-CHOP plus placebo, stratified by IPI score (3 vs 4, 5) and by region (USA vs China). This will be an event driven study. In the trial, 66 events are required to provide 90% power to detect a HR of 0.45 for OS in patients who are DGM1+, based on a stratified log-rank test at one-sided alpha of 0.025. Statistical significance will be achieved with an estimated HR ≤0.62. It is expected that 15% of trial patients will be DGM1-. To achieve 66 events in DGM1+ patients, the trial would enroll approximately 235 patients with target 200 being biomarker positive and 35 being biomarker negative. The treatment difference in OS will be compared using the log-rank test stratified by the randomization factors. Secondary end points include CR, ORR and EFS in DGM1+ patients. CR and ORR will be analyzed using Cochran-Mantel-Haenszel method stratified by randomization factors; and EFS will be analyzed using the method for primary end point of OS.

Study status

ENGINE is a global study, with centers located in both USA and China. The study is currently recruiting with expected accrual completion in June 2020. The trial is sponsored by Denovo Biopharma. More information is available at ClinicalTrials.gov (NCT03263026).

Conclusion

While R-CHOP is curative for the majority of patients with DLBCL, up to 40% of high-risk patients will fail first-line therapy. Only a fraction of relapsed/refractory patients are curable with second-line therapy, and the heterogeneity of outcomes is poorly understood. Targeting signaling pathways active in DLBCL remains a promising avenue of research despite negative randomized trials to date. These trials may have failed partly due to the lack of useful predictive biomarkers to identify patients responsive to novel agents. The ENGINE study will use a de novo genomic biomarker DGM1, which was identified from the PRELUDE and S028 studies, and correlate the presence of the DGM1 biomarker with OS in high-risk DLBCL treatment naive patients who receive enzastaurin/R-CHOP as frontline therapy.