C/EBPβ/AEP Signaling Drives Alzheimer’s Disease Pathogenesis

Abstract

Alzheimer’s disease (AD) is the most common type of dementia. Almost two-thirds of patients with AD are female. The reason for the higher susceptibility to AD onset in women is unclear. However, hormone changes during the menopausal transition are known to be associated with AD. Most recently, we reported that follicle-stimulating hormone (FSH) promotes AD pathology and enhances cognitive dysfunctions via activating the CCAAT-enhancer-binding protein (C/EBPβ)/asparagine endopeptidase (AEP) pathway. This review summarizes our current understanding of the crucial role of the C/EBPβ/AEP pathway in driving AD pathogenesis by cleaving multiple critical AD players, including APP and Tau, explaining the roles and the mechanisms of FSH in increasing the susceptibility to AD in postmenopausal females. The FSH-C/EBPβ/AEP pathway may serve as a novel therapeutic target for the treatment of AD.

Introduction

Alzheimer’s disease (AD) is an age-dependent neurodegenerative disease and the most common type of dementia, leading to progressive memory loss along with neuropsychiatric symptoms and a decline in the activities of daily life. The pathological hallmarks of AD include the extracellular senile plaque, primarily composed of β-amyloid peptide (Aβ), and intracellular neurofibrillary tangles (NFTs), mainly consisting of hyperphosphorylated and cleaved forms of the microtubule-associated protein Tau [1]. In addition, chronic neuroinflammation and neuronal death are involved in AD pathology. Numerous hypotheses have been proposed to trigger AD pathogenesis [2–6], including the amyloid cascade hypothesis, the Tau hypothesis, the acetylcholine hypothesis, the neuroinflammation hypothesis, and the oxidative stress hypothesis. The female prevalence of AD is well documented, and ~2/3 of patients with AD are women [7]. However, it is unclear why women are susceptible to AD.

The metabolic dysfunction of Aβ and the abnormal phosphorylation of Tau promote the pathologies of AD and ultimately contribute to the clinical features of dementia. Aβ is a peptide consisting of 38 to 43 amino acids and is derived from the sequential cleavage of amyloid precursor protein (APP) by a group of secretases [8, 9]. APP is fragmented via non-amyloidogenic and amyloidogenic pathways. In the amyloidogenic pathway, β-secretase (BACE1) cleaves APP at M596, and the remaining C-terminal portion of APP is subsequently cleaved by γ-secretase at V636 and A638 to generate insoluble Aβ. In the non-amyloidogenic pathway, APP is cleaved by α- and γ-secretase and thus precludes Aβ generation. Aβ deposition leads to dendritic and axonal atrophy, followed by neuronal cell death [10]. With the exception of α-, β- and γ-secretases, other proteases have also been found to cleave APP and regulate its functions [11–13]. On the other hand, Tau is reported to mediate Aβ-induced neurodegeneration and plays a crucial role in AD [14]. Tau cleavage disrupts the function of microtubules in axons and leads to its aggregation and neurodegeneration. Therefore, understanding the molecular mechanisms underlying Aβ and tau pathology is essential for the development of new treatments for AD.

In recent years, we have reported that chronic neuroinflammation activates the transcriptional factor CCAAT-enhancer-binding protein (C/EBPβ), which activates the transcription of the cysteine protease asparagine endopeptidase (AEP). AEP cleaves both APP and Tau, and promotes their deposition, leading to AD pathology [15–17]. Hence, any triggers in aged women in whom the C/EBPβ/AEP pathway is selectively activated may result in AD pathologies, which explains higher susceptibility for AD onset in women. Here, we summarize the roles of the C/EBPβ/AEP pathway in AD pathogenesis, (Table 1) and how follicle-stimulating hormone (FSH) promotes C/EBPβ/AEP pathway activation and increases the susceptibility to AD in females after menopause.

Table 1.

The role of the C/EBPβ/AEP pathway in AD.

