ATX-304

Quick answer

ATX-304 (also called O-304) is an experimental small-molecule drug, not a peptide, designed to switch on AMPK, an enzyme engaged by exercise. A 2018 trial in 65 adults with type 2 diabetes used 1,000 mg a day for 28 days and found a modest fall in fasting blood sugar. In 2026 the developer presented two Phase 1 studies using 400 mg doses, reporting a higher resting metabolic rate than at baseline but minimal weight loss. It is not approved anywhere.

Ask about ATX-304

UPDATE, NOVEMBER 18, 2026. Confirmed: the developer presented new Phase 1 data on ATX-304 as Poster-111 at ObesityWeek on November 14 (Cambrian Bio announcement). The figures below are read from the poster panels. Not confirmed: a full poster file or paper, meaningful weight loss, or that a higher daily amount works better.

Research snapshot

CategoryCurrent information
Compound categorySmall-molecule AMPK activator (not a peptide)
Primary research interestBlood sugar and insulin resistance in type 2 diabetes; resting energy use, liver fat and body composition in obesity
Highest available evidenceSmall published Phase IIa trial (65 adults, 28 days, 2018), plus developer-presented Phase 1 data from 2026: a 23-adult Phase 1b and a separate multiple-ascending-dose study, known only from conference abstracts and a poster
Human research availableYes, limited. One peer-reviewed trial; developer-authored Phase 1 data known only from conference abstracts and a poster
Development statusEarly clinical. Two Phase 2 studies (REWIRE-1 and REWIRE-2) announced; no start dates or registry listing found as of November 18, 2026
Regulatory statusNot approved anywhere. No ClinicalTrials.gov listing found; one EU registry record (CTIS 2023-505967-36-00)

Technical identity

Technical propertyInformation
Primary nameATX-304 (laboratory code O-304, also written O304)
Alternative namesO-304, O304, ATX304 (misspelling), pan-AMPK activator O-304; the sodium salt is listed as “O-304 sodium”
Peptide sequenceNot applicable. ATX-304 is a small molecule with no amino acids
Amino-acid lengthNot applicable
Molecular formulaC16H11Cl2N3O2S (free form, PubChem)
Molecular weight380.2 g/mol (free form, PubChem). A separate sodium-salt record lists 403.2 g/mol; see the note below the table
CAS Registry Number1261289-04-6 (free form); 2639972-26-0 (sodium salt)
PubChem CID50923806 (free form); 156276456 (sodium salt)
UNIISPS2TLH4CM
DrugBank IDNot confirmed in this research pass
InChIKeyWEDWLYRQKUTOAX-UHFFFAOYSA-N (free form)
Chemical name4-chloro-N-[2-[(4-chlorophenyl)methyl]-3-oxo-1,2,4-thiadiazol-5-yl]benzamide
Chemical modificationsNot applicable (not a peptide)
Compound classNon-peptide AMPK activator
Primary biological targetAMPK (AMP-activated protein kinase), reported to work by protecting the active form from being switched off by the enzyme PP2C
Developer or originatorBetagenon AB (Umeå, Sweden); now Amplifier Therapeutics, a Cambrian Bio company
Development statusInvestigational, early clinical

Salt-form note. The 2018 trial used a suspension; an EU registry record for an ATX-304 Phase 1B lists “Na Salt Tablets”; products sold online are described as free-form powder, sodium-salt powder or capsules. The free form and the sodium salt are different weights per milligram. PubChem’s sodium-salt record shows 403.2 g/mol, but swapping one hydrogen for sodium on the free form calculates to about 402.2, so regard the salt weight as unresolved and rely on a batch certificate of analysis rather than a database figure. No study has compared exposure between the forms.

What is ATX-304?

ATX-304 is a synthetic small molecule: a chlorinated benzamide built around a ring containing sulfur and nitrogen. It is not a peptide, has no amino-acid sequence, and appears on peptide sites mainly because it is sold alongside them as a “research compound” and grouped with other so-called exercise mimetics. The compounds it is most often compared with are SLU-PP-332, another exercise-mimetic research compound, and BAM15, a mitochondrial uncoupler that raises energy use by a different route.

It started life as O-304 at Betagenon AB, a Swedish biotech in Umeå, which published the first mouse and human data in 2018. The developers wanted a drug that turns on AMPK, the enzyme that acts like a low-fuel gauge inside cells. When energy runs low, for example during exercise or fasting, AMPK switches on and tells the cell to take in sugar and burn fat. Many metabolic drugs and several lifestyle habits touch this system. O-304 was built to be “pan,” meaning it should work across the different tissue-specific forms of AMPK rather than only one.

The 2018 paper in JCI Insight (Steneberg and colleagues) covered mice and one human trial. In obese mice, O-304 prevented the rise in blood sugar and insulin on a high-fat diet, improved blood flow in small vessels, and raised energy use without lowering food intake. The same paper described a Phase IIa trial, named TELLUS, in 65 adults with type 2 diabetes already taking metformin. After 28 days at 1,000 mg a day, fasting glucose and blood pressure fell a little more than on placebo, and blood flow in the calf muscle improved. The authors were mostly employees or shareholders of Betagenon, which is worth keeping in mind.

Development then moved to a new owner. Amplifier Therapeutics was launched in 2023 under Cambrian Bio to take the compound forward, and it is now called ATX-304. An EU registry record lists a two-part Phase 1B of ATX-304 sodium-salt tablets with 72 participants planned, running from 2023 to May 2025. In June 2026, Cambrian presented a Phase 1b result at the American Diabetes Association meeting (abstract 1782-P): 23 adults with obesity and prediabetes were randomized 2 to 1 to receive 400 mg by mouth once daily or placebo for 8 weeks, with an optional open-label extension to 16 weeks. The abstract reports higher adiponectin, less liver and visceral fat on MRI, and a higher resting metabolic rate; a company release on the same study adds lower triglycerides and minimal weight loss. How the 23-person study relates to the 72-person registry record is not stated publicly.

Two Phase 2 studies, REWIRE-1 and REWIRE-2, have been announced, one aimed at higher exposures and muscle and lipid measures, the other at showing weight loss. No start dates have been given and neither is listed on ClinicalTrials.gov as of November 18, 2026.

In short, ATX-304 is rare among compounds on this site: it has a peer-reviewed human trial. It is also still very early, and almost all of what the internet says about how much to take is not based on that trial.

How does it work?

