|Home | About | Journals | Submit | Contact Us | Français|
Environmental factors are likely to interact with genetic determinants to influence prostate cancer progression. The Agricultural Health Study has identified an association between exposure to organophosphorous pesticides including chlorpyrifos, and increased prostate cancer risk in pesticide applicators with a first-degree family history of this disease. Exploration of this potential gene-environment interaction would benefit from the development of a suitable animal model. Utilizing a previously described mouse model that is genetically predisposed to prostate cancer through a prostate-specific heterozygous PTEN deletion, termed C57/Luc/Ptenp+/−, we used bioluminescence imaging and histopathological analyses to test whether chronic exposure to chlorpyrifos in a grain-based diet for 32 weeks was able to promote prostate cancer development. Chronic exposure to chlorpyrifos in the diet did not promote prostate cancer development in C57/Luc/Ptenp+/− mice despite achieving sufficient levels to inhibit acetylcholinesterase activity in plasma. We found no significant differences in numbers of murine prostatic intraepithelial neoplasia lesions or disease progression in chlorpyrifos versus control treated animals up to 32 weeks. The mechanistic basis of pesticide-induced prostate cancer may be complex and may involve other genetic variants, multiple genes, or nongenetic factors that might alter prostate cancer risk during pesticide exposure in agricultural workers.
Prostate cancer is the most commonly diagnosed visceral neoplasm and second leading cause of cancer death in U.S. males.1 Intensive efforts are underway to identify both the genetic and environmental factors underlying its progression from an indolent, subclinical disease to a life-threatening one. A number of studies have shown that farmers as a group have an elevated risk of prostate cancer possibly due to occupational exposures to various chemicals including pesticides and herbicides. The Agricultural Health Study (AHS), a very large prospective cohort study of farmers in Iowa and North Carolina, has reported an interesting association between exposure to a group of organophosphorous pesticides (OPs) and prostate cancer risk specifically in cases with a first-degree family history of this disease.2 In that report, four of the six pesticides are chemically related OP insecticides: chlorpyrifos (CPF), coumaphos, fonofos, and phorate. Additional studies provide detailed analysis of exposure levels to these agents including CPF.3 Furthermore, these associations were not observed among those without a family history of the disease. Thus, these findings suggest a potential gene-environment interaction experienced by agricultural workers to alter prostate cancer risk.
A potential candidate for this interaction may involve the phosphatase and tensin homologue deleted on chromosome ten (PTEN), phosphatidylinositol 3-kinase (PI3K) and protein kinase B (AKT) signaling pathway (PTEN-PI3K-AKT signaling pathway). Deregulation of the PI3K-AKT pathway is increasingly implicated in carcinogenesis (for review, see Ref. 4) and somatic mutations in PTEN, a tumor suppressor that acts as a brake on this pathway, have been frequently described in prostate cancer patients.5-7 However, whether OPs are able to cooperate with PTEN deletions and PI3K-AKT signaling to facilitate prostate carcinogenesis has not been investigated to date.
CPF, coumaphos, fonofos, and phorate are chemically related OPs that are widely used agricultural pesticides worldwide. CPF, for example, is used on major crops such as corn and soybeans that are abundant in the Midwestern states. Although discontinued for residential use in 2002 due to potential health concerns, the EPA estimates about three million kilograms of CPF is applied annually for agricultural purposes and is present in hundreds of commercial formulations (http://www.epa.gov/oppsrrd1/REDs/factsheets/chlorpyrifos_fs.htm). OPs are metabolically activated to an oxon form that is an irreversible inhibitor of cholinesterases and are thus potent neurotoxins.8 Extensive investigations have explored the potential for background exposure to CPF to cause adverse neurodevelopmental outcomes, but this remains unclear.9 Toxicology studies have generally shown that CPF is not genotoxic, mutagenic, or carcinogenic using standard assays (reviewed in Ref. 9). However, standard rodent cancer bioassays are not designed to detect prostate cancer, and it is likely these standard toxicologic assays do not account for interactions between these agents and genetic or epigenetic factors in humans that might combine to increase prostate cancer risk.
