Evaluating sex as a biological variable in in vitro blood-brain barrier models: insights from primary mouse brain endothelial cells.

Background
Sex bias and omission hereof is prominent in preclinical and clinical biomedical research favoring the male sex [1, 2], with neuroscience being one of the research fields with the most prominent male bias [3]. The ultimate consequence has been the withdrawal of drugs due to subsequent female specific adverse effects [4, 5]. Furthermore, it is well documented that the prevalence and manifestation of neurological diseases, such as stroke and Alzheimer’s disease, differ between men and women [6, 7]. It is therefore essential to include both sexes, to ensure the development of effective and safe brain specific drugs for both males and females with neurological diseases [3, 8]. Development of efficient brain targeting drugs is challenged by the blood-brain barrier (BBB), a function of the smallest capillaries in the brain [9]. Brain endothelial cells (BEC) are a core component of the BBB. The BECs are tightly interconnected by tight junction proteins limiting paracellular diffusion. BECs also express nutrient -and efflux transporters as well as receptors and metabolizing enzymes. Together these properties ensure tight regulation and maintenance of brain homeostasis in addition to protecting the brain from bloodborne toxins and pathogens [9]. In order to understand the complex interplay of the various transporters and receptors contributing to the BBB functionality, and in the early development of brain targeting drugs, in vitro BBB models are essential [10].
In vitro models are often composed of BECs of different origins, e.g. human pluripotent stem cells, immortalized human or mouse BECs or primary BECs of rodent, porcine or bovine origin [10]. The in vitro models resemble the phenotype of the BBB to different extents independently of their origin [10]. However, the majority of the in vitro models do not include sex as a variable although some studies have indicated that there might be sex-dependent differences in the human and rodent BBB, both in the healthy and in the diseased brain [11–13]. To date, a few in vitro studies have suggested sex dependent differences in the barrier integrity of pluripotent stem cells differentiated into brain endothelial-like cells [14–17]. However, sex differences in BBB models based on primary cultures of BECs from sexually mature mice have not been investigated.
With this study, we therefore sought to characterize the phenotype of cultured primary mouse brain endothelial cells (PMBEC) and brain capillaries from which the PMBECs were cultured, both isolated from sexually mature female and male mice (8 weeks of age). We cultured PMBECs in the transwell setup and investigated the barrier integrity as well as the expression, localization and function of essential tight junction proteins, efflux transporters, carriers and receptors to investigate whether the PMBEC in vitro model and brain capillaries displayed sex dependent differences in selected BBB traits.

Methods

Materials
All chemicals and materials were purchased from Merck Life Science (Darmstadt, Germany) unless otherwise stated.

Animals
8-week-old female and male C57Bl/6JOlaHsd mice were purchased from Envigo (Envigo, Vernay, Nederland). The mice were housed at the In Vivo Pharmacology Research Unit, University of Copenhagen. They were housed in a 12 h light/dark cycle and had access to water and chow ad libitum. The animal studies were conduct in compliance with the Danish National Council for Animal Welfare under the license number: 2021-15-0201-01030.

Isolation of brain capillaries
The mice were weighed (Additional file 1A) prior to being euthanized by cervical dislocation. The method has previously been described in Hinca et al. [18]. Briefly, the cortices were isolated and weighed followed by homogenization in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, MA, USA) with 1% penicillin/streptomycin (P/S) (Gibco, MA, USA) with a 40 mL Dounce Tissue Grinder set. The homogenate was mixed 1:1 with 32% w/v dextran (Mw 60,000) in DMEM added 1% P/S and stored on ice for two minutes before centrifugation for 20 min at 2500 x g and 4 °C. The supernatant was removed, and the pellet was resuspended in DMEM added 10% fetal bovine serum (FBS), 1% P/S and 1% non-essential amino acids (Gibco, MA, USA) (DMEM + 10% FBS) and centrifuged for 10 min at 800 x g and 4 °C. The capillary pellet was resuspended in DMEM + 10% FBS and filtered through a 41 μm pore sized filter. The capillaries were flushed off the filter with DMEM + 10% FBS and centrifuged for five minutes at 500 x g and room temperature. The capillary pellet was either used for protein isolation (see the protein isolation and western blot section) or for endothelial cell culture (see below).