Model	Age	Sex	The role of CEBPβ/AEP pathway in AD	Reference
Wild-type, AEP−/−, 5xFAD, 5xFAD/AEP−/− APP/PS1, and APP/PS/AEP−/− mice	2, 3, 4, 5, 6, 10, 14, 17 months	male	AEP cleaves APP at N373 and N585 residues, promoting amyloid plaque formation	[15]
C57BL/6J, Tau P301S, Tau P301S/Lgmn−/− and Lgmn−/− mice	1, 2, 4, 6, 8, 13 months	male	AEP cleaves Tau at N368 and N255 residues, inducing NFT formation	[16]
Wild-type, AEP−/−, 5XFAD, 5XFAD/AEP−/−, APP/PS1, 3XTg and 3XTg/AEP−/− mice	4, 8, 9, 10, 12 months	male	AEP cleaves BACE1 at N294 residue and augments the BACE1 enzymatic activity, accelerating Aβ production.	[33]
C57BL/6J, 3xTg and 5xFAD mice	5 months	both genders	SRPK2 phosphorylates AEP and induces AD onset	[38]
3×Tg, 3×Tg/C/EBPβ+/−, and 3×Tg/AEP−/− mice	4, 8, 12, 17 months	both genders	C/EBPβ and AEP increase with age in AD	[53]
Wild-type, 3xTg, 5xFAD, Cebpb−/− mice	2, 3, 6, 9, 10, 12 months	both genders	C/EBPβ regulates AEP expression and mediates AD pathogenesis	[17]
Wild-type, Tau P301S, SNCA-Tg, Lgmn−/−, BDNF flox/flox, and TrkB flox/flox mice	3, 5, 6, 9, 10, 16 months	both genders	BDNF inhibits AEP via AKT phosphorylation, while BDNF deficiency activates the JAK2/STAT3 pathway, and upregulation of C/EBPβ/AEP signaling.	[39, 54]
5xFAD	5, 6 months	Both genders	TrkB agonist activates BNDF/TrkB pathway, inhibits AEP activation, and prevents AD pathogenesis	[43–45]
Wild-type, 5xFAD, APP/PS1, AEP−/− and 5xFAD/AEP−/−, BDNF flox/flox, TrkB flox/flox mice	6, 9 months	both genders	TrkB receptor cleavage by AEP at N365 and N486/489 residues abolishes its phosphorylation of APP, promoting AD pathogenesis	[58]
Wild-type, 3xTg, Netrin-1flox/flox, and APP/PS1 mice	6 months	both genders	AEP cleaves UNC5B at N467 and N547 residues, enhances subsequent caspase-3 activation, additively augmenting neuronal cell death	[65]
Germ-free 3xTg and 5XFAD mice	4, 6, 8, 12, 17 months	female	Gut dysbiosis induces chronic inflammation via PUFA metabolism, and stimulates C/EBPβ-AEP activation, resulting in insulin/IGF-1 signal dysfunction in AD mouse models	[76]
3xTg, 5xFAD, C57BL/6J mice	6 months	both genders	C/EBPβ/AEP activation in the gut induces amyloid pathology and NFTs, and these pathological changes translocate along the vagus nerve into the brain, promoting AD onset	[77]
Wild-type, Thy1-ApoE4-human-C/EBPβ, and Thy1-human-C/EBPβ transgenic mice	3, 6, 8, 9, 15 months	both genders	Neuronal ApoE4 promotes C/EBPβ activation and induces AD pathology.	[84]
C/EBPβ+/+, C/EBPβ+/−, C/EBPβ−/−, 3xTg, 3xTg/ C/EBPβ+/−, 3xTg/ C/EBPβ−/−, 5xFAD mice	6 months	both genders	C/EBPβ is a transcription factor for ApoE, modulating ApoE4's role in AD	[83]
Wild-type, 3xTg, 3xTg/C/EBPβ+/−, APP/PS1, APP-knock in mice	7 months	female	FSH induces AKT–SRPK2, ERK1/2 phosphorylation, and promotes C/EBPβ/AEP pathway activation, leading to AD onset.	[96]

APP, amyloid precursor protein; ApoE, apolipoprotein E; BACE1, β-secretase 1; SRPK2, serine-arginine protein kinase 2; PUFA, poly-unsaturated fatty acid; NFTs, neurofibrillary tangles.

Chronic Inflammation Activates the C/EBPβ/AEP Pathway, Leading to AD Pathologies

AEP Activation in the Brain

Mammalian AEP, also called legumain or δ-secretase, is a lysosomal cysteine protease. It is mainly activated by acidosis and specifically cleaves its substrates at the C-terminus of asparagine residues [18–20]. Previously, we reported that AEP is activated by acidosis after ischemia or excitotoxicity. Activated AEP cleaves SET, the nuclear protein inhibiting DNase, at the N175 residue and leads to neuronal cell death [21]. Physiologically, AEP is bound and inhibited by cystatin C in lysosomes. However, the cystatin C levels in the CSF and brain are lower in AD patients than in control individuals, resulting in AEP activation [22–24]. Cystatin C is a secreted protein. It binds to soluble Aβ and inhibits its oligomerization, preventing neurodegeneration in AD [25]. Therefore, these findings suggest that AEP plays a role in the pathogenesis of AD.

