madman
Super Moderator
ABSTRACT
Background: An organic etiology underpinning post-finasteride syndrome, a constellation of persistent sexual, neuropsychiatric, and somatic symptoms reported by men exposed to 5-alpha-reductase inhibitors (5ARIs), is debated. Persistent changes in neurosteroid levels or androgen receptor expression have been implicated.
Aim: To determine whether differences in gene expression, especially in relevant biologic pathways, exist between patients reporting post-finasteride syndrome symptoms and healthy controls.
Methods: This was a single-center, prospective case-control study taking place between March 2013 and September 2018. Men 18 years and older being evaluated for sexual dysfunction (study) or circumcision (control) were eligible for inclusion. Twenty-six men with a history of 5ARI use reporting symptoms consistent with post-finasteride the syndrome were included in the patient group. Twenty-six men consented to inclusion in the control group.
Outcomes: The primary outcome measure is gene expression data for genes affecting neurosteroid levels and androgen receptor activity from penile skin cells.
Results: Gene expression of cells from penile skin samples from twenty-six men of median age 38 years (IQR, 33-42) in the study group was compared with that from twenty-six men of median age 41 years (IQR, 35-62) in the control group (P = .13), with 1,446 genes significantly over-expressed and 2,318 genes significantly underexpressed in study patients. Androgen receptor expression was significantly higher in study patients compared to controls (9.961 vs 9.494, adjusted P value = .01). Serum levels of androgen receptor activity markers 5a-androstanediol (0.950 ng/mL [0.749-1.587] vs 0.949 [0.817-1.337], P = .34) or 3a-androstanedione (3.1 ng/mL [1.925-5.475] vs 6.7 [3.375-11.4], P = .31) revealed no significant differences. No significant differences were found between the number of trinucleotide repeats (21.5 [20-23.75], 22 [19-25], P = .94).
Clinical Implications: In this study, we present evidence of gene expression correlating with observed biologic differences in patients with post-finasteride syndrome; providers who prescribe 5ARIs should be aware and advise their patients accordingly.
Strengths & Limitations: Strengths of this study include the evaluation of multiple proposed etiologies for the post-finasteride syndrome. The study is also strengthened by the fact that not all data matched the initial hypotheses, qualifying the argument for the existence of PFS. Limitations include potential selection bias arising from more severe phenotypes seeking care; lack of gene expression data prior to 5ARI exposure; lack of non-penile tissue samples supposedly involved, and a lack of mechanistic data to imply causality.
Conclusion: This study is the first to consider and demonstrate gene expression differences in patients with PFS as a potential etiology of sexual dysfunction.
INTRODUCTION
5-alpha reductase inhibitors (5ARIs) prevent the reduction of steroids including testosterone, progesterone, androstenedione, epitestosterone, cortisol, aldosterone, and deoxycorticosterone by various isoforms of the enzyme 5-alpha reductase.1 Inhibition of testosterone reduction in particular results in a decrease in levels of dihydrotestosterone (DHT), and can therefore result in higher testosterone levels.2 This effect explains the utility of 5ARIs in certain androgen-related disorders. Although 5ARIs can be helpful in cases of benign prostatic hyperplasia (BPH) or androgenic alopecia (AGA), they come with several known side effects affecting sexual function, including erectile dysfunction (ED) and low libido.3 These effects may persist even after cessation of medication.4
*The full range of consequences of 5AR inhibition is not yet known, and other side effects are increasingly reported. In addition to ED and low libido, sexual effects including penile atrophy diminished ejaculatory volume and force, and an increased incidence of Peyronie’s disease have also been reported.5-7 Adverse effects outside the sexual domain include physical, neurological, and psychiatric concerns. This constellation of symptoms occurring in men who are exposed to 5ARIs and experience these symptoms even after discontinuation of treatment is known as post-finasteride syndrome (PFS).8 Previously, we cataloged the symptoms reported by patients who took 5ARIs for androgenic alopecia as part of PFS.9 In addition to penile vascular abnormalities and voiding dysfunction, these included changes in body composition, disturbances in memory and attention, and increased rates of depression and anxiety. Of note, subjective changes in body composition including muscular atrophy and increased fat deposition have not been confirmed by physical examination. Outside of our study population, others have reported possible effects of 5AR inhibition on metabolism (including increased insulin resistance, alteration of fat distribution, and cardiovascular disease) and bone physiology.10
To date, the existence and biological basis of PFS has been questioned, with some positing a functional or psychogenic etiology.11 One study reported persistent sexual symptoms in patients on a placebo treatment, while another noted higher reported rates of side effects in patients who were explicitly educated on them.12-14 However, there is evidence to support an organic etiology. One possibility is decreased levels of neurosteroids in patients with PFS. One study measured plasma and cerebrospinal fluid levels of neurosteroids, including progesterone, testosterone, and their metabolite hormones, in patients with PFS, finding them to be significantly lower compared to healthy controls even years after discontinuation of 5ARI therapy.15 While this would explain the emergence of sexual, cognitive, and psychiatric symptoms as well as their long-term persistence, it remains unclear why levels would remain low if 5A reduction was not actively inhibited. In addition, changes to neurosteroid levels alone may not adequately explain other PFS symptoms such as body composition changes and genitourinary issues.
