Minoxidil response was never one number. It is a balance of five things and the field measures one. Here are the eight experiments that would measure the rest, and what each one would finally answer.
Minoxidil is the most-used hair-loss drug in the world, and its foundational experiments were never run. We covered the gaps in a companion piece on what we still do not know about minoxidil. This is the other half of that story: not the questions, but the experiments that would actually answer them.
Start with the reframe, because it organizes everything. Minoxidil response was never one number. It is a balance of five things: how much drug reaches the follicle, how much of it gets switched on, how much cofactor the activating enzyme has, how fast the drug gets switched back off, and whether the target it acts on is even there. The field measures one of those five (the activating enzyme, on a plucked hair) and calls it your response. Here is how you measure the rest.
The short answer: The reason minoxidil is still poorly understood is not that it is mysterious, it is that the key measurements were skipped. Eight experiments would close the gap: measure the active metabolite in blood, map the activating enzyme in a human follicle, weigh activation against deactivation in the same follicle, test the target channel directly, ask whether the cofactor is limiting, re-validate the response test with real chemistry, check for a backup enzyme, and read the genes, including the potassium-channel genes, that may predict who responds. None are exotic. They are the experiments that should have been run decades ago.
Yes, with a purpose-built method. Minoxidil is a prodrug (McCall 1983); the form that actually grows hair is minoxidil sulfate (Buhl 1990), and it has never been detected in human blood. Every human measurement quantifies the inactive parent drug (Fleishaker 1989). The reason is not, as often claimed, that the sulfate is wildly unstable. It is that the body barely makes it systemically (the liver mostly makes a glucuronide instead; Gottlieb 1972) and that it falls apart during ordinary sample preparation, reverting to the parent drug before it is counted. One published skin method even monitored the mass of minoxidil while labeling it "minoxidil sulphate," which is exactly what happens when the sulfate group falls off in the instrument.
The experiment: a dedicated assay that catches the sulfate intact, on its own mass, in a cold, non-acidic, enzyme-inhibited sample, before it can convert. What it settles: whether the active drug can reach the follicle through the bloodstream at all. If it cannot, then oral minoxidil, like topical, must be activated locally at the follicle, which reframes why oral works and supports the case for smoother-release oral dosing.
Nobody knows, because the enzyme has never been localized in a human follicle. The only images are from rats (Dooley 1991). The obstacle is specific: SULT1A1 is roughly 93 to 96 percent identical to its cousins SULT1A2 and SULT1A3, and antibodies cannot tell them apart. The leading commercial antibody cross-reacts completely with SULT1A3.
The experiment: read the RNA instead of the protein. RNA probes can distinguish the near-identical enzymes because their genetic sequences differ even where the proteins do not, and then confirm the protein in the exact cells the RNA points to, using a mass-spec method built around a peptide unique to SULT1A1. What it settles: which follicular cell actually does the activating, which in turn tells you whether the standard plucked-hair test is even sampling the right cells.
Possibly the off-switch, and no one is looking. The scalp switches minoxidil on with SULT1A1, but human follicle cells also make sulfatases that strip the sulfate right back off, turning the active drug inactive again (Ichida 2020). The prime suspects are steroid sulfatase (the gene STS), which is the sulfatase concentrated in the dermal papilla (Hoffmann 2001), and arylsulfatase A (ARSA), though no one has yet confirmed which enzyme actually cleaves minoxidil sulfate. Every responder study reads only the forward, activating rate.
The experiment: measure both directions, activation and deactivation, in the same follicle, and compute the balance. What it settles: whether some non-responders are not enzyme-poor at all, but are deactivating the drug as fast as they make it, a group that no activation-boosting product could ever help.
This is the cleanest test in the set. Minoxidil sulfate ultimately works by opening one specific potassium channel in the follicle (Shorter 2008). Drugs like diazoxide open that same channel directly, without needing the enzyme at all.
The experiment: take confirmed non-responders and apply a direct channel opener that skips the enzyme entirely. What it settles: if their follicles respond to the direct opener, the target works fine and their block was upstream, at activation, they simply could not switch minoxidil on. If they do not respond even to the direct opener, the failure is the target itself, and no amount of activation would ever have helped. Either way, it is the first clean split of minoxidil non-response into an activation problem versus a target problem.
Untested, and mechanistically plausible. The activating enzyme cannot work without a sulfate donor called PAPS. A follicle could be rich in enzyme but starved of cofactor, and it would look identical to a low-enzyme non-responder on every test we have.
The experiment: measure the cofactor machinery across follicular compartments, then test whether topping up the sulfate supply raises activation without changing the enzyme level at all. What it settles: whether the true rate limit is the enzyme or its fuel, which determines whether "boost the enzyme" was ever the right target.