Going beyond associating an environmental exposure with a disease outcome to understanding the mechanistic basis for this association requires experimental approaches that may be difficult or impossible in humans. Therefore, the purpose of this study was to determine the effect of one of the most commonly used OPs worldwide, CPF, on prostate cancer development in a biologically relevant animal model. To this end we utilized a mouse model that is (i) genetically predisposed to prostate cancer based on a prostate-specific heterozygous deletion in the tumor suppressor Pten (C57/Luc/Ptenp+/−) and (ii) allows for bioluminescence imaging (BLI) of prostate tumorigenesis, facilitating the noninvasive detection of pathological changes in the prostate.10 In this study, we examined the effects of chronic dietary CPF exposure on prostate cancer development in C57/Luc/Ptenp+/− mice and we demonstrate the feasibility of using this mouse model for such a study.
All animal procedures were performed with approval from the University of Iowa Animal Care and Use Committee and by the authors’ Institutional Review Board. C57/Luc/Ptenp mice have been described previously.10 Age-matched cohorts of C57/Luc/Ptenp mice were generated using Jackson Laboratories JAX Speed Expansion Service. Mice were received from Jackson labs at five weeks old and genotyped for expression of PbCre4+ as previously described.10 At eight weeks of age, animals underwent BLI and were randomized into two groups to be placed onto their respective diets (10 mice for CPF and eight mice for control chow).
All BLI was performed in an IVIS100 imaging system (Caliper Life Sciences, Hopkinton, Massachusetts) as described previously.10 Briefly, mice were anesthetized in a chamber of 3% isoflurane and maintained under 3% isoflurane on a 37°C heated stage for the duration of the imaging and imaged 8 min following D-luciferin injection using a 20 cm field of view and an exposure time varying from 10 to 30 s. Bioluminescence intensity values were calculated by measuring photon flux (photons/second) in a region of interest surrounding the bioluminescence signal with the region of interest edge defined as 5% maximum signal using Living Image software (version 3.2).
CPF and control diets were formulated by Harlan Laboratories Teklad Diet using an original NIH-31 open formula irradiated base diet 7017. The NIH-31 open formula autoclavable diet is formulated for maintenance, growth, reproduction, and lactation of rats and mice and is used at the NIH as the standard reference diet for biological and biomedical research. 7017 is supplemented with additional vitamins to ensure nutritional adequacy after autoclaving. CPF solubilized in corn oil (2% final concentration) was added to the 7017 base diet to achieve a final concentration of 0.025% CPF (250 ppm) and oven dried for 2 h at 60°C. A control diet containing no CPF was formulated under identical conditions. Formulated diets were analyzed for CPF concentration by high-pressure liquid chromatography with a reference standard purchased from Chem Service at the University of Iowa hygienic laboratory.
Prostates were fixed in 4% paraformaldhyde overnight, transferred to 30% ethanol, embedded in paraffin, and stained with Hematoxylin & Eosin (H&E). Specifically, 5μM prostate sections were prepared stepwise through the entire lobe and consecutively stained and analyzed in order to capture changes throughout the entire length of the ducts. Prostate sections were imaged using an Olympus BX-61 microscope (Olympus). Prostate pathology was assessed according to the criteria established by Shappell et al.11
Activity of acetylcholinesterase (AChE) in plasma harvested from C57/Luc/Ptenp mice was calculated using a QuantiChrom acetylcholinesterase assay kit (Bioassay Systems, Harvard) according to manufacturer’s instructions. Briefly, plasma was isolated from whole blood via centrifugation (10 min at 13,000 rpm), following a submandibular cheek bleed of C57/Luc/Ptenp mice; 5 µl of plasma was used for subsequent AChE activity analysis.
All statistical analyses were performed using Student’s t-test.