Another proposed etiology for the development of PFS is as a response to androgen deprivation mediated by upregulation in androgen receptor (AR) expression. The AR is expressed in multiple tissues across organ systems including reproductive, genitourinary, nervous, musculoskeletal, cardiovascular, and immune.16 Given that gene overexpression has been implicated in numerous disease states, AR overexpression in response to an androgen-deficient state may negatively affect multiple tissues throughout the body.17-18 Upregulation in AR expression in penile skin has been demonstrated in patients with persistent sexual side effects following finasteride discontinuation.19 Overexpression of AR leading to changes in neurosteroid synthesis and action has been suggested as an explanation for certain behavioral phenotypes in mice.20 In addition, polymorphisms in CAG (polyglutamine-encoding) repeat length has been inversely associated with transcriptional function.21 Together, overexpression combined with less functional activity may be responsible for certain PFS symptoms, possibly through effects on neurosteroid levels but also through induced tissue toxicity.
RNA microarray studies have been used to assess the conditions of an organism and ascertain whether potential disease states can be attributable to differential gene expression.22 If we can identify processes that affect neurosteroid levels in a way that explains the findings of lowered levels in CSF; significant differences in AR expression or function; or any other biological processes that are differentially expressed, then the case for a biological etiology of PFS is strong. In the present study, we compared gene expression data and parameters relating to AR expression and the function between men with a history of 5ARI use and reported PFS symptoms with healthy controls. We hypothesize that gene expression in men with PFS symptoms will demonstrate differential expression compared to healthy controls.
Microarray Analysis
Penile skin samples from 52 patients yielded an analysis of 46,204 gene symbols. After false discovery rate correction, 1,446 genes were significantly overexpressed and 2,318 genes were significantly under-expressed in study patients. The fold change for overexpressed genes ranged from 1.105 to 5.453, while for underexpressed genes it ranged from -1.095 to -11.236. Of particular interest, we found AR expression to be significantly higher in study patients compared to controls (adjusted P-value = .01).
Pathway Analysis
Pathway analysis on both the over-and under-expressed gene sets yielded fourteen and fifteen clinically relevant clusters, shown in Figure 1 and Figure 2, respectively. Of note, the levels of significance are much higher for the under-expressed genes, indicating it is more likely that these biological systems are affected by differential gene expression in our study population. As we present these results, we will refer to clusters derived from the over-expressed gene set as “up-regulated,” while clusters derived from the under-expressed gene set will be termed “downregulated.”
Sexual Effects: Penile Vascular and Soft Tissue Health
We previously reported that the majority of our study patients suffered from confirmed penile vascular abnormalities on penile Doppler ultrasound, including arterial insufficiency, possible ED, and venous leak. In this study, we found that pathways including tissue migration, angiogenesis, and vasculature development were up-regulated. Conversely, pathways including blood vessel development, angiogenesis, and positive regulation of epithelial and endothelial cell migration and proliferation were down-regulated.
In addition to vascular issues, soft tissue components of penile anatomy may be affected in PFS. Peyronie’s disease, along with changes in penile length and testicular size, is reported in patients who have taken 5ARI. We found that pathways involving extracellular matrix regulation (including the “matrisome”), cell junction organization, and connective tissue health were down-regulated.
Genitourinary Effects: Voiding
We previously presented evidence that the immune system involvement and inflammation, bolstered by the low DHT-environment seen in 5ARI use, may lead to the development of benign prostatic hyperplasia and subsequent voiding symptoms in PFS. In this study, we observed that pathways controlling Tcell development, proliferation, and function, along with pathways involving cytokine signaling, were upregulated in the study patients.
Neurological Effects: Neuro-steroids
Although neuro-steroids levels should return to normal following discontinuation of 5ARI, we identified persistent changes in gene expression in study group patients that may explain the preservation of changes in levels of progesterone, testosterone, and their metabolites reported by Caruso et al.15 Upregulated clusters include “organic hydroxy compound biosynthetic process,” “regulation of small molecule metabolic process,” and response to a steroid hormone. Processes involving the regulation of cholesterol synthesis were particularly involved. Pathway analysis did not reveal a significantly down-regulated pathway involved in steroid metabolism or regulation.