Plain-English explanation

Think of AMPK as a cell’s low-fuel warning light. When the light is on, the cell pulls in more sugar and burns more fat. Cells normally keep the light on by tagging AMPK with a small chemical marker, and another enzyme, PP2C, removes the marker to turn the light off again. ATX-304 is reported to shield the marker from PP2C, so the light stays on longer. It does not flip the switch itself. That is also why the developers describe the drug as an energy-system activator rather than an appetite drug: in mice it raised energy use while the animals ate as much or more.

Technical explanation

In the 2018 paper, O-304 suppressed PP2C-mediated dephosphorylation of the threonine-172 site on recombinant human AMPK trimers without inhibiting PP2C itself, protected phosphorylated AMPK even in the presence of excess ATP, and raised phosphorylated AMPK in cells without lowering cellular ATP. It acted additively with ADP and did not activate AMPK allosterically, so the authors conclude that it mimics ADP, not AMP, in its effect on AMPK. It is described as a pan-AMPK activator because it raised phosphorylated AMPK in cells carrying different subunit combinations.

Pharmacology as reported: orally available in rodents with a long plasma half-life, and it does not cross the blood-brain barrier. In people, the paper reports that steady-state plasma levels were not reached until about day 14, which matters for how long a trial must run before the full effect can show. The 2025 mouse abstract and the developer now describe ATX-304 as a combined AMPK and mitochondrial activator. A June 2026 developer abstract (1788-P) fills in part of that step: in isolated mouse muscle mitochondria the compound raised oxygen use and lowered membrane potential, consistent with increased proton leak, but in intact cells it did not lower membrane potential or ATP, and the authors say its AMPK effect is not a result of an energy drop. That is laboratory work reported by the developer, and the abstracts reviewed give no human mitochondrial measurements.

Potential benefits and research applications

Blood sugar and insulin resistance (small human trial, modest effect)

Being investigated for: lowering fasting glucose and insulin resistance in type 2 diabetes. How it might work: AMPK activation increases glucose uptake in muscle without insulin. Evidence: in TELLUS, in the subgroup whose starting fasting glucose was between about 7 and 13.3 mmol/L, fasting glucose fell by 0.60 mM on ATX-304 versus 0.10 mM on placebo, with nominal significance (about P = 0.01 between groups). Insulin resistance (HOMA-IR) also fell within the ATX-304 group. Strength: has been shown in a small trial, with caution. The subgroup analysis was post hoc because the study enrolled by HbA1c, not by fasting glucose. The authors themselves called the glucose effect “relatively modest” and said 28 days may be too short.

Blood vessel perfusion and blood pressure (human and animal, early)

Is being investigated for: better blood flow in small vessels and lower blood pressure. Evidence: in TELLUS, average blood pressure fell by 5.8 mmHg (top number) and 3.8 mmHg (bottom number) on ATX-304 versus small rises on placebo (1.2 and 0.9 mmHg), and MRI of the calf showed improved perfusion measures. In mice and dogs, blood flow and blood pressure moved in the same direction. Strength: may support, preliminary. The MRI curve-fitting was also post hoc, and the groups were small.

Resting energy use, liver fat and visceral fat (developer-authored abstract, 2026)

Is being investigated for: a higher resting metabolic rate and less fat around organs. Evidence: the June 2026 Phase 1b abstract (1782-P) reports that resting metabolic rate rose “up to 33%” from baseline and that adiponectin rose and liver and visceral fat fell on MRI over time in 23 adults. The company release on the same study gives the resting metabolic rate rise as 8% and adds lower triglycerides. Strength: has not been established beyond developer-authored conference abstracts and a poster. The November 2026 poster gives the group average as +8% at week 8 in 15 adults on ATX-304 against +3% in 8 on placebo, and its individual-participant lines appear to top out near +33%, so “up to 33%” most likely describes the best single response and not the average. A separate 4-week study on the same poster shows +6% at 400 mg once daily and +13% at 400 mg twice daily against -9% on placebo, with the twice-daily interval including zero (see the ObesityWeek subsection below). In June the two figures were not reconciled: the abstract does not say what “up to” refers to, such as a time point, a person or an average, and neither June source gave group sizes per arm or starting values. Weight loss was reported as minimal in the release and on the poster. A related energy-use topic with human data is covered in SANA and creatine thermogenesis.

Weight loss and body composition (mouse evidence only)

Is being investigated for: weight loss without eating less. Evidence: in the 2018 mouse work, animals on ATX-304-containing food stayed lighter than animals on a high-fat diet while eating at least as much, and this held at 30 °C where mice do not need to make extra body heat. A 2025 conference abstract in obese mice reported larger weight loss with ATX-304 plus semaglutide than with either alone, with the loss coming from fat rather than muscle. Strength: has not been shown in people. The 2026 human data report minimal weight loss at the doses tested, and human weight loss is the question REWIRE-2 is meant to answer. For context on what an approved weight-loss medicine achieves in people, see tirzepatide; for the fat-versus-muscle question behind the mouse result, see what happens to muscle on GLP-1 drugs.

Heart effects (animal evidence only)

Evidence: in mice and rats, ATX-304 activated AMPK in the heart, raised stroke volume by about 20% in mice, and did not increase heart or brain weight in rats given up to 600 mg/kg a day for 6 months. The authors report no ECG abnormalities or effect on left-ventricular mass. Strength: animal only, with the rat finding being a safety check rather than a benefit.

“Exercise in a pill” claims online

Has been reported anecdotally: endurance, fat loss and recomposition. Has not been established. No published study measured exercise capacity or body composition in people taking it, and claims of this kind are usually borrowed from the mouse work.

What dosage information circulates?

Figures in this section summarize amounts and schedules reported in published research or circulating online. Their inclusion documents what is reported and does not establish that a regimen is verified, safe, effective, or appropriate.

