Cordian Beyer, Institute of Neuroanatomy, RWTH Aachen University, 52074 Aachen, Germany (e-mail: firstname.lastname@example.org).
Sex steroids have been demonstrated as powerful compounds to protect neurones and neural tissue from neurotoxic challenges and during neurodegeneration. A multitude of cellular actions have been attributed to female gonadal steroid hormones, including the regulation of pro-survival and anti-apoptotic factors, bioenergetic demands and radical elimination, growth factor allocation and counteracting against excitotoxicity. In recent years, immune-modulatory and anti-inflammatory characteristics of oestrogen and progesterone have also come under scrutiny. To date, each of these physiological responses has been considered to be partially and selectively integrated in the mediation of steroid-mediated cell protection and tested in suitable animal models and in vitro systems. To what extent these individual effects contribute to the overall neural protection remains sketchy. One idea is that a battery of cellular mechanisms operates at the same time. On the other hand, interactions and the control of the brain-intrinsic and peripheral immune system may play an additional and perhaps pioneering function in this scenario, notwithstanding the importance of secondary adjuvant mechanisms. In the present review, we highlight neuroprotective effects of oestrogen and progesterone in two different disease models of the brain, namely acute ischaemic and demyelination damage, which represent the most common acute and degenerative neurological disorders in humans. Besides other inflammatory parameters, we discuss the idea that chemokine expression and signalling appear to be early hallmarks in both diseases and are positively affected by sex steroids. In addition, the complex interplay with local brain-resident immune-competent cells appears to be controlled by the steroid environment.
The role of the gonadal sex steroids oestrogen and progesterone or derivatives of both steroids either reaching the brain via peripheral circulation or formed intrinsically within the brain as neuroprotective factors in the central nervous system (CNS) is well-established (1–5). A large number of studies has confirmed this assumption and, additionally, have pinpointed putative molecular, cellular, and systemic mechanisms of steroid action that might account for such a beneficial role of both steroid hormones under neuropathological circumstances (6–9). Clinical reports have corroborated those animal and cell studies, showing that female sex steroids can attenuate the symptoms associated with brain ischaemia and Parkinson’s diseases (10). In the present review, it is not possible to recapitulate and list the many different physiological effects of oestrogen and progesterone and their involvement during reparative, restoring and protecting processes in the CNS in detail, although several excellent recent review articles are available that cover this topic (10–15). Nevertheless, the major routes and mechanistic effects of oestrogen- and progesterone-mediated cellular protection need to be mentioned briefly.
There is convincing evidence that in particular oestrogen can directly protect neurones from apoptosis by interacting with intracellular cell death cascades. Thus, oestrogen either stimulates pro-survival or anti-apoptotic pathways such as Bcl-2, nuclear factor-κB, caspase-3 and Akt, thereby implementing a protective intracellular environment (16–20). The scientific community, however, might agree that such effects on their own might not be sufficient to rescue a tattered neurone and its circuit from death given the complex and fatal situation in vivo. Other supplemental molecular mechanisms including the stimulation of brain glia (i.e. oligodendrocytes, astrocytes, and microglia) might operate at the same time to provide a supportive extra-neuronal ambiance for damaged neurones. A promising mechanism described for oestrogen and progesterone is their potency to actively regulate the energy demand and bioenergetics of impaired neurones (2, 12, 21, 22). Of particular interest is the fact that, although still under debate, the mitochondrial compartment might contain oestrogen binding proteins and receptors (ER) and progesterone-responsive proteins or progesterone receptors (PR) (23, 24). This would allow steroids to directly and rapidly stabilise mitochondrial function and influence their structural composition, which is negatively affected in almost every disease model of the brain, being either primarily causative or secondarily affected. Other important mechanisms of steroid-promoted cell protection include the synthesis of growth factors [i.e. insulin-like growth factor (IGF)-1, brain-derived neurotrophic factor, glial cell line-derived neurotrophic factor, amongst others) (6, 13, 25–27), the influence on the local brain vascular system, endothelial cells, capillary formation, and the integrity of the blood–brain barrier (BBB) (28, 29), as well as glutamate excitotoxic signalling and metabolism (5, 30). Finally, the deeper focus of the present review is the more recently established role of steroid hormones as regulators in the inflammatory context. They have been demonstrated to dampen inflammatory immunological responses upon damage by down-regulating peripheral immune active cells such as lymphocytes (31, 32), as well as to keep local brain-intrinsic immune-competent cells and inflammatory signalling under control (1, 9, 17, 33, 34).
