Abstract

   The emergence of glucagon-like peptide-1 agonists represents a notable
   advancement in the pharmacological treatment of obesity, yet
   complementary approaches are essential. Through phenotypic drug
   discovery, we developed promising nitroalkene-containing small
   molecules for obesity-related metabolic dysfunctions. Here, we present
   SANA, a nitroalkene derivative of salicylate, demonstrating notable
   efficacy in preclinical models of diet-induced obesity. SANA reduces
   liver steatosis and insulin resistance by enhancing mitochondrial
   respiration and increasing creatine-dependent energy expenditure in
   adipose tissue, functioning effectively in thermoneutral conditions and
   independently of uncoupling protein 1 and AMPK activity. Finally, we
   conducted a randomized, double-blind, placebo-controlled phase 1A/B
   clinical trial, which consisted of two parts, each with four arms: (A)
   single ascending doses (200–800 mg) in healthy lean volunteers; (B)
   multiple ascending doses (200–400 mg per day for 15 days) in healthy
   volunteers with overweight or obesity. The primary endpoint assessed
   safety and tolerability. Secondary and exploratory endpoints included
   pharmacokinetics, tolerability, body weight and metabolic markers. SANA
   shows good safety and tolerability, and demonstrates beneficial effects
   on body weight and glucose management within 2 weeks of treatment.
   Overall, SANA appears to be a first-in-class activator of
   creatine-dependent energy expenditure and thermogenesis, highlighting
   its potential as a therapeutic candidate for ‘diabesity’. Australian
   New Zealand Clinical Trials Registry registration:
   [159]ACTRN12622001519741.

   Subject terms: Pharmacology, Metabolic disorders, Obesity, Metabolic
   syndrome, Molecular medicine
     __________________________________________________________________

   In this study, the authors describe SANA, a nitroalkene derivative of
   salicylate, as a potential activator of creatine-dependent energy
   expenditure and thermogenesis in adipose tissue. Preclinical and
   clinical data from this paper also suggest that SANA improves glucose
   homeostasis and promotes weight loss in mice and humans.

Main

   Obesity and its associated comorbidities, including type II diabetes
   and cardiovascular disease^[160]1, are a global epidemic, making them
   substantial drivers of healthcare burdens worldwide. Although obesity
   is preventable through the adoption of healthy nutrition and lifestyle
   changes, once the obesity phenotype is established, it becomes
   difficult to reverse through behavioural changes alone. The recent
   emergence of glucagon-like peptide-1 (GLP-1) agonists, such as
   semaglutide, has shown that pharmacological treatment for obesity and
   its related comorbidities is now a realistic possibility^[161]2.
   However, alternative or complementary approaches to GLP-1-based
   therapies may help improve long-term outcomes and minimize side
   effects.

   Among the promising molecules for treating metabolic diseases are
   nitrated unsaturated fatty acids^[162]3. These compounds have
   demonstrated a range of beneficial properties, including positive
   metabolic effects in preclinical models of obesity^[163]3–[164]6, due
   to their pleiotropic effects. The main pharmacological action of nitro
   fatty acids stems from their reactive nitroalkene group, which affects
   several signalling pathways^[165]3. Recently, we demonstrated that this
   nitroalkene group can be transferred to other molecular scaffolds with
   known pharmacological effects that differ structurally from fatty
   acids^[166]7–[167]9. By repositioning the nitroalkene group onto these
   scaffolds, we established three key findings: (1) The beneficial
   signalling properties of the nitroalkene group are preserved in the
   designed molecules; (2) the scaffold can be used to enhance the
   pharmacological properties of the nitroalkene group; (3) in some cases,
   emergent properties arise that could not have been
   predicted^[168]7–[169]10.

   Building on these concepts, we developed several nitroalkene-containing
   small molecules. Using phenotypic drug discovery, we screened these
   molecules for activity in obesity and related diseases. In the founding
   generation of molecules, based on a vitamin E scaffold designed for the
   treatment of atherosclerosis and obesity-induced inflammation, we
   observed improvements in systemic inflammation, glucose intolerance and
   atherosclerosis during obesity^[170]8–[171]10.

