Liu Yu

The blood-brain barrier (BBB) is an important physiological barrier that protects the brain by preventing entry of harmful substances into the central nervous system (CNS). While this barrier is effective at blocking most drugs from disrupting the tightly controlled environment of the CNS, certain classes of non-neurologic drugs can cross the BBB, potentially causing neurological side effects. Developing safer alternatives to these drugs requires a reliable method for predicting the BBB non-penetrance, whilst maintaining a high level of drug-like properties to the molecules. Existing regression methods for BBB penetration are inaccurate for this task, tending to overestimate logBB values due to imbalances in available datasets. To address this, we have developed a LightGBM-based model with a custom multi-task loss function that learns from both regression and classification data. Improving performance across classification and regression metrics, our model outperforms existing models in identifying non-penetrant molecules and achieves regression accuracy approximately equal to experimental uncertainty. Based on this, we introduce the BBB Exclusion (BBBX) Score, a scoring method which stratifies compounds by predicted BBB penetrance and confidence. To interpret model outputs, SHAP analysis was used to identify features contributing most to model outputs. We use these features to define threshold values associated with increased likelihood of BBB Exclusion whilst maintaining high oral bioavailability and drug-like properties required for developing a safe, non-CNS drug.

Artificial lipid vesicles are an elegant way of producing auto-assembled customizable chambers, and they have been widely studied for entrapping small molecules in the context of drug delivery. However, their use for microorganism entrapment remains largely unexplored. One obstacle is the generation of giant vesicles (GVs) large and rigid enough to contain potentially mobile micron-size organisms in a way that results in efficient cell encapsulation. To address this challenge we developed a modified version of the previously described gel-assisted swelling method, in which immobilized GVs are formed by hydrating a lipid film deposited on agar with a heated solution. We used solutions containing Magnetospirillum magneticum strain AMB-1 cells to test for bacterial entrapment. Systematically exploring POPC:DOPS:cholesterol ternary mixtures to vary the surface charge and rigidity of the GVs’ membrane allowed the determination of specific conditions leading to the efficient entrapment of these highly mobile cells. Lipid order measurements with the fluorescent membrane probe Laurdan show that entrapment is associated with a significant decrease in membrane rigidity between the swelling and entrapment steps. The high sensitivity of successful entrapment on DOPS concentration further suggests that membrane fluctuations play a major role in allowing the micron-size bacteria to push through the membrane and break in and out of the GVs.

Immune protection associated with consuming colostrum-based peptides is effective against bacterial and viral insults. The goal for this study was to document acute changes to immune surveillance and cytokine levels after consuming a single dose of a nutraceutical blend in the absence of an immune challenge. A double-blind, randomized, placebo-controlled, cross-over pilot study involved healthy participants attending two clinic visits. Blood draws were performed pre-consumption and at 1, 2, and 24 h after consuming a blend of bovine colostrum- and hen’s egg-based low-molecular-weight peptides (CELMPs) versus a placebo. Immunophenotyping was performed by flow cytometry, and serum cytokines were measured by multiplex cytokine arrays. Consumption of CELMPs triggered increased immune surveillance after 1 h, involving monocytes (p < 0.1), natural killer (NK) cells (p < 0.1), and natural killer T (NKT) cells (p < 0.05). The number of NKT cells expressing the CD25 immunoregulatory marker increased at 1 and 2 h (p < 0.1). Increased serum levels of monocyte chemoattractant protein-1 (MCP-1) was observed at 2 and 24 h (24 h: p < 0.05). Selective reduction in pro-inflammatory cytokines was seen at 1, 2, and 24 h, where the 2-h reduction was highly significant for IL-6, IFN-γ, and IL-13. The rapid, transient increase in immune surveillance, in conjunction with the reduced levels of inflammatory markers, suggests that the CELMP blend of natural peptides provides immune benefits of use in preventive medicine. Further studies are warranted in chronic inflammatory conditions.

Immune cells express Pattern Recognition Receptors (PRRs) to recognize potentially pathogenic microbial forms. Nutraceutical compounds can induce immune cell activation through PRRs. The nutraceutical immune blend (IB), QuickStart™, contains botanical and yeast-derived ligands for PRRs, along with vitamin C and zinc. We evaluated immune-activating effects of the IB and its ingredients in vitro. Human peripheral blood mononuclear cells were treated with either the IB or single ingredients: elderberry extract, the proprietary Saccharomyces cerevisiae fermentate EpiCor™ (Sacc), the plant-based hemicellulose preparation Natramune (PDS-2865)™ (Hemi), vitamin C (VitC), or zinc gluconate (Zinc). The IB triggered sequential waves of immune activation. Initial cytokine induction by the IB at 2 h involved the immune-activating cytokines IL-6, IL-8, MIP-1α, and TNF-α, and the stem cell-mobilizing growth factor G-CSF, as did Sacc and Hemi. The 24 h immune-activation by the IB included increases in IL-1β, IL-17A, IP-10, GM-CSF, Basis FGF, PDGF-BB, and the anti-inflammatory cytokine IL-10. Increased CD69 expression by the IB was also seen for VitC and Sacc. Increased CD25 expression by the IB on monocytes was also seen for Sacc. The IB triggered rapid immune activating events of higher magnitude than the single ingredients, involving immune-activating cytokines and restorative growth factors. Clinical research is warranted to evaluate rapid immune-modulating events upon consumption.