In the C57/Luc/Ptenp mouse model, PTEN deletion is coupled to activation of firefly luciferase from the ROSA26 promoter in the same prostate epithelial cell, enabling BLI of prostate cancer tumorigenesis.10 Because bioluminescence intensity is tightly correlated with prostatic epithelial proliferation, BLI facilitates the noninvasive detection of pathological changes in the prostate. We have previously demonstrated that mice heterozygous for PTEN expression in the prostate (C57/Luc/Ptenp+/−) develop focal murine prostatic intraepithelial neoplasia (mPIN), which can be distinguished from normal prostate glands by small increases in bioluminescence signal intensity. Furthermore, in this model, mPIN lesions exhibit activation of PI3K-AKT signaling but never progress to invasive adenocarcinoma.10 We therefore reasoned that Ptenp+/− mice would provide a suitable model to determine whether chlorpyrifos exposure might accelerate the development of mPIN or promote the development of prostatic adenocarcinoma.
To determine the effect of chronic exposure of CPF on prostate cancer progression, we utilized a dietary route of administration of CPF. CPF has been evaluated in a number of chronic dietary exposure studies in mice at concentrations ranging from 0 to 250 ppm, corresponding to 31.7–55.1 mg/kg/d (reviewed in Ref. 9). In these studies, there was no increase in mortality associated with the diet at the highest exposure levels. Therefore, we formulated CPF chow at a target level of 250 ppm, which after testing by the University of Iowa hygienic laboratory was found to be ~220 ppm. We also formulated a control diet containing no CPF under identical conditions for consistency. Next, we generated cohorts of C57/Luc/Ptenp+/− mice and randomized them into two groups at eight weeks of age based on bioluminescence signal intensity. Mice were then placed on either a CPF diet or control diet ad libitum and monitored for general health status and prostate cancer progression via BLI (Fig. 1) and histology (Figs. 2 and 3). We found that mice fed CPF were not averse to the diet and consumed similar levels of food compared to control diet–fed mice (data not shown). However, we did detect an initial decrease in body weight of mice fed CPF at two weeks postdiet [Fig. 1(a)]. Following this initial decrease, CPF-fed mice became accustomed to the diet and consistently gained body weight for the duration of the study such that average body weights at the 32 week endpoint were almost identical [Fig. 1(a)]. CPF is metabolically converted by oxidative desulurfation into chlorpyrifos-oxon (CPO) and the insecticidal action of chlorpyrifos stems from inhibition of butyrylcholinesterase and acetylcholinesterase (AChE) by CPO, resulting in severe cholinergic toxicity.8 CPO is then inactivated by paraoxonase 1, and importantly, this enzyme exhibits relatively high activity in the C57Bl/6 strain.12,13 Therefore, to determine whether the exposure levels of CPF in C57/Luc/Ptenp mice is sufficient to elicit this response, we measured the activity of AChE in plasma collected from either CPF-fed or control diet–fed animals [Fig. 1(b)]. We found that the activity of AChE was reduced threefold in mice fed CPF compared to control diet–fed mice [Fig. 1(b)]. These data demonstrate that CPF is appropriately metabolized in vivo in C57/Luc/Ptenp mice and that exposure levels are sufficient to significantly inhibit AChE. We also monitored CPF-fed animals for clinical signs of toxicity throughout the duration of the study and found that 1 of 10 animals treated with CPF began to display signs of neurotoxicity at 28 weeks post–diet initiation.