   In this work, we show an unprecedented metabolic effect of a
   nitroalkene-based small molecule, SANA (5-(2-nitroethenyl)salicylic
   acid), using the ancient drug salicylate as a scaffold. Salicylate, the
   precursor of aspirin, was used for more than a century to diminish
   inflammation and relieve pain. It inhibits nuclear factor-kappa B
   signalling and inflammation^[172]11. When administered at elevated
   doses (2–5 g per kg per day), it activates AMPK and improves metabolic
   dysfunction in obese mice, including modest effects in weight
   gain^[173]12–[174]15. Indeed, salicylate was tested in clinical trials
   for diabetes, revealing improvement in basal glycaemia, glycosylated
   haemoglobin and C-reactive protein^[175]16–[176]18.

   Our data demonstrate that SANA promotes weight loss and protects
   against diet-induced obesity (DIO), insulin resistance,
   hypertriglyceridaemia and liver steatosis in rodent preclinical models.
   Notably, these effects are markedly greater than those previously
   described for salicylate^[177]12–[178]14. Mechanistically, we found
   that SANA stimulates non-shivering thermogenesis in both brown and
   white adipose tissue, an unexpected property of this molecule. The
   effect of SANA was independent of uncoupling protein 1 (UCP1) and AMPK
   activation and was lost after creatine depletion. Furthermore, Ckmt1
   (encoding ubiquitous mitochondrial creatine kinase) knockout (KO) mice
   demonstrated impaired thermoregulation after SANA treatment, which was
   rescued by thermoneutral housing. Importantly, the safety,
   tolerability, pharmacokinetics and pharmacodynamics of SANA (referred
   to as MVD1 for clinical development) following both single-dose and
   multiple-dose administrations were evaluated in a phase 1A/B,
   first-in-human, randomized, double-blind, placebo-controlled clinical
   trial. Up to 400 mg a day, SANA demonstrated an overall favourable
   safety and tolerability profile. In individuals treated with the
   highest dose of SANA for 15 days, we observed an improvement in
   homeostatic model assessment for insulin resistance (HOMA-IR) and
   weight loss, coupled with a favorable safety profile. Taken together,
   we propose that SANA is not just an improved version of salicylate but
   a molecule with unexpected properties, such as its effects in energy
   homeostasis, and appears to be potentially a viable therapeutic
   strategy for treating obesity and metabolic syndrome.

Results

SANA prevents DIO

   The salicylate-based nitroalkene, SANA, was obtained by a one-step
   synthetic route with ~93% yield (Extended Data Fig. [179]1a). SANA
   showed similar biochemical (that is, Michael addition reactions with
   low-molecular-weight thiol-containing compounds (Extended Data Fig.
   [180]1a) and signalling properties to other nitroalkene compounds in
   vitro and in vivo (Extended Data Fig. [181]1b–j). Pharmacokinetic
   analysis in mice showed maximum plasma concentration (Tmax) of SANA
   0.5 h after oral administration (p.o., 400 mg per kg body weight), with
   a maximum observed plasma concentration (Cmax) of 4850 ng ml^−1. SANA
   was no longer detectable in plasma after 24 h (Extended Data Fig.
   [182]2a). Chronic administration of SANA (400 mg per kg per day) to
   normal mice for 8 weeks showed no signs of renal or liver toxicity
   (Extended Data Fig. [183]2b–f). Biodistribution analysis performed by
   administration of ^11C-SANA followed by positron emission tomography
   and a magnetic resonance imaging (MRI) scan revealed that SANA followed
   an enterohepatic circuit, being excreted mainly through the kidneys and
   intestine (Extended Data Fig. [184]2g–i). Importantly, SANA was also
   detectable in several tissues including inguinal white adipose tissue
   (iWAT; Extended Data Fig. [185]2j).

