Roflumilast

Roflumilast is the FDA-approved oral PDE4 inhibitor for severe COPD (Daliresp, 2011) and the topical FDA-approved PDE4 inhibitor for plaque psoriasis (Zoryve, Arcutis, 2022 — different formulation). Two small human cognitive RCTs (Van Duinen 2018, Gilleen 2018) show a real episodic-memory / cognition signal at 250-500 mcg/day, validating the cAMP→CREB→BDNF→LTP thesis in humans. The same catalytic-site, non-subtype-selective binding that delivers the cognitive effect also delivers nausea, diarrhea, and weight loss in 10-20% of users — the rolipram problem. For Dylan: SKIP. BPN14770 (allosteric PDE4D NAM, primate-specific UCR2 binding pocket) is the cleaner version of this mechanism and is already on the WATCH-LIST.

Roflumilast is a second-generation catalytic-site PDE4 inhibitor. Like rolipram before it, it binds the catalytic pocket of all four PDE4 isoforms (PDE4A, B, C, D) — the same pocket where cAMP itself docks for hydrolysis. It is more potent than rolipram (low-nM IC50) and more drug-like (longer half-life, smoother PK), but it is not subtype-selective, and that is the central pharmacological fact that defines its profile.

Step-by-step:

1. PDE4 hydrolyzes cAMP → AMP. Inhibiting PDE4 raises cAMP everywhere PDE4 is the dominant degradation enzyme.
2. In inflammatory cells (neutrophils, eosinophils, macrophages, T-cells), elevated cAMP suppresses pro-inflammatory cytokine release (TNFα, IL-8, leukotriene B4) and reduces cell trafficking. This is the COPD/psoriasis story — chronic airway inflammation in COPD, T-cell-driven inflammation in psoriasis.
3. In CNS neurons (cortex, hippocampus), elevated cAMP activates PKA → phosphorylates CREB at Ser133 → transcribes BDNF, Arc, c-Fos → facilitates late-phase LTP → improves memory consolidation. Same canonical "memory-enhancement molecular pathway" as the rolipram / BPN14770 / forskolin story.
4. In the area postrema (the brain's "vomiting center" — circumventricular organ outside the BBB), PDE4D is the dominant emetic isoform (Robichaud 2002 knockout work). Catalytic-site inhibitors hit PDE4D here at the same plasma concentrations that produce the desired effects — hence emesis is the dose-limiting toxicity. Roflumilast was approved at 500 mcg/day specifically because that is the highest dose that most patients can tolerate; even at 500 mcg, ~10-20% drop out from GI effects in COPD trials.
5. Weight loss (median ~2 kg over 1 year in COPD trials) appears mechanism-driven, not solely a side-effect of nausea — likely reflects elevated cAMP in adipocytes (lipolysis) and altered energy balance. It is consistent enough across trials to appear in the label.

Why roflumilast is the "older sibling" of BPN14770:

The BPN14770 file on this wiki goes deep on the mechanistic case for allosteric PDE4D NAMs — they bind a regulatory site (UCR2 / Phe196) that exists only in primate PDE4D, partially inhibit the enzyme (~50% max), and crucially do not bind the conformer associated with emesis. Roflumilast is the opposite design philosophy: full catalytic-site inhibition of all four PDE4 isoforms. The cognitive effect is real because it raises CNS cAMP. The emetic ceiling is also real because it raises area postrema cAMP at the same dose. There is no clean therapeutic window in between for chronic cognitive enhancement. This is exactly the gap that BPN14770 was engineered to close.

Plain English: Roflumilast turns up cAMP throughout the body by jamming the off-switch on PDE4. In the lungs and skin (where it's approved), this dampens inflammation. In the brain, this gives a real but small memory-consolidation lift via CREB/BDNF/LTP. In the gut and the brain's vomiting center, this also makes you nauseous and gives you diarrhea. The cognitive benefit is genuine; the GI cost is also genuine. It's the rolipram problem with better PK.