In this nested case–control investigation of OC compounds and TGCT risk, we observed elevated prediagnostic serum concentrations oxychlordane, trans-nonachlor, and p,p- ′ DDE among TGCT cases compared with controls. We also observed significantly higher levels of PCBs 99, 138, 153, 167, 183, and 195 among seminoma cases versus controls as well as significantly lower levels than controls of PCBs 44, 49, and 52.
Chlordane is a cyclodiene insecticide commonly used in agriculture and termite control from 1945 until around 1970, at which time countries began banning or restricting its use. (Norway instituted a ban in 1968.) Our finding that TGCT cases had higher levels of oxychlordane and trans-
nonachlor than controls offers further support to previously published evidence suggesting that chlordanes affect the risk of TGCT development. We observed a particularly strong and statistically significant association with oxychlordane for seminoma cases. In the STEED study, statistically significant associations with TGCT were observed for cis
-nonachlor and trans
-nonachlor, whereas oxychlordane was associated with seminoma risk (McGlynn et al. 2008
). A significant association with cis
-non-achlor was also observed in the Swedish study (Hardell et al. 2003
). In addition, the case mothers in that study had significantly higher levels of cis
-nonachlor, and the sum of chlordanes compared with control mothers, suggesting that the antenatal period may be a particularly important time window of exposure for influencing TGCT development. No associations with chlordane congeners were observed in the Washington State study (Biggs et al. 2008
). The biologic basis for a chlordane effect on TGCT development is unclear. High doses of chlordane have been reported to have deleterious effects on testicular tissues, with reductions observed in testis weight, seminiferous tubule diameter, and spermatogenesis (Al-Omar et al. 2000
; Balash et al. 1987
). Findings from in vitro
studies suggest that the influence of chlordane on sex hormone signaling may be complex. There is evidence that chlordane is a weak agonist for human estrogen receptors α and β (Kojima et al. 2004
; Lemaire et al. 2006
) and a weak antagonist for the human androgen receptor (Kojima et al. 2004
; Lemaire et al. 2004
). However, chlordane may also exert antiestrogenic effects by inhibiting aromatase activity through estrogen-related receptor α1 antagonism (Yang and Chen 1999
′-DDE is the primary metabolite of the first modern synthesized insecticide, DDT. DDT was widely used in agriculture and vector control in the mid- to late 20th century until concerns regarding its environmental persistence and possible health effects led several developed countries to ban its use in the 1970s and 1980s (including Norway, in 1970). Our observed association with p,p
′-DDE is consistent with findings from the STEED study, the largest and only other prospective study of OC compounds and TGCTs (McGlynn et al. 2008
). No associations with this compound were observed in the case–control studies conducted in Washington State (Biggs et al. 2008
) and Sweden (Hardell et al. 2003
), although the latter study, with its small size, was underpowered to detect associations of moderate magnitude. Research suggests that high exposure to p,p
′-DDE may increase TGCT risk through the induction of antiandrogenic effects, as this compound has been shown to be a competitive antagonist for the androgen receptor (Danzo 1997
; Kelce et al. 1995
; Xu et al. 2006
Considerable interest has focused on the possible health effects of PCBs, a class of OC compounds used widely in the manufacturing of electrical equipment and other industrial applications until the institution of bans in the 1970s (banned in Norway in 1980). Experimental evidence suggests that some PCBs may exert estrogenic and possibly anti-androgenic effects and induce cytochrome p450 activity (Bonefeld-Jorgensen et al. 2001
; DeCastro et al. 2006
; Portigal et al. 2002
; Soto et al. 1995
; Wolff et al. 1997
). In our study, we did not observe a clear association between PCB levels and TGCT risk. However, seminoma cases had significantly elevated concentrations of PCBs 99, 138, 153, 167, 183, and 195 compared with controls. We also observed that seminoma cases had significantly lower levels of PCBs 44, 49, and 52 than controls. The associations with semi-noma risk observed for the moderate chlorination, mixed-type induction, and Wolff 1A, 2 and 2B groupings appear to reflect the associations with the aforementioned congeners.
We are not aware of any clear biologic basis for the observed differences in the direction of association with seminoma risk for these two groups. In fact, members of each group (PCBs 99, 138, 153, 183, 44, 49, and 52) have demonstrated similar biologic effects, namely, phenobarbital-type cytochrome p450 induction activity and estrogenicity. (Bonefeld-Jorgensen et al. 2001
; DeCastro et al. 2006
; Wolff et al. 1997
). The only other published findings regarding PCBs and TGCTs come from the Swedish case–control study (Hardell et al. 2003
); in that study, no difference in PCB levels between cases and controls was observed. In maternal comparisons, however, mothers of seminoma and nonseminoma cases had significantly higher levels of total PCBs and several individual congeners, including PCBs 99, 138, 153, 183, and 195 (PCB 167 was not measured). Maternal levels of PCB 52 were not found to be associated with reduced cancer risk (PCBs 44 and 49 were not measured). In summary, the published epidemiologic evidence, although limited, suggests that some PCB congeners may be associated with TGCT risk.
An important strength of this study is the use of serum samples collected prior to diagnosis, thus eliminating the possibility that our measurements were affected by the disease state. Given the rarity and young age at onset of TGCTs, most cohort studies are not feasible for investigations of this cancer. The Janus cohort is one of the few cohorts in the world with a large enough number of subjects recruited as young adults and with stored serum to make prospective serologic investigations of TGCTs feasible. Moreover, the time period during which the samples were collected (1972–1978) is particularly relevant for studying the health effects of OC compounds, as these chemicals were still in use or had only recently been banned at the time of collection.
The most notable limitation of this study is its small sample size, which greatly limited the statistical power to detect associations of moderate magnitude. As a consequence, we cannot rule out the existence of modest effects on TGCT risk from among the other investigated OC compounds. In addition, given the large number of measured compounds and generally borderline significance level of our observed findings, we cannot rule out the possibility that some of the observed associations may have arisen due to chance, although it is worth noting that associations with oxychlordane, trans-nonachlor, and p,p′-DDE have been previously reported. Moreover, some of the observed associations may have resulted from the moderate to strong correlations between many compounds. Finally, we were unable to investigate the relative importance of specific time windows of exposure to OC compounds, most notably the antenatal and early childhood periods, on TGCT risk.
The mean age at diagnosis and proportion of cases with seminoma histology in this study were higher than those observed in population-based cancer registries. This is likely a reflection of the older age distribution of cohort participants at baseline relative to the general population. Nonetheless, the seminoma findings should be interpreted with caution, in light of their small sample size and the fact that they were not the primary end point of this study.
In conclusion, this study provides additional but qualified epidemiologic evidence that serum levels of p,p
′-DDE, oxychlordane, and trans
-nonachlor compounds—and more speculatively, some PCB congeners (44 and 167 in particular)—may be associated with risk of TGCTs, including seminomas. These chemicals, although long banned in developed countries, are still a concern because of their continued presence in certain food sources (Fattore et al. 2006
; Freijer et al. 2001
; Schecter et al. 2001
) and the continued use of OC pesticides in developing countries (Gorman and Tynan 2003
; Mathur 1993
; Santilo et al. 1997
). Additional investigations into the relationship between TGCTs and exposure to OC compounds, particularly in early life, are warranted.