Extended Data Fig. 1. Physicochemical and biological characterization of the
canonical nitroalkene effects of SANA.

   [186]Extended Data Fig. 1
   [187]Open in a new tab

   (a) Synthesis (left) and electrophilic properties (right) of SANA.
   Spectra of SANA (10 μM) recorded every 60 s after incubation with
   β-Mercaptoethanol (BME, 100 μM) or reduced glutathione (GSH, 100 μM).
   (b) Left: cytotoxicity of SANA 24hs treatment in THP-1 macrophages,
   measured by MTT assay. n = 13 per group Right: dose-response curve
   fitting yielded a IC50 = 0.17 mM, using an absorbance at 570 nm. n = 6
   per group. (c-d) Effect of SANA on NF-κB/p65 subunit translocation into
   nuclei of THP-1 macrophages. (c) Quantification of nuclear fluorescence
   intensity (NFI) relative to total fluorescence intensity (TFI). n:
   Control (DMSO) = 26; LPS-DMSO = 33; LPS-Salycilate=56; LPS-SANA = 55.
   (d) Representative pictures. Cells were treated for 2 h before
   activation with LPS (1 μg/mL) for 30 minutes. Control: cells without
   LPS treatment. Bar = 10 μm. (e-g) SANA inhibition of neutrophil
   recruitment in Zebrafish. (e) Acute inflammation model assay. Larvae at
   3 dpf were pre-treated with DMSO, Ibuprofen, Salycilate or SANA for
   2 h. Tail fins were wounded with a scalpel, larvae were maintained with
   the drugs and neutrophils in the injury were imaged and counted 4 h
   post transection. (f) Representative pictures. Dotted rectangles
   indicate the wounded region (ROI) where neutrophils were counted. (g)
   Quantification of recruited neutrophils after treatment. n: Control
   (DMSO) = 35; Ibuprofen 20 µM = 17; Salicylate 200 µM = 19; SANA
   200 µM = 20. (h) LPS-induced inflammatory model. C57BL/6 J mice were
   injected IP with SANA (100 mg/kg), Salicylate (100 mg/kg) or vehicle,
   1 h before IP injection of LPS (10 mg/kg) or vehicle. (i-j) ELISA
   quantification of IL-1β in plasma and peritoneal lavage. n: Control=9;
   LPS = 8; SANA = 9; LPS-SANA = 9; Salicylate=8; LPS-Salicylate=9. Data:
   mean ± SEM. Statistical analyses: one-way ANOVA followed by Bonferroni
   post hoc for multiple comparisons (c, i, j). Kruskal-Wallis test (b).

   [188]Source data

Extended Data Fig. 2. Pharmacokinetics, toxicity, and tissue distribution of
SANA.

   [189]Extended Data Fig. 2
   [190]Open in a new tab

   (a) Pharmacokinetics of SANA in mice. Mice were given a single dose of
   SANA (400 mg/kg) and plasma concentration was measured by LC-MS/MS over
   time. n = 3. (b-f) Toxicity effects of SANA. Mice were fed ND and
   administered with SANA (400 mg/kg/day, PO) for 8 weeks. (b) Liver and
   kidney macroscopic appearances and (c) Hematoxylin & Eosin liver and
   kidney staining. (d) ALT and AST in plasma/serum from mice at week 8.
   n = 11 per group. (e-f) Panel of liver (E) and renal (F) function
   markers (TP, ALB, GLO, TBIL and BUN). n = 12 per group. (g-j)
   Biodistribution analysis performed by 11C-SANA (9 MBq/animal) followed
   by PET-MRI scan. PET-scanning was performed the first 60 minutes after
   injection. (g) Representative picture of mice injected with 11C-SANA
   and followed by PET-MRI scans over time that were used to perform the
   analysis. (h) bladder and stomach, (i) kidney and liver, and j) iWAT,
   brain, and muscle. (h-j) n = 3. (k) ANCOVA analysis of ambulatory
   activity of mice treated with HFD or HFD + SANA (400 mg/kg/day, PO) at
   week 8. n = 8 per group. Data: mean ± SEM. Statistical analyses:
   unpaired two-sided Student’s t-test (d, e, f). ANCOVA (k).

