Transcriptomic analysis reveals suppression of steroidogenic acute regulatory protein in gender-specific differences in Alzheimer’s disease

Abstract

Alzheimer’s disease (AD)-related dementia preferentially impacts two-thirds of women and one-third of men.

The steroidogenic acute regulatory (StAR) protein mediates the biosynthesis of neurosteroids that sustain diverse neuronal activities.

Aging, involving neurosteroidal imbalance, is the predominant risk factor for AD causing dementia.

Transcriptomic analysis, including clinical cognitive diagnosis (cogdx) stages, displays marked attenuation of StAR in brains of AD women than those of AD men, compared with cognitively normal (Non-AD) subjects.

The present data provide the first evidence and new insights into the mechanism exemplifying the suppression of StAR in gender-specific differences in AD.

1. Introduction

The steroidogenic acute regulatory (StAR) protein (also known as StAR-related lipid transfer domain-1, STARD1), by governing the transport of intra-mitochondrial cholesterol, mediates the rate-limiting step in steroid biosynthesis in pertinent tissues, which involves endocrine, autocrine, or paracrine mechanisms [1–3]. Endocrinosenescence, disrupting hormonal homeostasis, is an inevitable phenomenon in aging that is the principal risk factor for Alzheimer’s disease (AD). The lateonset sporadic AD is the most common neurodegenerative disorder of the elderly population that accounts for ~70 % of dementia cases, in which etiology and disease-causing mechanisms are complex and multifactorial [4–6]. Neuropathologically, AD is hall-marked by the accumulation of amyloid-β (Aβ) and phosphorylated Tau (P-Tau) in the hippocampus and neocortex regions due to a variety of harmful events, including occluded neurosteroidogenesis and mitochondrial dysfunction [4,7,8]. We reported that suppression of the StAR protein is inversely correlated with the build-up of Aβ and P-Tau [7]. Therefore, these opposing events exacerbate neuronal imbalance and result in a progressive loss of synaptic transmission and neurons, which gender-specifically activate AD-related dementia. Noteworthy, AD impacts two-thirds of women and one-third of men, debating the influence of sex differences of longer and shorter lifespans for the former and latter categories, respectively [4,9]. A question concerns whether AD pathology is impacted by StAR, thus, sex neurosteroids, involving progesterone and estrogen (menopause) in women, and testosterone (andropause) in men, respectively.

All neurosteroids are made from cholesterol that is derived from de novo synthesis, lipoprotein mediated cholesteryl esters (CEs), and hydrolysis of CEs in lipid droplets; however, the conversion of CEs into free cholesterol is a critical event in StAR-mediated neurosteroid biosynthesis [1,9,10]. Regulation of cholesterol in the central nervous system (CNS) is independent of peripheral tissues, as this lipid molecule is impermeable to the blood brain barrier. Whereas de novo synthesis of neurosteroids, influenced by StAR, in the CNS is crucial for various neuronal activities, including mitochondrial integrity and synaptic transmission, malfunction in these processes is detrimental to neurosteroidal inequality for triggering AD-related dementia [7,11,12]. Concomitantly, we have demonstrated that postmortem AD brains exhibit suppression of the StAR protein, along with neurosteroid biosynthesis, compared with age-matched Non-AD counterparts [7]. While the decrease of StAR in AD brains is in agreement with another study [12], contradictory findings involving the increase in StAR protein expression have been reported [13,14]. However, the latter studies lacked validation and reproducibility, as they neither determined any neurosteroids (considering StAR’s involvement in the regulation of steroid/neurosteroid biosynthesis) nor assessed their correlation to Aβ and P-Tau accumulation. Additionally, there were no information on clinical cognitive diagnosis (cogdx) and/or other stages on postmortem AD brain samples analyzed for StAR, and whether those specimens were from subjects treated with drugs that convalesce cognitive impairment. Regardless of these perplexing scenarios, no studies today have determined gender-specific expression of StAR, leading to a key step in precision medicine for personalized treatment of dementia in AD patients.

