The choroid plexus (CP) modulates brain homeostasis through several processes including cerebral spinal fluid (CSF) secretion and expression of various proteins (Lun et al., 2015; Hofman and Chen, 2016). Specifically, the CP secretes numerous neurotrophic factors implicated in neurogenesis, vascularity regulation, maintenance of brain function, and signaling molecules including neuroinflammatory cytokines (Arnaud and Di Nardo, 2016; Mazucanti et al., 2019). The ubiquitous multifunctional role of the CP was highlighted in a recent study by Mazucanti et al. (2019), indicating that insulin is produced in the CP and modulated by serotonin. Structurally, the CP is a monolayer of epithelial cells connected by tight junctions situated within the ventricles that regulate the passage of solutes, proteins, and inflammatory cells to and from the CSF (Lun et al., 2015). Fenestrated capillaries are separated from the epithelial cells by stroma, and allow filtration of fluids, thereby forming the blood-CSF-barrier (BCSFB) (Liddelow, 2015).

According primarily to postmortem and animal studies, during normal aging, the CP undergoes several alterations, including morphological changes to the epithelial cells (Shuangshoti and Netsky, 1970; Serot et al., 1997, 2001; Marques et al., 2013), accumulation of calcifications (Yalcin et al., 2016), increased T helper type 2 (Th2) response (Baruch et al., 2013), dysfunction in CSF production and clearance (May et al., 1990; Chiu et al., 2012), and iron deposition (Joseph-Mathurin et al., 2013). Moreover, the CP is highly vascularized and contains energy-demanding epithelial cells, rendering it vulnerable to hypoxia, ischemia and hypoperfusion, potentially leading to cell death, increased permeability and dysfunction of the BCSFB (Chen et al., 2009), and reduced secretion of trophic factors and CSF generation. These changes have been associated with neurodegenerative diseases including Alzheimer’s disease (AD) (Kadel et al., 1990; Cornford et al., 1997; Ennis and Keep, 2006; Krzyzanowska and Carro, 2012; Marques et al., 2013; Kaur et al., 2016; Hubert et al., 2019). The CP is a small region and difficult to distinguish from the surrounding tissue, requiring high resolution imaging and careful postprocessing manipulation including CP volume segmentation; therefore, despite its mediation of several critical processes, the effect of normal aging on CP microstructure and perfusion has not been well studied in-vivo in humans.

Previous magnetic resonance imaging (MRI)-based studies have demonstrated that subjects with psychiatric disorders (Lizano et al., 2019), complex regional pain syndrome (Zhou et al., 2015), multiple sclerosis (Ricigliano et al., 2021) and stroke (Egorova et al., 2019), have larger CP volumes compared to healthy controls, whereas studies evaluating CP size in participants without any sign of pathology have only looked at thickness. Indeed, although Madhukar et al. (2012) and İmamoğlu et al. (2013) showed that CP thickness was consistent in pediatric and adult cohorts, changes in CP volume in normal aging has not been characterized. Moreover, to the best of our knowledge, MRI-based microstructural and perfusion differences in the CP with age or sex have only been documented by two studies. In this pioneering work, Alicioglu et al. (2017) assessed the mean diffusivity (MD), measured using diffusion tensor imaging (DTI) MRI, in human CP across various age ranges and found that the eldest participants exhibit significantly higher MD than the younger participants; this result was interpreted as an increase in diffusion in CP with age indicating a loss of structural integrity resulting in increased leakiness of the epithelial cell junctions. In contrast, Bouzerar et al. (2013) reported a decrease in capillary permeability with age. Additionally, the authors observed no significant difference in CP CBF with age using dynamic susceptibility contrast enhanced MRI (Bouzerar et al., 2013). The discrepancy in results may be due to cerebral lesions present in patient data or the limited age range or size of the cohorts studied.

Several quantitative MRI techniques have been used to evaluate cerebral tissue microstructure in-vivo with aging and neuropathology (Grieve et al., 2007; Hasan et al., 2010; Alisch et al., 2021). Longitudinal and transverse relaxation times (T₁ and T₂) depend on water mobility as well as macromolecular tissue composition including iron content; therefore, changes in T₁ or T₂ are directly associated with cerebral microstructural tissue changes (Deoni, 2010). Further, fractional anisotropy (FA) and MD, derived from DTI, can be used to probe cerebral tissue microstructural integrity based on water content and mobility (Alexander et al., 2007; Grech-Sollars et al., 2015). Finally, cerebral blood flow (CBF) in the CP can be reliably determined using arterial-spin labeling (ASL) MR imaging (Evans et al., 2020; Zhao et al., 2020). In the current study, we investigated volumetric differences with age as well as microstructural and perfusion characteristics of the CP in a large cohort of well-characterized cognitively unimpaired adults (n = 155) across the extended age range of 21–94 years. Our main goals were to investigate the effect of age and sex on relaxation times, DTI indices and CBF, to investigate the effect of age and sex on CP microstructure and function, and to develop further insights into CP maturation and degeneration with aging.