Neurological Effects: Nervous Tissue Maintenance
Separately from the effects of neuro-hormones, processes controlling nerve cell health appear to be impacted by 5ARI exposure. An upregulated cluster identified as “kinase and transcription activation” further specified as involving processes controlling neurogenesis and immune responses are upregulated in study patients. However, transforming growth factor-beta and bone morphogenic protein signaling, along with several pathways including neuron differentiation, neurogenesis, and axonogenesis, are pathways that are down-regulated. In addition, the actin cytoskeleton organization cluster was down-regulated. Notably, down-regulated pathways exhibit much higher levels of significance.
Physical Effects: Musculoskeletal and Metabolic Changes
Many of our patients reported musculoskeletal complaints, including fatigue, muscle atrophy, and joint pain, along with skin changes and visual disturbances. Pathway analysis revealed upregulation in pathways affecting skin development, epidermis development, and “establishment of the skin barrier”. Ossification was also upregulated. Down-regulated clusters included skeletal system development, muscle structure development, sensory organ development, sensory organ (visual) development, and connective tissue development.
Altered steroid metabolism can affect metabolic health, and may lead to issues with insulin resistance, increased fat deposition, and cardiovascular disease.10 We observed upregulation in pathways affecting insulin (within a cluster called “regulation of establishment of protein localization”) including positive regulation of peptide secretion, regulation of insulin secretion, and response to carbohydrates. Similarly, we observed upregulation in pathways controlling homeostasis and responses to stress, which both were found to be related to metabolic processes on further analysis of relevant annotations. Down-regulated clusters included circulatory system process (involving processes like cardiac muscle contraction and ion transport); “cellular response to growth factor stimulus” specifically referring to the TGF-beta signaling pathways involving bone morphogenic protein, which regulate osteogenesis, growth, and homeostasis;28 and “transmembrane receptor protein tyrosine kinase signaling pathway,” which included pathways involved in the cellular response to insulin.
Androgen Receptor
Data evaluating AR parameters are reported in Table 3. Per microarray data, AR was overexpressed in study patients. However, neither the number of CAG repeats nor the levels of 5- alpha-androstanediol and 3-alpha-androstanedione glucuronide were significantly different between the study and control groups. For study patients, the median length of CAG repeats was 21.5 (IQR, 20-23.75), while for controls, the median length was 22 (IQR, 19-25). The median level of 5-alpha-androstenediol was 0.950 ng/ml (IQR, 0.749-1.587 ng/ml) in study patients and 0.945 ng/ml (IQR, 0.817-1.337 ng/ml) in controls. The median level of 3-alpha-androstanedione glucuronide was 3.1 ng/ml (IQR, 1.925-5.475 ng/ml) in study patients and 6.7 ng/ml (IQR, 3.375-11.4 ng/ml) in controls.
DISCUSSION
In the present study, we explored the hypothesis that PFS is underpinned by a biological mechanism triggered by the use of 5ARIs and the resultant androgen-deficient state. We identified genes that are differentially expressed in penile skin tissue between men with a history of 5ARI use and PFS symptoms and healthy controls, identifying biological pathways that were upregulated or inhibited and may be relevant in the development of PFS symptoms. We proposed that these effects are mediated largely through effects on neurosteroid levels or changes in AR expression leading to downstream effects on sexual, physical, neurological, and psychiatric functioning. In addition, we identified a number of other pathways involving differentially expressed genes that may also contribute to PFS symptoms.
Neurosteroids
When 5A-reduced steroids such as DHT are 3A-reduced, they become 3-alpha,5-alpha neurosteroids. These molecules are steroid hormone metabolites that affect inhibitory or excitatory function in the central nervous system through action on neuronal membrane receptors.29 In addition, they can modify gene expression through interaction with intracellular steroid receptors in nervous tissue.30 They play a role in several important functions including mental health (depression, anxiety, stress), cognition (learning, memory), and nervous system plasticity.31-35 Both inhibitory and excitatory neurosteroids ultimately have antidepressant, anxiolytic, neuroprotective, neurogenic, and cognition-enhancing effects.36-37 5ARIs may induce cognitive and psychiatric side effects through decreased neurosteroid levels by preventing steroid hormones from metabolizing through reduction.
