In this study, we found that the sexually biased transcriptome was distinct in proestrus
and diestrus females. Surprisingly, while relatively few genes were sexually biased,
we observed a profound reorganization of the rat mPFC transcriptome throughout the
estrous cycle with 10–14 times more DEG between proestrus and diestrus than between
males and females in either estrous cycle stage. While both female groups exhibited
alterations in cellular communication when compared to males, proestrus females displayed
widespread up-regulation of genes and exon usage related to synaptic neurotransmission,
which represented the preponderant sex?×?estrous cycle alteration. Furthermore, proestrus
females showed a specific down-regulation of Egr1 levels, along with variations in its binding at synaptic function-related genes.
This association was particularly strong for the genes being also differentially expressed,
revealing a direct involvement of Egr1 in the specific transcriptomic signature of proestrus.
The advances in genome-wide analyses tools such as microarrays and RNA-seq have allowed
a better understanding of sex differences in gene expression in various tissues and
organisms. As such, and contrary to somatic tissues such as the liver, adipose, or
muscle tissue, a limited sex bias is observed in the brain 16]–18], 21]. In the rat mPFC, we found 67 sexually biased genes (0.43 % of all detected genes),
in line with the human PFC where the sex bias was only 0.14–0.24 % (representing 22–35
genes, respectively) 17], 18]. Such small differences can appear limited in view of the profound differences in
mPFC physiology 6], and could result from the high cell-type heterogeneity within cortical tissues 22]. Nevertheless, given the relatively wide sex bias observed in genes related to protein
degradation and translation, an amplification at the protein level cannot be ruled
out. Notably, we show that, despite not affecting the number of DEG, the estrous cycle
does impact the biological significance of the sexually biased transcriptome. Interestingly,
while confirming sex differences in translation, ECM organization, and mitochondrial
function described in the mouse and human brain (including PFC) 16], 17], 23], we detailed partly opposite regulations in proestrus and diestrus females when compared
to males or to each other. Indeed, interaction with the ECM was higher in proestrus,
and mitochondrial function and translation were enriched in diestrus. In line with
our observations, such up-regulation of mitochondrial function has already been reported
in the rat mPFC, where cytochrome c oxidase activity was greater in diestrus than
estrus 24]. It is particularly interesting to note that in addition to the multitude of variations
observed within females, the majority of genes were up-regulated in proestrus, but
down-regulated in diestrus when compared to males (Fig. 4, cluster 1). As a result, this variability is likely to mask a substantial number
of sex differences specific to either stage of the estrous cycle. Variations in synaptic
function, for instance, are among such processes whose enrichment was revealed after
discriminating between proestrus and diestrus, and illustrate the importance of accounting
for the estrous cycle in experimental designs 3].
Given the extent of gene regulation observed within females, we favored an analysis
of biological pathways and processes over gene-based or transcript-based investigations
in order to capture an integrative understanding of the sex and estrous cycle effects,
and thus identified a multi-level up-regulation of processes related to synaptic transmission
in proestrus. Accordingly, without consideration of the estrous cycle, female rats
display fewer neurons and glia, smaller structure volume, lower spine density, as
well as shorter and less branchy apical dendritic arbors than males in the mPFC 25]–27]. However, as illustrated by our study, such differences are affected by the estrous
cycle and depend on cyclic ovarian hormone fluctuations, with higher spine density
in cortical neurons dendrites in proestrus than estrus or diestrus 28], 29]. In other brain regions such as the hippocampus, these regulations are paralleled
by variations in synaptic activity and plasticity 10], with male rats generally displaying slightly higher basal activity than females
30], 31]. Within females, however, the estrogen surge in proestrus increases the excitability
of CA1 and CA3 pyramidal neurons over other estrous cycle stages, including diestrus
31], 32]. It is interesting to note that such effect is dependent on the slow rise of estrogen
the preceding day 33], which thus strengthens the importance of cyclicity in hormonal fluctuations. Although
similar regulations have been reported in other regions, the effects of sex and the
estrous cycle on the mPFC electrophysiology remain unclear. Gamma-aminobutyric acid
(GABA) binding, as well as mPFC contents of serotonin, dopamine, and their respective
metabolites, differ between male and female rodents and vary across the estrous cycle
34]–36], whereas glutamatergic transmission is higher in proestrus than in diestrus 37]. Through widespread modulations of these systems, ovarian hormones represent the
main candidates in mediating such variations 10]. Interestingly, estrogen treatment in ovariectomized rats induces, in the mPFC, a
reorganization of genes and pathways common with our study, including neurotransmission,
signal transduction, transport, transcription, ECM, and cell adhesion 15]. Furthermore, in our study we found 399 genes (41 % of DEG) with known regulation
by estrogens or progestins between proestrus and diestrus, 63 of which (16 %) are
related to synaptic function at multiple levels, including synaptic assembly, neurotransmitter
release and metabolism, and ion channels, as well as post-synaptic receptors and their
downstream signaling (Additional file 14: Table S7; Additional file 15: Table S8). Combined with the enrichment of processes related to neurotransmission
at the structural (cell–cell junction, ECM, focal adhesion, actin cytoskeleton), receptor
(G protein-coupled receptor signaling, intracellular signaling pathways), and transporter
(slc-mediated transmembrane transport, transmission synapse) levels in proestrus,
these observations strongly support substantial sex differences in synaptic structure,
function, and plasticity governed by hormonal fluctuations across the estrous cycle.