Spore-forming probiotic bacteria, including Bacillus coagulans, are resilient and produce a variety of beneficial metabolites. We evaluated the immune-modulating effects of the novel probiotic strain Bacillus coagulans JBI-YZ6.3, where the germinated spores, metabolite fraction, and cell wall fraction were tested in parallel using human peripheral blood mononuclear cell cultures under both normal and lipopolysaccharide-induced inflamed culture conditions. The expression of CD25 and CD69 activation markers was evaluated via flow cytometry. Supernatants were tested for cytokines, interferons, chemokines, and growth factors using Luminex arrays. The germinated spores were highly immunogenic; both the cell wall and metabolite fractions contributed significantly. Under normal culture conditions, increased levels of immune activation were observed as increased expressions of CD25 and CD69 relative to natural killer cells, suggesting an increased ability to attack virus-infected target cells. On monocytes, a complex effect was observed, where the expression of CD25 increased under normal conditions but decreased under inflamed conditions. This, in combination with increased interleukin-10 (IL-10) and decreased monocyte chemoattractant protein-1 (MCP-1) production under inflamed conditions, points to anti-inflammatory effects. The production of the stem cell-related growth factor granulocyte colony-stimulating Factor (G-CSF) was enhanced. Further research is warranted to characterize the composition of the postbiotic metabolite fraction and document the characteristics of immunomodulating agents secreted by this probiotic strain.

The Nerium oleander extract PBI 05204 (PBI) and its cardiac glycoside constituent oleandrin have direct anti-viral properties. Their effect on the immune system, however, is largely unknown. We used an in vitro model of human peripheral blood mononuclear cells to document effects under three different culture conditions: normal, challenged with the viral mimetic polyinosinic:polycytidylic acid Poly I:C, and inflamed by lipopolysaccharide (LPS). Cells were evaluated for immune activation marks CD69, CD25, and CD107a, and culture supernatants were tested for cytokines. Both PBI and oleandrin directly activated Natural Killer (NK) cells and monocytes and triggered increased production of cytokines. Under viral mimetic challenge, PBI and oleandrin enhanced the Poly I:C-mediated immune activation of monocytes and NK cells and enhanced production of IFN-γ. Under inflammatory conditions, many cytokines were controlled at similar levels as in cultures treated with PBI and oleandrin without inflammation. PBI triggered higher levels of some cytokines than oleandrin. Both products increased T cell cytotoxic attack on malignant target cells, strongest by PBI. The results show that PBI and oleandrin directly activate innate immune cells, enhance anti-viral immune responses through NK cell activation and IFN-γ levels, and modulate immune responses under inflamed conditions. The potential clinical impact of these activities is discussed.

Lipopolysaccharide (LPS) is a key surface component of Gram-negative bacteria, populating the outer layer of their outer membrane. A number of experimental studies highlight its protective role against harmful molecules such as antibiotics and antimicrobial peptides (AMPs). In this work, we present a theoretical model for describing the interaction between LPS and cationic antimicrobial peptides, which combines the following two key features. The polysaccharide part is viewed as forming a polymer brush, exerting an osmotic pressure on inclusions such as antimicrobial peptides. The charged groups on LPS (those in lipid A and the two Kdo groups in the inner core) form electrostatic binding sites for cationic AMPs or cations. Using the resulting model, we offer a quantitative picture of how the brush component enhances the protective role of LPS against magainin-like peptides, in the presence of divalent cations such as Mg²⁺. The LPS brush tends to diminish the interfacial binding of the peptides, at the lipid headgroup region, by about 30%. In the presence of 5 mM of Mg²⁺, the interfacial binding does not reach a threshold value for wild-type LPS, beyond which the LPS layer is ruptured, even though it does for LPS Re (the simplest form of LPS, lacking the brush part), as long as [AMP] ≤ 20 μM, where [AMP] is the concentration of AMPs. At a low concentration of Mg²⁺ (≈1 mM), however, a smaller [AMP] value (≳2 μM) is needed to reach the threshold coverage for wild-type LPS. Our results also suggest that the interfacial binding of peptides is insensitive to their possible weak interaction with the surrounding brush chains.