The regional effect of serum hormone levels on cerebral blood flow in healthy nonpregnant women

Abstract

Sex hormones estrogen (EST) and progesterone (PROG) have received increased attention for their important physiological action outside of reproduction. While studies have shown that EST and PROG have significant impacts on brain function, their impact on the cerebrovascular system in humans remains largely unknown. To address this, we used a multi-modal magnetic resonance imaging (MRI) approach to investigate the link between serum hormones in the follicular phase and luteal phase of the menstrual cycle (MC) with measures of cerebrovascular function (cerebral blood flow [CBF]) and structure (intracranial artery diameter). Fourteen naturally cycling women were recruited and assessed at two-time points of their MC. CBF was derived from pseudo-continuous arterial spin labeling while diameters of the internal carotid and basilar artery was assessed using time of flight magnetic resonance angiography, blood samples were performed after the MRI. Results show that PROG and EST had opposing and spatially distinct effects on CBF: PROG correlated negatively with CBF in anterior brain regions (r = -.86, p < .01), while EST correlations were positive, yet weak and most prominent in posterior areas (r = .78, p < .01). No significant correlations between either hormone or intracranial artery diameter were observed. These results show that EST and PROG have opposing and regionally distinct effects on CBF and that this relationship is likely not due to interactions with large intracranial arteries. Considering that CBF in healthy women appears tightly linked to their current hormonal state, future studies should consider assessing MC-related hormone fluctuations in the design of functional MRI studies in this population.

Keywords: arterial spin Labelling; cerebral blood flow; estrogen; menstrual cycle; neurovascular coupling; progesterone; time-of-flight.

FIGURE 1

Sagittal view of T1‐weighted image with overlay of maximum intensity projection of high resolution TOF slab depicting location of the ICA and BA arteries. Black box depicts segmented ICA and BA in sagittal view. The blue box depicts the region where ICA lumen diameters where averaged, the black box denote regions where BA lumen diameters were extracted and averaged

FIGURE 2

Idealized MC is shown in (a). Green line represents expected change in serum EST during the cycle, blue line represents expected change in serum PROG during the cycle. Gray rectangles represent approximately when the participants sessions were scheduled during the FP and LP. EST_FP and EST_LP in (b) and serum PROG_FP and PROG_LP are shown in (c). The relationship between EST_LP and PROG_LP is shown in ([D]; r = .43, p = .13)

FIGURE 3

Correlation results between EST_FP and PROG and CBF. Compared to chance (gray distribution is shuffled correlations), the majority of voxels tended to correlate positively between EST_FP and CBF_FP (green distribution[A1] Dashed gray line represents r = 0, dashed black line is r = 0.55 which was used for clustering threshold). The correlation was strongest in the posterior portions of the brain (A.2) and were weakly correlated. PROG correlated negative with CBF compared to chance (B1; gray distribution is the shuffled correlations, dashed gray line represents r = 0, dashed black line is r = −0.56, which was used for clustering threshold) The correlations were strongest in the anterior circulation (B2) and were strongly, negatively correlated with CBF

FIGURE 4

Controlling for serum EST in LP does not significantly alter the association between PROG_LP and CBF_LP (a). PROG_LP correlated with more GM‐voxels than EST_LP. Percentage of EST_LP clusters were not more than what would be expected by chance ((b); Black‐dashed line represents random chance, set as 2.5% or 2 SD of the shuffle distribution)

FIGURE 5

Correlations of EST_FP are primarily located in the perfusion territory of the PCA (a). PROG clusters are primarily located in the anterior perfusion territories (ACA and MCA; b). EST_FP clusters in red (c) and PROG_LP clusters in dark blue (d) are shown in relation to a 3D vascular tree from Bernier et al., 's vascular atlas. Gray dashed line in (a) and (b) represent 2.5% threshold denoting significance. Green regions in (c) and (d) indicate anterior perfusion territories (ACA and MCA) and light blue in (c) and (d) represent PCA perfusion territory