To determine whether CPF was having an effect on prostate cancer progression, we performed BLI biweekly on CPF-fed and control diet–fed C57/Luc/Ptenp+/− animals [Figs. 1(c), 4(c), and 4(d)]. In both treatment groups, we detected a continuous increase in bioluminescence signal intensity consistent with asynchronous expression of PbCre4+ and the development of focal mPIN as previously reported.10 However, we did not detect any significant difference in bioluminescence intensity between CPF-fed and control diet–fed mice throughout the duration of the study, indicating that chronic exposure to CPF did not have an effect on epithelial proliferation or mPIN development in C57/Luc/Ptenp+/− animals for up to 32 weeks. To correlate our BLI findings to histopathology and to determine whether there were any underlying changes to microscopic prostate cancer progression such as invasion, we analyzed hematoxylin and eosin (H&E) stained prostate sections from CPF-fed and control diet–fed animals [Figs. 2(a) and 2(b)]. Specifically, prostate sections were prepared stepwise through the entire lobe and consecutively stained and analyzed in order to capture changes throughout the entire length of the ducts. The majority of prostate glands in both treatment groups appeared normal and were composed of a single layer of epithelial cells constrained within a basement membrane. However, we did detect discrete focal mPIN lesions and accompanying fibrosis in anterior prostate [Fig. 2(a)] and dorsal, lateral and to a lesser extent ventral prostate [Fig. 2(b)] in all of the animals analyzed, as previously reported.10 mPIN lesions were characterized by proliferation and stratification of epithelial cells displaying nuclear atypia according to the criteria established by Shappell et al.11 Quantitation of mPIN lesions from serial prostate sections of CPF-fed and control diet–fed animals revealed no significant difference in the numbers of mPIN lesions in CPF-fed versus control diet–fed C57/Luc/Ptenp+/− animals (Fig. 3). Furthermore, we did not detect any disease progression beyond mPIN in either of the treatment groups. Taken together, these data demonstrate that chronic exposure to CPF under the conditions used within this study is not able to promote the development of prostate cancer in a genetically predisposed animal model, C57/Luc/Ptenp+/− mice, and highlight that the ability of CPF to alter prostate cancer risk may involve other genetic alterations either in combination or independent of PTEN status.
How genetic determinants of prostate cancer may interact with pesticide exposure to alter prostate cancer risk is an important question in cancer epidemiology. In this study, we sought to determine the effect of chronic exposure to one the most commonly used insecticides worldwide, CPF, on prostate cancer development in mice genetically pre-disposed to prostate cancer via a prostate specific heterozygous PTEN deletion (C57/Luc/Ptenp+/−). Using both BLI and histopathological analyses, we found that chronic exposure to CPF in the diet for 32 weeks did not promote prostate cancer development in C57/Luc/Ptenp+/− mice despite being able to significantly inhibit AChE activity in plasma.
Initiated in 1993 and continuing to date, the AHS is the largest prospective cohort study undertaken to evaluate the relationship between agricultural exposures and disease.14 It involves 89,658 pesticide applicators and their spouses in Iowa and North Carolina enrolled between 1993 and 1997, representing 82% of the targeted population. Among the many findings in this study, prostate cancer was linked (significant interaction odds ratios) to exposure to six pesticides, among 50 evaluated, in patients with a first-degree family history of prostate cancer.2 Following the initiation of our studies, Koutros et al. reported the interaction among pesticide use, 8q24 variants, and prostate cancer risk using a nested case control study and found that insecticides, particularly OPs, were significantly linked to increased prostate cancer risk (estimated odds ratio).15 This was the first study to describe that the association between 8q24 SNPs and prostate cancer is modified by pesticide use. Several GWAS have identified genetic variants in the chromosomal region 8q24 in prostate cancer patients.16-19 Interestingly, the genetic variants of 8q24 are located in a region with no known protein coding gene but are located ~200 kb from the MYC gene, a well-established oncogene. Studies have demonstrated that pesticides can influence the expression of c-Myc,20 however, whether c-Myc acts as a putative oncogene responsible for pesticide-induced prostate cancer remains unknown.