Figure 2 shows the relationship between the CP volume or the LV volume, after correcting for the total intracranial volume, and age for all participants (n = 155). The CP and LV volumes showed significant quadratic effects of age (Table 1), age², with both plots showing a non-linear increase in CP or LV volume with age (Figure 2 and Table 1). The effect of sex on the CP and LV volumes was statistically non-significant (Table 1).

Figure 3 shows the relaxometry measures, T₁ and T₂ (n = 146), the DTI measures, FA and MD (n = 136), and CBF (n = 88) for all participants within the CP as a function of age. Multiple linear regression analysis showed that the effect of age was statistically significant on all of these MRI measures (Table 1). Furthermore, besides T₂, the quadratic effect of age on all MR measures was statistically significant, with FA (p < 0.01) and CBF (p < 0.05) exhibiting non-linearly decreasing trends with age, while T₁ (p < 0.01) and MD (p < 0.01) exhibiting non-linearly increasing trends with age (Table 1 and Figure 3). Finally, sex effect on MD and CBF was statistically significant, with women exhibiting higher CBF values and lower MD values as compared to men.

Correlations between relaxometry or DTI measures and CBF were assessed using multiple linear regression analyses while accounting for the effect of age. As shown in Figure 4 and Table 2, CBF was significantly correlated with FA exhibiting a linearly increasing trend with increasing CBF, while T₁, T₂, and MD exhibiting linearly decreasing trends with increasing CBF (Table 2 and Figure 4).

In this cross-sectional study, we assessed sex and age-related microstructural and functional differences that occur in the CP using a combination of robust MRI measures in a large cohort of cognitively unimpaired subjects (n = 155) spanning 21–94 years of age. We found significant positive association between CP volume and age. We attribute this novel observation to the functional and morphological alterations that occur within the CP as well as to ventriculomegaly and, perhaps, inflamm“aging,” which are common features of normal brain aging. Structurally, the CP is known to undergo numerous changes including accumulation of calcifications and lipofuscin deposits as well as thickening of the stroma (Shuangshoti and Netsky, 1970; Serot et al., 2001; Yalcin et al., 2016). Functionally, CSF production and turnover decrease with age in the CP (Preston, 2001; Stoquart-ElSankari et al., 2007). Thus, an increase in CP volume may serve as a compensatory or protective mechanism to morphological disruptions and the diminished ability to produce CSF (Redzic et al., 2005). The CP also plays a critical role in producing growth factors (e.g., insulin, IGF1, and IGF2) necessary for normal brain development and regulating its own growth and maintenance of epithelial cell health. While it has been observed that epithelial cell growth of the CP slows with age in younger animals (McDonald and Green, 1988; Liddelow et al., 2010), previous reports have demonstrated that enlarged ventricles and increasing inflammation due to stroke or direct tissue injury upregulated CP-derived growth factors resulting in volume proliferation of CP epithelial cells (Li et al., 2002; Barkho and Monuki, 2015). Further, previous studies using similar segmentation methods have found a positive association between increasing LV size and CP volume when comparing healthy controls to a disease state (Egorova et al., 2019; Lizano et al., 2019; Tadayon et al., 2020). Due to the multifunctional role of the CP, it is difficult to discern the dominant source and purpose of volume enlargement during normative aging, which necessitates further investigations.

Our finding of higher MD values with age is consistent with Alicioglu et al. (2017). This age-related increase in MD is further supported by our findings that FA significantly decreased with age and that T₁ and T₂ significantly increased with age. MD describes the overall diffusion and motion of water molecules in the brain, with higher MD values reflect more water mobility (Alexander et al., 2007; Salat, 2014). Similarly, T₁ and T₂ both depend on water content with higher values are also associated with more water mobility (Deoni, 2010). FA, on the other hand, is used to describe anisotropy with low values associated with less restricted movement of molecules (Alexander et al., 2007; Salat, 2014), with lower values of FA with age here are likely due to microstructural alterations or damage to the CP epithelium, leading to increased isotropy of water diffusion. Taken together, we provisionally attribute these observed changes in relaxometry and DTI measures primarily to structural alterations to the CP microstructure and, potentially, to a decreased structural integrity of the BCSFB with advanced age. Indeed, it has been observed in animals that the BCSFB becomes leakier with increasing age, specifically for compounds that were low and medium in molecular weight (Chen et al., 2009). Possible sources for structural alterations in the BCSFB stem from morphological disruptions, as previously described, as well as increasing inflammation. Indeed, previous work showed that aging was associated with a shift toward the Th2 response, which consequently resulted in compromised epithelial tight junctions that are vital in maintaining the structural integrity of the BCSFB (Baruch et al., 2013). However, additional investigations, especially dynamic contrast-enhanced-based measures as well as histological studies, are required to support our conjecture of a potential association between decreased structural integrity of CP and higher BCSFB permeability.