We identified genes involved in pathways that may affect neurosteroid levels which were upregulated. Considered alone, this would appear to be at odds with the findings reported by Caruso et al.15 However, it is possible that this particular alteration in gene expression may represent a feedback loop attempting to drive synthesis to replace depleted hormone levels, including progesterone, testosterone, and their reduced neuro-active metabolites, in response to 5AR inhibition. However, to better understand the likely effect of 5ARI exposure in our population of PFS patients, we searched for other gene expression changes that could further elucidate the mechanisms involved. We considered three ways for neurosteroid action to be affected: (i) changes in pathways affecting steroid hormone metabolism; (ii) changes in pathways affecting neuronal membrane receptors; and (iii) changes in pathways affecting the androgen receptor, which we found to be significantly over-expressed in PFS patients. Using a list of genes of interest across these categories that we constructed prior to conducting the experiment, we cross-referenced our microarray results to identify whether genes involved with these receptors were differentially expressed in our study population.
Androgen Receptor
The AR itself was over-expressed in PFS patients. Overexpression of normal genes has diverse pathologic consequences across a variety of tissues, leading to certain neurodegenerative diseases, fibrosis, diabetes, and cancer.17 Since patients in this study did not have an elongated CAG polymorphism, we can consider their AR gene to be normal. Overexpression of AR in penile tissue may be responsible for sexual symptoms experienced by PFS patients.19 If AR is also overexpressed in other tissues, for example, nervous tissue, it may play a role in cognitive and psychiatric symptoms as well.
*Depending on the interaction with the previously discussed genes, AR activity could be increased, decreased, or balanced out in PFS patients. If it is higher than in controls, this may be a result of the body attempting to correct for the decreased level of neurosteroids available. If it is lower than in controls, it could be another contributor to antiandrogenic symptoms in PFS. The fact that AR expression was elevated in PFS patients suggests a chronic androgen deficient, or activity-deficient, state. However, the lack of a difference in androgen activity markers indicates the opposite conclusion. Since expression data came from specific tissue, it is possible that there are differing site-specific expression patterns in patients. Ultimately, whether the differentially expressed genes affecting AR activity play a role in PFS remains unclear given the lack of mechanistic data, and requires investigation pointed in this direction.
Other Contributing Pathways
In addition to the effects of decreased neurosteroid activity, changes to steroid levels, and AR overexpression, it is possible that changes to other biologic mechanisms may be present in PFS patients contribute to the syndrome. Results of pathway analysis indicated the involvement of biological processes affecting the sexual, genitourinary, neurological, musculoskeletal, cardiovascular, metabolic, and immune systems. From our data, it is unclear to what extent neurosteroids and AR are responsible for changes in expression for genes affecting these pathways.
*Our data revealed differential expression of many genes in pathways controlling gene expression itself. This may be due in part to the fact that steroid hormones act by inducing changes in gene expression or by AR activity, and may also explain the nature of the persistent effects of 5ARI exposure. Namely, exposure to 5ARI could lead to permanent changes in genetic expression through unknown mechanisms which cause the changes we were able to measure and analyze. However, further work would need to be done to explore these mechanisms and is beyond the scope of the current study.
Strengths and Weakness
There are several strengths to this study, including the use of advanced gene expression and pathway analysis and gene expression data from a large cohort of cases and controls. This study is the first to evaluate multiple proposed etiologies for PFS and provide biological evidence for each. In addition, this study is strengthened by the fact that not all data matched the initial hypothesis, qualifying the argument for the existence of PFS as a unified clinical entity. This study also has weaknesses and limitations. We base the objective existence of symptoms partially on subjective survey data. Selection bias may arise from only the more severe phenotypes seeking care. Gene expression data prior to 5ARI exposure is unavailable for comparison and thus we are less certain that differences in expression between study and control groups are attributable to 5ARI exposure. Further, we lack tissue samples from supposedly involved tissues (ie CNS tissues), ultimately limiting the conclusions that can be drawn from both the microarray and AR expression data. Finally, there exists a lack of mechanistic data to imply causality for proposed mechanisms; gene expression changes alone do not necessarily translate to concordant changes in protein expression or protein activity.
CONCLUSION
This study is the first to consider gene expression differences in men with PFS in explaining the etiology of this condition. Given gene expression per se is not mechanistic and does not imply causality, experiments with downstream processes of protein expression and activity should be undertaken to provide mechanistic data and clarify the results of this work. Further investigation should also explore upstream processes including the mechanisms regulating gene expression in the setting of PFS and identify risk factors for individuals, with a potential focus on genetic risk factors. At this time, patients should be informed regarding possible side effects of 5ARI that may persist even following discontinuation of treatment as part of their counseling.