Some of the fundamental questions that need to be addressed in future steroid-brain protection research are which of the numerously reported physiological steroid effects play a decisive and indispensable mechanistic role during neuroprotection and which are only supportive and dispensable but nevertheless active. Do different neurotoxic and neurodegenerative challenges require different modes of steroid action? Finally, which cellular signalling processes and cell–cell interactions are activated by protective steroid hormones. Equally important, there is the need to elaborate whether the natural ebbing of sex steroid hormone production and the resulting drop in hormone plasma and tissue levels during menopause is causally linked to the susceptibility, onset, course, severity and observed sex differences of distinct degenerative and/or acute damages to the brain. This review article does not attempt to answer all these questions and cover the broad spectrum of protective steroid action. Instead, we rather exemplify some issues on three established animal and disease models: the demyelisation, cuprizone and experimental autoimmune encephalomyelitis (EAE) models, which partially and, with reservations, mimic demyelisation comparable to multiple sclerosis (MS) (35), and transient middle cerebral artery occlusion (tMCAO), serving as a model for focal brain ischaemia (36). To follow these disease models, we will highlight some of the above points concerning the interference and fine tuning of local immune responses and the complex working together of brain resident cells with the inflammatory machinery.
Hypoxic challenge in the brain and the protective influence of sex steroids
Female sex hormones are considered to protect against cerebrovascular events (9, 37, 38). Pre- and postmenopausal women differ in their incidence rate of stroke (39). Numerous experimental studies using tMCAO or global ischaemia mouse and rat animal models have provided convincing evidence that both steroid hormones reduce the outcome of neuronal degeneration, tissue damage and neurological deficits under ischaemic conditions in the forebrain (40–43). We have recently shown that gonadal steroids have an individual neuroprotective capacity with a slightly better neuropathological outcome after progesterone compared to oestrogen and a better behavioural recovery after oestrogen (data not shown) (9). Overall, the combined application of both hormones was found to achieve a maximum neuroprotective result after tMCAO (1 h) followed by a short-term (24-h) recovery period (9) (Fig. 1). Although female ovariectomised rats usually displayed smaller infarct volumes than males, we observed a comparable degree of steroid-dependent protection in both sexes (Fig. 1a). By combining triphenyltetrazolium chloride staining for visualising vital tissue and neuronal marker NeuN-immunohistochemistry, we were able to perfectly demarcate the penumbra as the ‘tissue at risk’ of the infarcted cerebral cortex. This allowed us to dissect defined tissue samples from the penumbra and further subject these samples to selective quantitative protein and gene expression analysis (44). In the salvageable penumbra, the chemotactic cytokine ligands (CCL)-2 and 5 and interleukin (IL)-6 are, in addition to other well-described cytokines such as tumour necrosis factor (TNF)-α and IL=1β (not shown), massively increased after tMCAO (Fig. 1b). The administration of oestrogen/progesterone abolished the induction of IL-6 and CCL2 (also known as monocyte chemotactic protein-1) and CCL5 (also known as regulated upon activation, normal T-cell expressed and secreted), whereas the expression of CCL3 (also known as macrophage inflammatory protein-1α) was increased further. In addition, the stroke-dependent increase of microglial and lymphocyte markers, as well as the number of microglial cells in the penumbra, was diminished by both steroid hormones (Fig. 2). The soluble immune mediators, chemokines and cytokines, have been generally implicated in contributing to the initiation, propagation and regulation of inflammatory responses in several neurological diseases such as stroke and MS (45–47). These small peptides, usually in the range 8–14 kDa, are formed within the brain matrix by glial and nerve cells. Besides its deleterious potential, inflammation is considered as a cardinal host defence response to tissue ischaemia. Chemokines act as chemotactic cytokines and attract various types of leukocytes to sites of brain tissue destruction, regulate their trafficking, and participate