AEP Cleaves APP and Promotes Aβ Production

The generation of neurotoxic Aβ peptide from APP is a crucial step in the pathogenesis of AD. We reported that AEP is activated in an age-dependent manner and cleaves APP at both the N373 and N585 residues. The APP fragment levels generated by AEP are related to the activation of AEP. Notably, the AEP-generated APP fragment (586–695) is more easily cleaved by β- and γ-secretases to generate Aβ, while the APP fragment (1–373) is toxic to neurons. These results indicate that AEP possesses secretase activity, therefore, we renamed AEP as δ-secretase. Knockout of AEP from APP/PS1 and 5XFAD mice alleviates synaptic loss, Aβ deposition, and cognitive dysfunctions [15]. Collectively, these results suggest that AEP contributes to AD pathology by regulating Aβ pathology.

AEP Mediates Tau Pathology in AD

NFTs, composed of truncated and phosphorylated tau, is another pathological hallmark of AD in addition to Aβ. Tau fragmentation disturbs its microtubule assembly activity and leads to tau aggregation and neurotoxicity [26]. Tau is a substrate of multiple endogenous proteases, including caspases [27, 28], calpains [29, 30], thrombin [31, 32] and AEP [16]. We have reported that AEP cleaves tau at both the N368 and N255 residues. The effect of AEP on tau cleavage is independent of caspases or calpains. Of note, the AEP-generated Tau (1-368) is abundant in the brain of AD patients but barely detectable in the control brain. Moreover, Tau (1–368) is more prone to aggregate and is highly neurotoxic. Knockout of AEP from Tau P301S mice partially alleviates tau deposition and memory loss [16].

AEP is the only reported protease that simultaneously cleaves both APP and Tau, leading to AD pathogenesis. We further found that AEP binds to BACE1 and cleaves it at the N294 residue in an age-dependent manner. Interestingly, the truncated BACE1 enzymatic domain (1-294) exhibits increased secretase activity and accelerates Aβ production, promoting AD pathogenesis and cognitive dysfunctions in an APP/PS1 AD mouse model [33]. Of note, the Tau (1-368) fragment cleaved by AEP binds and activates signal transducer and activator of transcription 1 (STAT1), one of the main transcription factors for BACE1, and facilitates BACE1 transcription and Aβ production [34]. Thus, we showed that δ-secretase not only directly cleaves APP but also enhances the activity of β-secretase in AD pathogenesis. Inhibiting AEP activity by compound #11, a small molecular inhibitor of AEP, reduces Tau and APP cleavage, ameliorating AD pathology and cognitive dysfunctions in Tau P301S and 5XFAD mice [35].

Post-translational Modification of AEP

AEP is synthesized as an inactive proenzyme consisting of a caspase-like catalytic domain and an LSAM (legumain stabilization and activity modulation domain) prodomain in the C-terminus, which are linked by an activation peptide. For activation, AEP undergoes pH-dependent (acidosis) autoproteolytic activation, whereby the C-terminal and N-terminal propeptides are released. Activated AEP is usually located in lysosomes but is also found extracellularly and is even translocated to the cytosol or nucleus [19, 36, 37]. The molecular mechanisms by which these subcellular translocations are regulated are not fully understood. We have found that phosphorylation of AEP regulates its subcellular localization and enzyme activation [38, 39]. Brain-derived neurotrophic factor (BDNF) is one of the most important neurotrophins that regulates neuronal development, differentiation, and survival in both the peripheral and the central nervous systems. BNDF binds to the TrkB receptor and triggers the activation of numerous downstream signaling cascades, including PI3K/Akt, Ras/Raf/MAPK, and PLC-γ1 [40, 41]. We have shown that Akt phosphorylates AEP on residue T322 upon BDNF treatment. Phosphorylation of AEP on residue T322 promotes its lysosomal translocation and inhibits its activation. BDNF levels in the brain decrease with age and neurodegenerative diseases, such as AD and Parkinson’s disease (PD). BDNF deprivation blocks phosphorylation of the AEP T322 residue and triggers AEP activation and cytosolic translocation, where it cleaves Tau at the N368 residue [39]. The Tau N368 fragment subsequently binds to the TrkB receptor and blocks neurotrophic signals, promoting neuronal cell death [42]. Therefore, treatment of AD mouse models with a TrkB agonist diminishes AD pathologies and improves cognitive function by inhibiting AEP activity [43–45].

In contrast, we found that serine-arginine protein kinase (SRPK2), a cell cycle-regulated kinase, selectively phosphorylates AEP at the S226 residue, and accelerates its autocatalytic cleavage, triggering the cytoplasmic translocation and activation of AEP. AEP is highly phosphorylated at the S226 residue in AD brains, and this is correlated with elevated SRPK2 activity. Overexpression of a phosphorylation mimetic mutation (S266D) in AD mouse models promotes AEP activation and downstream APP and Tau cleavage, facilitating AD pathogenesis and cognitive impairment. However, a non-phosphorylatable AEP mutant (S226A) attenuates AD pathologies [38]. Remarkably, we showed that activated AEP cleaves SRPK2 at the N342 residue in human AD brains, and increases its nuclear translocation as well as kinase activity [46]. SRPK2 is known to play an important role in pre-mRNA splicing by phosphorylating SR (serine/arginine)-splicing factors and is abnormally activated in tauopathies including AD [47–49]. The fragment of SRPK2 cleaved by AEP phosphorylates the SR proteins including the SC35 and ASF/SF2-splicing factors, accelerating the pathological imbalances in 3R- and 4R-Tau by augmenting Tau exon 10 inclusion, leading to cognitive decline in human Tau transgenic mice [46]. Hence, our findings show that AEP and SRPK2 crosstalk may play an important role in various neurodegenerative diseases.