   [191]Source data

   When tested in DIO models, SANA showed unexpected properties for a
   nitroalkene-based small molecule. Indeed, SANA dramatically protected
   mice from DIO (Fig. [192]1a–f), despite not affecting normal growth
   (Fig. [193]1c), total food consumption (Fig. [194]1f), stool mass and
   calories (Fig. [195]1g) and digested energy (Fig. [196]1h) and activity
   (Extended Data Fig. [197]2k). Importantly, SANA showed a maximum effect
   on obesity at doses where the scaffold salicylate was completely
   ineffective (Fig. [198]1d,e). Protection against DIO by SANA was due to
   decreased fat accumulation in all adipose depots analysed (Fig.
   [199]1i–k), parallelled by an increase in the percentage of lean mass
   in the SANA-treated mice (Fig. [200]1l). SANA showed no effect on body
   weight on mice fed with normal chow (Fig. [201]1m,n).

Fig. 1. SANA protects against DIO.

   [202]Fig. 1
   [203]Open in a new tab

   a, Mice fed HFD, HFD + salicylate (SAL) or HFD + SANA (400 mg per kg
   per day, mixed in HFD, p.o.) for 8 weeks. b, Weight gain. n: HFD = 30;
   HFD + SAL = 29; HFD + SANA = 30; pooled from three experiments. c,
   X-ray image (left) and tibia length (right). n = 9 per group. d,e,
   Weight gain under normal diet (ND), HFD or HFD + SANA (d) or HFD + SAL
   (e) at different doses administered p.o. n: HFD = 8; HFD + SANA 200 mg
   per kg per day = 6; HFD + SANA 300 mg per kg per day = 8; HFD + SANA
   400 mg per kg per day = 10; ND = 8 in d, and HFD = 11; HFD + SAL 200 mg
   per kg per day = 5; HFD + SAL 300 mg per kg per day = 9; HFD + SAL
   400 mg per kg per day = 10; ND = 8 in e. f, Cumulative food intake.
   n = 4 per group. g, Stool production (left) and caloric content (right)
   at week 8. n: HFD = 9; HFD + SAL = 11; HFD + SANA = 11 and HFD = 7;
   HFD + SAL = 6; HFD + SANA = 6, respectively. h, Digested energy
   calculated as the percentage of gross energy intake at week 8. n = 6
   per group. i, Representative images; arrows indicate perigonadal fat.
   j, Total fat mass by EchoMRI. n = 8 per group. k, Fat depot
   quantification: interscapular brown (iScap_BAT), inguinal (i_WAT),
   subscapular (sScap_WAT) and perigonadal (p_WAT) adipose tissue. n:
   iScap_BAT: HFD = 10; HFD + SAL = 14; HFD + SANA = 11; i_WAT: HFD = 10;
   HFD + SAL = 14; HFD + SANA = 12; sScap_WAT: HFD = 10; HFD + SAL = 14;
   HFD + SANA = 12 and p_WAT: HFD = 10; HFD + SAL = 13; HFD + SANA = 10.
   l, Total lean mass by EchoMRI. n = 8 per group. m, Mice fed ND or
   ND + SANA (400 mg per kg per day, p.o.) for 8 weeks. n, Weight at start
   and end of treatment n: ND = 7; ND + SANA = 8. Data are the
   mean ± s.e.m. Statistical tests used were two-way analysis of variance
   (ANOVA) followed by Tukey’s multiple-comparisons test (b, d, e, f and
   n), repeated-measures one-way ANOVA (c) and one-way ANOVA (g, h, j, k
   and l). One-way ANOVA was followed by Bonferroni post hoc for multiple
   comparisons.

