Progesterone and contraceptive progestin actions on the brain: A systematic review of animal studies and comparison to human neuroimaging studies

Hormonal contraceptives (HC) have been approved for more than 60 years now with an estimated 150 million users worldwide (UnDeAS, 2019). By mimicking the effects of endogenous progesterone, they target the hypothalamic-pituitary-gonadal axis and downregulate endogenous hormone production (e.g. Kuhl, 2011). Thus, while their primary target is in fact situated in the brain, potentially widespread effects on other neuronal and behavioral outcomes beyond the reproductive axis have only been recognized in the past decade (e.g. Pletzer & Kerschbaum, 2014). Behavioral side effects of hormonal contraceptives include both positive and negative alterations in mood (for reviews see Fruzzetti & Fidecicchi, 2020; Robakis et al., 2019; Sundstrom-Poromaa & Seglebadh, 2012; Böttcher et al., 2012; Schaffir et al., 2016), and cognition (last systematically reviewed by Warren et al., 2014). Regarding mood alterations, approximately 4-10% of women develop depressive symptoms during the first few months of treatment, which is a major reason for discontinuation of contraceptive treatment (Kelly et al., 2010). In contrast, mood-stabilizing effects have been reported in long-term users of hormonal contraceptives (Jarva & Oinonen, 2007). Regarding cognitive changes, results suggest an overall improvement of verbal memory, while effects on other cognitive abilities were inconsistent and potentially dependent on the specific contraceptive formulation (Warren et al., 2014).

Particularly with regards to emotional side effects, it is of high relevance to understand, what determines women’s individual sensitivity towards emotional and cognitive alterations during contraceptive treatment in order to be able to individually tailor contraceptive approaches to minimize side effects. However, the neuronal alterations underlying these behavioural changes are not well understood. Human neuroimaging studies have identified a number of brain areas relevant to emotional and cognitive processing that appear to be sensitive to hormonal contraceptive use (compare section 1.3.). For these areas studies reported changes in gray matter volume or cortical thickness, brain activation or connectivity to other brain areas (systematically reviewed by Broennick et al., 2020). But the cellular and molecular mechanisms potentially underlying these changes in neuroimaging parameters have received less attention (see Porcu et al., 2019). In order to fully understand what happens in individual women’s brains during treatment with specific contraceptive progestins, it appears that a systematic interdisciplinary approach has to be the way forward. Human neuroimaging studies and preclinical animal studies have to inform each other in order to clearly delineate the brain systems responsive to contraceptive steroids. Given the replication crisis and power failures in the neurosciences due to massive multiple comparisons problems (Button et al., 2013), synchronizing approaches in terms of brain areas seems to be the logical and also the simplest first step. A second step is to focus further studies on those neural systems that most consistently respond to endogenous ovarian hormones. Accordingly, the first goal of the current review is to provide an overview of our current state of knowledge regarding cellular and molecular alterations in the brain and map out, which changes concern those brain areas that consistently respond to contraceptive treatment in human neuroimaging studies. The second goal is to obtain a systematic overview of the cellular and molecular responses to endogenous progesterone – alone or in combination with estradiol. In combination, these overviews should highlight those brain systems most likely to respond to contraceptive treatment and thereby provide a roadmap for contraceptive research in the upcoming years.

To that end, we have identified four mechanisms at the cellular and/or molecular level that may potentially underly changes in structural and functional neuroimaging parameters. These include: (i) changes in the number of neurons, i.e. neurogenesis, (ii) changes in the number of synapses between neurons, i.e. synaptogenesis, (iii) changes in the efficacy/velocity of electro-chemical information transfer between neurons, i.e. myelination, and (iv) changes in neurotransmitter signalling that may occur at multiple levels, including the synthesis, release, degradation or receptor binding of a specific neurotransmitter. All of these mechanisms have been extensively studied with respect to their modulation by endogenous sex hormones, specifically estradiol and – to a lesser extent – also progesterone and testosterone. Via a systematic review, we seek to understand in which brain areas contraceptive progestin administration – alone or in combination with estrogens – is related to alterations in these four mechanisms and whether similarities to human neuroimaging studies arise. We furthermore seek to identify how these mechanisms respond to endogenous progesterone – alone or in combination with estrogens – and compare responses between endogenous progesterone and contraceptive progestins.

