Testosterone and prostate cancer - a paradigm shift
27 June 2014
"Bye-bye androgen hypothesis, welcome saturation model"
A new era of testosterone and prostate cancer: from physiology to clinical implications. Khera M, Crawford D, Morales A, et al., Eur Urol 2014; 65(1): 115-23.
A long-held belief is that testosterone stimulates development of prostate cancer (PCa) and/or accelerates its growth. This summary gives an overview of an in-depth review of current literature regarding the relationship of serum testosterone and PCa and the effect of testosterone therapy on PCa progression and recurrence.1 Key studies which have refuted the old belief that testosterone has harmful effects on the prostate are presented, along the new testosterone-prostate paradigm known as the saturation model.
What is known - the androgen hypothesis
The idea that testosterone has detrimental effects on the prostate, the so called "androgen hypothesis" arose from two small studies in the 1940s in which men with metastatic PCa demonstrated clinical and biochemical improvement with androgen deprivation via castration or estrogen treatment, and conversely rapid PCa progression with testosterone administration.2, 3 Rather than concluding that exogenous testosterone attenuates the effects of surgical castration, the authors concluded that prostate cancer is activated by testosterone.18 Notably, these observations were made in a special population (castrated men) and are therefore not relevant to testosterone therapy in hypogonadal men.4
Medical students and doctors have since been taught that high testosterone levels promote the development of prostate cancer, that low testosterone is protective, and that the administration of testosterone to a man with existing prostate cancer is like "pouring gasoline on a fire."1 This fear is also the most common reason for doctors' reluctance to prescribe testosterone therapy, even in hypogonadal men19,20, which unnecessarily deprives many hypogonadal men of clinical benefits.
Although the dramatic effects of androgen deprivation therapy (ADT) in PCa are indisputable21, a large body of current evidence fails to support the concept that increasingly high levels of testosterone or DHT lead to ever-greater growth of benign or malignant prostate tissue (see below). It is critical to note that the androgen hypothesis was accepted prior to the discovery of the androgen receptor and PSA (prostate specific antigen), and before the availability of reliable serum testosterone assays. It should therefore not be surprising that some predictions of the androgen hypothesis would turn out to be false when submitted to rigorous scientific investigation.
What this study adds - the saturation model
It has been conclusively demonstrated that PCa risk is unrelated to endogenous serum androgen concentrations5, and several studies show no correlation between endogenous testosterone and PSA or prostate volume.22,23 Thus, men with higher endogenous testosterone are at no greater risk for PCa than men with lower serum testosterone.
The incidence of PCa during long-term (up to 20 years) testosterone therapy has been demonstrated to be equivalent to that expected in the general population.8 In healthy men, administration of supra-physiologic doses of testosterone (weekly injections of 500-600 mg of testosterone enanthate to healthy volunteers for up to 16 weeks) resulted in no increase in PSA nor prostate volume.24,25 In hypogonadal men treated with testosterone, levels of PSA typically rise up to levels of eugonadal men, but stay within the normal range.26 This elevation in PSA and prostate volume commonly occurs during the initial 3-6 months after initiation of testosterone therapy, and then stabilize, even with continued testosterone therapy.27-30
In line with these findings, two meta-analyses of testosterone therapy intervention studies, which focused on analyzing potential adverse effects of testosterone therapy, did not find any significant differences in prostate outcomes between testosterone therapy vs. placebo treated men.6, 7 Specifically, the most recent meta-analysis published in 2010 demonstrated no difference in the rates of prostate cancer, the need for prostate biopsy, international prostate symptom score (IPSS), increase in PSA, or total number of prostate-related adverse events when comparing the testosterone group with the placebo group.7
To explain these finding, the androgen hypothesis has been replaced by the saturation model.31 The saturation model explains the paradoxical observations that prostate tissue is exquisitely sensitive to changes in testosterone levels at low concentrations, but becomes insensitive to changes in androgen concentrations at higher levels.31 This response is consistent with the observation that testosterone exerts its prostatic effects primarily via binding to the AR, and that maximal testosterone-AR binding is achieved at testosterone levels well below the physiologic range.31
Changes in testosterone levels below the point of maximal testosterone -AR binding can elicit substantial changes in prostate cancer growth, as seen with castration, or with testosterone administration to castrated or hypogonadal men. In contrast, once maximal testosterone -AR binding is reached, further increasing testosterone levels results in little further effect. Thus, there is a threshold where increasing testosterone levels reach a limit (the saturation point) beyond which there is no further induction of androgen-driven changes in prostate tissue growth, see figure 1.
Figure 1: The saturation model. Increasing testosterone levels result in increased prostate tissue growth, as reflected by PSA elevation, until a limit is reached (the saturation point), beyond which further elevations in testosterone levels will not cause more prostate tissue growth.
From: Khera M, Crawford D, Morales A, Salonia A, Morgentaler A. A new era of testosterone and prostate cancer: from physiology to clinical implications. Eur Urol. Jan 2014;65(1):115-123.
Maximal testosterone-AR binding (i.e., saturation) occurs at fairly low androgen concentrations. In clinical practice, the saturation point appears to be approximately 8 nmol/L (230 ng/dL). However, there is inter-individual variation in the saturation point. Other physiologic mechanisms may contribute, as well. For example, in aging men with low testosterone levels, 6 months of testosterone therapy normalized serum testosterone levels but had little effect on prostate tissue androgen levels, suggesting the presence of local regulatory mechanisms.32 It should be noted that different tissues likely have different saturation points.
This explains why dramatic changes in PSA are noted when testosterone levels are treated into or out of the castration range, and when hypogonadal men get testosterone therapy treatment, whereas minimal or no PSA changes occur when higher doses are administered to most men. Thus, according to the saturation model, the initial modest PSA elevation and prostate growth within the reference range is a normal physiologic response to testosterone therapy in hypogonadal men. Therefore, testosterone-induced prostate growth and modest PSA elevations should not preclude hypogonadal men from testosterone substitution therapy.26
The long-held belief that prostate cancer risk is related to high testosterone levels (the androgen hypothesis) is not supported by clinical data. The saturation model and paradigm change that it brings to old inaccurate reasoning is that testosterone has a finite ability to stimulate prostate cancer growth.
The saturation model explains the paradoxical observations that prostate tissue is sensitive to changes in testosterone levels at low concentrations, but becomes insensitive to changes in testosterone levels at higher levels. Men with high testosterone levels are not at increased risk of developing prostate cancer, low testosterone levels provide no protection against the development of prostate cancer, and some men with untreated prostate cancer have received testosterone therapy without evidence of prostate cancer progression.33
Current evidence indicates that maximal testosterone-stimulated prostate cancer growth is achieved already at low sub-optimal testosterone levels. Provocative new research suggests that it is not high serum T that is problematic for PCa, but to the contrary that it is low serum T that is associated with worrisome cancer features and outcomes, in that androgens promote less aggressive PCa phenotypes and inhibit metastasis of established PCa.