KW-6356 / Sipagladenant

What KW-6356 is

KW-6356, also called sipagladenant, is an investigational selective adenosine A2A receptor antagonist and inverse agonist developed by Kyowa Kirin for Parkinson's disease. Laboratory work found high A2A affinity, slow dissociation, and insurmountable antagonism. Those receptor findings do not establish cognitive enhancement in healthy people.

Phase 1 established pharmacokinetics, not focus

Healthy men received single doses from 1 to 60 mg and repeated doses of 6 or 24 mg once daily for up to 14 days. The controlled studies reported tolerability at those studied exposures and the long half-life range used in this experiment's carryover planning. They did not test focus, reaction time, or performance during sleep loss.

The Parkinson's results were mixed

In an early trial of untreated Parkinson's disease, motor scores improved numerically by 5.37 points with 3 mg and 4.76 points with 6 mg, versus 3.14 points with placebo. Neither active dose significantly separated from placebo. A later 502-person trial as an add-on to levodopa did meet its primary motor endpoint, with improvements of 7.8 points at 3 mg and 7.0 points at 6 mg, versus 5.2 points with placebo. The numerical pattern does not prove that 3 mg is superior to 6 mg.

Why development stopped

Kyowa Kirin discontinued development in July 2022 after reviewing the global regulatory landscape, development hurdles, and expected time to market. The company did not describe the decision as a safety failure.

Why sleep pressure was the test condition

Adenosine signaling contributes to accumulating sleep pressure. Mouse knockout work supports a necessary role for A2A receptors in caffeine-induced arousal, and human caffeine research links ADORA2A variation with vigilance and recovery sleep during prolonged wakefulness. These findings made an A2A-based wakefulness hypothesis reasonable. They did not show that KW-6356 preserves focus or performs better than caffeine.

The direct human evidence gap

No controlled human study identified here tested KW-6356 for focus, vigilance, or decision quality during sleep deprivation. A small crossover study in 20 rested healthy men tested the related drug istradefylline, not KW-6356. One social-information measure changed; the other cognitive and cardiovascular measures did not significantly change. This experiment therefore began with an open question, not an expected positive result.

Material documentation

The supplied one-page qNMR report lists KW-6356, an assumed molecular weight of 397.4, and 99.9% purity. Its displayed arithmetic recalculates to 99.909%. The page does not include the full spectrum, an orthogonal identity test, a finished-solution concentration assay, or visible linkage between the tested powder and this vial. For that reason, this page treats the recorded amounts as label-equivalent exposures based on the vial concentration.

Sources

• Ohno et al., Molecular Pharmacology, 2023: receptor pharmacology, slow dissociation, and insurmountable antagonism.
• Tayama et al., Clinical Pharmacology in Drug Development, 2023: Phase 1 safety and pharmacokinetics in healthy men.
• Maeda et al., Parkinsonism & Related Disorders, 2023: early untreated Parkinson's trial.
• NCT03703570: levodopa-adjunct Phase 2b study record.
• Kyowa Kirin, October 21, 2020: Phase 2b primary-endpoint announcement.
• Kyowa Kirin, July 15, 2022: development discontinuation notice.
• Hook et al., European Neuropsychopharmacology, 2023: istradefylline cognition study in rested healthy men.
• Huang et al., Nature Neuroscience, 2005: A2A receptors and caffeine-induced arousal in mice.
• Bodenmann et al., Sleep, 2012: ADORA2A variation, caffeine, vigilance, and recovery sleep after prolonged wakefulness.
• Van Dongen et al., Sleep, 2003: cumulative objective impairment and subjective sleepiness during sleep restriction.

Final verdict

KW-6356 was detrimental to my health and performance overall. Lower exposures produced a potentially useful objective vigilance signal, but the effect arrived too late and lasted too long. Even when I took it at 03:00, it disrupted the next sleep cycle more than 16 hours later. The resulting sleep loss eventually impaired the performance I was trying to preserve.

Research question

Could KW-6356 preserve objective psychomotor vigilance during a sleep-deprived day without creating an unacceptable cost in the following sleep period?

Recovery sleep sequence

Exposure and outcome ledger

Sleep Performance is the wearable's sleep score, not cognitive performance. PVT is the separate objective reaction-time measure.

What the PVT data show

My seven-morning baseline PVT mean was 348.11 ms. The morning after the original 3 mg exposure, I recorded 316.19 ms after about four hours of sleep, the fastest confirmed mean in the 22-session same-instrument series. After the October 6 dose and only 3:36 of sleep, the October 7 PVT was 352.78 ms, essentially baseline.

After the 5 mg escalation and 2:27 of sleep, the October 8 PVT mean slowed to 397.32 ms. That was 14.1% slower than baseline and the slowest confirmed mean in the same-instrument series. The median slowed from a 331 ms baseline reference to 408 ms. The impairment appeared as broad slowing rather than a large increase in extreme lapses.

The October 6 pre-dose check

Before taking 2.5 mg at 03:00, a differently formatted PVT export recorded a 414.57 ms mean, a 396 ms median, two lapses, and three false starts. The next morning's 352.78 ms result was much faster despite less sleep. Because the October 6 test used a different application or protocol format, I treat that change as supporting context rather than a clean same-instrument comparison.

Why the lower-dose signal did not rescue the result

Objective vigilance was preserved on the morning after 2.5 mg, but the useful wakefulness was not present when I wanted it during the first half of the prior day. It became much more noticeable later, then continued into the sleep period. Preserving one short morning test does not make the compound useful when the same exposure damages the recovery required for the following day.

Carryover was the failure mode

Phase 1 reported mean terminal half-lives from 18.4 to 43.1 hours. At 16 hours, simple half-life arithmetic suggests that roughly 55% to 77% of the parent exposure could remain. Even at 24 hours, the estimate is about 40% to 68%. The October 7 escalation therefore added 5 mg while substantial October 6 exposure could still have been present. Taking KW-6356 at 03:00 did not make it clear enough for the next sleep cycle.

Decision

I stopped KW-6356 on October 8. Sleep increased to 4:19 on the first night after stopping, and Sleep Performance recovered from 31% to 50%, although it had not yet returned to the pre-test level. I am closing the experiment as intolerable rather than testing another dose. For my schedule and objective, the duration of action made the compound a net liability.

Safety boundary

Sleep-deprived driving and other safety-sensitive tasks are not test conditions. Feeling alert cannot be used as evidence that impairment is absent. Severe or acute symptoms belong in medical care, not in an experiment ledger.