Transcriptional Regulation of AEP

The mRNA and protein levels of AEP in the brain increase with age [15, 16]. To explore how aging regulates AEP expression, we chose age-related transcription factors as indicated in the GenAge database for in-depth investigation. Finally, we found that CCAAT-enhancer-binding protein (C/EBPβ), an inflammation-regulated transcription factor, acts as a crucial transcription factor for AEP [17]. C/EBPβ belongs to the basic-leucine zipper DNA-binding protein family and regulates the expression of various genes essential for memory formation, neuroprotection, inflammation, and activation of microglia [50–52]. Recently, we found that the C/EBPβ/AEP axis is activated with age in different brain regions of AD mouse models, combined with downstream APP and Tau expression and proteolytic cleavage, mediating AD pathologies [17, 53]. Overexpression of C/EBPβ in AD mouse models increases AEP expression, facilitates AD pathogenesis, and worsens cognitive dysfunctions, whereas depletion of C/EBPβ decreases the expression of AEP, APP, and Tau, inhibits the downstream APP and Tau cleavage, reducing AD pathologies and restoring cognitive function. Knockout of AEP from 3xTg mice alleviates AD pathogenesis [17, 53]. Hence, C/EBPβ is the main transcriptional regulator of AEP. The C/EBPβ/AEP axis plays an important role in AD pathogenesis.

BDNF/TrkB neurotrophic signaling regulates neuronal development, differentiation, and survival. We have found that there is a molecular association between BDNF/TrkB signaling and the C/EBPβ/AEP axis in AD pathogenesis. For instance, deprivation of BDNF/TrkB from primary cultured neurons activates the JAK2/STAT3 pathway, resulting in upregulation of the transcription factor C/EBPβ, coupled with overexpression of AEP, APP, and Tau cleavage and neuronal loss [54]. Moreover, BDNF stimulates TrkB receptors to bind and phosphorylate APP at the Y687 residue, which promotes APP accumulation in the trans-Golgi network and diminishes its amyloidogenic cleavage, reducing Aβ production. It has been reported that the levels of BDNF decline, while AEP activity increases with age and in AD patients [55–57]. Notably, AEP cleaves TrkB at the N365 and N486/489 residues and abolishes its neurotrophic activity, decreases p-APP Y687, and alters its subcellular trafficking, inducing AD pathogenesis. Blockade of the TrkB cleavage attenuates AD pathology in 5xFAD mice, rescuing learning and memory [58]. Therefore, these results demonstrate that the reduction of BNDF/TrkB signaling elicits AEP upregulation and AD-like pathology by activating C/EBPβ. Consequently, the C/EBPβ/AEP axis and the BDNF/TrkB pathway modulate each other during AD pathogenesis.

In addition to repression of BDNF/TrkB signaling, C/EBPβ also binds to the promoter of netrin-1 and suppresses its mRNA transcription. Netrin-1 is a diffusible factor for axon guidance [59, 60]. It is a multifunctional secreted molecule implicated in tissue patterning and neuronal activity [61]. In the peripheral organs, netrin-1 is expressed in the intestinal epithelium and modulates inflammation [62]. Both BDNF and netrin-1 are dramatically decreased in the brain and the gut of PD patients. Conditional knockout of these trophic factors in the gut elicits dopaminergic neuronal loss, constipation, and motor dysfunctions. Interestingly, the C/EBPβ levels are inversely correlated with BDNF and netrin-1 in PD patients [63]. It is worth noting that a decrease in netrin-1 triggers the activation of AEP, which subsequently cleaves one of the netrin-1 receptors, UNC5C, in both AD and PD, exacerbating the pathologies [64, 65]. Hence, gut inflammation induces C/EBPβ activation that leads to both BDNF and netrin-1 reduction and triggers PD (Figure 1). Conceivably, C/EBPβ-mediated biological events might be early diagnostic biomarkers for AD and PD.

Fig. 1.

