Members of the glutathione and ABC-transporter families are associated with clinical outcome in patients with diffuse large B-cell lymphoma

Abstract

Standard chemotherapy fails in 40% to 50% of patients with diffuse large B-cell lymphoma (DLBCL). Some of these failures can be salvaged with high-dose regimens, suggesting a role for drug resistance in this disease. We examined the expression of genes in the glutathione (GSH) and ATP-dependent transporter (ABC) families in 2 independent tissue-based expression microarray datasets obtained prior to therapy from patients with DLBCL. Among genes in the GSH family, glutathione peroxidase 1 (GPX1) had the most significant adverse effect on disease-specific overall survival (dOS) in the primary dataset (n = 130) (HR: 1.68; 95% CI: 1.26-2.22; P < .001). This effect remained statistically significant after controlling for biologic signature, LLMPP cell-of-origin signature, and IPI score, and was confirmed in the validation dataset (n = 39) (HR: 1.7; 95% CI: 1.05-2.8; P = .033). Recursive partitioning identified a group of patients with low-level expression of GPX1 and multidrug resistance 1 (MDR1; ABCB1) without early treatment failures and with superior dOS (P < .001). Overall, our findings suggest an important association of oxidative-stress defense and drug elimination with treatment failure in DLBCL and identify GPX1 and ABCB1 as potentially powerful biomarkers of early failure and disease-specific survival.

Figure 1

Role of the glutathione pathway and ABC-transporter family in drug elimination and defense from ROS-mediated oxidative stress. Shaded diamonds represent intracellular drug and drug metabolites. Glutathione (GSH) is represented by clear pentagons. The rate-limiting step in its de novo biosynthesis is the enzyme glutamate-cysteine ligase (GCL). The balance of drug/drug metabolites and intracellular GSH is a significant determinant of drug levels and ROS stress. Glutathione peroxidase 1 (GPx1) reduces hydrogen peroxide generated after exposure to certain drugs, by oxidizing GSH to its disulfide form GSSG. Glutathione S-transferases (GST) conjugate GSH to drugs and drug metabolites, thus facilitating drug-conjugate elimination and efflux. GSTs also function as peroxidases to a lesser extent than GPx1 (not shown for simplicity). The ABC transporters ABCB1 (MDR1), ABCC1 (MRP1), and ABCC2 (MRP2) among others are involved in ATP-dependent efflux of drugs or drug-GS conjugates.

Figure 2

CART for the dOS rate at 2 years. The cutoffs represent median-adjusted expression values of the genes indicated. Circles represent branch points, and squares are terminal nodes. The P values at each split are calculated for the binomial distribution.

Figure 3

Long-term outcome by CART classification. The group with low GPX1 and low ABCB1 expression levels was defined by the 2-year dOS rate and has no early failures. Tick marks reflect censoring.

Figure 4

Effect of GPX1 expression and the LLMPP signature score on dOS. Patients are grouped in quartiles by GPX1 expression, with quartile 1 having the lowest and quartile 4 the highest level. The association with dOS is presented for the entire cohort (A), patients with the activated B-cell (ABC) subtype (B), and patients with non-ABC disease subtypes (C). Censoring not shown for simplicity.

Figure 5

Clinical outcome among patients in the primary and validation datasets. DOS is presented for the primary and validation datasets (A). DOS is compared among patients in the lowest (B), middle (C), and highest (D) tertile of GPX1 expression in each dataset. Tick marks reflect censoring.