Reported use or research objectiveRoute reportedAmount reportedFrequency reportedReported durationEvidence or source category
Fasting glucose in type 2 diabetes on metformin (TELLUS, 65 adults)Suspension (route not spelled out in the text reviewed)1,000 mg per dayDaily total; doses per day not stated28 daysPublished human clinical trial
Resting metabolic rate, liver fat in adults with obesity and prediabetes (Phase 1b, 23 adults)Oral400 mgOnce daily8 weeks, then optional open-label extension to 16 weeksPublished human clinical trial (conference abstract only; authors are company employees)
Resting metabolic rate, triglycerides, adiponectin (multiple-ascending-dose Phase 1, 4 weeks; 12 on ATX-304, 6 on placebo)Not stated on the poster panels reviewed400 mg per dose in each arm (the twice-daily arm totals 800 mg a day by arithmetic only)Once daily in one arm; twice daily in the other4 weeksPatent or developer documentation (conference poster by company employees, not peer reviewed)
Diet-induced obesity in miceOral gavage100 mg per kg per dayDaily6 to 8 weeksPublished animal research
Diet-induced obesity in miceMixed into food0.4, 0.8 or 2 mg per gram of foodContinuous6 to 7 weeksPublished animal research
General metabolic use, “exercise mimetic”Oral capsule100 to 300 mg per day (about 200 mg called the sweet spot)1 to 3 capsules dailyNo cycling givenCommunity reported
Body recomposition, glucose, liverOral50 to 150 mg per day1 to 3 times daily8 to 12 weeks continuousCommunity reported
General metabolic use (capsule label range)Oral capsule with food200 to 400 mg per day1 or 2 times daily“8 to 12 week cycles,” no basis givenVendor reported
“Phase IIa dose” restated onlineOral500 mgTwice daily28 daysOrigin unclear (the paper reports 1,000 mg per day)

Explanation of the reported figures

The only figure with a real source in a peer-reviewed paper is 1,000 mg per day. It comes from the 2018 TELLUS trial, which ran 28 days. The paper gives the daily total and says it was given as a suspension; the text reviewed does not say how many separate doses made up the day or name the route. Websites that say “once daily” or “500 mg twice daily” are filling that gap on their own, and the two versions disagree with each other.

The second sourced figure is 400 mg once daily by mouth, from ADA abstract 1782-P (June 2026). It is a conference abstract written by company employees, not a peer-reviewed paper, and its text gives no pharmacokinetic numbers.

A third sourced set is two arms of a 4-week study, 400 mg once daily and 400 mg twice daily, from the developer’s ObesityWeek 2026 poster (November 14, 2026). The poster labels the arms 400 mg QD and 400 mg BID, so the twice-daily arm is 800 mg a day only by arithmetic. Social posts about the poster restated the arms as 400 mg and 800 mg per day. It is a poster by company employees, not a peer-reviewed paper, the panels reviewed do not state the route, and the study did not set out to establish a dose.

The figures people say to actually use are lower and have no trial behind them. The widely read ranges are 100 to 300 mg per day, 200 to 400 mg per day, and 50 to 150 mg per day. None is a trial arm. One of the most-read community ranges was written before the 400 mg Phase 1b results existed, so it cannot have come from that study, and its author says it was scaled from preclinical work and personal experience. Another page calls its 50 to 150 mg schedule community practice and offers three different run-in schedules on the same page. These are widely repeated online, but their original source could not be verified. Several pages agree on a 100 mg capsule strength, but agreement among pages that sell or promote capsules is not independent confirmation.

Where these sit against what was studied: every community range sits below the 1,000 mg per day that was actually given for 28 days. The 200 to 400 mg range tops out at the 400 mg Phase 1b arm. Nothing below 400 mg per day has a human study behind it that this entry could find, so there is no sourced human result for the 50 to 300 mg range. Time to steady state at lower daily amounts has not been published; at 1,000 mg per day it was about 14 days.

Disagreements. Route: everything is oral except one templated page that lists injection amounts and stacks for what is an oral small molecule. That page’s citations look like placeholders and its mechanism does not describe ATX-304, so it is regarded as describing the wrong compound. Food: some pages say empty stomach, others say with food, and one contradicts itself. Cycling: from “not needed” to 4 to 12 weeks with a taper. Form: the 2018 trial used a suspension, the registered later study lists sodium-salt tablets, and capsules online are described as free-form or salt powder. Whether a milligram of each form delivers the same exposure is unknown.

Weight-based figures. The mouse work used milligrams per kilogram and milligrams per gram of food. No study converted these to a human amount. One chemical supplier’s catalog gives its own conversion of a dietary dose into mg per kg; it is not supported by a study and this entry does not carry it forward.

What remains unknown. Which dose works best in people, whether anything below 400 mg per day has an effect, how the tablet compares with the suspension, how long the effect lasts, and whether the 2018 glucose result holds up in a larger trial.

Detailed dosage and protocol evidence

Amounts studied in human research

Research objectiveAmount studiedFrequencyDurationRouteStudy populationSource
Safety and fasting plasma glucose1,000 mg per day (single dose level)Daily total; doses per day not stated28 daysSuspension; route not spelled out65 adults with type 2 diabetes on metformin for at least 3 months, HbA1c 6.5% to 9.0%Steneberg et al., JCI Insight 2018 (TELLUS)
Resting metabolic rate, liver fat, visceral fat400 mgOnce daily8 weeks versus placebo, then optional open-label extension to 16 weeksOral23 adults with obesity and prediabetes, randomized 2 to 1 to ATX-304 or placeboADA abstract 1782-P, Diabetes 2026;75(Supplement_1); Cambrian Bio release, June 18, 2026
Resting metabolic rate, adiponectin, triglycerides (multiple-ascending-dose study)400 mg per dose (arms: 400 mg QD and 400 mg BID)Once daily; twice daily4 weeks, versus placeboNot stated on the poster panels reviewed12 on ATX-304 (6 per arm) and 6 on placebo; population not stated on the panels reviewedCambrian Bio, ObesityWeek 2026 Poster-111, November 14, 2026 (not peer reviewed)
Earlier Phase I safety studyNot reportedNot reportedNot reportedNot reportedNot reportedMentioned in Steneberg 2018 with no data
Phase 1B, two-part (registry record)Not publishedNot publishedRegistered window Sept 15, 2023 to May 7, 2025Sodium-salt tablets (registry)72 planned (48 Netherlands, 24 Germany)EU CTIS 2023-505967-36-00

ObesityWeek 2026: Dose Response, RMR Confidence Intervals and the Weight-Loss Gap

On November 14, 2026, at ObesityWeek, the developer presented Poster-111, “Positive Metabolic Effects of ATX-304, an AMPK and Mitochondrial Activator, from Two Phase 1 Studies.” It places the 23-adult translational study first shown in June next to a separate multiple-ascending-dose (MAD) Phase 1 study. That is one new presentation of two studies, not two new independent replications, and it is a poster, not a peer-reviewed paper. The developer’s November 12 announcement confirms the poster and its title; the figures below were read from the poster panels as printed. A technical research guide on the same poster walks through the dose-response charts in more detail.

