This was originally a really long toot.
I’ve been looking at papers on antihistamines and what makes them work better like, molecularly, not just doing clinical trials and seeing what happens. As a primer, allergens trigger the release of histamines in the body, which then bind to various histamine receptors, and the H₁ receptors are the important ones that trigger rhitinis (runny nose, etc.) and uticaria (itchiness). The antihistamines you can get for allergies are H₁-receptor antagonists, meaning they bind to the H₁ receptors and prevent histamine from triggering them, as well as producing some sort of antagonist effect (?). Simons and Simons 2011 is a pretty good overview of how it works.
Sedative effects
Histamine receptors are also found in the brain, and triggering them causes the usual sedative effects like drowsiness and impairment. Antihistamines are divided into 1st-gen and 2nd-gen, where 1st-gen antihistamines cross the blood–brain barrier (BBB) and have a sedative effect, and 2nd-gen antihistamines “don’t”. In reality, a lot of 2nd-gen antihistamines do cross the BBB, only to a lesser extent; Yanai et al. 2017 quantify this by measuring the H₁ receptor occupancy percentage in the brain by typical clinical doses of various antihistamines. Moderately sedating antihistamines are considered to be between 20% and 50%, while highly sedating ones over 50%, and nonsedating ones are below 20%. In particular, fexofenadine and bilastine are especially nonsedating.
Whether an antihistamine crosses the BBB or not depends on its lipophilicity, where more lipophilic molecules are more likely to cross it. Molecules with a -COOH or -NH₂ group are more lipophobic, supposedly because with the -COOH group the molecule is zwitterionic and that prevents crossing (citation needed?). The paper I cited above shows that the 2nd-gen antihistamines have -COOH or NH₂ groups, but doesn’t explain what differences among the molecules produce the different levels of sedation
Duration of action
The effectiveness of antihistamines seems to be related to the molecule’s binding affinity Kd = kon/koff, where kon is a “kinetic association rate constant” and koff is a “kinetic dissociation rate constant”. I’m not sure what kon relates to (onset time?), but 1/koff is the residence time, how long the antihistamine stays bound to the histamine receptor. Akimoto et al. 2021 correlate molecular properties of antihistamines to koff. Firstly, koff is positively correlated with entropy and negatively correlated with enthalpy.
From previous work by Hishinuma et al. 2014, they’ve picked five properties: “sum of degrees” (number of non-hydrogen atoms), eletrostatic potential, water-accessible surface area (lipophobicity = hydrophilicity?), hydrogen bonding acceptor count (??), and ovality which, you guessed it, is how oval-shaped a molecule is. Anyway, most of that doesn’t matter, because none of them except for electrostatic potential was statistically significant.
They did find that koff was correlated to some metric called vsurfCW2, which is:
a 3D molecular field descriptor weighted by capacity factor 2, the ratio of the hydrophilic surface to the total molecular surface
I guess it’s neat they can computationally predict well the residence time, although I was hoping it would be a property I could intuitively understand.
Molecular shape of antihistamines
Most antihistamines share the same molecular structure given below (the X can be C or N or CO).
I’ve come across a few papers that try to test what specific aspects of this shape affect binding affinity and residence time:
- Bosma et al. 2019 look at desloratadine and rupatadine, both of which are tricyclic (the two rings are joined) and have another amine ring at the end. They attach a bunch of different groups to the N of the amine with either one spacer carbon or no spacer, and they found that the ones with a spacer have a higher kinetic rate index, which appears to be associated with longer residence times. The idea is that having the spacer increases residence time, and explains why rupatadine’s (which has a spacer) is longer than desloratadine’s (which just ends in an H).
- Wang et al. 2021 find that joining the two rings into a tricyclic prolongs residence time. They also tried different links between the tricyclic and the amine, whether it’s an ethylene (double bond) or an ethyl (single bond…?).
- Kuhne et al. 2024 look at attaching groups to the “upper” aromatic ring (how are they oriented?), which they call olopatadine mimics, as well as attaching groups to the N of the amine, which they call levocetirizine mimics.
Relations between antihistamines
In terms of the antihistamines you find in pharmacies, you can group them together into families:
- piperazines (middle ring has two Ns): cetirizine is a racemic mixtrue of dextrocetirizine, which isn’t an antihistamine, and levocertirizine, which is.
- piperidines (middle ring has one N): desloratatine, the tricyclic I mentioned earlier, is a metabolite of loratadine and rupatadine. Notable other ones include fexofenadine and bilastine. There’s a bunch more listed, but they seem to be obscure eye drops.
The overview Simons and Simons paper also lists a bunch under “other”, which don’t even have the same generic shape: azelastine, emedastine, epinastine, olopatadine. They’re all eye drops, and it turns out there’s a bunch of antihistamines that are specifically only eye drops.