We observed a strong contribution of the neuronal cell type (both “pyramidal neuron,â€
and “interneuronâ€) to the profile of gene expression between proestrus and diestrus
females (Additional file 11: Figure S6; Additional file 12: Figure S7). In addition to further supporting the preponderance of synaptic transmission
in regulation by the estrous cycle (Fig. 4), this finding strengthens existing observations that both excitatory and inhibitory
transmissions are affected by the estrous cycle or ovarian hormones 34], 37]. Similarly, the suggested contribution of pericytes, vascular smooth muscle, and,
to a lesser extent, endothelial cells to the regulations by sex are in line with the
known sex differences in brain morphology as well as the cerebrovascular system and
endothelial cell function and reactivity 5], 25]–27], 38]. Notably, sex hormones, and particularly estrogen, strongly regulate cerebrovascular
and endothelial cell function and reactivity 38]–40], and alter the expression of genes related to vascular transport 15], which could thus explain the more pronounced enrichment of the mural and endothelial
cell types in the DEG between proestrus and males than between diestrus and males.
Similarly, we observed a suggested contribution of oligodendrocytes to the regulations
by sex or the estrous cycle, although at different maturation stages, in accordance
with the known sex differences in oligodendrocytes number, white matter volume, myelin
sheaths, and myelinated fibers in the rodent brain 41]. Given the crucial roles played by each of these cell types in regulating neuronal
signal transmission, these observations strengthen evidence for the widespread nature
of the regulation of synaptic transmission by sex and the estrous cycle.
Among the genes affected by sex or the estrous cycle, we detected several key transcription
factors likely to explain part of the transcriptomic and biological pathway regulations.
For instance, it is interesting to note that the transcription factors Fosb, Maff, Bcl6b, and Klf4—all down-regulated in females when compared to males—exhibit high expression in endothelial
and mural cells in the mouse brain 22], 42], and regulate endothelial cell function and ECM components 43]–46] in line with the enrichment of these cell types in the sexually biased gene expression
profiles. Furthermore, Maff and Klf4 are also induced in response to nerve-growth factor in rat PC12 cells alongside several
other immediate early genes, such as Egr1,2,4 or Id1, that are all differentially expressed in proestrus females when compared to males
(Additional file 2: Table S1), which highlights their involvement in neuronal function. Nevertheless,
among these regulations, we detected an interesting alteration of the immediate early
gene Egr1. Indeed, Egr1 was down-regulated in females, and was identified as a main contributor to the down-regulation
of transcription-related processes between proestrus females and males. Furthermore,
Egr1 was revealed as a main candidate transcription factor associated with sexually biased
genes in proestrus females by an in silico analysis (Additional file 13: Table S6), which was further detailed by our ChIP-seq analysis by underlying its
particular involvement in the transcriptional signature of proestrus females, with
a direct association with synapse-related genes.
Despite its widely accepted involvement in and regulation by synaptic activity, the
exact targets of Egr1 and their respective connection to the control of synaptic functions
remain unclear 47]. Indeed, in the cortex, Egr1 is induced by synaptic activity or major signaling factors such as Elk-1, NF-?B,
Egr1 itself, or the mitogen-activated protein kinase (MAPK) pathway 47]–49]—which, notably, was up-regulated in proestrus. Moreover, although its regulation
by ovarian hormones in the mPFC remains to be characterized, Egr1 is at the center of a gene regulation network induced by estrogen in the mouse mammary
gland 50], and while estrogen up-regulates Egr1 mRNA in the mouse uterus via activation of the MAPK pathway, co-treatment with progesterone
dampens this effect 51]. Because tissues used in our study were collected in the early afternoon of proestrus,
the down-regulation of Egr1 mRNA in proestrus may result from the early rise in progesterone levels in this stage
of the cycle. Nevertheless, proestrus females exhibited a widespread differential
binding of Egr1 to its transcriptional targets when compared to males or diestrus
females, suggesting enhanced Egr1-mediated transcriptional regulations despite lower
mRNA levels. We could thus identify a proestrus-specific alteration of transcriptional
regulators including Egr1 itself and some of its targets previously associated, in different systems, with
pathways and processes regulated by the estrous cycle in the rat mPFC. Indeed, Egr1 regulates ECM composition, mitochondrial function, apoptotic processes, signal transduction,
and transcription, through the transcriptional control of a variety of genes 52]–54], of which several are differentially expressed in proestrus. Furthermore, in vitro
evidence for a role of Egr1 in the control of synaptic functions in neurons exist, as its overexpression in rat
PC12 cells affects the expression of 135 genes—the majority being down-regulated—related
to synaptic function, including neurotransmitters, signal transduction, presynaptic
vesicular trafficking, synapse formation and assembly, and protein translation and
degradation 54]. In our study, we report a similar enrichment from the genes showing a differential
Egr1 binding in proestrus, especially among DEG (Table 3). Combined to the over-representation of Egr1 among the transcription factors associated with the gene expression profile of proestrus
females (Additional file 13: Table S6), these data support a direct control of synapse-related genes by Egr1
throughout the estrous cycle in the rat mPFC. It is important to note, however, that
the contribution of Egr1 is likely not exclusive because Egr1 can form heterodimers
with other transcription factors involved in neuronal function and activity, such
as Fos or Jun 47], 55], 56]. In addition to extending the range of Egr1 targets, this highlights an additional
layer of complexity in transcriptional regulations by sex and the estrous cycle.