   [204]Source data

SANA protects mice from DIO-associated metabolic damage

   While assessing the metabolic consequences of DIO, we found that the
   livers of mice fed a high-fat diet (HFD) and treated with SANA had
   normal macroscopic appearance, indistinguishable from the livers of
   lean, age-matched control mice (Fig. [205]2a). Histological analysis of
   livers confirmed these findings, revealing that mice treated with SANA
   showed no signs of liver steatosis (Fig. [206]2b). Examination of
   livers revealed SANA treatment abrogated DIO-mediated elevation in
   total liver weight induced by a HFD (Fig. [207]2c). Analysis of liver
   transaminases in plasma showed complete protection from obesity-related
   liver damage (Fig. [208]2d). Assessment of glucose management after DIO
   revealed that SANA protected mice from elevated blood glucose levels
   (Fig. [209]2e). Furthermore, SANA improved impaired glucose tolerance
   and made it indistinguishable from lean age-matched controls and
   significantly better than obese mice treated with salicylate
   (HFD + SAL; Fig. [210]2f,g). Obese mice treated with SANA showed lower
   insulin levels (Fig. [211]2h) and a decrease in glucose-induced insulin
   secretion (Fig. [212]2i). This was reflected in increased insulin
   sensitivity as further confirmed (Fig. [213]2j,k) by HOMA-IR test (Fig.
   [214]2l). Leptin and free fatty acid levels were also improved in
   SANA-treated mice (Fig. [215]2m,n). Importantly, complete
   solubilization and pH-dependent stabilization of SANA, followed by
   once-a-day subcutaneous (s.c.) injection (20 mg per kg per day)
   provided similar protective effects on weight gain, fasting blood
   glucose levels and glucose tolerance, with a similar pharmacokinetics
   profile to oral administration (200 mg per kg per day; Extended Data
   Fig. [216]3a–f) and similar levels of SANA in adipose tissue at steady
   state (Extended Data Fig. [217]3g). High-performance liquid
   chromatography (HPLC) and mass spectrometry analysis from plasma showed
   that metabolization of SANA in vivo led to formation of a saturated
   form of the molecule 5-(2-nitroethyl)salicylic acid (named metabolite
   1, M1; Extended Data Fig. [218]3h,i). Treatment of mice with M1 had no
   effect on DIO and glucose management, confirming the need of the
   reactive nitroalkene group for the metabolic effect of SANA (Extended
   Data Fig. [219]3j,k). AMPK activation by salicylate and SANA occurred
   at similar doses both in vivo and in vitro (Extended Data Fig.
   [220]3l,m), and SANA protected against DIO and glucose intolerance in
   AMPK (α1)-KO mice (Extended Data Fig. [221]3n,o) indicating that the
   metabolic effects of SANA are AMPK independent.

Fig. 2. SANA protects against glucose intolerance and liver steatosis in
response to DIO.

   [222]Fig. 2
   [223]Open in a new tab

   a–d, Liver’s macroscopic appearance (a); H&E staining (b); liver
   weight, n: ND = 11; HFD = 10; HFD + SAL = 12; HFD + SANA = 12 (c); and
   liver alanine transaminases (ALT) and aspartate aminotransferase (AST)
   in plasma/serum, n: ND = 5; HFD = 9; HFD + SAL = 9; HFD + SANA = 7 (d)
   in mice fed with ND, HFD, HFD + SAL or HFD + SANA (400 mg per kg per
   day, p.o.) for 8 weeks. e–g, Fasting glucose (e) n = 10 per group;
   glucose tolerance test (GTT) (f), n: ND = 9; HFD = 10; HFD + SAL = 12;
   HFD + SANA = 11; and GTT area of the curve (AOC) (g), n: ND = 9;
   HFD = 9; HFD + SAL = 11; HFD + SANA = 11 at week 8. h, Fasting insulin.
   n: ND = 17; HFD = 14; HFD + SAL = 12; HFD + SANA = 17. i,
   Glucose-stimulated insulin secretion (GSIS) at week 8. n: HFD = 7;
   SAL, = 8; SANA = 8. j,k, Insulin tolerance test (ITT) (j) and ITT AOC
   (k) at week 8. n: HFD = 16; HFD + SAL = 18; HFD + SANA = 15 (l).
   Insulin resistance calculated by HOMA-IR. n: HFD = 9; HFD + SAL = 8;
   HFD + SANA = 9. m, Plasma leptin levels at week 8. n: ND = 16;
   HFD = 16; HFD + SAL = 13; HFD + SANA = 17. n, Plasma non-esterified
   fatty acid (NEFA) levels at week 8. n: ND = 10; HFD = 12;
   HFD + SAL = 9; HFD + SANA = 9. Data are the mean ± s.e.m. Statistical
   analyses used were one-way ANOVA followed by Bonferroni post hoc for
   multiple comparisons (c–e, g, h, j and l–n) and two-way ANOVA followed
   by Tukey’s multiple-comparisons test (k).