Figures below summarize amounts and schedules reported in published research or circulating online. Listing a figure documents what is reported; it is not a recommendation and does not show a regimen is safe, effective, or appropriate for any person.

Three evidence categories, kept apart

  • ADA abstract 1782-P (June 2026): human data, 23 adults, the translational Phase 1b. A conference abstract written by company employees.
  • ADA abstract 1788-P (June 2026): a laboratory study in mouse mitochondria and cells. It contains no human data.
  • ObesityWeek Poster-111 (November 14, 2026): a new public presentation of dose and time-response charts that combines the translational study with the MAD study. Not peer reviewed.

MAD study, week 4: resting metabolic rate and triglycerides

The MAD study had 12 people on ATX-304 (6 on 400 mg once daily, 6 on 400 mg twice daily) and 6 on placebo, read out at week 4. The panels reviewed do not state the route or the population. The dose labels describe what each arm received in this experiment; they are not a protocol.

MeasureArm as labeled on the posterMean change from baseline95% confidence intervalHow to read it
Resting metabolic rate400 mg once daily (6 people)+6%0% to +13%Lower end of the interval touches zero
Resting metabolic rate400 mg twice daily (6 people)+13%-15% to +41%Interval includes zero by a wide margin
Resting metabolic ratePlacebo (6 people)-9%-19% to 0%Placebo mean fell
Triglycerides400 mg once daily-27%Not reportedLipids from this study were not assessed for statistical significance
Triglycerides400 mg twice daily-39%Not reportedSame footnote; the placebo bar on the same chart also falls below baseline

The poster states that the resting-metabolic-rate group averages are arithmetic means with 95% confidence intervals. Where it marks significance on the translational result, that comes from a prespecified ANCOVA on least-squares means that is not shown.

How to read the 13%

The 13% is the mean of six people in the twice-daily arm. Their individual lines on the poster run from below baseline to far above it, with the axis broken to fit the highest one, and the 95% interval runs from -15% to +41%, so it includes no change. It is not an established placebo-adjusted improvement, and it does not show that the higher daily amount works better than the lower one. The once-daily arm’s +6% interval starts at zero. With six people per group, one outlier moves a mean a long way, and the placebo group’s own fall of 9% adds noise to any raw comparison with it. The poster labels the arms 400 mg QD and 400 mg BID; the 800 mg per day figure used in some social posts is arithmetic, not a label.

Triglycerides and the cross-study trap

The poster reports triglycerides down 27% and 39% in the two MAD arms at week 4, with the footnote that these lipid results were not assessed for statistical significance. The placebo bar on the same chart also sits below baseline. The translational study had a different design, length and group size, so its numbers should not be lined up against the MAD arms as if they were one dose comparison.

Translational study and the open-label extension

MeasureGroupResultIntervalNotes
Resting metabolic rate, week 8400 mg once daily (15 people)+8%+1% to +14%Marked P below 0.01 versus baseline (prespecified ANCOVA, least-squares means not shown)
Resting metabolic rate, week 8Placebo (8 people)+3%-6% to +12%No significance mark
Triglycerides, week 16Open-label extension, ATX-304 (7 people)About -31%Not reportedPost hoc, versus baseline, no placebo group at this point
Adiponectin, week 16Open-label extension, ATX-304 (7 people)About +90% as drawnNot reportedPost hoc, versus baseline; the chart carries no numeric label

Week 16 is the end of the optional open-label extension, and the poster calls it a post-hoc analysis of participants treated with ATX-304 for 16 weeks. That is a small subset (7 people in the charts) compared with their own baselines, with no placebo group, so it is weaker evidence than the week 8 comparison. The poster calls the adiponectin rise dose-dependent, although the two MAD arms sit close together on the chart.

Where the earlier “up to 33%” fits

The June abstract said resting metabolic rate rose “up to 33%.” The poster’s individual lines for the translational study top out at about that level, while the group average is +8%. That points to the best single response and not the average, although the poster does not label it that way. The +8% is the group figure.

The weight-loss gap

The poster’s own summary says the metabolic effects occurred without meaningful weight loss and describes liver and visceral fat reductions “despite minimal weight loss at this tested dose.” That is the main limit on the evidence. Higher adiponectin, lower triglycerides and a higher resting metabolic rate are laboratory and imaging measures, not weight loss. The poster says future studies at higher exposures will test whether clinically relevant weight loss can be reached, which matches the planned REWIRE-1 and REWIRE-2 studies. The developer expects ATX-304 could complement appetite-suppressing medicines, but that is an expectation, and no combination has been tested in people.

Safety and what the poster does not show

The poster reports that treatment-emergent adverse events were mostly mild and occurred at similar frequencies on ATX-304 and placebo, and that the resting-metabolic-rate rise came without changes in core body temperature or 24-hour heart rate at these doses. Exposure was short: 4 weeks, 8 weeks, and an extension to 16 weeks in a few people. There are no long-term safety data and no human measure of muscle preservation. No product is approved, no combination with GLP-1 medicines, SANA, BAM15 or other compounds has been tested in people, and nothing here shows that research-use-only catalog material sold online matches the clinical ATX-304 material.

Laboratory background: why dose and mechanism questions keep coming up

The June laboratory abstract (1788-P) used isolated mouse mitochondria and cells at up to 20 µM and concluded that AMPK activation by ATX-304 is not secondary to a fall in ATP. That answers a question about how the drug acts in cells; it did not test amounts in people and does not replicate the clinical readouts above. A 2025 mouse study of fatty liver disease from Umea and Maastricht universities (Holm et al., JCI Insight) reported less body fat mass, lower blood cholesterol, and less liver fat and fibrosis, with marked differences between regions of the liver. Only its abstract was reviewed for this update.

What could not be verified

  • The full poster file: only the panels covering adiponectin, triglycerides, resting metabolic rate and the summary were reviewed. The liver fat, visceral fat and de novo lipogenesis panels (Figures 5 to 7) were not. The summary says lower hepatic de novo lipogenesis was seen in Phase 1b participants whose plasma levels were above 100 µg/mL.
  • The route, formulation, population and wider design of the MAD study beyond its 4-week readout.
  • How the MAD study relates to the 23-adult study and to the 72-person EU record, so the group sizes should not be added together.
  • Whether the MAD resting-metabolic-rate result was also analyzed against placebo; the panels show only changes from baseline.