Alongside synaptic changes, oxidative phosphorylation and ribosome-associated genes
were enriched in females, in line with previous transcriptomic and enzymatic observations
16], 17], 23]. While both these processes essential to synaptic plasticity vary throughout the
estrous cycle and are regulated by ovarian hormones in various systems 24], 57]–61], we revealed a diestrus-specific up-regulation of their related genes in the mPFC.
It is important to note, however, that this suggested down-regulation of mitochondrial
function and translation in proestrus appears in contradiction with the classically
reported enhancing effects of estrogen 61]. Nevertheless, while estrogen and progesterone both stimulate mitochondrial function
when analyzed separately, progestins can antagonize estrogens’ effects in the female
rat brain 61], and extensively down-regulate ribosome-associated genes in other systems 59], 60]. This down-regulation in proestrus could thus result from the early rise in progesterone
at this stage of the cycle. Similarly, the enhancing effects of estrogens—rising in
the morning of proestrus—on these processes are thought to represent rapid non-genomic
effects 62], which could therefore be on the decline in the afternoon of proestrus. Nevertheless,
genomic and non-genomic effects of sex hormones are not exclusive but may rather act
in concert 63] and result in complex transcriptomic variations throughout the estrous cycle. Non-genomic
effects, for instance, were proposed to initiate a rapid enhancement of synaptic plasticity
through MAPK-dependent and Akt-dependent signaling and actin cytoskeleton remodeling
that would be further stabilized in the event of sustained synaptic activity 62]. Because MAPK, Akt, and actin remodeling pathways are enriched in proestrus, it is
tempting to hypothesize that their gene expression profile would thus prepare the
female mPFC for the hormonal surge in proestrus. Interestingly, following a small-scale
study of protein expression by western blotting, we could confirm the profiles of
regulations of the majority of targets assessed (Additional file 3: Figure S2c, d), suggesting that the alteration of translation-related processes
observed between sexes and estrous cycle stages represents an additional component
of the regulation of synaptic transmission, as previously suggested 57]. Nevertheless, we cannot rule out the existence of compensatory mechanisms at the
protein level affecting different targets or biological pathways.
Females in proestrus thus exhibit a widespread transcriptomic reorganization, partially
under the control of Egr1, that suggests differences in synaptic activity in the mPFC
when compared to either males or diestrus females. Because the integration of afferent
signals by the mPFC is critical in controlling both perception of the environment
and the corresponding behavioral response 7], these transcriptomic variations could underlie sex and estrous cycle differences
in perception and response to anxiogenic environments. Interestingly, sex-dependent
and estrous cycle-dependent variations in anxiety levels and perception of aversive
elements are reported in both women and female rodents, although are variable between
strains and experimental paradigms 9], 10], 64]. In addition, estrogen, progesterone, and their metabolites can alter anxiety levels
through modulation of dopaminergic, serotoninergic, and GABAergic systems 10], whose mRNA levels vary between proestrus and diestrus. Interestingly, in premenstrual
dysphoric disorder, anxiety and depressed mood occur around the onset of menstruation
and present with different sensitivities of serotonin and GABA receptors 10]. Furthermore, our transcriptomic regulations overlap with those recently identified
as critical regulators of anxiety behaviors in the male mouse mPFC and several neuropsychiatric
disorders 65], supporting the association of our DEG with depression, behavioral diseases, and
bipolar disorders (Additional file 7: Figure S4). Notably, alterations in genes related to synaptic assembly and transmission,
cell communication, mitochondrial function, protein translation and degradation, or
neurotransmitter systems are recurring features reported in a variety of brain regions
upon cognitive decline or neuropsychiatric disorders 66]–69], with which Egr1 is associated 66], 67]. Altogether, these clinical and pre-clinical data suggest a critical role for transcriptomic
regulations in the adult rat mPFC by the estrous cycle in modulating the organism’s
interaction and perception of its environment.