INTRODUCTION

To date, genetic characterization of prognosis for soft tissue sarcoma has focused on somatic alterations in the tumors. However, as in other cancers, a number of candidate germline characteristics may affect survival. Candidates from molecular pathways important to the pathogenesis of soft tissue sarcoma include detoxification, DNA repair, and metabolic pathways.

Glutathione S-transferases (GSTs) interact with both endogenous and exogenous compounds, such as free radicals generated by normal physiologic processes and polycyclic aromatic hydrocarbons omnipresent in the environment, to produce less reactive metabolites.¹ GSTM1 and GSTT1 each have a null form that results in loss of the enzyme. GSTM1 has been associated with improved survival in acute myeloid leukemia² but not in advanced colorectal cancer,³ nor in children with acute lymphoblastic leukemia.⁴

DNA repair gene variants have been identified as prognostic markers in acute lymphocytic leukemia⁴ and in colorectal cancer.^5,6 Three (of many) DNA repair genes have variants that are common and have the potential to influence survival from soft tissue sarcoma: x-ray cross-complementing 1 (XRCC1), XRCC3, and Xeroderma pigmentosum group D (XPD). XRCC1 plays a role in the base excision repair pathway, interacts with DNA polymerase beta, polyadenosine ribose polymerase, and DNA ligase III. It has a structure characteristic of proteins involved in cell cycle checkpoint functions and responsive to DNA damage.⁷ XRCC3 participates in DNA double-strand break/recombination repair and is a member of the family of Rad-51–related proteins that participate in homologous recombination to maintain chromosome stability and repair DNA.^8,9 XPD is involved in the nucleotide-excision repair pathway and repairs lesions such as bulky adducts.¹⁰

Variants of genes, such as the aryl hydrocarbon receptor (AhR), that regulate the transcriptional expression of metabolic enzymes may also affect survival. The AhR belongs to the basic helix-loop-helix/PAS family of transcription factors and is a key regulator of the transcriptional expression of cytochromes P4501A1, 1A2, and 1B1.¹¹ Activation of AhR to a DNA binding form by exogenous xenobiotic ligands, such as 2,3,7,8-tetrachlorodibenzo-p-dioxin or polycyclic aromatic hydrocarbons (PAH), leads to the potent induction of these P450 enzymes, and human exposure to dioxin and dioxin-like chemicals has been associated with the etiology of soft tissue sarcoma.¹²

We sought to investigate the role that polymorphisms in detoxifying genes (GSTM1 and GSTT1), DNA repair genes (XRCC1, XRCC3, and XPD), and a transcription factor that regulates metabolic enzyme expression (AhR) may play in survival of individuals with soft tissue sarcoma.

PATIENTS AND METHODS

We identified adult patients (N = 120) admitted with a new diagnosis of soft tissue sarcoma between 1983 and the end of 1994 to the Department of Surgery at Memorial Sloan-Kettering Cancer Center (MSKCC; New York, NY), who had frozen normal tissue saved from their first operation. Clinical records supplied information on tumor characteristics. Vital status and cause of death were traced through February 2002. The MSKCC institutional review board approved this study conducted at MSKCC and the University of Turin in Italy.

DNA was extracted from frozen normal tissue. Polymorphisms in GSTM1 and GSTTI were detected with polymerase chain reaction (PCR)–based assays, first with gene specific primers (detection of one or two alleles) and then by multiplex PCR (verification of nulls in the presence of a housekeeping gene, ANDRR). Specific primers used were as follows: GSTM1 [forward 5′-tgccctacttgattgatggg-3′; reverse 5′-ctggattgtagc agatcatgc-3′], GSTT1 [forward 5′-TTCCCTTACCCATCATGACC−3′; reverse 5′-ACATTCCCAGCCTCACCTTA- 3′], ANDRR [forward 5′-GTGCGCGAAGTGATCCAGAA-3′; reverse 5′-TCTGGGACGCAACCTCTCTC-3′]. PCR followed by enzymatic digestion was used for the genotyping of the XRCC1-Arg399Gln, XPD-Lys751Gln, and XRCC3-Thr241Met polymorphisms.¹³ PCR followed by sequencing was used for genotyping the AhR-Arg554Lys polymorphism. Specific primers used were as follows: [forward 5′-ACCAGCCTCAGGATGTGAAC-3′; reverse 5′- GAATCTTGGACATACGTCAG-3′].

Survival was calculated as the time from diagnosis to death from sarcoma or until last contact if the patient was known to be alive. The Cox proportional hazards model was used to estimate the hazard ratio for variables singly and in combination. Tests for Hardy-Weinberg equilibrium were not significant.

RESULTS

Fifty-six patients were alive at the end of follow-up (median, 7.6 years), whereas 64 were deceased (median time to death, 2.4 years). Seventy-seven percent of subjects were white, with the other 23% distributed evenly among black, Hispanic, Asian, and Indian patients. The AhR variants were also more prevalent among nonwhite patients (39.3% v 19.5%). The median age at diagnosis was 55 years (range, 23 to 88 years; median age for male patients, 59.4 years; median age for female patients, 51.5 years). Anatomic sites were classified as lower extremity (n = 50), retro-interabdominal (n = 40), upper extremity (n = 10), visceral (n = 7), thoracic (n = 7), trunk (n = 5), and head and neck (n = 1). Histology was classified as liposarcoma (n = 44), leiomyosarcoma (n = 25), malignant fibrous histiocytoma (n = 21), fibrosarcoma (n = 12), and other (n = 18). These histologies are quite similar to the entire group (n = 1,975) of soft tissue sarcoma patients seen at MSKCC between 1983 and 1994.