Amounts studied in animal research

Research objectiveSpecies or modelAmount studiedFrequencyDurationRouteSource
Glucose, insulin, with or without metforminDiet-induced obese mice100 mg per kg per dayDaily8 weeksOral gavageSteneberg 2018
Glucose, amyloid, beta cellsObese mice carrying human islet amyloid polypeptide100 mg per kg per dayDaily6 weeksOral gavageSteneberg 2018
Hind-leg blood flowDiet-induced obese mice40 mg per kg per dayDaily8 weeksOral gavageSteneberg 2018
Dose response on glucose, AMPK, weightCBA mice on high-fat diet0.4, 0.8 and 2 mg per gram of foodContinuous6 to 7 weeksIn foodSteneberg 2018
Heart weight safetyRats100, 300 and 600 mg per kg per dayDaily6 monthsOral gavageSteneberg 2018
Weight, body composition, alone and with semaglutideDiet-induced obese miceNot stated in the abstractNot statedNot statedNot statedSchneider et al., J Endocr Soc 2025 (abstract)
ToxicologyRat and dogMentioned, not itemized in the text reviewedNot reportedNot reportedNot reportedSteneberg 2018

Practitioner and community-reported protocols

No Reddit, forum, telehealth or clinic reports with ATX-304 dosing could be retrieved. That does not show that none exist. Everything below comes from educator newsletters, product pages, supplier listings and aggregator pages. Sources are described by type and are not named, per site policy.

Reported objectiveAmount reportedFrequencyDurationRouteSource typeSource
General metabolic, AMPK activation100 to 300 mg per day (about 200 mg “sweet spot”; 100 mg capsules)1 to 3 capsules daily, morning, empty stomachNo cycling or titration givenOralEducator newsletter with a product linkAggregate pattern; not named
Stack with another exercise-mimetic compound200 to 400 mg per day (author’s “practical experience”)DailyCycling “not needed”OralEducator newsletterAggregate pattern; not named
Recomposition, glucose, liver, “healthspan”50 to 150 mg per day (50 mg once daily as a start; 100 as 50 + 50; 150 as 3 x 50)1 to 3 times daily8 to 12 weeks continuous; 12 weeks or more for liver goalsOralAnonymous protocol pageAggregate pattern; not named
General metabolic200 mg capsule; 1 or 2 per day (200 to 400 mg per day)1 or 2 times daily, with food“8 to 12 week cycles,” no basisOral capsuleSupplement-style product pageAggregate pattern; not named
Ingredient use50 to 200 mg per day for supplements; 100 to 400 mg per day for pharmaceutical R&DDivided dosesNot statedOralBulk-ingredient supplier pageAggregate pattern; not named
Trial restatement500 mgTwice daily28 daysOralCommercial guide pageAggregate pattern; attributed to the Phase IIa but the paper reports 1,000 mg per day
Single comment“Went up to 300 mg,” then lethargy and brain fogNot statedNot statedNot statedMember comment in a paid community; product named only as “ATX”Identity unconfirmed

Conflicts and unanswered questions

  • Number of doses per day in TELLUS is not stated in the text reviewed. “Once daily” and “500 mg twice daily” both circulate. Neither is confirmed.
  • Route in TELLUS is not named; everything about the compound points to a swallowed medicine, but the paper’s text reviewed says only “suspension.”
  • Dose conversion: nothing connects a mouse mg per kg figure to a human amount. Do not read the animal rows as human doses.
  • Salt versus free form: exposure per milligram is untested across the suspension, the registered tablet and capsules sold online.
  • Registry numbers in the 2018 paper (NCT00508287, NCT01167881) point to unrelated trials. The TELLUS record could not be located on ClinicalTrials.gov.
  • Relationship between studies: the poster presents a translational study and a MAD study, and the panels reviewed do not say how the MAD participants relate to the 23-adult study or the 72-person EU record.
  • No pharmacokinetic table was available for any human dose. The 2018 supplementary figures show exposure but could not be read for this entry.

Important published research

Steneberg et al., JCI Insight 2018 (PMID 29925691; PMC6124394). Mouse studies plus the TELLUS Phase IIa: 65 adults with type 2 diabetes on metformin, randomized, double-blind, placebo-controlled, 28 days, 1,000 mg per day. Findings: fasting glucose fell 0.60 mM on ATX-304 versus 0.10 mM on placebo in the post hoc subgroup with starting fasting glucose of about 7 to 13.3 mmol/L; systolic and diastolic blood pressure fell 5.8 and 3.8 mmHg versus small rises on placebo; calf MRI perfusion measures improved. Limitations the authors state: short and small, one dose level, the placebo group was heavier, the glucose drop was modest, steady state was not reached until about day 14, and the suspension formulation was to be replaced. Most authors are Betagenon employees, shareholders or consultants. In plain English: a real but small and early signal, from the developer’s own team.

ADA abstract 1782-P and Cambrian Bio release, June 2026. Phase 1b in 23 obese prediabetic adults, randomized 2 to 1 to 400 mg by mouth once daily or placebo for 8 weeks, with an optional open-label extension to 16 weeks. The abstract describes the compound as “safe and well tolerated” and reports, compared with baseline, a rise in total adiponectin, falls in liver and visceral fat on MRI in a time-dependent way, and a resting metabolic rate by indirect calorimetry that rose “up to 33%”. The company release on the same study gives resting metabolic rate up 8% (p < 0.01), lower liver fat by MRI (p < 0.05), lower visceral fat (p < 0.05), lower triglycerides (p < 0.01), higher adiponectin (p < 0.01), minimal weight loss, mostly mild side effects similar to placebo, and no rise in core temperature or 24-hour heart rate. Limitations: a conference abstract written by company employees, results stated against baseline, the 33% and 8% figures not reconciled, no per-arm group sizes or starting values in June (the November poster later gave 15 on ATX-304 and 8 on placebo), not peer reviewed. In plain English: encouraging for how the body burns energy, not yet evidence of weight loss.