Endothelial cell culture
The capillary pellet was resuspended and incubated in digestion mixture (1050 U/mL trypsin TRL, 2000 U/mL collagenase type III, 3400 U/mL DNAse I in DMEM + 10% FBS) (Worthington Biochemical Corporation, NJ, USA) for one hour at 37 °C. An equal volume of DMEM + 10% FBS was added to the digested capillaries, and the mixture was centrifuged for five minutes at 1000 x g and room temperature. The pellet was resuspended in DMEM + 10% FBS and added to a collagen IV and fibronectin (0.01 mg/mL) coated T25 flask (Corning incorporated, NY, USA) and placed in an incubator (37 °C, 5% CO2) for three hours. Then the media was changed to puromycin (4 µg/mL) containing DMEM + 10% FBS. Two days later the media was changed to DMEM + 10% FBS. At day four, the endothelial cells were detached from the T25 flask using trypsin-EDTA before resuspension in DMEM + 10% FBS and centrifugation for 10 min at 1000 x g and room temperature. The cell pellet was resuspended in DMEM + 10% FBS and the cell number was determined using a MultiCount disposable counting slide. Endothelial cells (1.4*105 cells/cm2) were seeded on the apical side of collagen IV and fibronectin (0.01 mg/mL) coated T12- or T24-transwell inserts (Greiner Bio-One, Austria, pore size 0.4 μm)) for bidirectional transport studies or isolation of protein, or in 6 well plates for isolation of RNA to yield sufficient RNA. Three days after, the media were changed to differentiation media (DMEM high glucose added 50mM 2-[(2-Hydroxy-1,1-bis(hydroxymethyl)ethyl)amino]ethanesulfonic acid (TES), 4.9 µg/mL RO-20-1724, 153.8 µg/mL cAMP and 0.2 µg/mL dexamethasone). Two days after the PMBECs were ready for experiments and transendothelial electrical resistance (TEER) was measured with an EVOM Manual TEER measurement system (World Precision Instrument, Fl, USA) prior to experimental start.

Transendothelial transport assays
PMBECs seeded on T24-transwell inserts were used for bidirectional transport experiments. 1 µCi/mL of [14C]-mannitol (PerkinElmer, Ma, USA) and [3H]-digoxin (Revvity, MA, USA) in differentiation media were added to the apical or basolateral compartment, and the PMBEC monolayers were incubated at 37 °C with horizontal shaking at 90 rpm. Samples were taken from the receiver compartment 15, 30, 45, 60, 90 and 120 min after spiking with [14C]-mannitol or [3H]-digoxin. The withdrawn volume was immediately replaced with fresh differentiation media. The samples were transferred to scintillation tubes (Revvity, MA, USA) and Ultima Gold™ scintillation solution (PerkinElmer, MA, USA) was added followed by analysis on a Tri-Carb 2910 TR liquid scintillation analyzer (PerkinElmer, MA, USA). The apparent permeabilities (Papp) were calculated using Eq. 1 where J is the flux (nmol/cm2 · min), C0 is the concentration at time 0, Q is accumulated amount of compound in receiver at time t and A is the area of the Transwell filter. To estimate the flux, the accumulated amount of compound in the receiver compartment at sampling time was plotted against time which gives the flux curve. The flux was calculated as the slope of the straight part of the flux curve.
The calculated permeabilities from basolateral (B) to apical (A) compartment and vice versa were then used to calculate the efflux ratios according to Eq. 2.

Gene expression studies
RNA was isolated from PMBECs cultured in 6 well plates according to manufacturer’s protocol (NucleoSpin RNA Plus mini kit, Macherey-Nagel, Germany) and the RNA concentrations were determined using a Nanodrop 1000 Spectrophotometer (Thermo Fisher Scientific, MA, USA). Purified RNA (0.2 µg) was reversed transcribed to cDNA using a High-Capacity cDNA Reverse Transcription Kit according to manufacturer’s protocol (Applied Biosynthesis, Thermo Fisher Scientific, MA, USA) and an MJ Research PC-200 Gradient Thermal cycler (Marshall Scientific, NH, USA). This was followed by running a quantitative polymerase chain reaction (qPCR) on a LightCycler 96 (Roche Diagnostics, Switzerland) using a SYBR-green based detection method (Roche Diagnostics, Switerland). The cycling conditions were one cycle of preincubation at 95 °C for 10 min followed by 45 cycles at 95 °C for 10 s, 55 °C for 10 s and 72 °C for 20 s and ending with one cycle at 95 °C for 10 s, 65 °C for 60 s and 97 °C for one second. The primers used and their determined efficiencies are listed in Table 1. The 2−ΔCt method was used to calculate the relative gene expression normalized to the geometric mean of the housekeeping genes Ywhaz and Hprt1 [19].