Background: An organic etiology underpinning post-finasteride syndrome, a constellation of persistent sexual, neuropsychiatric, and somatic symptoms reported by men exposed to 5-alpha-reductase inhibitors (5ARIs), is debated. Persistent changes in neurosteroid levels or androgen receptor expression have been implicated.
Aim: To determine whether differences in gene expression, especially in relevant biologic pathways, exist between patients reporting post-finasteride syndrome symptoms and healthy controls.
Methods: This was a single-center, prospective case-control study taking place between March 2013 and September 2018. Men 18 years and older being evaluated for sexual dysfunction (study) or circumcision (control) were eligible for inclusion. Twenty-six men with a history of 5ARI use reporting symptoms consistent with post-finasteride the syndrome were included in the patient group. Twenty-six men consented to inclusion in the control group.
Outcomes: The primary outcome measure is gene expression data for genes affecting neurosteroid levels and androgen receptor activity from penile skin cells.
Results: Gene expression of cells from penile skin samples from twenty-six men of median age 38 years (IQR, 33-42) in the study group was compared with that from twenty-six men of median age 41 years (IQR, 35-62) in the control group (P = .13), with 1,446 genes significantly over-expressed and 2,318 genes significantly underexpressed in study patients. Androgen receptor expression was significantly higher in study patients compared to controls (9.961 vs 9.494, adjusted P value = .01). Serum levels of androgen receptor activity markers 5a-androstanediol (0.950 ng/mL [0.749-1.587] vs 0.949 [0.817-1.337], P = .34) or 3a-androstanedione (3.1 ng/mL [1.925-5.475] vs 6.7 [3.375-11.4], P = .31) revealed no significant differences. No significant differences were found between the number of trinucleotide repeats (21.5 [20-23.75], 22 [19-25], P = .94).
Clinical Implications: In this study, we present evidence of gene expression correlating with observed biologic differences in patients with post-finasteride syndrome; providers who prescribe 5ARIs should be aware and advise their patients accordingly.
Strengths & Limitations: Strengths of this study include the evaluation of multiple proposed etiologies for the post-finasteride syndrome. The study is also strengthened by the fact that not all data matched the initial hypotheses, qualifying the argument for the existence of PFS. Limitations include potential selection bias arising from more severe phenotypes seeking care; lack of gene expression data prior to 5ARI exposure; lack of non-penile tissue samples supposedly involved, and a lack of mechanistic data to imply causality.
Conclusion: This study is the first to consider and demonstrate gene expression differences in patients with PFS as a potential etiology of sexual dysfunction.
INTRODUCTION
5-alpha reductase inhibitors (5ARIs) prevent the reduction of steroids including testosterone, progesterone, androstenedione, epitestosterone, cortisol, aldosterone, and deoxycorticosterone by various isoforms of the enzyme 5-alpha reductase.1 Inhibition of testosterone reduction in particular results in a decrease in levels of dihydrotestosterone (DHT), and can therefore result in higher testosterone levels.2 This effect explains the utility of 5ARIs in certain androgen-related disorders. Although 5ARIs can be helpful in cases of benign prostatic hyperplasia (BPH) or androgenic alopecia (AGA), they come with several known side effects affecting sexual function, including erectile dysfunction (ED) and low libido.3 These effects may persist even after cessation of medication.4
*The full range of consequences of 5AR inhibition is not yet known, and other side effects are increasingly reported. In addition to ED and low libido, sexual effects including penile atrophy diminished ejaculatory volume and force, and an increased incidence of Peyronie’s disease have also been reported.5-7 Adverse effects outside the sexual domain include physical, neurological, and psychiatric concerns. This constellation of symptoms occurring in men who are exposed to 5ARIs and experience these symptoms even after discontinuation of treatment is known as post-finasteride syndrome (PFS).8 Previously, we cataloged the symptoms reported by patients who took 5ARIs for androgenic alopecia as part of PFS.9 In addition to penile vascular abnormalities and voiding dysfunction, these included changes in body composition, disturbances in memory and attention, and increased rates of depression and anxiety. Of note, subjective changes in body composition including muscular atrophy and increased fat deposition have not been confirmed by physical examination. Outside of our study population, others have reported possible effects of 5AR inhibition on metabolism (including increased insulin resistance, alteration of fat distribution, and cardiovascular disease) and bone physiology.10
To date, the existence and biological basis of PFS has been questioned, with some positing a functional or psychogenic etiology.11 One study reported persistent sexual symptoms in patients on a placebo treatment, while another noted higher reported rates of side effects in patients who were explicitly educated on them.12-14 However, there is evidence to support an organic etiology. One possibility is decreased levels of neurosteroids in patients with PFS. One study measured plasma and cerebrospinal fluid levels of neurosteroids, including progesterone, testosterone, and their metabolite hormones, in patients with PFS, finding them to be significantly lower compared to healthy controls even years after discontinuation of 5ARI therapy.15 While this would explain the emergence of sexual, cognitive, and psychiatric symptoms as well as their long-term persistence, it remains unclear why levels would remain low if 5A reduction was not actively inhibited. In addition, changes to neurosteroid levels alone may not adequately explain other PFS symptoms such as body composition changes and genitourinary issues.