in their activation (48, 49). Chemokine receptors are widespread within the brain and can be localised not only to peripheral inflammatory cells, but also to neurones, oligodendrocytes, microglia and astroglia. Similarly, the synthesis of chemokines in the CNS has been attributed to all types of glial cells. The family of (human) chemokines comprises approximately 50 proteins. The knowledge of which of these chemokines are present in the brain under ischaemic attacks and in which chronological sequence they appear is only fragmented. Nevertheless, the appearance and function of distinct subtypes of chemokines and cytokines during the course of stroke and ischaemic degeneration appears to be more perspicuous. Despite other reported functions and a great overlap in their cell-specific action, CCL2 appears to be important for brain resident immune cells, thus controlling microglia migration and recruitment to areas of inflammation, their activation and function, and has a significant role in mediating post-traumatic secondary cell damage (50–52). Less valuable information is available on the role of CCL5, which preferably induces the directional migration of monocytes and T lymphocytes and transmigration across the BBB, and further potentiates the dysfunction of the cerebral microvascular system (47, 53, 54); CCL3 usually contributes to the recruitment and stimulation of polymorphonuclear leukocytes in the acute inflammatory state; its role in the activation of microglia is uncertain, although unpublished data from our group indicate an involvement in early microglia regulation (55–58). As noted abobe, sex steroids abolished the tMCAO-mediated induction of CCL2 and CCL5. Therefore, we assume that steroid-dependent neuroprotection is, at least partially, the result from a delayed or diminished activation of microglia and immigration of peripheral immune cells. Although we do not have direct evidence, gene expression studies in the penumbra support this view and show a significantly reduced expression of the pan-macrophage marker Iba-1, the activated microglia marker, and CD3 as part of the receptor complex implicated in T-cell activation, suggesting reduced lymphocyte infiltration after hormone application (9). The immunosuppressive role of oestrogen during experimental stroke has recently been demonstrated (59). Hormone replacement in ovariectomised female mice normalised the aberrant ischaemia-induced expression of subsets of cytokines and chemokines, corresponding well with our data and introducing additional cytokine/chemokine targets for oestrogen. Another recent study using lipopolysaccharide (LPS)-induced neuroinflammation shows that ERα, ERβ and ligand-independent steroid signalling protects against tissue damage and counteracts the dramatic inflammatory cytokine/chemokine response (60). In addition to the above discussed chemokines, Chiappetta et al. (61) revealed that oestrogen-mediated neuroprotection after tMCAO is accompanied by an early modulation of IL-1β production. Besides scattered information from the uterus (62, 63), no clear direct evidence is provided that progesterone interferes with chemokine expression and signalling in the brain. Nevertheless, progesterone appears to actively influence local brain inflammation after traumatic injury by suppressing the induction of transforming growth factor-β2, without affecting TNF-α in both permanent and focal transient ischaemia (64) and attenuates BBB dysfunction through regulation of the expression of matrix metalloproteinases (65).
With respect to the role of a combined treatment with both steroid hormones and protection from traumatic brain injury, the two steroids might differ in their mechanistic action to prevent cell damage. The stimulation of microglia and subsequently of peripheral immune-competent cells after ischaemic insults is considered to provide detrimental effects on the surrounding neurones (66, 67). There is good evidence that oestrogens show a prominent anti-inflammatory activity that specifically targets microglia (1, 17, 32, 33). Our findings extend this view and suggest that oestrogens diminish chemokine production probably by local astroglia, thereby affecting microglia attraction and function (9). These regulatory effects on early inflammation after stroke might predate secondary processes initiated by oestrogen and progesterone which, similalarly important, might stabilise damaged neuronal functions, abolish oxidative stress, ameliorate local neovascularisation and capillary function, and prevent neuronal cell death.