Schematic model for activation of the C/EBPβ/AEP pathway triggering the pathogenesis of Alzheimer’s disease (AD) and Parkinson’s disease (PD). Chronic inflammation or oxidative stress in the gut separately triggers activation of the C/EBPβ/AEP pathway in AD and PD models. C/EBPβ is a transcription factor that enhances AEP, ApoE, APP, Tau, α-synuclein, and MAO-B transcription. Activated AEP cleaves numerous substrates, including APP, Tau, and α-synuclein, promoting AD and PD pathologies. Moreover, AEP also cleaves TrkB receptors and the UNC5B receptor and inactivates BDNF or netrin-1 signaling, which further activates the C/EBPβ/AEP pathway and facilitates AD and PD pathologies. DSS, 1% dextran sodium sulfate; α-SNCA, α-synuclein transgenic mice; MAO-B, monoamine oxidase B; DOPAL, 3,4-dihydroxyphenylacetaldehyde.

Chronic Inflammation Triggers C/EBPβ/AEP Activation in AD Pathogenesis

Chronic neuroinflammation mediated by astrocytes and microglia has been identified as another pathological hallmark of AD [66–68]. Neuroinflammation in AD is activated by central stimuli, such as extracellular Aβ deposition [6], and peripheral stimuli such as endotoxins [69]. C/EBPβ is involved in neuroinflammation. It is not only a promoter of inflammatory mediators but can also be induced by the classical pro-inflammatory triad of IL-1β, IL-6, and TNF-α [70–73]. Hence, there is a feedback loop between neuroinflammation and C/EBPβ in the brain. The factors that trigger chronic inflammation activate the C/EBPβ/AEP pathway, mediating AD pathologies.

Converging evidence indicates that there is a bidirectional gut-brain axis. The imbalance of pro-inflammatory and anti-inflammatory bacteria induced by gut dysbiosis is associated with systemic inflammatory states in patients with cognitive dysfunctions and brain amyloidosis [74]. In our most recent study, we found that gut dysbiosis occurs in 5xFAD and 3xTg-AD mouse models in an age-dependent manner. Microbiota from aged 3xTg mice but not aged wild-type mice accelerates AD pathology in young 3xTg mice, accompanied by C/EBPβ/AEP pathway activation [75]. Moreover, antibiotic treatment represses the C/EBPβ/AEP pathway and alleviates AD pathologies. To investigate the roles of gut dysbiosis in triggering inflammation in the brain and its contribution to AD pathogenesis, we generated germ-free 3xTg mice and re-colonized the germ-free mice with fecal samples from AD patients. Notably, the gut microbiota from AD patients promotes C/EBPβ/AEP pathway activation, AD pathologies, and cognitive dysfunctions in ex-germ-free 3xTg mice. Remarkably, the gut microbiota enhances the pro-inflammatory pathway for poly-unsaturated fatty acid metabolism and then triggers C/EBPβ/AEP pathway activation, resulting in insulin/IGF-1 signal dysfunction in AD mouse brains [76]. Therefore, these findings suggest that modulation of peripheral inflammation caused by gut dysbiosis might be a promising therapeutic approach for treating AD.

To explore the roles of the gut-brain axis in the development of AD-like pathologies and to monitor the role of C/EBPβ/AEP signaling in these processes, 3xTg mice were chronically administered 1% dextran sodium sulfate (DSS), triggering gut leakage, or colonic injection of Aβ or Tau fibrils, or brain lysates from AD patients. We found that C/EBPβ/AEP signaling is activated in the gut of AD patients and 3xTg mice, initiating the formation of Aβ and Tau fibrils that spread to the brain. DSS treatment promotes gut leakage and facilitates AD-like pathologies in both the gut and brain of 3xTg mice in a C/EBPβ/AEP-dependent manner. Vagotomy selectively blunts this signaling, attenuates Aβ and Tau pathologies, and restores learning and memory. Injection of recombinant Aβ or Tau fibrils or brain lysates from AD patients into the colon of mice induces the spreading of pathology into the brain via the vagus nerve, triggering AD pathology and cognitive dysfunction [77]. In addition, C/EBPβ has also been found to mediate both α-synuclein and MAO-B mRNA transcription in PD pathogenesis [78]. Furthermore, AEP cleaves α-synuclein at the N103 residue and enhances the propagation of α-synuclein fibrils from the gut to the brain [78, 79]. As a result, inflammation activates C/EBPβ/AEP and initiates AD- and PD-associated pathologies in the gut, which are subsequently transmitted to the brain via the vagus nerve (Fig. 2).

Fig. 2.

The gut is the early scene for the initiation of AD/PD pathology. Schematic representation of how gut dysbiosis contributes to amyloid pathology, NFTs, or Lewy body formation associated with C/EBPβ/AEP pathway activation in AD and PD. Gut dysbiosis induces chronic inflammation in 3xTg mice and oxidative stress in α-synuclein transgenic mice, leading to activation of the C/EBPβ/AEP pathway. Activated AEP cleaves APP, Tau, and α-synuclein, and induces the formation of amyloid plaques, NFTs, and Lewy bodies in the gut. Finally, the pathological changes propagate along the vagus nerve into the brainstem, from where they spread to other brain regions and promote AD and PD pathology.

