Blood-Brain Barrier Penetration

The logarithm of a compound's partition coefficient between octanol and water. It is the single most critical determinant of passive BBB permeability.

• Optimal Range: A LogP between 1.5 and 2.7 is generally favored for CNS penetration.
• Mechanism: High lipophilicity increases partitioning into the lipid bilayer of endothelial cells, but excessive values (LogP > 5) lead to sequestration in adipose tissue and high plasma protein binding, reducing the free fraction.
• LogD vs. LogP: The distribution coefficient (LogD) at physiological pH (7.4) is often more predictive as it accounts for ionization.

The sum of the atomic weights of all atoms in a molecule. It serves as a primary proxy for molecular size and complexity.

• Threshold: The widely accepted upper limit for optimal brain penetration is < 400 Da, though successful CNS drugs can reach up to 500 Da.
• Mechanism: Larger molecules diffuse more slowly through the tight junctions and lipid membranes. A higher MW also increases the likelihood of being a substrate for efflux transporters like P-glycoprotein (P-gp).
• Paracellular Route: The paracellular aqueous pathway is highly restrictive, effectively excluding molecules larger than ~200 Da unless actively transported.

The total number of hydrogen bond donors (HBD) and acceptors (HBA) on a molecule. This is a key component of the Rule of 5 and a dominant negative predictor for BBB penetration.

• Desolvation Penalty: To enter the hydrophobic membrane interior, a molecule must shed its water shell. Breaking hydrogen bonds with water requires significant energy, making highly polar compounds energetically unfavorable for passive diffusion.
• Thresholds: The sum of nitrogen and oxygen atoms (a rough HBA count) should be ≤ 5. The number of NH and OH groups (HBD) should be ≤ 3.
• Polar Surface Area (PSA): A closely related metric; a PSA < 60–70 Ų is a strong predictor of good brain penetration, directly quantifying the capacity for hydrogen bonding.

A calculated sum of the surface areas of polar atoms (primarily oxygen, nitrogen, and their attached hydrogens) in a molecule. It is a superior descriptor to simple HBD/HBA counts.

• Predictive Power: TPSA is inversely correlated with passive brain permeability. It elegantly combines molecular size and hydrogen bonding capacity into a single dynamic property.
• Mechanism: A low TPSA minimizes the energetic cost of desolvation required for membrane permeation.
• Clinical Relevance: This parameter is heavily weighted in multiparameter optimization (MPO) scores used by medicinal chemists to design CNS drug candidates.

The acid dissociation constant (pKa) determines the fraction of a drug that is ionized at a given pH. The BBB preferentially permits the passive diffusion of neutral, uncharged species.

• pH Trapping: The "pH-partition hypothesis" states that only the unionized form of a drug can cross the membrane. Once inside the CNS (pH 7.4), a weak base may become ionized and trapped.
• CNS vs. Non-CNS Drugs: CNS-active drugs tend to have a lower number of basic centers and a higher proportion of neutral molecules compared to peripheral drugs.
• Ampholytes: Zwitterionic compounds can sometimes cross via specific transporters, but their passive permeability is generally low.

The number of rotatable bonds is a simple count of single, non-ring bonds connecting heavy atoms. It is a measure of molecular flexibility and conformational entropy.

• Entropic Penalty: A highly flexible molecule has many degrees of freedom. Immobilizing it within the rigid, ordered environment of a lipid membrane incurs a significant entropic cost, reducing permeability.
• Threshold: Keeping the number of rotatable bonds ≤ 8 is a common design guideline for CNS candidates.
• Rigidification: Medicinal chemistry strategies like introducing rings or double bonds can reduce rotatable bond count and improve brain exposure.

The logarithmic ratio of a compound's concentration in the brain to its concentration in the blood. It is the primary quantitative endpoint for blood-brain barrier penetration models. A LogBB value greater than 0 indicates preferential brain distribution, while a value less than 0 suggests limited penetration. In silico models often predict LogBB from molecular descriptors like polar surface area (PSA) and lipophilicity (LogP).

The Central Nervous System Multiparameter Optimization score is a desirability function developed by Pfizer to guide medicinal chemists in designing brain-penetrant candidates. It aggregates six key physicochemical properties into a single score from 0 to 6:

• ClogP: Calculated partition coefficient
• ClogD: Calculated distribution coefficient at pH 7.4
• MW: Molecular weight
• TPSA: Topological polar surface area
• HBD: Number of hydrogen bond donors
• pKa: Most basic center's pKa A score ≥ 4 is generally associated with a higher probability of CNS target engagement.

An in vitro cell-based assay using Madin-Darby Canine Kidney cells transfected with the human MDR1 gene, which encodes P-gp. This assay measures the apparent permeability (Papp) of a compound in both apical-to-basolateral and basolateral-to-apical directions. The resulting efflux ratio (ER) is a gold-standard experimental input for training machine learning models to predict whether a compound will be actively excluded from the brain.

The principle that only the unbound (free) fraction of a drug in plasma and brain tissue is available to interact with a pharmacological target. For CNS drugs, achieving a high Kp,uu (the ratio of unbound brain concentration to unbound plasma concentration) is often more critical than total brain-to-plasma ratio. Predictive models must account for both plasma protein binding and brain tissue binding to estimate this pharmacologically active concentration.

A widely used 2D molecular descriptor calculated as the sum of the surface contributions of polar atoms (primarily oxygen and nitrogen, including attached hydrogens). TPSA is inversely correlated with passive brain penetration. A common rule of thumb is that a TPSA value below 60-70 Ų is a prerequisite for significant passive BBB permeability, while a value above 140 Ų generally precludes it. It is a cornerstone feature in almost all blood-brain barrier classification models.