Relevance of sex steroids in experimental demyelination animal models
MS, a chronic inflammatory and demyelination disease, is the most common disorder of the CNS in young adults. Brain inflammation is associated with lesions appearing in plaques within the white matter. Recent histopathological studies, however, have convincingly shown that grey matter regions are equally affected (68). MS is more common in women than men (69). The disease typically becomes clinically apparent between 20 and 40 years of age, although it can start earlier or later in life. Post-mortem tissue is widely used to study molecular and cellular mechanisms involved in the development of MS lesions. General histopathological hallmarks of MS lesions are inflammation, demyelination and axonal degeneration (68). The pathophysiology of MS appears complex, with demyelination and axonal degeneration contributing to inflammatory neurodegenerative processes. Despite the well-established view of an inflammatory neurodegenerative disease, the ability of current immune-modulatory therapies for MS, such as interferon-β, to prevent long-term disability is debatable (70). Axonal injury occurs very early during the course of MS and accumulates with disease progression. It is not only restricted to focal demyelinated lesions in the white matter, but also affects the normal appearing white matter and the grey matter (71–73). No established treatments that directly reduce nervous system damage or enhance its repair are currently available. Although the actual cause of MS remains to be elucidated, lesions in early disease stages contain relevant amounts of T-cells and, likely, fewer plasma cells. However, activated microglia or macrophages by far outnumber cells of the acquired immune system (74). As MS progresses, the number of T- and plasma cells dramatically decrease, and they are rarely found within lesions in the progressive phase of the disease (68). On the basis on the neuroprotective potency of sex steroids, oestrogen and progesterone, either alone or in combination, might be promising therapeutic tools to prevent neurodegenerative processes in MS. Because of the controversial and decreasing importance of T- and B-cells during MS progression, the modulation of brain-intrinsic inflammatory sequences in general and by female sex steroids have come into the focus of research. Below, we summarise the new findings of protective steroidal effects in different demyelination animal models.
Over the last decade, numerous studies have shown that oestrogen treatment administered before or after disease onset ameliorates EAE in mice (75–78). For an in extenso discussion of this particular topic, we refer to a previous review (1). Tiwari-Woodruff et al. (79) compared and differentiated effects between ERα and ERβ activation during MOG33–55 peptide-induced EAE in C57BL6 mice. Treatment with an ERα-ligand (PPT), if administered 1 week before active EAE induction, ameliorated symptoms of disease in both wild-type and ERβ but not ERα- deficient mice (80). By contrast, the administration of the ERβ-ligand diarylproprionitrile (DPN) had no significant influence on the early disease course (79). However, DPN demonstrated a significant protective effect later in disease (i.e. lower peak disease scores). When DPN was administered to ovariectomised ERβ-deficient animals with EAE, the treatment was no longer protective, demonstrating that ERβ activation accounts for protective effects at late disease stages. Suppression of clinical symptoms by PPT treatment was paralleled by a reduced inflammatory infiltrate in the spinal cord white matter, both at early and later time-points. By contrast, DPN-treated EAE mice did not show reduced inflammation in the white matter at any time point. Other studies suggest that ERα-mediated dampening of EAE-induced spinal cord inflammation might not directly result from a defective priming of auto-antigen-specific CD4+ T-cells (78, 81). Nonlymphocytic cells such as macrophages and dendritic cells are considered to be primarily responsive to oestrogen treatment in EAE (31, 82). Below, we will point out later that brain-resident immune-active cells might also be targeted by sex steroids.