The best predictor of survival in univariate analysis was histologic grade (Table 1). Median survival was 3.6 years for high-grade cancers and 6.9 years for low-grade cancers (log-rank test = 19.1; P < .0001). Survival was similar in the two sexes (median, 5.5 years for men and 6.3 years for women; log-rank test = 0.69; P = .40). No differences in survival were associated with the three DNA repair polymorphisms or the null genotypes of GSTM1 and GSTT1. Survival curves for these genotypes were largely overlapping, with P values (log-rank test) of .9.

DISCUSSION

We found that after accounting for the important clinical factors of grade, size of tumor, age, and anatomic site, a single nucleotide polymorphism in codon 554 of exon 10 (1661G>A) in AhR significantly and adversely affected survival (Fig 1). In multivariate analysis, we combined the heterozygote form (20%) with the homozygous mutant form (4.2%) because of small numbers. Had we treated them separately, the corresponding hazard ratios would have been 2.1 (95% CI, 1.2 to 3.8; P = .02) and 7.1 (95% CI, 2.0 to 24.9; P < .01). Subjects received differing forms of therapy (radiation or chemotherapy). As the AhR polymorphism was not associated with any or all forms of therapy, it is unlikely that the variant form was associated with survival because of drug resistance.

It is notable that the AhR variants are more common among nonwhite patients. Wong et al¹⁴ measured wide variation for the prevalence of this polymorphism among different populations, ranging from a high prevalence among Africans of 0.53 to 0.58 to a low prevalence among German white patients of 0.07. However, race was not significantly associated with survival in either the univariate or the multivariate model, indicating that the AhR variant is most likely affecting survival independent of race.

The phenotypic changes associated with this single nucleotide polymorphism in the human AhR are unclear. This polymorphism is located within the transactivation domain of the AhR, and thus changes in the regulation of downstream gene expression might be expected. In support of this idea, it has been demonstrated that this polymorphism does not affect the ability of the AhR to bind a xenobiotic ligand and to be activated to a DNA binding form.¹⁵ However, three studies involving white individuals have demonstrated that this polymorphism increases CYP1A1 and 1A2 activity, which would be consistent with a functional change in transactivation of gene expression. Notably, in all three cases, expression of a single allelic polymorphism (ie, heterozygote) was sufficient to observe the change in enzymatic activity.

Healthy nonsmoking volunteers exhibited significantly elevated CYP1A1 enzyme activity, as assessed by ethoxyresorufin-O-deethylase assay in peripheral-blood lymphocytes.¹⁶ These data would suggest that this polymorphism may increase the basal activity of CYP1A1, because the increase in CYP1A1 enzyme activity was not associated with exposure to a known exogenous xenobiotic ligand. In a second study, healthy young women smokers exhibited significantly higher CYP1A2 activity, as assessed by urinary caffeine metabolite ratios.¹⁷ These data would suggest that the polymorphism may increase the inducibility of CYP1A2 when AhR is activated by PAH found in cigarette smoke. This conclusion is supported by a third study in which peripheral-blood lymphocytes from healthy nonsmoking subjects exhibited significantly higher ethoxyresorufin-O-deethylase activity when exposed to a PAH in vitro.¹⁸

Induction of CYP1A activity is associated with bioactivation of procarcinogens to genotoxic electrophilic intermediates,¹⁹ and thus an AhR polymorphism that enhances basal or inducible CYP1A activity could theoretically increase the production of genotoxic metabolites and the incidence of cancer. In support of this idea, a constitutively active AhR leads to an increased incidence of stomach tumors in mice,²⁰ whereas loss of AhR expression results in resistance to PAH-induced skin tumors in mice.²¹ These studies would suggest that overactivation of AhR promotes the development of cancer, whereas loss of AhR may protect against the development of cancer.

It must also be noted, however, that conflicting evidence exists as to the functional effects of the codon 554 polymorphism. Studies involving both French and Japanese individuals expressing this AhR polymorphism have failed to uncover any phenotypic changes in CYP1A regulation or any association with lung cancer, commonly associated with elevated pulmonary CYP1A2 activity.^22,23 The nucleotide change in codon 554 results in a conservative amino acid substitution of a Lys for an Arg. Although the wild-type Arg residue is conserved among human, rat, and mouse AhR sequences, the polymorphic Lys residue is found at this codon in other mammalian species, including guinea pig and hamster. Thus a phenotypic functional change associated with this conservative amino acid substitution might not be expected. More recently, it has been demonstrated that two additional polymorphisms in the human AhR exhibit apparent linkage disequilibrium with the codon 554 polymorphism.¹⁷ Thus it is possible that the codon 554 polymorphism might not itself be functionally significant, but rather could be linked in white individuals to other polymorphisms that contribute to a functional change.

In sum, the major finding of this study is that germline codon 554 AhR polymorphism affects the survival of individuals with soft tissue sarcoma. However, further study is needed to reveal the functional mechanism by which this polymorphism contributes to the poor prognosis of this cancer.

Authors’ Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.