Hormonal contraceptives consist of either a synthetic progestin (progestin only), or a synthetic progestin in combination with an estrogen (see Table 1). Figure 1 provides an overview of synthetic progestins predominantly used for hormonal contraception. They are devided in four main groups, the Pregnanes (progesterone derivates), Gonanes (19-nortestosterone derivates), the spironolactone derivate Drospirenone and the fourth-generation progestin Dienogest, an aromatized 19-nortestosterone derivative. Derivates of progesterone i.e. medroxyprogesterone, chlormadinon and cyproterone, synthesized via 17-hydroxy-progesterone, were already used 60 years ago in first-generation contraceptives and are nowadays experiencing a revival due to their anti-androgenic properties (e.g. Davtyan, 2012). The most recent progestin nomegestrol, is derived from progesterone via 19-norprogesterone. All Pregnanes are administered as acetate and are known for their anti-androgenic properties. The 19-nortestosterone derivative levonorgestrel is the key progestin contained in second-generation contraceptives and the standard of care in many countries due to its low risk for thrombo-embolic events. Derivates of levonorgestrel (gestoden, desogestrel, etonorgestrel, norelgestromin and norgestimate) are used in third-generation contraceptives. All Gonanes have androgenic properties, though the latter are lower in third-generation progestins compared to Levonorgestrel.

Progestin only contraceptives include 1) progestin only pills (POPs), which contain either desogestrel or drospirenone 2) IUDs, which contain levonorgestrel 3) contraceptive implants, which contain etonorgestrel 4) or injections, which contain medroxyprogesterone acetate. In combined oral contraceptives (COCs), as well as rings and patches, 1-4 and various other progestins are administered in combination with an estrogen. The most commonly used estrogen is ethinylestradiol, although newer COCs contain the bio-identical estrogens estetrol or estradiol-valerate, which is quickly metabolized to estradiol upon ingestion.

Synthetic progestins have been thouroughly studied and characterized regarding their pharmacological profile and binding affinity to various nuclear steroid hormone receptors (compare Table 2). All progestins are designed to activate nuclear progesterone receptors and are anti-estrogenic due to the downregulation of estrogen receptor expression and endogenous estradiol production (e.g. Kuhl, 2011). This is relevant, since endogenously, progesterone actions are primed by estrogens, given that estrogen receptors upregulate the expression of progesterone receptor genes (Quadros et al., 2008). Accordingly, important differences may arise between progestin only formulations and combined estrogen-progestin formulations that supplement for the reduced endogenous estradiol production.

However, while synthetic progestins mimic the actions of endogenous progesterone via their binding to intracellular progesterone receptors, their actions may still differ from endogenous progesterone in various aspects. First, synthetic progestins differ in their binding affinities as well as agonistic and antagonistic potential towards androgen receptors, glucocorticoid receptors and mineralocorticoid receptors (compare Table 2; see Kuhl, 2011 for a review). Accordingly, synthetic progestin actions may also arise from their androgenic or anti-androgenic activities, as well as glucocorticoid or mineralocorticoid activities, and these actions may differ between different synthetic progestins.

Second, progesterone actions on the brain are not only excerted via nuclear progesterone receptors. The first, and best documented alternative pathway concerns progesterone metabolites, in particular the 5α-reduced metabolite allopregnanolone. Allopregnanolone is a positive allosteric modulator of the GABA-A receptor, which is potentially responsible for the anxiolytic, anti-depressive and sedative properties of progesterone (Pinna, 2020). However, synthetic progestins are not metabolized to allopregnanolone and differentially affect the levels of allopregnanolone in the brain (systematically reviewed in the current manuscript, compare Table 3). Nevertheless, some progestins, particularly anti-androgenic progestins, posess 5α-reduced metabolites with similar allosteric effects on the GABA-A receptor as allopregnanolone (Picazo et al., 1999).

