Secoisolariciresinol diglucoside is a blood-brain barrier protective and anti-inflammatory agent: implications for neuroinflammation

Isolation of primary human brain microvascular endothelial cells (BMVEC) was accomplished as previously described from normal brain resection tissue following surgery for treatment of intractable epilepsy [20] and provided by Michael J. Bernas and Dr. Marlys H. Witte (University of Arizona, Tucson, AZ). BMVEC were characterized and maintained as previously described from our laboratory [21]. Primary human monocytes were isolated and purified from the University of Nebraska Medical Center (Department of Pharmacology and Experimental Neuroscience, Omaha, NE) by counter current centrifugal elutriation as described [22] from HIV-1 and hepatitis B seronegative donors. BMVEC or monocytes were treated with different concentrations of SDG (0, 1, 2, 5, 10, or 50 μM) in accordance with previous studies [7, 17, 23].

Monocytes were used within 24 h of isolation in adhesion assays following 16-h treatment with TNFα (20 ng/ml) in the absence or presence of SDG. All treatments were removed prior to labeling of the monocytes with calcein-AM (Life Technologies) as described [24]. Calcein-labeled monocytes were added to BMVEC monolayers stimulated by 16-h treatment with TNFα (20 ng/ml). Adhesion of monocytes to BMVEC was measured by fluorescence using a Synergy 2 plate reader (Biotek Instruments, Winooski, VT) from triplicate determinations and is expressed as fold difference in monocyte attachment (mean ± SEM) compared to baseline (untreated control).

An in vitro model of BBB transendothelial migration was used to measure monocyte migration through BMVEC monolayers as previously described [24]. Monocytes and BMVEC were treated with TNFα and/or SDG as described above. Human recombinant MCP-1 (CCL2, 30 ng/ml) was added to the lower chamber of the BBB construct to promote migration of monocytes through the BMVEC monolayer. Migration was quantitated with ImageJ software (NIH, Bethesda, MD) from triplicate determinations and is expressed as fold difference in monocyte migration (mean ± SEM) compared to baseline (untreated control in the absence of MCP-1).

Surface expression of vascular cell adhesion molecule 1 (VCAM-1) was measured by flow cytometry (FACS) as described [21]. BMVEC were pretreated in the absence or presence of SDG for 1 h followed by treatment with TNFα or IL1β (100 ng/ml) for 4 h. FITC-conjugated antibody to VCAM-1 (CD106) was from BD Biosciences (San Jose, CA). Actin cytoskeleton rearrangements were measured in monocytes treated in the absence or presence of the very late antigen 4 (VLA-4)-specific ligand, LDV peptide (12 nM), and SDG (10 or 50 μM, 4 h) as described [25]. HUTS-21 antibody, specific for the activated conformation of VLA-4, was detected by HUTS21 (R&D Systems), was measured by FACS as described [25, 26]. LFA-1 conformation, detected by MEM-148 Ab, was measured as described [27]. Quantitation of integrin conformational activation was performed where the mean fluorescence intensity (MFI) of activated non-treated cells was assigned a value of 100 and a value of 0 was assigned to the MFI of cell autofluorescence [25, 28, 29]. All other calculations were done utilizing the regression curve calculation tool of Prism v5 software (GraphPad Software Inc., La Jolla, CA). Fibrillar (F) and globular (G) forms of actin were quantitated with Acti-Stain Alexa-488 (Cytoskeleton Inc., Denver, CO) and DNase-1-Alexa 594 (Life Sciences), respectively. The F/G actin ratio was calculated by dividing mean fluorescence intensity (MFI) of F-actin by MFI of G-actin and the levels of F/G actin in non-stimulated, non-treated cells were assigned a value of 1. A FACS Canto II flow cytometer and FlowJo software version 8.7 (Tree Star, Ashland, OR) were used to acquire and analyze data as described [25, 29]. Data from at least 10,000 recorded events per treatment condition are presented as MFI (mean ± SEM from at least three experiments).

All in vivo experiments were approved by the Temple University Institutional Animal Care and Use Committee in accordance with guidelines based on the National Institutes of Health (NIH) guide for care and use of laboratory animals and ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines (www.​nc3rs.​org.​uk/​arrive-guidelines). SDG was administered to mice by oral administration using a non-forceful feeding technique, which consisted of hand feeding mice (10 weeks old, male) that had been previously trained to accept up to 30 μl volume of treatment solution. SDG doses were tailored to individual mouse weights and were freshly prepared 10 min before being offered to the mice. Mice were pretreated with SDG (4 mg/mouse) 2 h before i.c. administration of TNFα (0.5 μg/mouse) as an inflammatory insult. Administration of SDG formulations delivering a daily dose of 4 mg SDG/mouse (200 mg/kg) has been found to be protective in murine models of asbestos-induced inflammation [7] as well as radiation-induced inflammation [18, 30].