Whereas genome-wide association studies have considerably advanced our understanding of genetic anomalies involved in AD, the mechanisms underlying the disease pathogenesis remain obscure. Consequently, gender-specific log-transformed CPM data from the Religious Orders Study and Memory and Aging Project (ROS/MAP) were downloaded via the AD Knowledge Portal (https://www.synapse.org/), an open-access resource of NIA’s Translational Research Program involving AD and AD-related dementias (ADRDs) [15]. Analysis of bulk neuron tissue RNA sequencing (RNA-Seq) data was performed for diverse factors and processes, pertaining to the dorsolateral prefrontal cortex (DLPFC) brain region (634 samples), which sustains many cognitive activities, including decision making, memory, and executive function. The results revealed striking suppression of StAR (FDR < 0.001, False Discover Rate) in brains of AD women compared with Non-AD subjects, but unaltered between AD and Non-AD men (Table 1, top panel). Besides, genomic profiles of many cholesterol modifying and AD pathology inflicted factors/processes, i.e., HSD17B1 (influences estrogen synthesis), SREBF2 (mediates cholesterol trafficking and lipid biosynthesis), ACAT1 (controls protein and fat metabolism), HMGCR (participates in the biosynthesis of cholesterol), VDAC1 (an outer mitochondrial membrane interacts with StAR and impacts mitochondrial function), BACE1 (cleaves APP for the generation of Aβ peptides), and PSEN2 (a gamma secretase complex involved in the generation of APP), were significantly decreased (highlighted in yellow) in AD women’s brains compared with their Non-AD counterparts. Moreover, LIPE (catalyzes the hydrolysis of various substrates for controlling lipolysis), CYP27A1 (mediates the rate-limiting step 27-hydroxycholesterol biosynthesis), STARD3 (participates in neurosteroid biosynthesis), TSPO (participates in cholesterol transport and neurosteroidogenesis), CYP11A1 (catalyzes the conversion of cholesterol to pregnenolone), SREBF1 (a SREBF isoform impacts lipid homeostasis) and ABCA1 (controls cholesterol and phospholipid homeostasis), ApoE (influences lipid transport and metabolism), and PSEN1 (a PSEN isoform involved in APP processing) were unaltered in brains of AD and Non-AD women and men. Interestingly, these factors were not altered between AD and Non-AD men’s brains. Importantly, genomic profiles of APP and MAPT/Tau were markedly decreased in the brains of AD women, but not in AD men, compared with Non-AD subjects (Table 1). The suppression of APP and Tau genes, in contrast to the accumulation of Aβ and P-Tau [7,8], denotes transcriptional and translational regulation of these AD pathological markers. Statistical analyses for expression of various genes between AD and Non-AD women and men were computed and presented with FDR values. These analyses indicate that a plethora of factors/processes, especially StAR, are accountable for gender-specific variations in AD/ADRD risks.

Table 1.

Analyses of RNA-Seq data in the DLPFC region for gender-specific expression of various factors, including StAR (highlighted in red), involved in cholesterol trafficking and balance, neurosteroid biosynthesis, and AD pathogenesis, (N=sample numbers).

Gender-specific bulk neuron tissue RNA-Seq data for the DLPFC samples (ROS/MAP project) pertaining to 634 subjects with various clinical cognitive diagnosis (cogdx) stages, ages 67–90 years. Genomic expression profiles for a variety of factors and processes analyzed were: LIPE (hormone-sensitive lipase, a multifunctional enzyme), CYP27A1 (a cytochrome P450 oxidase, sterol 27-hydroxylase), StAR/STARD1 (the primary regulator of neurosteroid biosynthesis), STARD3 (a late endosomal StAR-related lipid transfer domain-3), CYP11A1 (a mitochondrial cytochrome P450 family 11 subfamily A member 1 enzyme), HSD17B1 (hydroxysteroid 17β dehydrogenase 1 enzyme), TSPO (an outer mitochondrial membrane translocator protein), SREBF1 (sterol regulatory element-binding transcription factor 1), SREBF2 (sterol regulatory element-binding transcription factor 2), ACAT1 (acetyl-coenzyme A acetyltransferase 1 enzyme), HMGCR (3-hydroxy-3-methylglutaryl coenzyme A reductase enzyme), ABCA1 (an ATP-binding cassette transporter A1), BACE1 (beta-site amyloid precursor protein cleaving enzyme 1), ApoE (apolipoprotein E), PSEN1 (presenilin-1), PSEN2 (presenilin-2), APP (amyloid-β precursor protein), and MAPT/Tau (microtubule-associated protein Tau). These events/factors influence cholesterol trafficking, metabolism, and homeostasis, thus, neurosteroidogenesis and AD pathogenesis. Statistical analyses were performed using the statistical software R (Version 4.3.3, Vienna, Austria; R Foundation for Statistical Computing). The eBayes function in the limma package was used to estimate the posterior probabilities of gender-based gene expression profiles for AD and Non-AD subjects. Posterior probabilities of differential expressions were calculated by shrinking the estimated gene-specific variances towards a common value. Data presented in the table summarize key descriptive statistics, including mean, and fold changes with 95 % confidence intervals for the targeted genes across different groups, stratified by gender. Statistical comparisons are reported for gender-specific FDR (False Discovery Rate) values adjusted for multiple testing with Benjamini-Hochberg correction, with significance defined as FDR < 0.05, which were highlighted in yellow. Ages were centered at 90 for statistical analysis, in which sample number varies across different AD and Non-AD subjects. Panels A (AD and Non-AD women) and B (AD and Non-AD men) reveal stage-based cogdx scores for women and men (shown under a red rectangle), respectively. Data presented in boxplots demonstrate cogdx dependent expression of StAR (designated as Non-AD = cogdx 0 through 5, based on the severity, where 4 and 5 cogdx groups were combined as 4/5 due to sample limitations) in AD and Non-AD women and men. Boxplots were generated using ggplot2 in R for evaluating distribution of StAR gene expression across different groups, in which boxes represent the interquartile range (IQR), with whiskers extending to 1.5 times the IQR. Group comparisons were statistically verified using a non-parametric test (either Kruskal-Wallis or Mann-Whitney U), and FDR values were provided on plots, where applicable. Panels C-E illustrate schematic representation of gender-specific alterations of sex neurosteroids, especially progesterone and estrogen (menopause, C), and testosterone (andropause, E), respectively. This model is extrapolated based on higher suppression of StAR (represented by arrows) in brains of AD women than those of AD men, compared with Non-AD subjects. Panel D shows a model in which StAR transports free cholesterol (cholesterol) from the outer mitochondrial membrane (OMM) to inner mitochondrial membrane (IMM), where cholesterol is converted to the first neurosteroid, pregnenolone, by the action of the cytochrome P450 side chain cleavage/CYP11A1 enzyme. Pregnenolone exits the mitochondria and it is then converted to various sex neurosteroids (C and E). Consequently, StAR-driven suppression of neurosteroids plays an important role in gender-based variations in dementia, representing higher prevalence of AD in women in comparison to lower prevalence of AD in men. OMM, outer mitochondrial membrane; IMM, inner mitochondrial membrane.