Finally, both membrane-associated progesterone receptors (mPR) and progesterone receptor membrane components (PGMRC) have been identified and are likely responsible for some fast-acting cellular changes in response to progesterone (reviewed by Zhu et al., 2008). In particular, some of the neuroprotective effects of progesterone have been attributed to its interaction with mPRs and PGMRCs (e.g., Sun et al., 2016). To the best of our knowledge, it is yet unclear, whether the contracptive progestins listed in Table 2 also interact with mPRs and PGMRCs.

Accordingly, results from progesterone administration studies do provide a promising starting point for contraceptive progestin administration studies, but differences may occur when progesterone actions are not excerted via intracellular progesterone receptors. Thus, when assessing contraceptive progestin actions on the brain, it is important to (i) distinguish between the different types of progestins, (ii) distinguish between synthetic progestins administered alone or in combination with estrogens, and (iii) compare the synthetic progestin actions to progesterone actions to distinguish between actions dependent on nuclear progesterone receptors and actions dependent on other mechanisms.

A recent meta-analysis from 2020 identified 33 neuroimaging studies on hormonal contraceptive actions on the brain, though some reported on the same samples (Broennick et al., 2020). Since then, five additional neuroimaging studies (Petersen et al., 2021, Chen et al., 2021, Wen et al., 2021, Menting-Henry et al., 2022; Noachtar et al., 2022), one PET (Larsen et al., 2022) and two EEG studies have been published on this subject. Results of those studies are mixed, which is most likely due to the following reasons: Most importantly, 33 out of the 38 total neuroimaging studies are cross-sectional studies comparing current users of hormonal contraceptives (mostly COC) to current non-users of hormonal contraceptives. It has been extensively discussed that those designs underly a variety of sampling biases (Pletzer & Kerschbaum, 2014; Broennick et al., 2020; Schuster et al., 2022). Even if demographic differences in age, education and socio-economic status are controlled for, long-term users usually show little to no side effects or beneficial effects on their contraceptive, while non-users usually have a history of adverse side effects on various hormonal contraceptives. Accordingly, cross-sectional studies are just as likely to capture pre-existing differences in the brain that contribute to the vulnerability to develop side effects or changes attributable to the previous use of hormonal contraceptives, as they are to capture changes attributable to the current contraceptive use. Accordingly, the results of cross-sectional studies should be discussed separately from the results of longitudinal or placebo-controlled studies. Comparisons may be cautiously drawn, if previous contraceptive use or previous adverse effects during contraceptive use are taken into consideration by cross-sectional studies.

Furthermore, different progestins may differ in their neuronal outcomes and should therefore be studied separately regarding their effects on the brain. Unfortunately, 24 out of 33 cross-sectional studies and one out of five longitudinal studies available did not assess or report the progestin used in hormonal contraceptives. Accordingly, results are hard to interpret. Therefore, we will only summarize results of longitudinal neuroimaging studies and cross-reference them with those cross-sectional studies, that specify the type of progestin used.

