Abstract
I read with interest the article by Sistare et al. (2011), “An Analysis of Pharmaceutical Experience with Decades of Rat Carcinogenicity Testing: Support for Proposal to Modify Current Regulatory Guidelines” (Toxicologic Pathology, volume 39, pp. 716–44), which appeared in the Regulatory Forum. Although I am in agreement with the general principles of this well-written article, it is necessary to point out that one of my papers referenced in this article, and the conclusions stated to be from it, are incorrect. Indeed, the statement in Sistare et al. (2011) is the complete antithesis of my arguments as given in Harvey (2005) and elsewhere.
Sistare et al. (2011) quote Harvey (2005) and state, “In the rat, prolonged hyperprolactinaemia, as induced by anti-dopaminergic compounds, can result in the development of pituitary and mammary gland tumours. Because of major differences in hormonal and reproductive physiology between rodents and humans, this is considered to be a rodent specific mechanism (Gopinath 1995; Harvey 2005).” In fact, the Harvey (2005) reference is incorrectly described and quoted, and indeed this statement in Sistare et al. (2011) is the antithesis of my arguments on the subject of prolactin, mammary carcinogenesis in rodents, and relevance to human risk extrapolation and breast cancer, as developed in Harvey (2005) and other publications (Harvey, Everett, and Springall 2006, 2008; Harvey and Everett 2006; see also Peveler et al. 2008). As my arguments have been so significantly misattributed, I feel that it is necessary to both correct this and provide brief background on the subject.
The whole purpose of my article (Harvey 2005, as cited) was to point out that prolactin-induced mammary carcinogenesis was not species specific to rodents and in fact was a mechanism that was highly relevant to humans. The epidemiology and human clinical experience over the past decade clearly supports the view that prolactin has a major role in human breast cancer development and progression in women. Data originally from the U.S. Nurses Health Study shows increased breast cancer risk in postmenopausal women associated with higher prolactin levels (Hankinson et al. 1999). The original findings have been further researched and extended to premenopausal women (e.g., Tworoger et al. 2004; Tworoger and Hankinson 2006, 2008; Tworoger, Sluss, and Hankinson 2006; Tworoger et al. 2007; Eliassen, Tworoger, and Hankinson 2007), and it is now generally agreed that prolactin is associated with human breast cancer to the point that upper quartile reference range serum prolactin concentrations are associated with an approximate doubling of risk compared with lower quartile concentrations. These findings of a role of prolactin in human breast cancer dispel the myth that prolactin-induced mammary carcinogenesis is species specific to rodents as employed in carcinogenicity studies and have gained wide acceptance (e.g., Clevenger 2003; Clevenger et al. 2003; Clevenger, Gadd, and Zheng 2009; Jacobson et al. 2011). Furthermore, among other developments since the initial results from the U.S. Nurses Health Study was published, prolactin has been shown to be correlated with mammographic density (Greendale et al. 2007), a known important correlate of breast cancer risk.
The biological and molecular basis for the action of prolactin in breast cancer is also well researched, and prolactin has a proliferative, mitogenic, and antiapoptotic action; modulates cell cycle regulators; and stimulates human tumor cell growth (e.g., Vonderhaar 1989; Das and Vonderhaar 1997; Ben-Jonathan et al. 2002; Schroeder, Symowicz, and Schuler 2002; Liby et al. 2003; Perks et al. 2004). The data are so clear for a role of prolactin in human breast cancer that the prolactin receptor is being evaluated as a new therapeutic target for breast cancer treatment (Goffin et al. 2005, 2006; Jacobson et al. 2010; Carver, Arendt, and Schuler 2009; Scotti et al. 2008; Howell et al. 2008; Clevenger et al. 2008). Collectively, the epidemiological, clinical, biological, and pharmacological data show that prolactin acts as a tumor promoter in humans and rodents alike, through similar mechanisms. This has now been recognized in clinical guidance: the potential role of prolactin in breast cancer, and the risk that antipsychotic dopamine antagonist drugs that induce hyperprolactinaemia may have on patients, has led to the production of clinical recommendations (Peveler et al. 2008) in which hyperprolactinaemia has been acknowledged as a breast cancer risk factor (see also Bostwick, Guthrie, and Ellingrod 2009). As antidopaminergic (dopamine antagonist) drugs were specifically mentioned by Sistare et al., (2011), it is worth pointing out that there is evidence that they have been associated with increased breast cancer risk, too (Wang et al. 2002).
Sistare et al. (2011) go on to state that “a prolactin-mediated mechanism effect was concluded to be the mechanism for tumorigenesis for rat TP compounds 15, 16, 45, 47 and 50.” Although apparently not the most common of endocrine mechanisms of tumorigenesis in their analysis, prolactin-induced mammary tumorigenesis is not uncommon either, and it is therefore important, when providing an authoritative position paper, that the authors are not only fully aware of the literature but also correctly quote the limited references that they do cite. The evidence is now strong and consistent that there is a significant involvement of prolactin in human breast cancer, and of most importance, the mechanism is analogous to prolactin-induced mammary tumorigenesis in rodents. Therefore, rodent carcinogenicity bioassays are relevant and predictive of this response in humans. Contrary to the contention of Sistare et al. (2011) that mammary tumors induced in rodents by prolactin are considered species specific, current evidence suggests that this claim can no longer be supported in toxicology. For more up-to-date reviews of the toxicological and risk assessment considerations associated with hyperprolactinaemia, see Harvey, Everett, and Springall (2006, 2008).
Finally, there are many risk factors for human breast cancer (estrogen window as influenced by age at menarche, menopause, and parity; obesity; age; diet; hormone replacement therapy), and the many different contributing etiological factors provide differing risk weightings and predispose to different tumor phenotypes. Not all authors consider prolactin to be a carcinogenic risk factor (Goodman and Bercovich 2008), and although very much a minority opinion, it may be that there are subtle interactions between various hormonal and molecular mechanisms such that hyperprolactinaemia may be a risk in a subpopulation of women whereas other cohorts are resistant. However, the evidence to date is consistent that elevated prolactin is associated with breast cancer risk in women, and this should be the foremost consideration in interpreting the outcome of regulatory carcinogenicity studies, until such time as further research does fully identify the high-risk subpopulations and combinations of all etiological factors and interactions at work.