The ε4 allele of the apolipoprotein E (ApoE4) is the strongest genetic risk factor for late-onset AD. ApoE is primarily synthesized by glial cells in the brain, but it can also be expressed in neurons under stress. Neuronal ApoE4 is a neurotoxic driver of neurodegeneration [80, 81]. We found that ApoE4, combined with 27-hydroxycholesterol, elicits the upregulation of inflammatory cytokines (IL-6, IL-1β, and TNF-α) and promotes C/EBPβ activation [82]. C/EBPβ subsequently escalates APP, MAPT, ApoE4, and BACE1 transcription and induces APP and Tau cleavage by elevating AEP, promoting AD pathologies [83]. Based on these findings, we hypothesize that the constant activation of C/EBPβ by ApoE4 might trigger AD pathologies in the absence of any human APP or Tau mutation. Consequently, we generated a double transgenic mouse model that expresses both human ApoE4 and C/EBPβ driven by the neuron-specific promoter Thy1 [84]. Notably, the Thy1-ApoE4/C/EBPβ double transgenic mouse model exhibits amyloid deposits, Tau aggregates, and neurodegeneration in an age-dependent manner via endogenous mouse machinery [84].

FSH Activates C/EBPβ/AEP Signaling and Triggers AD

The Role of FSH in the Reproductive System

FSH is a gonadotropin synthesized and secreted by the anterior pituitary gland. Its secretion is regulated by the hypothalamic gonadotropin-releasing hormone [85]. FSH is a heterodimeric glycoprotein consisting of an α-subunit, which is common to other glycoprotein hormones, and a specific β-subunit. FSH binds to and activates the FSH receptor (FSHR), which is a member of the 7-transmembrane G-protein-coupled receptor (GPCR) family [86, 87]. The role of FSH in the reproductive system is well established: it acts on the female and male reproductive systems by binding to FSHRs in granulosa and Sertoli cells. In females, FSH controls the menstrual cycle, stimulates the growth and maturation of the granulosa cells of the ovarian follicle, and controls estrogen secretion [88]. In males, it promotes the growth and maturation of Sertoli cells and induces the production of androgen-binding proteins [89]. Knockout of the FSH β-subunit or the FSHR genes in mice results in significant reproductive defects in both sexes [90, 91].

FSH Beyond Fertility

In recent years, more and more studies have focused on the functions of FSH in extra-gonadal tissues, as FSHRs are expressed in various extra-gonadal tissues, such as bone, fat, malignant tissues, monocytes, and neurons [92–96]. During the menopausal transition, many women experience significant physiological changes, including bone loss and increased visceral adiposity (Table 1).

FSH and Bone Loss

A series of studies have shown a relationship between serum FSH and bone loss among peri-menopausal women. A survey of post-menopausal and peri-menopausal women aged 42-60 years showed that increased serum FSH levels are associated with 2.78 greater odds for osteoporosis (OR: 2.59, 95% CI: 1.49-5.42) [97]. Multi-center multi-ethnic cohort SWAN data indicated that bone turnover markers and bone mass density in peri-menopausal and early post-menopausal women are negatively related to serum FSH, and independent of serum estradiol [98, 99]. In vitro and in vivo studies have shown that FSH acts on bone both directly and indirectly. First, FSH binds directly to bone FSHRs on osteoclasts, stimulating MEK/Erk, NF-κB, and Akt activation and increasing osteoclastogenesis and bone resorption [92, 100–103]. Moreover, enhanced FSH secretion indirectly stimulates osteoclast formation by inducing pro-osteoclasts to secrete TNF-α, IL-1β, and IL-6 and expands the number of bone marrow osteoclast precursors [104, 105]. Blocking FSH with a specific antibody prevents bone loss by inhibiting bone resorption and stimulating bone formation [92, 106].

FSH and Body Fat

FSH levels show a strong correlation with body fat in post-menopausal women. A longitudinal study in the Michigan cohort of SWAN showed that increased FSH levels are positively associated with body fat and waist circumference in women during the menopausal transition [107]. In addition, a cross-sectional study of infertile women aged 20-35 years showed that serum FSH levels are positively correlated with central obesity, including waist circumference and waist/hip ratio [108]. FSH directly binds to the Gα_i-coupled FSHR in murine adipocytes and 3T3-L1 cells (a pre-adipocyte model), stimulates the expression of fat genes, such as Fas, Lpl, and Pparg, and inhibits Ucp1 activation in brown adipocytic Thermo cells, inducing lipid biosynthesis [93]. Another study found that FSH regulates fat accumulation and redistribution in aged males and females through the Gα_i/Ca²⁺/CREB pathway [109]. FSH blockade with a specific Fshβ antibody in mice increases the expression of brown fat genes in white adipose tissue, including Ucp1, Cox7, Cidea, and Cox8a. The antibody also robustly increases UCP1 expression and mitochondrial content in the brown adipose tissue and enhances thermogenesis [93]. These results indicate that FSH blockade reduces diet/menopause-induced fat mass.