Schneider et al., ADA abstract 1788-P, Diabetes 2026;75(Supplement_1). Laboratory study from the developer’s team using isolated mouse muscle mitochondria, mouse embryonic fibroblasts, HEK-293T cells and human hepatocytes, at up to 20 µM (the solubility limit). In isolated mitochondria, ATX-304 raised oxygen use and lowered membrane potential and reactive oxygen species, consistent with increased proton leak. In intact cells, membrane potential was not lowered, ATP in human hepatocytes was not reduced after 48 hours, and AMPK phosphorylation rose to a similar extent in normal cells and in cells carrying an AMP-insensitive AMPK mutant. The authors conclude that AMPK activation is not secondary to a change in ATP. Limitations: isolated-mitochondria and cell work only, a conference abstract written by company employees, not peer reviewed. In plain English: it supports the idea that the drug turns AMPK on directly instead of by draining cell energy, in cells in a dish and not in people.

Schneider et al., Journal of the Endocrine Society 2025 (ENDO abstract, PMC12545902). Diet-induced obese mice given ATX-304 alone, with semaglutide, and after semaglutide was stopped. Reported weight reductions of about 21% (alone) and 27% (combined), loss from fat rather than lean tissue, and less regain after semaglutide withdrawal. Limitations: mouse study, abstract only, company-affiliated authors, ATX-304 amount not stated. In plain English: an interesting idea that has not been tested in people.

EU CTIS record 2023-505967-36-00. Phase 1B of ATX-304 sodium-salt tablets, sponsor Betagenon AB, 72 planned participants, registered window September 15, 2023 to May 7, 2025. No results posted. Relationship to the 23-person study is not stated.

Cambrian Bio, ObesityWeek 2026 Poster-111, November 14, 2026. Two Phase 1 studies: a translational study (400 mg once daily for 8 weeks with an optional open-label extension; 15 on ATX-304, 8 on placebo) and a multiple-ascending-dose study (400 mg once or twice daily for 4 weeks; 12 on ATX-304, 6 on placebo). Resting metabolic rate rose by an average of 8% at week 8 in the translational study and by 6% and 13% in the two MAD arms at week 4, with the twice-daily interval running from -15% to +41%. Triglycerides fell 27% and 39% in the MAD arms (not tested for significance) and about 31% at week 16 in the extension, and adiponectin rose. Limitations: a poster, not peer reviewed; small groups; the extension was post hoc and open-label; weight loss was minimal. In plain English: a closer look at how the body responded, still no evidence of weight loss. Detail is in the ObesityWeek subsection above.

Holm et al., JCI Insight 2025;10(7). A mouse study of ATX-304 in progressive fatty liver disease from Umea University and Maastricht University, doi 10.1172/jci.insight.179990. The abstract reports reduced body fat mass, lower blood cholesterol and less liver fat and fibrosis, with marked local differences within the liver, alongside a shift toward fatty acid oxidation and less lipid synthesis. Limitations: mouse work, abstract-level review only for this update, and no human data.

Timing, duration and research variables

  • Duration: TELLUS ran 28 days; the Phase 1b ran 8 weeks with an optional extension to 16 weeks. The multiple-ascending-dose study on the November 2026 poster read out at week 4. No study has run longer in people.
  • Time to effect: the 2018 paper reports steady-state blood levels at about 14 days, and its authors argue that 28 days may be too short to see the full glucose effect.
  • Frequency and timing: the daily total is known for TELLUS; the number of doses and the relationship to meals are not. The November 2026 poster compared once-daily and twice-daily 400 mg arms, but only over 4 weeks with 6 people per arm. The community pages disagree on empty stomach versus food.
  • Continuous versus intermittent: all human research was continuous. The 8 to 12 week cycles and breaks suggested online have no study behind them.
  • Persistence of effect: not tested after stopping in people.
  • Variables: formulation (suspension, tablet, capsule), salt versus free form, background metformin, and body weight differences between groups in the 2018 trial could all change results.

Combinations and related research

Formally studied combinations. In people, only the background use of metformin in TELLUS (all patients were on it). In mice, ATX-304 with metformin gave a larger HOMA-IR fall than ATX-304 alone, and a 2025 abstract tested it with semaglutide.

Mechanistically plausible. Pairing with an incretin medicine to preserve lean mass or reduce regain is the idea in the 2025 mouse abstract. No human combination study has been reported.

Practitioner-reported. None identified.

Community-reported. Stacks with SLU-PP-332, with MOTS-c and with metformin appear on educator and protocol pages. Popularity here is not evidence of safety or benefit, and the combinations have not been studied. See SLU-PP-332 and MOTS-c for what is known about those compounds individually.

Side effects, risks and limitations

Human studies. The 2018 paper describes ATX-304 as well tolerated and reports no ECG abnormalities or effect on left-ventricular mass, but the text reviewed gives no adverse-event counts, discontinuations or laboratory safety tables. The June 2026 abstract (1782-P) calls the compound “safe and well tolerated” without adverse-event counts. The company release says side effects were mostly mild and similar to placebo and that core body temperature and 24-hour heart rate did not rise, again with no numbers. The November 2026 poster says treatment-emergent adverse events were mostly mild and occurred at similar frequencies on ATX-304 and placebo, still without counts in the panels reviewed and after short exposure.

Animal findings. Rats given up to 600 mg per kg per day for 6 months showed no increase in heart or brain weight. In mice, heart glycogen fell with dose. Acute blood-pressure lowering was seen in dogs.

Mechanism-based concerns. A compound that raises energy use could raise heart rate or temperature at higher exposures; REWIRE-1 plans to test higher exposures, so that safety question is open. AMPK has many tissue roles and long-term effects of chronic activation are not established.

Interactions. The authors considered an overlap with metformin’s own AMPK effect and judged it less likely to explain their results. No formal interaction studies are published.

Unknown. Long-term safety, safety in people without diabetes beyond 16 weeks, pregnancy, and use with other drugs.

Anecdotal. One community comment links a higher amount to lethargy and brain fog; the post never names the compound in full, so it cannot be attributed to ATX-304.

Quality problems with products sold online. Material marketed as ATX-304 or O-304 is unregulated research-chemical stock. Listings advertise purity claims and certificates of analysis of varying completeness; one listing’s certificate was both “included” and “coming soon.” The same compound is sold as free form and as sodium salt, one supplier page describes the compound as something unrelated to AMPK, and nothing here shows that the identity, purity, concentration or sterility of any listing matches its label. The guide to reading a peptide COA explains what a certificate does and does not show.

