Using a population-based sample of female NHL cases diagnosed from 1995 to 2000 in Connecticut and followed through mid 2008, we identified nine SNPs from 5 metabolism genes (CYP2E1, GSTP1, GSTT1, NAT1 and NAT2) that were associated with overall survival of NHL overall or histological subtypes. Five of these nine SNPs were further found to be associated with disease-free survival of NHL overall or histological subtypes. Our study offers the first analysis of the impact of genetic variations in the CYP and NAT genes on the survival of NHL by subtype, and our results should be confirmed by other studies.
CYPs, GSTs and NAT are all important enzymes involved in metabolism of exogenous and endogenous compounds. The family of CYPs activates environmental carcinogens to electrophilic metabolites capable of binding to DNA. The GST family of enzymes detoxifies the reactive compounds by converting them to inactive, water soluble metabolites. NAT catalyzes aromatic and heterocyclic amines via N- or O-acetylation thus involved in both activation and detoxification of numerous drugs and carcinogens. The roles of their genetic polymorphisms in cancer risk have been the subject of numerous studies. Evidence, although not so extensive, did show that genetic variations in CYP, GST and NAT might be associated with risk of NHL overall and/or subtypes. For example, polymorphisms in CYP1A1
were found to be associated with risk of DLBCL [18
] and CLL [19
]; distributions of CYP2E1
genotypes were found to be different among NHL patients and healthy controls [20
]; single reports also suggest some relevance of genetic variants in CYP1B1, CYP2C9,
and risk of NHL or its subtypes [19
]. Multiple studies reported higher risk of NHL [24
], DLBCL [18
], and MZBL [25
] associated with GSTT1
null; two GSTP1
polymorphisms were found to be associated with DLBCL risk in two studies [18
]; studies also showed that GSTM1
null is associated with increased risk of NHL [19
]. A large study found increased risk of NHL associated with NAT1*10/*10
genotype and intermediate and rapid NAT2
]. These findings supported that the genetic variations in such metabolic pathways may play a role in lymphomagenesis.
We observed a favorable effect of carrying A allele in CYP2E1
(rs2070673) on NHL survival, especially pronounced in CLL/SLL patients. As far as we know, this is the first time that the relationship of CYP genetic variations and NHL survival is studied. As a member of CYP family expressed in liver and extrahepatic tissues, CYP2E1 metabolizes a variety of nitrosamines and low-molecular-weight chemicals including drugs, solvents, toxins and environmental pollutants [30
]. It is known that CYP2E1 mediates metabolism of several drugs such as acetaminophen, chlorzoxazone, and enflurane, but its full potential has not been well characterized, especially with drug candidates in development [31
]. Our results suggest that CYP2E1 might play a role in the metabolism of drugs used in CLL/SLL treatment, such as Fludarbine, Rituximab, Cyclophosphamide, Alemtuzumab, Bendamustine, Chlorambucil.
We observed that GSTT1
null was associated with a 59% reduced risk of death and a 52% reduced risk of relapse, secondary cancer or death in FL patients. This is in line with previous studies on lymphoma prognosis: Stanulla et al. found that GSTT1
null was associated with a reduced risk of relapse in childhood acute lymphoblastic leukemia patients [32
], and Hohous et al. found that GSTT1
null was more prevalent in low-stage HL patients than in high-stage HL patients. However, a recent study of 89 FL by Hohaus et al. found that GSTT1
null was associated with worse event-free survival [33
]. Several possible reasons could explain the discrepancy between our finding and theirs: first, since different drugs are used in treating low-stage and high-stage cancer patients, GST may play different drug-metabolism roles in these two types of patients. There are more low-stage FL patients in our study than in theirs (52% vs. 7% Stage I); when we limited our analysis to FL patients with tumor stage higher than I, no survival difference for GSTT1
deletion was observed (P-value for log-rank test: 0.8447 for OS and 0.9856 for DFS). Another possible reason is that given the evidence of sex-related differences in the expression of various GST isozymes [34
] and the facts that GST could be induced by various hormones [36
may play different roles in different genders. While our study subjects were all female, the gender composition of their study was about half and half.
We also observed impacts of polymorphisms on NAT1
genes on the prognosis of NHL overall, DLBCL and CLL/SLL. There has been few studies linked NAT
genetic variations to NHL survival or prognosis. Active NAT1 overexpression was shown to enhance cell growth and etoposide resistance [38
], and NAT1
genotype was found to be associated with the risk of relapse or death among children with neuroblastoma [39
]. Although expression of NAT2 was recently found to alter the toxicity of CB1954, the component of an attractive therapy for tumor treatment [40
polymorphism studies among patients with breast cancer [41
], gastric cancer [42
] and colorectal cancer [43
] did not find any prognostic significance. Our study in NHL patients was the first one that observed NAT2
polymorphisms’ prognostic effects in cancer patients.
In order to investigate the generalizability of the study results, we compared its overall survival curve with that of 13,899 Female NHL patients aged 21-84 diagnosed during 1996-2000 at 17 Surveillance, Epidemiology and End Results (SEER) registries [44
]. The two survival curves were parallel throughout the follow-up period except during the first half year, during which 15.8% of SEER patients died, while none of the patients died in our study. Considering 167 out of 1,122 (14.9%) identified cases were not able to be enrolled in the CT study because they died before interview [45
], the overall survival of our case series is comparable to the survival observed by SEER. Our results might not apply to the most aggressive NHL cases with short-term survivals since there were limited cases in our study who died within a half year.
We utilized CTR to abstract follow-up information. Being the oldest tumor registry in the U.S. and functioning as one of the SEER registries since 1973, CTR is a reliable source for vital status, recurrence and occurrence of secondary cancer. According to the recently submitted SEER database (Nov 2007) [46
], among those microscopically confirmed female NHL patients diagnosed in 1996-2000 in CT and aged 21-84, 99.2% were actively followed by CTR through 12/31/2004. Several published studies using CTR as follow-up method have shown its validity [47
]. Through CTR, we were also able to obtain the information on tumor stage, B-symptoms, and initial treatment and adjusted them as a priori
confounders in our analysis. However, the treatment information collected by CTR is not so detailed and comprehensive, which limited our analysis because further exploration of the metabolism interactions of the investigated genes and specific drugs was not possible.
Compared to most of the published clinical reports on NHL survival, our study has a relatively larger sample size, which provided power to detect differences among NHL subtypes, especially for the two most common subtypes: DLBCL and FL. Although we did not find any significant effects of metabolism genetic polymorphisms on prognosis and survival for MZBL and T-cell lymphoma patients, it could be due to the small number of patients with these subtypes in our study. We also cannot rule out the possibility that some significant findings were due to chance.
In conclusion, this study shows that genetic polymorphisms in metabolism pathway genes of CYP, GST, and NAT may affect NHL prognosis and survival. They might be useful factors in prediction of NHL prognosis and treatment selection. A better understanding of the underlying biological mechanisms should be pursued.