Another proposed etiology for the development of PFS is as a response to androgen deprivation mediated by upregulation in androgen receptor (AR) expression. The AR is expressed in multiple tissues across organ systems including reproductive, genitourinary, nervous, musculoskeletal, cardiovascular, and immune.16 Given that gene overexpression has been implicated in numerous disease states, AR overexpression in response to an androgen-deficient state may negatively affect multiple tissues throughout the body.17-18 Upregulation in AR expression in penile skin has been demonstrated in patients with persistent sexual side effects following finasteride discontinuation.19 Overexpression of AR leading to changes in neurosteroid synthesis and action has been suggested as an explanation for certain behavioral phenotypes in mice.20 In addition, polymorphisms in CAG (polyglutamine-encoding) repeat length has been inversely associated with transcriptional function.21 Together, overexpression combined with less functional activity may be responsible for certain PFS symptoms, possibly through effects on neurosteroid levels but also through induced tissue toxicity.
RNA microarray studies have been used to assess the conditions of an organism and ascertain whether potential disease states can be attributable to differential gene expression.22 If we can identify processes that affect neurosteroid levels in a way that explains the findings of lowered levels in CSF; significant differences in AR expression or function; or any other biological processes that are differentially expressed, then the case for a biological etiology of PFS is strong. In the present study, we compared gene expression data and parameters relating to AR expression and the function between men with a history of 5ARI use and reported PFS symptoms with healthy controls. We hypothesize that gene expression in men with PFS symptoms will demonstrate differential expression compared to healthy controls.
Microarray Analysis
Penile skin samples from 52 patients yielded an analysis of 46,204 gene symbols. After false discovery rate correction, 1,446 genes were significantly overexpressed and 2,318 genes were significantly under-expressed in study patients. The fold change for overexpressed genes ranged from 1.105 to 5.453, while for underexpressed genes it ranged from -1.095 to -11.236. Of particular interest, we found AR expression to be significantly higher in study patients compared to controls (adjusted P-value = .01).
Pathway Analysis
Pathway analysis on both the over-and under-expressed gene sets yielded fourteen and fifteen clinically relevant clusters, shown in Figure 1 and Figure 2, respectively. Of note, the levels of significance are much higher for the under-expressed genes, indicating it is more likely that these biological systems are affected by differential gene expression in our study population. As we present these results, we will refer to clusters derived from the over-expressed gene set as “up-regulated,” while clusters derived from the under-expressed gene set will be termed “downregulated.”
Sexual Effects: Penile Vascular and Soft Tissue Health
We previously reported that the majority of our study patients suffered from confirmed penile vascular abnormalities on penile Doppler ultrasound, including arterial insufficiency, possible ED, and venous leak. In this study, we found that pathways including tissue migration, angiogenesis, and vasculature development were up-regulated. Conversely, pathways including blood vessel development, angiogenesis, and positive regulation of epithelial and endothelial cell migration and proliferation were down-regulated.
In addition to vascular issues, soft tissue components of penile anatomy may be affected in PFS. Peyronie’s disease, along with changes in penile length and testicular size, is reported in patients who have taken 5ARI. We found that pathways involving extracellular matrix regulation (including the “matrisome”), cell junction organization, and connective tissue health were down-regulated.
Genitourinary Effects: Voiding
We previously presented evidence that the immune system involvement and inflammation, bolstered by the low DHT-environment seen in 5ARI use, may lead to the development of benign prostatic hyperplasia and subsequent voiding symptoms in PFS. In this study, we observed that pathways controlling Tcell development, proliferation, and function, along with pathways involving cytokine signaling, were upregulated in the study patients.
Neurological Effects: Neuro-steroids
Although neuro-steroids levels should return to normal following discontinuation of 5ARI, we identified persistent changes in gene expression in study group patients that may explain the preservation of changes in levels of progesterone, testosterone, and their metabolites reported by Caruso et al.15 Upregulated clusters include “organic hydroxy compound biosynthetic process,” “regulation of small molecule metabolic process,” and response to a steroid hormone. Processes involving the regulation of cholesterol synthesis were particularly involved. Pathway analysis did not reveal a significantly down-regulated pathway involved in steroid metabolism or regulation.