Regulatory and developmental status

  • Approval: none anywhere. Not approved by the US FDA, the European Medicines Agency or any other regulator identified.
  • Phase: early clinical. A Phase IIa (2018), a Phase 1b (reported June 2026) and a multiple-ascending-dose Phase 1 (poster, November 2026) have been completed; Phase 2 studies named REWIRE-1 and REWIRE-2 have been announced by the developer without start dates.
  • Registries: no ClinicalTrials.gov record found for ATX-304, O-304 or the sponsors on November 8 and again on November 18, 2026. One EU CTIS record (2023-505967-36-00, Betagenon AB, Phase 1B). The two NCT numbers printed in the 2018 paper point to unrelated trials.
  • Owner: Amplifier Therapeutics, a Cambrian Bio company; originator Betagenon AB.
  • Sports: the World Anti-Doping Agency’s prohibited list (category S4, hormone and metabolic modulators) names activators of AMPK as a class, with AICAR as an example. ATX-304 is not named individually, and this entry did not confirm how WADA handles it. Check the current list before assuming anything.
  • Difference by country: none identified; it is investigational everywhere.
  • Where to watch: the clinical trial tracker lists registered studies for compounds on this site, and how this site grades evidence explains the labels used on this page.

Frequently asked questions

What is ATX-304?

An experimental small-molecule drug, also called O-304, that is designed to switch on the cell energy enzyme AMPK. It is not a peptide.

Is ATX-304 a peptide?

No. It has no amino-acid sequence. It is a small organic molecule sold alongside peptides as a research compound. O-304 is the same compound under its original code.

What dosage has been studied in people?

1,000 mg per day for 28 days in a 2018 trial of 65 adults with type 2 diabetes, and 400 mg once daily by mouth for 8 weeks, with an optional extension to 16 weeks, in a 2026 developer-authored conference abstract of 23 adults. A November 2026 developer poster added 400 mg once daily and 400 mg twice daily arms for 4 weeks in a separate study of 18 adults.

What dosage commonly circulates online?

Ranges of 50 to 150, 100 to 300 and 200 to 400 mg per day are the most repeated. None matches a published trial arm, and the sources do not trace them to one.

How often is it reported being taken?

Usually once to three times a day by mouth. The TELLUS paper gives only the daily total, so the schedules online are not confirmed. The November 2026 poster compared once-daily and twice-daily 400 mg arms for 4 weeks.

How long do reported protocols last?

Community pages say 8 to 12 weeks, sometimes longer. The studied durations were 28 days, 4 weeks (the multiple-ascending-dose arms) and 8 to 16 weeks. No study supports breaks or cycles.

Has it been studied in humans?

Yes, in one peer-reviewed Phase IIa, plus a Phase 1b and a multiple-ascending-dose study known only from developer-authored abstracts and a poster. All are small. In the November 2026 poster, the twice-daily arm’s 13% rise in resting metabolic rate had an interval running from -15% to +41%, so a better result at the higher daily amount was not shown.

What side effects have been reported?

The trials describe it as well tolerated, and the November 2026 poster says adverse events were mostly mild and similar to placebo, with no itemized adverse-event data in the sources reviewed. Long-term safety is unknown.

Is ATX-304 approved?

No. It is investigational, with Phase 2 studies announced but no start dates or registry listing found.

What remains unknown?

The best dose, whether low doses work, whether twice-daily dosing does better than once daily, how tablets compare with the suspension, long-term safety, and whether it causes weight loss in people.

Bottom line

ATX-304 is interesting because it is one of the few “exercise mimetic” compounds with a published human trial. The strongest evidence is a 28-day Phase IIa from 2018 in 65 adults with type 2 diabetes at 1,000 mg a day, which found a modest fall in fasting glucose and blood pressure, in a post hoc analysis, run by the developer’s own team. A second human dataset, 23 adults at 400 mg a day by mouth for 8 weeks, came from a developer-authored conference abstract in June 2026 and reports a higher resting metabolic rate (up to 33% from baseline in the abstract, 8% as the group average in the company release and on the November poster) and less liver fat, with minimal weight loss. A November 2026 poster added a 4-week multiple-ascending-dose study in which resting metabolic rate rose 6% at 400 mg once daily and 13% at 400 mg twice daily against a 9% fall on placebo, but the twice-daily interval ran from -15% to +41%, so it does not show that more is better. That is the whole sourced dosing record: 1,000 mg per day in 2018, 400 mg once daily for 8 weeks, and 400 mg once or twice daily for 4 weeks in the 2026 developer-presented studies.

The much lower figures that circulate online, 50 to 300 mg a day, are not trial arms and could not be traced to a source, and the schedules and cycles attached to them have no study behind them. Weight-loss claims remain speculative until REWIRE-2 reports. ATX-304 is investigational and not approved anywhere, no regulator has set a dose, and nothing here establishes that any amount is safe or effective for anyone outside a supervised trial. For a plain-language overview of the energy-expenditure research without the dosing detail, see the explainer on The Weight Loss Atlas. To compare this entry with others, use the comparison hub.

What circulates E5

Figures reported online. Not dosing, not verified, not endorsed.
Reported useRouteAmount reportedFrequencyReported length
General metabolic use, AMPK activationOral capsule100 to 300 mg per day1 to 3 capsules dailyNo cycling given
Stack with another exercise mimeticOral200 to 400 mg per dayDailyCycling said to be unneeded
Recomposition, glucose, liverOral50 to 150 mg per day1 to 3 times daily8 to 12 weeks continuous
General metabolic use (capsule label range)Oral capsule with food200 to 400 mg per day1 or 2 times daily8 to 12 week cycles, no basis given
Restated as the Phase IIa dose (origin unclear)Oral500 mgTwice daily28 days

Widely repeated online, but its original source could not be verified: oral ranges of 50 to 150, 100 to 300 and 200 to 400 mg per day, usually split into one to three doses and run for 8 to 12 weeks. None of these is a trial arm. The only amounts given to people are 1,000 mg per day for 28 days (2018, peer reviewed) and 400 mg once daily by mouth for 8 weeks (2026, conference abstract by the developer), plus 400 mg once and twice daily for 4 weeks (November 2026, developer poster). Pages that quote 500 mg twice daily attribute it to the Phase IIa, but the paper reports 1,000 mg per day.

Recorded as an observation about what is published elsewhere. No figure here is a dose, a protocol, or a recommendation, and nothing in this section is evidence that any amount is safe or effective.