Protein expression studies
Freshly isolated brain capillaries or PMBECs grown on T12-transwell inserts were lysed with RIPA buffer (10 mM Tris-Cl (pH 8.0), 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl) supplemented with cOmplete Protease Inhibitor Cocktail (Roche Diagnostics, Switzerland). Protein content was determined using the bicinchoninic acid assay. 20 or 18–24 µg of protein from brain capillaries or PMBECs, respectively, were loaded onto a 4–15% Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad, CA, USA) alongside a Precision Plus Protein WesternC Blotting Standards (Bio-Rad, CA, USA) and run at 200 V for ~ 30 min in a Tris-glycine-SDS buffer (Thermo fisher Scientific, MA, USA). Protein transfer was performed using a Trans-Blot Turbo Mini PVDF Transfer Pack (Bio-Rad, CA, USA) and the Trans-Blot turbo transfer system (Bio-Rad, CA, USA) at 25 V for three minutes. Blots were blocked in 5% skim milk in Tris-buffered saline with 0.1% Tween-20 (TBST), prior to overnight incubation with primary antibody at 4 °C (Table 2). Blots were washed in TBST, before one hour of incubation with horseradish peroxidase-conjugated secondary antibody (Table 2) at room temperature. The blots were developed using an Amersham ECL Prime Western Blotting Detection Reagent (Cytiva, MA, USA) and imaged with the FluorChem Q system (ProteinSimple, Bio-Teche, MN, USA). Band intensities were quantified using Fiji (ImageJ 1.54p, Java 1.8.0_322) and normalized to loading control (β-actin or GAPDH).

Immunocytochemistry
PMBEC monolayers grown on T24-transwell filters were fixed in 4% paraformaldehyde (Thermo Fisher Scientific, MA, USA) in phosphate buffered saline (PBS) for 10 min followed by permeabilization with 1% Triton X-100 in PBS for five minutes. Blocking buffer (2% bovine serum albumin in PBS) was added to the filters for 30 min. Subsequently, the filters were incubated with primary antibodies diluted in blocking buffer for two hours at room temperature (Table 3). After three washing steps in blocking buffer, Alexa Fluor 488-conjugated goat anti-rabbit or goat anti-mouse secondary antibody was diluted in blocking buffer with Hoechst or DAPI (both 1:1000 dilution) (Thermo Fisher Scientific, MA, USA) and added to the filters for 30 min. The filters were washed in blocking buffer before being mounted on microscope slides using Immu-Mount (Epredia, NH, USA). A Leica DMi8 fluorescence microscope with a 63x oil objective (Leica Microsystems, Wetzlar, Germany) was used to capture the stainings and the images were subsequently processed in Fiji (ImageJ 1.54p, Java 1.8.0_322).

Data analysis
Data was visualized and statistical analyses were conducted in GraphPad Prism (10.5.0 for Windows) (GraphPad Software Inc., CA, USA). Data is reported as mean with standard deviation unless otherwise stated. The bi-directional transport studies and RNA expression experiments were conducted across three biological replicates with three technical replicates. N denotes biological replicates whereas n denotes technical replicates. Western blots contain three or four biological replicates. Levels of significance were tested with Students t-test. A p-value under 0.05 was considered as statistically significant.

Results and discussion

Establishment of primary in vitro BBB models from sexually mature female and male mice
In vitro BBB models based on PMBECs cultured from brain capillaries isolated from sexually mature female and male mice were established to evaluate if potential sex-dependent differences were maintained in vitro (Fig. 1). Brain capillaries were isolated from mice cortices that were weighed and normalized to body weight. The apparent weight of the isolated cortices was comparable between female and male mice (Additional file 1B), however, when normalized to body weight, the nominal weight value of cortices of female origin was significantly larger than the cortices of male origin (Additional file 1C). Capillaries were isolated and seeded in culture flasks where PMBECs sprouted. The PMBECs were removed when they were near confluency and transferred to permeable supports, as described in the Methods section (Fig. 1A). A comparable number of PMBECs was obtained from female and male cortices (Fig. 1B). Visualization with F-actin confirmed that PMBECs from both female and male mice formed monolayers of spindle-shaped cells and staining for von Willebrand factor (vWF) confirmed an endothelial phenotype (Fig. 1C and Additional file 2A and B). The integrity of the monolayers was evaluated by measuring the TEER and assessment of permeability of the small paracellular flux marker mannitol. The TEER values were comparable between PMBEC monolayers derived from female and male mice (Fig. 1D). PMBEC monolayers of both origins formed tight barriers resulting in low Papp of mannitol (female: 4.32 × 10− 7 cm/sec ± 1.88 × 10− 7 and male: 2.82 × 10− 7 cm/sec ± 0.69 × 10− 7) while no sex-associated difference in mannitol permeability was observed (Fig. 1E). The observed TEER values were consistent with those reported in other studies [20–22], even though we observed quite some variation among TEER values from different batches, a phenomenon which is also well described in the literature [10, 22]. However, the permeability of mannitol was slightly lower than reported in other studies using primary rodent brain endothelial cells [18, 22, 23].