Neurological Effects: Nervous Tissue Maintenance
Separately from the effects of neuro-hormones, processes controlling nerve cell health appear to be impacted by 5ARI exposure. An upregulated cluster identified as “kinase and transcription activation” further specified as involving processes controlling neurogenesis and immune responses are upregulated in study patients. However, transforming growth factor-beta and bone morphogenic protein signaling, along with several pathways including neuron differentiation, neurogenesis, and axonogenesis, are pathways that are down-regulated. In addition, the actin cytoskeleton organization cluster was down-regulated. Notably, down-regulated pathways exhibit much higher levels of significance.
Physical Effects: Musculoskeletal and Metabolic Changes
Many of our patients reported musculoskeletal complaints, including fatigue, muscle atrophy, and joint pain, along with skin changes and visual disturbances. Pathway analysis revealed upregulation in pathways affecting skin development, epidermis development, and “establishment of the skin barrier”. Ossification was also upregulated. Down-regulated clusters included skeletal system development, muscle structure development, sensory organ development, sensory organ (visual) development, and connective tissue development.
Altered steroid metabolism can affect metabolic health, and may lead to issues with insulin resistance, increased fat deposition, and cardiovascular disease.10 We observed upregulation in pathways affecting insulin (within a cluster called “regulation of establishment of protein localization”) including positive regulation of peptide secretion, regulation of insulin secretion, and response to carbohydrates. Similarly, we observed upregulation in pathways controlling homeostasis and responses to stress, which both were found to be related to metabolic processes on further analysis of relevant annotations. Down-regulated clusters included circulatory system process (involving processes like cardiac muscle contraction and ion transport); “cellular response to growth factor stimulus” specifically referring to the TGF-beta signaling pathways involving bone morphogenic protein, which regulate osteogenesis, growth, and homeostasis;28 and “transmembrane receptor protein tyrosine kinase signaling pathway,” which included pathways involved in the cellular response to insulin.
Androgen Receptor
Data evaluating AR parameters are reported in Table 3. Per microarray data, AR was overexpressed in study patients. However, neither the number of CAG repeats nor the levels of 5- alpha-androstanediol and 3-alpha-androstanedione glucuronide were significantly different between the study and control groups. For study patients, the median length of CAG repeats was 21.5 (IQR, 20-23.75), while for controls, the median length was 22 (IQR, 19-25). The median level of 5-alpha-androstenediol was 0.950 ng/ml (IQR, 0.749-1.587 ng/ml) in study patients and 0.945 ng/ml (IQR, 0.817-1.337 ng/ml) in controls. The median level of 3-alpha-androstanedione glucuronide was 3.1 ng/ml (IQR, 1.925-5.475 ng/ml) in study patients and 6.7 ng/ml (IQR, 3.375-11.4 ng/ml) in controls.
DISCUSSION
In the present study, we explored the hypothesis that PFS is underpinned by a biological mechanism triggered by the use of 5ARIs and the resultant androgen-deficient state. We identified genes that are differentially expressed in penile skin tissue between men with a history of 5ARI use and PFS symptoms and healthy controls, identifying biological pathways that were upregulated or inhibited and may be relevant in the development of PFS symptoms. We proposed that these effects are mediated largely through effects on neurosteroid levels or changes in AR expression leading to downstream effects on sexual, physical, neurological, and psychiatric functioning. In addition, we identified a number of other pathways involving differentially expressed genes that may also contribute to PFS symptoms.
Neurosteroids
When 5A-reduced steroids such as DHT are 3A-reduced, they become 3-alpha,5-alpha neurosteroids. These molecules are steroid hormone metabolites that affect inhibitory or excitatory function in the central nervous system through action on neuronal membrane receptors.29 In addition, they can modify gene expression through interaction with intracellular steroid receptors in nervous tissue.30 They play a role in several important functions including mental health (depression, anxiety, stress), cognition (learning, memory), and nervous system plasticity.31-35 Both inhibitory and excitatory neurosteroids ultimately have antidepressant, anxiolytic, neuroprotective, neurogenic, and cognition-enhancing effects.36-37 5ARIs may induce cognitive and psychiatric side effects through decreased neurosteroid levels by preventing steroid hormones from metabolizing through reduction.