Intravital video microscopy (IVM) (via cranial window) was used to quantify in vivo leukocyte adhesion in the presence or absence of SDG treatment and inflammatory insult [31]. 5 days after implantation of the cranial window with adjacent cannula [31], mice were injected with rhodamine 6G (0.1%). Serial images were obtained through the cranial window 2 h after inflammatory insult using a Stereo Discovery V20 epifluorescence microscope (Carl Zeiss Microimaging, Thornwood, NY) as described [31]. The number of adherent leukocytes was quantitated as previously described [31‐33].

Animals were injected i.p. injection of 200 μl of 2% sodium-fluorescein (Na-F) in saline. The amount of Na-F evaluated as described [21] was measured using a Synergy 2 plate reader (BioTek). Fluorescent dye content was calculated using external standards, and the data are expressed as amount of tracer per mg of tissue.

It was documented before that SDG administration in vivo diminished end-organ injury in models of systemic inflammation and lung radiation injury. However, to date, there have not been studies assessing SDG effects on BBB in the setting of systemic inflammation or neuroinflammation. To address this question, we used a novel in vivo imaging technique, intravital microscopy through a cranial window, in our model of aseptic encephalitis following i.c. injection of TNFα, a cytokine overexpressed in many neuro-inflammatory conditions including multiple sclerosis, encephalitis, stroke, and HIV encephalitis [34‐36]). Our prior work indicated that this stimulus enhances both leukocyte adhesion to and migration across the BBB [19]. Indeed, i.c. administration of TNFα led to a 20-fold increase in leukocyte adhesion and enhanced migration of leukocytes across the barrier 2 h later (Fig. 1), consistent with previous reports. Oral administration of SDG attenuated adhesion of leukocytes to the endothelium by 50% and attenuated migration of leukocytes across the BBB by 64% (Fig. 1). Leukocyte adhesion resolved 24 h later and SDG feeding therefore did not show a difference compared with the vehicle-treated control group after 24 h. Of note, administration of SDG alone did not affect leukocyte adhesion or migration.

The systemic inflammatory response accompanied by leukocyte adhesion to brain endothelium results in BBB leakage [21]. Thus, we assessed BBB permeability in vivo using the low molecular weight tracer, sodium fluorescein (Na-F), after 4 h of LPS administration (Fig. 2). LPS administration led to increased BBB permeability (20%) that was prevented by SDG, pointing to barrier protective effects of SDG in vivo.

In order to test whether SDG can diminish neuroinflammation in human tissues, we modeled inflammatory responses using monolayers of primary human brain endothelial cells (BMVEC). We tested whether SDG could decrease adhesion of primary human monocytes to TNFα-stimulated brain endothelium. BMVEC monolayers were activated by TNFα in the presence of SDG. Primary human monocytes were placed on the BMVEC only after all treatments were removed and the medium changed. TNFα upregulated monocyte adhesion 2.3-fold and SDG treatment of BMVEC (10 or 50 μM) diminished immune cell adhesion by 87% (Fig. 3a). Alternatively, we treated primary human monocytes with SDG and showed a decrease in adhesion of 54 or 79% after treatment with 10 or 50 μM SDG, respectively. Using migration assays in an in vitro BBB model, we tested whether SDG treatment of endothelial cells could prevent monocyte passage across BMVEC monolayers using CCL2 as a relevant cytokine. Application of CCL2 to the lower chamber of BBB constructs increased monocyte migration 2.6-fold. Pre-treatment of BMVEC with SDG attenuated monocyte migration across endothelial monolayers by 30–46% (Fig. 3b). Next, we studied whether pretreatment of monocytes with SDG would decrease their migration across BBB models. Indeed, application of SDG to monocytes before migration assays resulted in complete reversal of migration to control levels (conditions without CCL2/TNFα) (Fig. 3c). There was no dose-dependent effect of SDG in migration assays.

To investigate the mechanism of diminished adhesion, we tested whether SDG treatment of activated BMVEC diminishes expression of the adhesion molecule VCAM-1. BMVEC stimulation with TNFα resulted in a 4-fold increase of VCAM-1 expression, and SDG reduced it by 35% (Fig. 4a). Similarly, IL1β increased VCAM-1 expression in BMVEC 2.3-fold, and SDG treatment diminished it by 33% (Fig. 4b).

Changes in monocyte cytoskeleton are essential for cell migration. To simulate changes in actin seen in inflammation, we stimulated primary human monocytes with LDV (mimicking interactions with adhesion molecules), which led to a 1.95-fold increase in the ratio of fibrilar/globular actin. These cytoskeletal changes in monocytes were completely blocked by SDG (Fig. 5a, b). Monocyte adhesion to activated endothelium is mediated by integrins, such as very late antigen 4 (VLA-4), whose active conformation is stimulated by inside-out or outside-in activation [28]. Expression of the active form of VLA-4 is associated with enhanced monocyte adhesion and migration. To test whether SDG treatment could change expression of the active form of VLA-4, we treated monocytes with the relevant stimulus, LDV peptide, mimicking VLA-4 interactions with VCAM-1, and fibronectin. LDV caused a 26-fold increase in expression of the activated form of VLA-4 and SDG reduced it by 35–42% (Fig. 6). Overall, our results indicate that SDG can attenuate inflammatory changes in BMVEC and monocytes and reduce adhesion/migration of monocytes across BBB models in vitro.