Gender-based analyses of RNA-Seq data on deterioration of StAR were further assessed with increased cogdx stages, since this cholesterol transporter regulates the biosynthesis of neurosteroids. An interesting aspect of the present finding is cogdx stage-dependent attenuation of StAR, which was robust at cogdx 3 and 4/5 in the brains of AD women compared with Non-AD (cogdx-0) subjects (panel A). However, genomic alterations of StAR were not significantly observed among men with cogdx stages (panel B). These data mirrored with age-related suppression of StAR and neurosteroid levels (3–18 months) in sex-specific humanized Aβ knock-in and Tau transgenic AD mouse models [8], parameters were more pronounced in females than those of male mice (Manna PR et al., manuscript in preparation). Based on these circum-stances, schematic models illustrating the plausible mechanisms underscoring StAR-driven suppression of sex neurosteroids involving menopause (progesterone and estrogen, panel C) and andropause (testosterone, panel E), promoting neurodegenerative vulnerabilities for higher and lower incidence of AD in women and men, respectively, are proposed. Panel D depicts the transport of StAR-mediated cholesterol (free cholesterol) from the OMM to the IMM where it is converted to pregnenolone (the precursor of all neurosteroids) by the CYP11A1 enzyme [7]. Pregnenolone exits the mitochondria and is metabolized to various neurosteroids in women and men. The mechanism accounting for gender-based suppression of StAR, thus sex neurosteroids, influenced by assorted signaling, disrupts neuronal homeostasis, resulting in dementia in AD/ADRD patients, and this premise has been exemplified in our findings with hippocampal neuronal cells and postmortem brains [3,7].

In conclusion, gender-based analyses of RNA-Seq data, pertaining to the DPLFC brain region, for diverse signaling processes, impacting cholesterol homeostasis and AD pathogenesis, provide the first evidence, and mechanistic insights into robust attenuation of StAR, which regulates sex neurosteroids, for the increased susceptibility of women to develop AD than in men, regardless of their varied lifespans. However, we recognize that sample numbers analyzed are different, which limit power for statistical robustness towards the disease-progressing events. Regardless of the mechanism involved, the suppression of StAR provoked neurosteroidogenic machinery, promoting neurodegeneration, is a persuasive occurrence for gender-specific differences in AD risks. Future studies will elucidate the underlying mechanisms involved in StAR-mediated upregulation of sex neurosteroids as novel therapeutic targets for preventing, delaying, or ameliorating dementia in AD women and men.

Abbreviations

StAR

steroidogenic acute regulatory protein

AD

Alzheimer’s disease

Non-AD

cognitively normal

Aβ

amyloid-β

APP

amyloid precursor protein

P-Tau

phosphorylated Tau

cogdx

clinical cognitive diagnosis

OMM

outer mitochondrial membrane

IMM

inner mitochondrial membrane

Data availability

Transcriptome analysis data generated in this study are included in this manuscript.