Sources

  1. Steneberg P, Lindahl E, Dahl U, et al. PAN-AMPK activator O304 improves glucose homeostasis and microvascular perfusion in mice and type 2 diabetes patients. JCI Insight. 2018;3(12):e99114. Checked against the full text on PubMed Central (PMC6124394). Source for the compound identity and mechanism, the mouse and rat amounts, and the TELLUS Phase IIa trial in type 2 diabetes (65 adults, 1,000 mg per day for 28 days, suspension). Most authors are Betagenon employees, shareholders or consultants. PMID 29925691; DOI 10.1172/jci.insight.99114
  2. Cambrian Bio. Cambrian Bio presents positive human translational data for ATX-304, the first AMPK network activator, at the American Diabetes Association 86th Scientific Sessions. Press release, June 18, 2026. Sponsor-reported findings from a conference presentation; group sizes per arm, starting values and the comparison for the resting metabolic rate change are not given. Cambrian Bio, June 18, 2026
  3. Thieroff-Ekerdt RI, Jacobs D, Hall J, Schneider EJ, Ainscough J, Peyer J, Heise T. 1782-P: Phase 1b study of AMPK/mitochondrial activator ATX-304 in prediabetic obese participants. American Diabetes Association 86th Scientific Sessions. Diabetes 2026;75(Supplement_1):1782-P. doi:10.2337/db26-1782-P. Conference abstract, not a peer-reviewed paper; several authors are employees of Cambrian BioPharma or Amplifier Therapeutics AB. Diabetes 2026;75(Suppl 1):1782-P
  4. Schneider EJ, Thieroff-Ekerdt RI, Peyer J, Gudiksen A, Pilegaard H, Ling NXY, Oakhill JS, Hall J. 1788-P: AMPK activation by ATX-304 is not secondary to changes in ATP levels. American Diabetes Association 86th Scientific Sessions. Diabetes 2026;75(Supplement_1):1788-P. doi:10.2337/db26-1788-P. Conference abstract, not a peer-reviewed paper. Laboratory work; several authors are employees of Cambrian BioPharma or Amplifier Therapeutics AB. Diabetes 2026;75(Suppl 1):1788-P
  5. Schneider EJ, et al. Weight loss and change in body composition in a DIO mouse model by the combined AMPK and mitochondrial activator ATX-304, alone, with semaglutide, and after semaglutide withdrawal. Journal of the Endocrine Society. 2025;9(Suppl 1):bvaf149.080. ENDO conference abstract. Mouse study with company-affiliated authors; the ATX-304 amount is not stated in the abstract. PMC12545902; J Endocr Soc 2025;9(Suppl 1):bvaf149.080
  6. European Medicines Agency. Clinical Trials Information System (CTIS) public register, trial 2023-505967-36-00: A Translational Phase 1B, 2-Part Study to Evaluate Pharmacokinetics, Safety and Pharmacodynamics of Treatment with ATX-304 Na Salt Tablets. Sponsor Betagenon AB; 72 planned participants (48 in the Netherlands, 24 in Germany); 15 September 2023 to 7 May 2025; status ended; summary of results deferred until November 2027. Registry listing with no results. Checked November 8, 2026 through the register search and the CTIS.eu mirror. EU CT number 2023-505967-36-00
  7. National Library of Medicine. PubChem compound records for O-304 (CID 50923806, free form) and O-304 sodium (CID 156276456). Database record used for the formula, molecular weight, CAS numbers and other identifiers in the identity table. The sodium-salt weight listed there (403.2 g/mol) differs from the roughly 402.2 calculated from the free form. Accessed November 8, 2026. PubChem CID 50923806; CID 156276456
  8. Cambrian Bio. Pipeline. Developer web page listing ATX-304 and describing planned Phase 2 studies REWIRE-1 and REWIRE-2. Checked November 8, 2026. Developer documentation, not an independent registry. Cambrian Bio pipeline page
  9. Cambrian Bio. Amplifier Therapeutics, a Cambrian Bio pipeline company, doses first patients in Phase 1B clinical trial and closes new financing from Future Ventures and RA Capital Management. Press release, October 17, 2023. Developer announcement of the start of the Phase 1B trial of ATX-304 in the European Union. Cambrian Bio, October 17, 2023
  10. Springer Nature AdisInsight. ATX 304 drug profile (alternative names ATX-304, O 304). The visible summary lists a Phase I obesity trial in the Netherlands and Germany under EU registry number 2023-505967-36-00 as completed on May 7, 2025. Full profile is paywalled; this site did not open the registry record itself. AdisInsight drug 800035145; EU registry number 2023-505967-36-00
  11. BioSpace. Amplifier Launches to Bring First AMPK Activator for Obesity to the Clinic. 2023. Source for Amplifier Therapeutics' formation, the reported Betagenon transaction, and the company's stated plan for a Phase Ib clinical trial; treated as company-announced press coverage, not an independently confirmed trial record. n/a
  12. German Sport University Cologne, Department of Biochemistry. S4 Hormones and metabolic modulators. Secondary summary of the World Anti-Doping Agency Prohibited List, used for the statement that activators of AMPK are named as a class. ATX-304 is not named individually. Accessed November 8, 2026. WADA Prohibited List, category S4
  13. Cambrian Bio. Cambrian Bio to present metabolic effects of ATX-304 at ObesityWeek 2026. Press release, November 12, 2026. Announces a poster of data from two Phase 1 studies at ObesityWeek, November 14 to 17, 2026. The poster itself (Poster-111, "Positive Metabolic Effects of ATX-304, an AMPK and Mitochondrial Activator, from Two Phase 1 Studies") is a conference presentation by the developer; the figures cited on this page were read from the poster panels as shown, and no full text or peer-reviewed paper has been published. Cambrian Bio, November 12, 2026
  14. Holm E, Vermeulen I, Parween S, Lopez-Perez A, Cillero-Pastor B, Vandenbosch M, Remeseiro S, Hornblad A. AMPK activator ATX-304 reduces oxidative stress and improves MASLD via metabolic switching. JCI Insight 2025;10(7). Mouse study from Umea University and Maastricht University. Abstract reviewed through PubMed; the full text could not be opened for this update. doi 10.1172/jci.insight.179990; PMID 40197369; PMC11981618

Live registry status

These are the registered trials this page cites, pulled from ClinicalTrials.gov and refreshed automatically. Status and dates are what the registry currently shows. Listing a trial here does not mean it produced a positive result.

Open these in the trial tracker

Research area: Metabolic and incretin.

Share thisFacebookXEmail