The expression and localization of essential tight junction proteins were comparable between female- and male-derived PMBECs and capillaries
To investigate whether the high barrier integrity of the PMBECs reflected comparable expression and localization of junctional complexes between sexes, we measured gene and protein expression of the tight junction proteins zonula occludens-1 (ZO-1, Tjp1), occludin (Ocln) and claudin-5 (Cldn5), in both freshly isolated capillaries and cultured PMBEC monolayers, using qPCR and western blotting. Gene and protein expression levels did not differ significantly between sexes in neither PMBECs nor brain capillaries (Fig. 2A-C and Additional file 3A and B). Immunostaining demonstrated comparable and correct localization of the tight junction proteins in female- and male-derived PMBECs (Fig. 2D and Additional file 4A-C). ZO-1 and claudin-5 were clearly localized to the cell junctions while only a weak staining of occludin could be detected here. Although culture-induced changes have been reported for primary bovine brain endothelial cells [24, 25], our data from the PMBEC monolayers and brain capillaries indicate that such effects did not mask sex dependent differences in tight junction mRNA or protein expression. Collectively, these findings support that female and male-derived PMBECs have a comparable barrier integrity as observed in Fig. 1D and E. In studies using pluripotent stem cell-derived brain endothelial-like cells, some researchers have also found comparable TEER values between monolayers of female and male origin [15, 16], while another reported higher TEER values in cell monolayers of female origin when co-cultured with astrocytes [14]. The picture is equally inconsistent in in vivo studies of BBB integrity in disease models such as mild traumatic brain injury [26–28] and stroke [29, 30]. In cases where a sex-associated difference was observed in vivo, it was typically accompanied by changes in tight junction protein expression [26, 27]. In a recent study by Mi et al. (2025), they observed a significant sex dependent difference in gene expression of claudin-5 and occludin in PMBECs and cortex samples from 3 month old mice (males > females) [31]. This is contradictory to our results (Fig. 2A), where we isolated PMBECs from 2-month-old mice, which potentially could explain the discrepancy. Taken together, these discrepancies, along with the limited and sometimes conflicting in vivo evidence, suggest that sex-dependent differences in BBB integrity may not be robust or consistently present, and could depend strongly on the specific model and experimental context.

Key transporters were expressed to comparable extents in PMBECs and capillaries of female and male origin
BECs express transporters and receptors to ensure transport of nutrients, hormones and vitamins [32] to the brain. To assess whether there were sex-dependent differences in the BBB at the level of key transporters, we analyzed the expression of glucose transporter 1 (GLUT1, Slc2a1), insulin receptor (InsR, Insr) and transferrin receptor (TfR, Tfrc) in PMBECs and brain capillaries. The mRNA expression levels of GLUT1, InsR and TfR were similar when comparing PMBECs of female and male origin (Fig. 3A). Immunocytochemistry also revealed similar localization of GLUT1, InsR and TfR (Fig. 3B and Additional file 5A-C). GLUT1 appeared to be localized to the cell membrane whereas the stainings for InsR and TfR were very faint making it difficult to firmly establish their cellular localization. Interestingly, in bovine primary brain endothelial cells, the InsR has previously been reported to be located intracellularly while the localization of the TfR was comparable to what was observed in Fig. 3B [33, 34].
The protein expression of GLUT1 and TfR was examined in capillaries and it did not differ between sexes (Fig. 3C and Additional file 3C), consistent with previous findings on mouse capillaries and PMBECs [31, 35]. By contrast, in studies on human induced pluripotent stem cell-derived monolayers GLUT1 expression was found to be higher in female-derived brain endothelial-like cells than in male-derived brain endothelial-like cells [14]. It should be noted that isolation of brain capillaries does not result in a pure isolation of the capillary endothelial cells [36], as astrocytic endfeet and pericytes remain attached to the endothelial capillaries and therefore also might contribute to the detected GLUT1 and TfR expression [36–41]. Nevertheless, our data indicate that GLUT1, InsR and TfR expression is comparable between female- and male-derived PMBECs and capillaries under baseline conditions.