Although GWAS have not implicated PTEN as a germ line risk allele for prostate cancer, abundant evidence indicates that somatic mutation in PTEN is a common event in prostate cancer progression.7,21-23 PTEN functions primarily as a lipid phosphatase that negatively regulates intracellular levels of phosphatidylinositol-3,4,5-trisphosphate (PIP3) and thereby functions as a tumor suppressor by inhibiting the PI3-AKT signaling pathway.24-26 Recent studies have indicated that this signaling pathway is altered in as many as 42% of primary tumors and 100% of metastases.27 Furthermore, prostate-specific deletion of PTEN in mice leads to prostate adenocarcinoma.28 Thus, we reasoned that activation of the PI3K-Akt signaling pathway, via deletion of PTEN, may act as a surrogate for the genetic association between OP exposure and prostate cancer risk. Because we have recently shown that disease progression is slow when the prostate-specific PTEN mutation is expressed in the C57Bl/6 background, and PTEN heterozygotes only sporadically develop pre-neoplastic mPIN lesions, we thought that this would be an ideal background to evaluate the effects of CPF on disease progression.10 Moreover, subclinical, latent prostate cancer is highly prevalent and prostate progression is generally slow, so that clinical presentation may be a product of progression, rather than initiation.29
We found that chronic exposure to CPF did not promote the development of prostate cancer in C57/Luc/Ptenp+/− mice up to 32 weeks. We terminated the study at 32 weeks based on the fact that we had anticipated changes in the bioluminescence signal to be evident at this time if significant disease progression had occurred based on our previous studies.10 Quantitative histologic analyses determined that CPF was not able to promote an increase in the number of mPIN lesions in prostates nor was it able to promote the progression from mPIN to microinvasive cancer or prostate adenocarcinoma. Although this exploratory study was not sufficiently powered to detect weak effects, with the current sample size the study would have 67% power to detect a cancer risk of 5% in controls versus 55% in chlorpyrifos using a one-sided Fisher’s exact test performed at the 5% level of significance, these data are the first to specifically evaluate the effect of CPF on prostate cancer in mice and provide initial evidence that the mechanistic basis of CPF-induced prostate cancer in humans may not solely involve PI3K-AKT activation. It is possible that the gene-environment interaction involving CPF is related to other genetic determinants such as myc as suggested by Koutros et al.15 Future experiments using relevant models such as the Hi-Myc mouse model of prostate cancer30 may aid in identifying a role for myc in this process. Alternatively, since we have only evaluated the active agent CPF here, it is possible that other constituents within the farm-grade formulations of OPs are involved in the observed gene-environment interaction. Furthermore, effects of chronic exposure to CPF may develop after longer latency such that mice may need to be exposed for much longer durations, as we have only evaluated its effects up to 32 weeks in this study.
Although our studies here are negative, we do not conclude that CPF or other OPs do not have effects on prostate cancer initiation or progression. We developed C57/Luc/Ptenp mice to assess how genetic and environmental factors might affect prostate cancer progression and demonstrate the feasibility of doing so with this model in this study. Prostate cancer is a complex disease and one can postulate a number of mechanisms by which OPs might influence disease progression. Toxicology studies have generally shown that CPF is not genotoxic or carcinogenic using standard assays (reviewed in Ref. 9). However, it should be noted that the mouse prostate is a difficult organ in which to observe carcinogenic effects since spontaneous prostate cancer is exceedingly rare and may exhibit long latency.31 However, CPF may have indirect effects on mutagenesis or carcinogenesis as they have been shown to induce reactive oxygen species by numerous studies (reviewed in Ref. 9). Other mechanisms by which organophosphorous chemicals may influence prostate cancer include endocrine disruption and interactions with the androgen receptor or potentially epigenetic effects.32-34
This work was supported by pilot grants from the University of Iowa Cancer and Aging Program (Grant No. 5 P20 CA103672), Center for Health Effects of Environmental Contamination, and Grants No. R21 CA137490, No. RC1 ES018097 (to MDH), No. P42 ES 013661(LWR), and No. P30 ES05605 (LWR). The contents are solely the responsibility of the authors and do not necessarily represent the official views of the granting agencies. We thank members of the Henry lab for comments on the manuscript, and Michael R. Miller for assistance with the figures.