It has never been checked properly. The plucked-hair "minoxidil response test" is a cheap color-change assay from a single group (Goren 2014), never independently validated with a definitive method. And there is a warning sign: one study found it running backwards for oral minoxidil, where lower measured enzyme activity was associated with a better response (Jimenez-Cauhe 2024).
The experiment: re-run the test with mass spectrometry, quantifying the actual active metabolite, in the follicular cells confirmed to matter (which is why this experiment has to wait for the localization work above). What it settles: whether the test is measuring genuine drug activation in the right cells, or a colorimetric proxy in the wrong ones.
Maybe, and it would change the whole picture. In a test tube, at least four different sulfotransferases can activate minoxidil (Anderson 1998), including a liver DHEA-sulfotransferase characterized directly with minoxidil as its substrate (Kudlacek 1997). SULT1A1 is the dominant one in the follicle, but no one has ever removed it and checked whether a backup quietly covers.
The experiment: knock down SULT1A1 in follicle culture and measure whether activation survives, and test the untested enzymes against minoxidil directly. What it settles: whether SULT1A1 is truly a single point of failure. If a backup exists, a "low-SULT1A1" scalp is not a lost cause, and blocking the enzyme with a supplement would not fully stop the drug.
The genetics field has focused almost entirely on the activating enzyme, where only one variant is loosely tied to response (Ramos 2021). But minoxidil ultimately acts on a potassium channel built from two genes, KCNJ8 and ABCC9, and people born with overactive versions of those genes have a condition whose hallmark is hair growing everywhere (Harakalova 2012). Minoxidil essentially mimics that condition for a few hours.
The experiment: a genotype panel spanning the whole pathway, not just the enzyme, tested against real response, with the channel genes and downstream effectors included alongside the enzyme, and androgen-pathway genes treated as severity covariates. What it settles: whether a pre-treatment genetic readout can predict response better than the current activity test, and whether the target side of the drug, not the enzyme side, is where the real genetic signal lives.
We did not write this list to admire the gap. Anagen's research is built around closing it, working through these measurements with the labs and analytical chemists who can run them, starting with the two that unlock the rest: catching the active metabolite in blood, and localizing the activating enzyme in a real human follicle. We intend to publish what we find, including the results that do not go our way, because the honest version of this science is the entire point. Minoxidil is going to keep being the first thing prescribed for hair loss. It should be the best understood drug in the field, not one of the least.
For the questions behind these experiments, see what we still do not know about minoxidil. For the practical responder story, see why minoxidil works for some people and not others. Not sure where you fall? Take the hair quiz or see treatment options.
Because it is a chemistry problem, not an oversight. Minoxidil sulfate is only a minor systemic metabolite (the liver mostly makes a glucuronide), it is a polar molecule that behaves badly on standard assays, and it reverts to the inactive parent drug during ordinary sample preparation. It is measurable now with a dedicated method that keeps it intact and detects it on its own mass, but no one has built one.
Because SULT1A1 is about 93 to 96 percent identical to its cousins SULT1A2 and SULT1A3, and antibodies recognize the shared protein surface, so they cannot tell the three apart (the leading antibody cross-reacts completely with SULT1A3). RNA probes target the genetic sequence, which differs between the near-identical enzymes, so they can localize SULT1A1 specifically where antibodies cannot.
It is an open, testable possibility. Human follicle cells make sulfatases that convert active minoxidil sulfate back to the inactive parent (Ichida 2020), and no study has measured that deactivation alongside activation in the same follicle. If someone deactivates the drug as fast as they make it, they would look like a low-enzyme non-responder on current tests but would not be helped by boosting the enzyme.
By hitting the target directly. Minoxidil works by opening a specific potassium channel, and other drugs open that same channel without needing the activating enzyme. If a confirmed non-responder's follicles grow when given a direct channel opener, then their activation was never the problem, and the failure is downstream at the target.
It has never been independently validated. The accuracy figures come from a single group's colorimetric assay, it has not been confirmed with mass spectrometry, and one study found it running backwards for oral minoxidil, where lower measured activity tracked with better response (Jimenez-Cauhe 2024). Re-checking it with real chemistry, in the right cells, is one of the eight experiments.
This article is for educational purposes only and is not medical advice. It describes proposed research and open scientific questions, which are labeled as such, and does not describe any product, treatment recommendation, or proven outcome. Oral minoxidil is used off-label for hair loss and should be undertaken only under the supervision of a licensed clinician; topical minoxidil is FDA-approved for that use. Always consult a qualified healthcare provider before starting, stopping, or changing any treatment.