FSH and AD

The female prevalence of AD among the aging population is well documented and approximately 2/3 of AD patients are women [110]. The cause of the gender difference in AD is unclear. The menopause transition, a midlife neuroendocrine transition state unique to females, may be an important factor associated with the onset of AD in women [111, 112]. An observational multimodality brain imaging study found that increased indicators of AD endophenotype in women, including hypometabolism and Aβ deposition, happen in peri-menopause [113]. The role of estrogen reduction after menopause in AD pathogenesis remains controversial because the estrogen-replacement therapy for AD in women after menopause show improvement [114], no change [115, 116], or worsening of cognitive function [117]. These results prompted us to think about whether other hormones besides estrogen deficiency promote AD pathogenesis in women.

The clinical observations during peri-menopause (years just prior to menopause) have shown that women experience a rapid rate of bone loss, visceral obesity, dysregulated energy homeostasis, and spikes of cognitive decline when estrogen levels are relatively unperturbed but FSH levels are rising [118–121]. High serum levels of FSH are highly associated with the onset of AD [122]. Notably, FSH strongly induces C/EBPβ expression in ovarian cells and Sertoli cells [123, 124]. C/EBPβ mediates the steroidogenic acute regulatory protein and prostaglandin-endoperoxide synthase 2 genes in ovarian granulosa cells upon FSH stimulation [124–126]. Furthermore, FSH also activates C/EBPβ transcriptional activity via cAMP in Sertoli cells [123]. Accordingly, the striking effect of FSH on C/EBPβ prompted us to investigate the role of FSH in AD pathogenesis.

In our latest study, we provided evidence of the effects of FSH on neurons, thus establishing FSH as an AD-promoting hormone [96]. First, FSHRs are expressed in the cortex and hippocampus of humans and mice. They are mostly expressed on neurons, with a little-to-no expression on glial cells [96]. Second, to explore whether FSH mediates AD pathologies in the 3xTg-AD mice model, we injected human recombinant FSH intraperitoneally (i.p.) into 2.5-month-old female and male 3xTg mice daily for 3 months. FSH strongly induces cognitive dysfunctions and escalates the levels of C/EBPβ and its downstream target AEP in both female and male 3xTg brains compared to controls. Notably, AEP is robustly activated, leading to APP and Tau fragmentation, and p-Tau elevation. Moreover, immunofluorescence revealed that FSH induces C/EBPβ/AEP activation and Aβ and p-Tau escalation mainly in neurons, confirming a primary neuronal action of FSH in AD pathogenesis. Likewise, FSH stimulates neuronal apoptosis in the hippocampus and cortex, with more plaque formation validated on hippocampal sections by thioflavin staining, and in the cortex and the hippocampal regions CA1 and dentate gyrus by silver staining. The plaques were more abundant in female than male mice. Hence, FSH administration increases the pathogenesis of AD in young female and male 3xTg mice, and female mice are more susceptible to the lesion than male mice [96]. To mimic the hormonal changes during the menopausal transition, when FSH is rising while estrogen is undisturbed, we replaced estrogen in ovariectomized 3xTg mice by implanting 17β-estradiol 90-day-release pellets before FSH injection (5IU daily) for 3 months. We found that high FSH induces C/EBPβ/AEP pathway activation, downstream APP and Tau cleavage, Tau phosphorylation, and Aβ and p-Tau accumulation in this estrogen-replenishment state. These data indicate that the effect of FSH on AD pathogenesis is estrogen-independent in female mice.

To further assess the role of FSH in AD pathogenesis, we used an anti-Fshβ antibody to block the action of Fshβ on FSHRs, by treating 3xTg mice with an anti-Fshβ antibody (200 μg per mouse, every 2 days, intraperitoneally (i.p.)) 4 days after ovariectomy for 8 weeks. Ovariectomy expectedly accelerates the formation of plaques, NFTs, and cognitive dysfunctions in 3xTg mice, combined with FSH elevation and estrogen reduction [96, 127, 128], associated with strong activation of the C/EBPβ/AEP pathway. As reported in previous studies, FSH-Ab does not disturb serum FSH, LH, or estrogen levels, but dramatically reduces ovariectomy-induced amyloid plaque and NFT formation, neuronal apoptosis, and dendritic spine reduction, and reverses the cognitive defects in 3xTg mice of AD [93, 106]. FSH-Ab also inhibits the activation of the C/EBPβ/AEP pathway induced by ovariectomy, APP and tau cleavage, and tau phosphorylation. Moreover, suppression of basal levels of serum FSH by FSH-Ab prevents Aβ accumulation in male APP/PS1 mice as well [96].