Histopathological changes are not confined to the white matter in human MS and MS-related animal models (83). As a consequence of T-cell-mediated inflammation, a decreased staining for neuronal markers accompanied by increased immunolabelling of microglia/monocytes can be observed in the grey matter of spinal cord in EAE mice. Because of its potent anti-inflammatory effects, PPT exposure reduced the grey matter pathology (80). Interestingly, attenuated demyelination and axonal loss occurred not only in PPT, but also in DPN-treated mice as determined by anti-myelin basic protein and anti-neurofilament immunohistochemistry (79). No loss in neuronal numbers was detected in both PPT- and DPN-treated animals. This study was the first to show that the administration of ERβ-ligands is neuroprotective, without any clear-cut evidence of an anti-inflammatory effect, and provided information on oestrogen-dependent neuroprotection irrespective of its capacity to affect inflammatory processes. At the molecular level, the activation of ERα by PTT induced sustainable changes in cytokine production by systemic immune cells (reduction of TNF-α, interferon-γ, IL-6 versus induction of IL-5), whereas the ERβ-ligand did not provoke such responses. In conclusion, the activation of ERβ signalling cascades might be a better therapeutic option for the progressive MS disease stage.
The potency of progesterone to improve disease severity in EAE is less established. Garay et al. (84) investigated the efficacy of progesterone to prevent axonal damage in the ventral spinal cord white matter in MOG40-45-induced EAE in female C57Bl/6 mice. Treatment with progesterone was started 1 week before EAE induction. In a previous study using the same model and experimental set-up, they were able to demonstrate that progesterone delays the onset and reduced the severity of the disease (85). From a histological viewpoint, progesterone reduced the inflammatory response and the occurrence of demyelination in the spinal cord during the acute phase of EAE. Thus, the effects of oestrogen and progesterone in this animal model are comparable in terms of modulating the onset of EAE. Along with the findings of reduced inflammation, progesterone-treated animals displayed less signs of axonal damage. Axonal loss, as determined by quantification of axonal density in semi-thin sections, and the accumulation of amyloid precursor protein-positive axonal spheroids as an index of axonal dysfunction, were less severe in progesterone-exposed compared to vehicle-treated animals. One should always keep in mind that axonal damage occurs during early EAE disease course, as demonstrated in several studies using different EAE protocols and species (86–88). Thus, a reduced axonal damage in the progesterone-treated group might be a direct consequence of reduced disease severity (i.e. reduced recruitment of inflammatory cells to the side of lesion) rather than a direct neuroprotective effect. Two other groups have reported cell protective effects of progesterone administration if treatment was initiated at the onset of disease (the day with a mean clinical score of one or higher). After the occurrence of disease symptoms, progesterone significantly impaired the cumulative disease index and peak disease score (89). Protective progesterone effects are not restricted to the spinal cord, as recently demonstrated by Yu et al. (90) in rats, where EAE was induced by guinea pig spinal cord homogenates. Future studies need to show which cellular and cell-regulatory mechanisms might account for the neuroprotective effects of progesterone and what role the anti-inflammatory properties essentially play.
To further dissect peripheral versus central hormonal effects in MS animal models, we used the toxic cuprizone demyelination mouse model (35). Feeding of cuprizone to young adult mice (0.2%, 5 weeks) induced a massive demyelination of distinct brain regions such as the corpus callosum (CC, an inter-hemispherical white matter tract), the basal ganglia (BG), cerebellum, and hippocampus (91–93). In this animal model, demyelination is considered to be the consequence of primary oligodendrocyte apoptosis. The simultaneous treatment of cuprizone-exposed mice with oestrogen or progesterone alone resulted in a moderate myelin protection. However, combined treatment with both steroid hormones prevented demyelination in the CC almost completely (94). Similar observations were made for the BG (unpublished results, M. Kipp and C. Beyer). Analysis of marker proteins specific for mature oligodendrocytes confirmed the immunohistochemical presence of myelination. T2-weighted magnet resonance images of cuprizone-fed animals showed a ventricle volume increase, a developing hydrocephalus, and an elevated signal intensity of the CC, indicating active demyelination (94). Such lesions and changes of signal intensity were not detectable after combined hormone treatment (94). The neuroprotective effects of oestrogen in the CC of this animal model were very recently reported by another group (95), who demonstrated that the protection of myelination and preservation of functional oligodendrocytes is accompanied by a delay in microglia accumulation and a reduction in the expression of pro-inflammatory signals (i.e. cytokines). It was concluded that oestrogen might mediate cell protection after a cuprizone challenge in a T-cell-independent manner, thus supporting our view of a selective brain-intrinsic modulation of immune responses, besides affecting the peripheral immune system. This observation is in line with previous findings from our group revealing a strong anti-inflammatory potential of combined oestrogen and progesterone application on cytokine synthesis of astrocytes and neuronal death rates under toxic conditions (Fig. 3) (1, 34, 96, 97). Our findings also imply that oestrogen and progesterone stimulate local astroglia proliferation in the damaged CC regions and cause the induction of IGF-1 expression (i.e. in activated astrocytes), which is a well-known mitogen for the recruitment of new oligodendrocyte precursors arising most likely from the subventricular zone (1, 35, 94). This is supported by the demonstration of a massive induction of platelet-derived growth factor receptor alpha expression after oestrogen and progesterone. Of particular interest, microglial responses during cuprizone challenge appear to be detrimental and valuable at the same time because remyelination might depend to some extent on the removal of cell debris and toxic compounds from the axonal surrounding.