We identified genes involved in pathways that may affect neurosteroid levels which were upregulated. Considered alone, this would appear to be at odds with the findings reported by Caruso et al.15 However, it is possible that this particular alteration in gene expression may represent a feedback loop attempting to drive synthesis to replace depleted hormone levels, including progesterone, testosterone, and their reduced neuro-active metabolites, in response to 5AR inhibition. However, to better understand the likely effect of 5ARI exposure in our population of PFS patients, we searched for other gene expression changes that could further elucidate the mechanisms involved. We considered three ways for neurosteroid action to be affected: (i) changes in pathways affecting steroid hormone metabolism; (ii) changes in pathways affecting neuronal membrane receptors; and (iii) changes in pathways affecting the androgen receptor, which we found to be significantly over-expressed in PFS patients. Using a list of genes of interest across these categories that we constructed prior to conducting the experiment, we cross-referenced our microarray results to identify whether genes involved with these receptors were differentially expressed in our study population.
Androgen Receptor
The AR itself was over-expressed in PFS patients. Overexpression of normal genes has diverse pathologic consequences across a variety of tissues, leading to certain neurodegenerative diseases, fibrosis, diabetes, and cancer.17 Since patients in this study did not have an elongated CAG polymorphism, we can consider their AR gene to be normal. Overexpression of AR in penile tissue may be responsible for sexual symptoms experienced by PFS patients.19 If AR is also overexpressed in other tissues, for example, nervous tissue, it may play a role in cognitive and psychiatric symptoms as well.
*Depending on the interaction with the previously discussed genes, AR activity could be increased, decreased, or balanced out in PFS patients. If it is higher than in controls, this may be a result of the body attempting to correct for the decreased level of neurosteroids available. If it is lower than in controls, it could be another contributor to antiandrogenic symptoms in PFS. The fact that AR expression was elevated in PFS patients suggests a chronic androgen deficient, or activity-deficient, state. However, the lack of a difference in androgen activity markers indicates the opposite conclusion. Since expression data came from specific tissue, it is possible that there are differing site-specific expression patterns in patients. Ultimately, whether the differentially expressed genes affecting AR activity play a role in PFS remains unclear given the lack of mechanistic data, and requires investigation pointed in this direction.
Other Contributing Pathways
In addition to the effects of decreased neurosteroid activity, changes to steroid levels, and AR overexpression, it is possible that changes to other biologic mechanisms may be present in PFS patients contribute to the syndrome. Results of pathway analysis indicated the involvement of biological processes affecting the sexual, genitourinary, neurological, musculoskeletal, cardiovascular, metabolic, and immune systems. From our data, it is unclear to what extent neurosteroids and AR are responsible for changes in expression for genes affecting these pathways.
*Our data revealed differential expression of many genes in pathways controlling gene expression itself. This may be due in part to the fact that steroid hormones act by inducing changes in gene expression or by AR activity, and may also explain the nature of the persistent effects of 5ARI exposure. Namely, exposure to 5ARI could lead to permanent changes in genetic expression through unknown mechanisms which cause the changes we were able to measure and analyze. However, further work would need to be done to explore these mechanisms and is beyond the scope of the current study.
Strengths and Weakness
There are several strengths to this study, including the use of advanced gene expression and pathway analysis and gene expression data from a large cohort of cases and controls. This study is the first to evaluate multiple proposed etiologies for PFS and provide biological evidence for each. In addition, this study is strengthened by the fact that not all data matched the initial hypothesis, qualifying the argument for the existence of PFS as a unified clinical entity. This study also has weaknesses and limitations. We base the objective existence of symptoms partially on subjective survey data. Selection bias may arise from only the more severe phenotypes seeking care. Gene expression data prior to 5ARI exposure is unavailable for comparison and thus we are less certain that differences in expression between study and control groups are attributable to 5ARI exposure. Further, we lack tissue samples from supposedly involved tissues (ie CNS tissues), ultimately limiting the conclusions that can be drawn from both the microarray and AR expression data. Finally, there exists a lack of mechanistic data to imply causality for proposed mechanisms; gene expression changes alone do not necessarily translate to concordant changes in protein expression or protein activity.
CONCLUSION
This study is the first to consider gene expression differences in men with PFS in explaining the etiology of this condition. Given gene expression per se is not mechanistic and does not imply causality, experiments with downstream processes of protein expression and activity should be undertaken to provide mechanistic data and clarify the results of this work. Further investigation should also explore upstream processes including the mechanisms regulating gene expression in the setting of PFS and identify risk factors for individuals, with a potential focus on genetic risk factors. At this time, patients should be informed regarding possible side effects of 5ARI that may persist even following discontinuation of treatment as part of their counseling.