FSH Induces AD Pathologies in a C/EBPβ/AEP-Dependent Manner

The C/EBPβ/AEP pathway plays a crucial role in triggering AD pathogenesis [17, 53, 129]. To investigate whether the C/EBPβ/AEP pathway mediates the biological actions of FSH, we used human neuroblastoma SH-SY5Y cells and primary cultured rat cortical neurons. FSH induces C/EBPβ/AEP pathway activation in a time-dependent manner in both cell types, which subsequently cleaving both APP and Tau into APP N585 and Tau 368 truncates, respectively. Moreover, FSH increases the concentrations of the pro-inflammatory cytokines IL-1β and IL-6 in a time-dependent manner, fitting with the upstream p-C/EBPβ activation pattern. In addition, to investigate whether FSH-stimulated AD pathologies are mediated by the C/EBPβ/AEP pathway in vivo, we used young female 3xTg/C/EBPβ+/− mice that were treated with recombinant human FSH daily for 3 months. FSH treatment elevates C/EBPβ/AEP pathway activation and induces AD neuropathology and cognitive disorders in 3xTg mice, and these are prominently diminished in 3xTg/C/EBPβ +/− mice. Since high endogenous FSH contributes to the acceleration of AD after ovariectomy, we also examined the effect of ovariectomy in 3xTg/C/EBPβ+/− mice. Indeed, C/EBPβ haploinsufficiency reduced the neuropathology and alleviated the cognitive dysfunction induced by ovariectomy. Taken together, these data support the primary role of the C/EBPβ/AEP pathway in mediating FSH-induced AD pathology [96].

FSH mediates multiple signaling pathways by binding to its FSHRs [86, 87]. To test whether FSH activates the C/EBPβ/AEP pathway via FSHRs, we knocked down the receptors with siRNA and found that activation of the C/EBPβ/AEP pathway stimulated by FSH is abolished in both in vitro and in vivo models. Consequently, APP N585 and Tau N368 fragments are greatly diminished when FSHR is depleted. Collectively, these data suggest that FSH stimulates C/EBPβ/AEP signaling activation via binding to its specific receptor. Furthermore, FSH has been found to bind to FSHRs, inducing the phosphorylation of Akt, ERK1/2, and SRPK2, leading to activation of the C/EBPβ/AEP pathway and promoting AD pathogenesis (Figure 3).

Fig. 3.

The mechanism by which FSH triggers AD pathogenesis in female AD mice. Schematic representation of the mechanism of action of FSH in AD pathogenesis. FSH binds to FSHRs on the plasma membrane of neurons and then induces AKT-SRPK2 and ERK1/2 phosphorylation and promotes C/EBPβ/AEP pathway activation. Active AEP, which is phosphorylated on S226 by SRPK2, cleaves APP and Tau, accelerating AD onset.

Conclusion

During the menopausal transition, women experience a sharp change in hormones, with serum FSH elevation and estrogen reduction. The risk for AD, osteoporosis, obesity, and cardiovascular pathology increases at that time. High levels of FSH have now been implicated in AD pathogenesis through activating the C/EBPβ/AEP pathway (Figure 4), in addition to osteoporosis and visceral adipocytes. Blocking the action of FSH with an antibody against the FSHβ subunit improves cognitive dysfunction, increases bone mass, and reduces body fat, serum cholesterol, and coronary atherosclerosis in a mouse model [92, 93, 96, 118, 130, 131]. At present, Dr. Zaidi's group has developed a first-in-class humanized antibody to FSH [132], which provides a basis for preclinical or clinical studies on the treatment of cognitive dysfunctions with a specific humanized FSH antibody. Moreover, the antagonism of the FSHR, a GPCR that is frequently targeted by numerous FDA-approved medicines, may provide an unprecedented strategy to alleviate cognitive dysfunctions. Most importantly, inhibition of AEP by its specific inhibitors at the early stage of the disease may slow down AD pathology onset and progression [35], which will pave the way for developing a disease-modifying pharmacological intervention for this devastating disease.

Fig. 4.

High FSH levels induce AD pathologies via C/EBPβ/AEP pathway activation. FSH is secreted by the anterior pituitary gland under the control of GnRH. FSH stimulates the secretion of estrogen from ovaries in women, while estrogen has a negative feedback effect on FSH secretion. During peri-menopause, FSH levels increase sharply, while estrogen levels are relatively unperturbed. The high FSH levels may induce C/EBPβ/AEP pathway activation and promote AD pathologies.

Conflict of interest

The authors declare that they have no conflicts of interest.