PMBECs of female and male origin displayed similar expression level and functionality of the efflux transporter P-glycoprotein but higher expression of BCRP in capillaries of female origin
Another key property of brain endothelial cells is the expression of functional efflux transporters that actively extrude compounds from BECs back into the blood [9, 10]. The gene expression and localization of breast cancer resistance protein (BCRP, Abcg2) and P-glycoprotein (P-gp, Abcb1a) were investigated in PMBECs from female and male mice (Fig. 4A and B and Additional file 6A and B) revealing similar expression levels and localization. Furthermore, the function of P-gp was examined with a bi-directional transport study using [3H]-digoxin, a well-known P-gp substrate [42]. The permeability of digoxin was larger in the basolateral to apical (B-A) direction compared to the apical to basolateral (A-B) direction independent of sex (Fig. 4C). The calculated efflux ratios were above 2 for both female-derived (9 ± 4) and male-derived (15 ± 10) PMBECs, which confirms active efflux of digoxin and shows that isolated PMBECs express functional P-gp. Functionality of BCRP in PMBECs was not investigated since a BCRP specific substrate is lacking. Estrone-3-sulphate has previously been used as a BCRP substrate, but it is also a substrate for multidrug resistance-associated protein 1 [43, 44]. Moreover, there was no difference in gene expression of BCRP between female- and male-derived PMBECs, thus it is unlikely that there would be a sex dependent difference in functionality. But protein expression of BCRP was investigated in capillaries. It was not possible to detect P-gp due to lack of a suitable antibody. Capillaries from female mice displayed higher expression of BCRP compared to capillaries from male mice (Fig. 4D, Additional file 3D and Additional file 7), but whether this sex dependent difference is also present in PMBECs is unknown, since it was not possible to detect BCRP in protein lysates from PMBECs. However, in contrast to GLUT1 and TfR, BCRP is not expressed in pericytes and astrocytes [45, 46], thus excluding the possibility from obtaining a BCRP signal from contaminating cells of the capillary preparations. Contradictory to our results, Banks et al. observed a higher BCRP expression in male mouse brain capillaries compared to females [47], while a study using rat capillaries [48] and a study using mouse cortex samples [31] reported no difference in BCRP expression when comparing male and female preparations. In the study using rat capillaries, a similar P-gp expression was observed between the two sexes [48]. In humans, a sex-dependent difference in P-gp function (males > females) in younger individuals has been observed, but this difference is lost with age [49, 50], potentially due to age-related loss of P-gp functionality [51], P-gp expression [31, 52] and/or barrier integrity [53]. We did not observe a difference in P-gp functionality between female- and male-derived PMBECs obtained from younger sexually mature animals (Fig. 4C). An important difference between our study and the above referenced human studies is the choice of P-gp substrate; digoxin or [11C]verapamil [50] and [18F]MC225 [49], respectively, since different substrates interact with P-gp in distinct ways which challenges direct comparison [54, 55]. Moreover, it is also known that there is species-specific P-gp activity [56, 57] which also could contribute to the discrepancy between our study (Fig. 4C) and the human findings [49, 50].

Conclusion
The present study represents, to our knowledge, the first systematic investigation of potential sex differences in both primary mouse brain endothelial cells and capillaries isolated from sexually mature female and male mice. Overall, the phenotype of female and male PMBECs and capillaries were highly comparable, with similar barrier integrity, expression and localization of tight junction proteins as well as nutrient transporters. In addition, we observed similar P-gp expression and functionality. One observation could indicate potential sex-associated differences: female-derived capillaries showed slightly higher BCRP protein expression. This difference was, however, not observed at mRNA level or in immunostainings upon analysis of BCRP in PMBECs. Overall, our findings suggest that under baseline conditions, sex does not markedly influence the phenotype of PMBECs. Future studies should include incubation of PMBECs with sex hormones, which could provide a more physiologically relevant model for studying sex dependent BBB properties. Additionally, sex difference may also be investigated in PMBECs isolated from disease models as some variations may only appear under certain pathologies. These approaches would allow a complete assessment of the relevance of this in vitro model in studying BBB sex differences.

Supplementary Information
Below is the link to the electronic supplementary material.