Interestingly, our MD measure exhibited significant differences between men and women, with men exhibiting higher MD values. Alicioglu et al. (2017) observed that women experience a steeper increase in MD after the age of 61 as compared to men. Further, several studies have shown that women exhibit higher CBF as compared to men (Devous et al., 1986; Rodriguez et al., 1988; Podreka et al., 1989; Gur and Gur, 1990; Esposito et al., 1996; Parkes et al., 2004; Liu et al., 2012, 2016; Alisch et al., 2021). Sex differences in CP microstructure and function would be expected. Indeed, in addition to reduced CSF turnover, it is suggested that a dysfunctional BCSFB contributes to greater protein content in the CSF during aging (Chen et al., 2009). In fact, previous reports found that men had a significantly higher CSF/serum albumin ratio (QAlb), a biomarker of BCSFB permeability, compared to women which demonstrates a greater loss of integrity in men (Parrado-Fernandez et al., 2018; Castellazzi et al., 2020). The current scientific literature on sex differences within the CP remain limited, but we speculate that sex hormones may play a role in BCSFB permeability (Santos et al., 2017). Further investigations are required to elucidate sex-related differences in CP’s microstructure and function.

We also observed an age-related decrease in CBF in the CP, likely reflecting changes in CP morphology or in the hormonal and neurotransmitter levels. CP blood flow is regulated by several molecules including arginine vasopressin, angiotensin II, dopamine, and serotonin. Arginine vasopressin and angiotensin II have an inhibitory effect on blood in the CP (Faraci et al., 1988; Maktabi et al., 1990), while dopamine and serotonin are both known to increase blood flow (Townsend et al., 1984; Faraci et al., 1989). Previous studies have showed that aging is associated with increases in arginine vasopressin (Frolkis et al., 2000) and that angiotensin II induces the release of vasopressin (Keil et al., 1975; Qadri et al., 1993), thus indicating a potential mechanism for the age-related decline in CP blood flow. Further, dysfunctional binding and transport of dopamine and serotonin, as observed in other parts of the brain (Meltzer et al., 1998; Karrer et al., 2017), may contribute to decreased CBF. Moreover, morphological examinations show that the arterial walls are thickened and that there is a reduction in blood vessel density during normal aging which leads to limited contact between the blood and the CP epithelial cell layer (Van Cauwenberghe et al., 2020). Finally, our results indicate that CBF was associated with both relaxometry and DTI measures, suggesting that low CBF is associated with the reduced structural integrity of the CP. Indeed, adequate CBF is paramount for nutrient and oxygen delivery and clearance of metabolic by-products; dysfunction of this ability is a characteristic feature of neurodegenerative diseases.

Although we examined a relatively large cohort and used advanced MR methodology, our study has limitations. Our dataset is cross-sectional so that the observed trends in the CP microstructure and function with age as well as the associations between CBF and the other MRI measures requires further validation through longitudinal studies. Such work, motivated by the present results, is underway. Furthermore, our analysis of functional and structural differences in CP with aging were limited to the LVs. We note that CSF partial volume effects may bias derived parameter values. More accurate automated segmentation methods, including the third and fourth ventricles, as well as higher resolution structural images are needed for a better evaluation. In addition, our DTI-related results must be interpreted with caution. Indeed, given the large fraction of free water in the CP, this could have introduced bias in derived parameter values. Although the free-water elimination DTI (FWE-DTI) approach has been used widely to distinguish free-water partial-volume effects from tissue’s diffusion in healthy aging and degenerative diseases (Pasternak et al., 2009), it has recently been shown that this method is unstable when applied to single-shell DTI data, requiring careful implementation (Golub et al., 2021). Further, we note that our derived DTI parameters values could change depending on the choice of the b-values. Several DTI protocols involve use of b-values of 0 and 1000 s/mm² while others incorporate lower b-values. Our choice of using lower b-values was driven by the desire to minimize non-Gaussian diffusion. Moreover, we used identical ASL experimental parameters for all subjects and acquired CBF images at a single post-labeling delay, implicitly assuming negligible effects of spatial variation in arterial transit time (ATT), the time of the arterial bolus to transit from the labeling plane to the imaging volume (Duvernoy et al., 1981; Qin et al., 2014). Although this is a reasonable assumption (Alsop et al., 2015), ATT may vary spatially and may differ between subjects due to arterial blood velocity differences; this could introduce a small bias in derived CBF values. A multiple post-labeling delay ASL technique may provide more accurate CBF determination (Qin et al., 2014). Finally, other factors such as inflammatory markers, blood pressure, and medications were not considered in this work.

We examined age-related differences in CP microstructure and function in a cohort of cognitively unimpaired participants across a wide age range using quantitative MRI. We showed that aging was associated with increases in volume and decreases in blood flow or structural integrity. This work lay the foundation for further investigation of the functional and structural changes of the CP with normal aging and neuropathology.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

This study including our MRI protocol was approved by the MedStar Research Institute and the National Institutes of Health Intramural Ethics Committees, and all examinations were performed in compliance with the standards established by the National Institutes of Health Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.

JE and MB: research design, results interpretation, and manuscript writing and editing. JA, JE, MK, CT, MA, ZG, and NK: analysis, results interpretation, and manuscript writing and editing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.