A Swedish double-blind randomized placebo-controlled trial included a task on emotional face matching (Gingnell et al., 2013), response inhibition (Gingnell et al., 2016), as well as a resting state scan (Engman et al., 2018). 34 women with previous negative affect during COC-use were recruited and received either placebo or a COC containing 0.30 μg EE and 0.15 mg LNG for one month. Alongside this study, Petersen et al. (2021) recruited 26 women with previous negative affect during COC-use to receive one month of placebo and one month of 0.30 μg EE and 0.15 mg LNG double-blinded and in randomized order and assessed cortical thickness as well as mood scores. Both studies reported a significant deterioration during COC-use compared to placebo. Furthermore, imaging studies emerged the prefrontal cortex as area of interest in the context of levonorgestrel treatment. During the emotional face matching task, activity in the left middle frontal and bilateral inferior frontal gyri was significantly reduced during COC-use compared to placebo (Gingnell et al., 2013). During response inhibition, activity in the right middle frontal gyrus was significantly reduced during COC-use compared to placebo (Gingnell et al., 2016). Finally, cortical thickness was significantly reduced in the right inferior frontal gyrus during COC-use compared to placebo (Petersen et al., 2021). In accordance with these longitudinal observations, reduced gray matter volumes in the bilateral middle frontal gyri and reduced cortical thickness in the orbitofrontal gyri were observed in cross-sectional studies comparing users of androgenic COC to naturally cycling women (Pletzer et al., 2015; Petersen et al., 2015). However, no differences between users of androgenic COCs and non-users were observed upon viewing of food stimuli, though samples size in this study was very small (Arnoni-Bauer et al., 2017). Taken together, the results suggest, that several prefrontal areas underly a down-regulation during treatment with androgenic COC, specifically LNG. These results may be related to decreased emotional responsiveness and mood worsening observed during treatment with LNG. However, it is yet unclear, whether these results are restricted to, or more pronounced in women with previous negative affect during COC use, as they may be particularly sensitive to the emotional effects of LNG. Accordingly, it is of particular interest for animal models, whether LNG affects neurogenesis, synaptogenesis or neurotransmitter content in frontal brain areas.

Finally, it is an interesting question, whether different progestins have different effects on frontal cortex volume or reactivity. For instance, frontal cortex reactivity appears to be increased upon injection with the anti-androgenic progestin MPA (Basu et al., 2016). An observation that is in line with results from a cross-sectional study in which the majority of women were using anti-androgenic COC (Bonenberger et al., 2013). Likewise, our restrospective study revealed that frontal cortex volumes were increased in a group of COC users, that were majorily treated with drospirenone containing COC (Pletzer et al., 2010). However, no differences between groups of mostly anti-androgenic COC use and non-users were observed in the prefrontal cortex upon viewing of erotic stimuli (Abler et al., 2013) or during a stress task (Chung et al., 2016). However, it has to be kept in mind that all of the above-mentioned cross-sectional studies included lass than 15 subjects per group. The findings of increased frontal cortex reactivity with anti-androgenic OC use are in accordance with an animal study demonstrating increased synaptogenesis in this area upon MPA treatment (Chisholm et al., 2012). Indeed, future research should be conducted focusing on the effect of anti-androgenic synthetic progestins contained in COC like chlormadinone acetate or drospirenone on neurogenesis or synaptogenesis in the frontal cortex.

Lisofsky et al. (2016) recruited 28 women, who were planning to start COC use, and collected structural and resting state scans prior to their COC use and after three months of COC use. 28 naturally cycling women, who did not start COC use served as a control group. As in Gignell et al. (2013) and Petersen et al. (2021), mood was significantly worsened during COC-treatment. A significant reduction in gray matter volumes of the left amygdala was observed during COC-treatment and accompanied by a significant reduction in resting state connectivity between the amygdala and the dorsolateral prefrontal cortex. In line with this observation, reduced amygdala reactivity to emotional stimuli in COC-users was observed in a cross-sectional study (Petersen & Cahill, 2015). Since neither study reported information on the specific contraceptive formulation, it is unclear whether these findings hold across contraceptive formulations. Reduced amygdala connectivity to cortical areas during combined EE/LNG use was observed by Engman et al. (2018) during the resting state. Furthermore, a recent cross-sectional study demonstrates differential modulation of amygdala-behavior relationships by androgenic and anti-androgenic COC (Menting-Henry et al., 2022). In line with contraceptive studies, reduced amygdala reactivity to emotional stimuli was also observed in phases of high endogenous progesterone (Derntl et al., 2008).