Studies in progress in our laboratory now additionally show that the synthesis and release of distinct chemokines, such as CCL2 and 3, during the early phases of demyelination and well before actual signs of demyelination might play a decisive role in disease progression. It may be assumed that astroglia and/or damaged oligodendrocytes account for this chemokine surge. The role of defective axons in this scenario is not at all clear but should not be neglected. The use of selective knockout mice for these chemokines will help to better understand the participation of these chemoattractant proteins in local microglia/macrophage attraction and activation during demyelination, and also whether gonadal steroid hormones are capable of modulating chemokine responses. Despite these uncertainties, it becomes more and more evident that the regulation of local brain-intrinsic inflammatory events is as important as the dampening of peripheral immune responses with respect to obtaining full neuroprotection by sex steroids.
Neuronal death and dysfunction in stroke and MS develop on a background of complex rapid and/or long-lasting physiological changes and pathophysiological processes that depend on alterations of mitochondrial properties and oxidative stress with respect to ischaemia, as well as on oligodendrocyte damage with respect to MS (1). Local inflammatory incidents involving central and peripheral immune-responsive and -active cells might precede or accompany neurodegeneration. In both diseases and in accordant animal models, gonadal steroids provide a striking reduction of tissue injury and preservation of functional integrity. Oestrogen and progesterone have been demonstrated in numerous basic, preclinical and clinical studies to interact and influence a wide variety of beneficial and potentially important therapeutic cellular and molecular activities. Each of them might play a particular role for the damaged tissue, and one might be lead to assume that the concurrence of a multitude of regulatory steps of physiological events is finally responsible for the observed neuroprotection. The importance of such an entanglement of mechanisms cannot be denied. Nevertheless, we and others consider there are key players (cells, molecules) in this scenario that need to be controlled to achieve an optimum of neuroprotection, possibly irrespectively of the disease. Chemokine–microglia interactions could be attributed to such an important function because they usually occur early during brain inflammation and determine the magnitude of cell damage. Recent studies have now elaborated that gonadal steroids regulate local chemokine availability, and most likely also signalling, in a number of neurodegenerative animal models. Although we are still are only beginning to understand the complexity of steroid–immune interactions in the brain, we can learn from the pathological processes appearing in other tissues and organs, such as the cardiovascular system, endometrium and cancer, where gonadal steroids play an active role in dampening or regulating local inflammatory episodes (98–100). It remains to be seen whether this novel perception might help to better understand sex steroid-mediated neuroprotection in the brain and thus lead to additional potential clinical implications.
Technical support provided by H. Helten and U. Zahn is acknowledged. The work was supported by IZKF BIOMAT (C.B.) and START (M.K.) of the Faculty of Medicine, RWTH Aachen University, B. Braun Melsungen AG (C.B.), the DFG (M.K.) and the Hertie-Foundation (M.K.).