The salience network is a large-scale brain network comprised of the anterior cingulate cortex (ACC) and bilateral Insula (Uddin et al., 2016). The ACC demonstrates increased connectivity to the precuneus and dorsolateral prefrontal cortex during EE/LNG treatment (Engman et al., 2018), while the insula showed decreased reactivity to emotional stimuli (Gingnell et al., 2013). The latter effect appeared to be related to mood deterioration. In contrast, a cross-sectional study reported heightened reactivity to traumatic film viewing in the insula and ACC of COC-users compared to non-users (Miedl et al., 2018). Furthermore, increased gray matter volumes of the ACC and insula were reported during the active compared to the inactive pill phase by deBondt et al. (2013), while Petersen et al. (2015) observed decreased cortical thickness of the ACC and insula in COC-users compared to non-users. Neither study did report information on the specific COC-formulations used, however given the timing and place of the studies, it may be speculated that majority of participants in deBondt et al. (2013) and Miedl et al. (2018) used anti-androgenic COCs, while majority of participants in Petersen et al. (2015) used androgenic COCs. In addition, reactivity of the ACC and amygdala to endogenous estradiol was abolished in COC-users (deBondt et al., 2013; deBondt et al., 2016). Accordingly, it would be of great interest to explore, whether androgenic and anti-androgenic COC differentially modulate the salience network. However, in a sample of mostly anti-androgenic COC-users compared to non-users, a study by Abler and colleagues observed decreased insula reactivity upon erotic stimulation (Abler et al., 2013), but increased insula reactivity during monetary reward processing (Bonenberger et al., 2013). Accordingly, it is possible that COCs alter the type of stimuli selected by the salience network, rather than the reactivity of the salience network per se.

No longitudinal study observed changes in hippocampal volumes or activation due to COC use, although the cluster of reduced gray matter in the left amygdala observed by Lisofsky et al. (2016) extended in the the parahippocampal gyrus. However, given that the hippocampus is the most plastic area of the human brain and appears to be particularly receptive to endogenous hormonal fluctuations in both human and animal studies (see Hillerer et al., 2019 for a review), it seems worthwhile to investigate the effects of different contraceptive formulations on the hippocampus. Unfortunately, results regarding these areas are rather mixed. While some studies report increased volumes of the hippocampus/parahippocampus with anti-androgenic COC use (Pletzer et al., 2010, Pletzer et al., 2015), the opposite effect was observed in studies using mixed samples of androgenic and anti-androgenic COC-users (deBondt et al., 2013; Lisofsky et al., 2016; Pletzer et al., 2019). However, it should be taken into account that the most pronounced hormonal effects in the hippocampus were observed in response to endogenous estradiol, which are reversed by endogenous progesterone (Hillerer et al., 2019). Accordingly, COC-effects in the hippocampus may be the result of how specific progestins modulate the effects of ethinylestradiol.

The goal of the current review was to address the effects of (i) endogenous progesterone and (ii) the contraceptive progestins listed in Table 1 alone or in combination with estrogens on the cellular and molecular mechanisms, potentially underlying changes in brain structure and function identified by human neuroimaging studies on contraceptive use. As outlined in the introduction potential mechanisms include (i) neurogenesis, (ii) synaptogenesis, (iii) myelination and (iv) neurotransmitter signalling. With respect to the latter, we restricted our search to six major neurotransmitter systems, i.e., acetylcholine, glutamate, GABA, serotonine, dopamine, and norepinephrine, which are related to cognitive and emotional functioning. Beyond the scope of this review, indirect actions of contraceptive progestins on neurotransmission may arise via their interactions with various endocrine and neuromodulatory systems, e.g., thyroid hormones, oxytocin or beta-endorphin. Results on endogenous progesterone are included to (i) compare the actions of synthetic progestins to endogenous progesterone and (ii) generate hypotheses for future studies on contraceptive progestin actions on the brain. While results on the whole brain will be extracted, a particular focus lies on cognitively relevant brain areas, as identified as responsive to contraceptive progestins in human neuroimaging studies.

留言 (0)

沒有登入
gif