   [224]Source data

Extended Data Fig. 3. Effect of SANA formulation, main metabolite and
evaluation of AMPK activation.

   [225]Extended Data Fig. 3
   [226]Open in a new tab

   (a-e) SANA (or SAL) were daily injected SC at 20 mg/kg. (a) Percent of
   weight on day 0. n = 5 per group. (b) Cumulative food intake. n = 2 per
   group (c) Basal glucose levels. n: HFD = 7; HFD + SAL = 11;
   HFD + SANA = 14. (d-e) GTT (d) and GTT AOC (e) at week 6. n: HFD = 4;
   HFD + SAL = 5; HFD + SANA = 5. (f) Pharmacokinetics of SANA (single
   dose of 20 mg/kg, SC). SANA in plasma was determined at different time
   points by HPLC. n = 3. (g) Representative chromatograms of steady state
   levels of different SANA formulations (400 mg/kg/day, PO and
   20 mg/kg/day, SC) in iWAT. Salicylic acid (S.A.) was used as an
   internal standard. (h-i) Chromatograms of organic extraction of plasma
   from control mice (H left) and SANA-treated mice (H right). M1
   indicates the peak assigned to SANA main metabolite
   5-(2-nitroethyl)salicylic acid by MS (i). (j) Percent of initial weight
   (day 0) of mice fed with HFD and treated with SANA or M1 (20 mg/kg/day,
   SC) for 8 weeks. n: HFD = 5; HFD + M1 = 5; HFD + SANA = 4. (k) Basal
   glucose levels at the onset of the experiment (0w) or after 8 weeks
   (8w) of HFD, HFD + SANA or HFD + M1 (20 mg/kg/day, SC). n = 5 per
   group. (l-m) Acute AMPK activation by SANA and SAL in vivo and in
   vitro. (l) Mice were treated with SANA, SAL (200 mg/kg, IP) or vehicle.
   AMPK activation was determined 1 h later by AMPK phosphorylation
   (Thr172) and ACC phosphorylation (S80) in the liver. n = 4 per group.
   (m) AMPK activation was measured in 3T3L1 adipocytes treated with
   100 μM of SAL or SANA. A769662 (10 μM) was used as positive control.
   n = 3 per group. (n-o) Effect of SANA (20 mg/kg/day, SC) in AMPK KO
   (ɑ1) mice in HFD. n = 7 per group. (n) Weight (left) and food intake
   (right), after 8 weeks. (o) GTT (left) and GTT AOC (right) at week 8.
   Data: mean ± SEM. Statistical analyses: two-way ANOVA followed by
   Tukey’s multiple comparisons test (a, b, k, n, o), one-way ANOVA
   followed by Bonferroni post hoc for multiple comparisons (c, e, j, l,
   m).

   [227]Source data

SANA is effective in treating obesity and metabolic damage

   Addressing DIO and metabolic complications once they are established
   may be more representative of the clinical need in humans. As such,
   SANA was administered to adult mice after obesity was established (5
   weeks of HFD). SANA-treated mice lost weight (Fig. [228]3a,b) while
   cumulative food intake was not affected (Fig. [229]3c), demonstrated
   diminished basal glycaemia (Fig. [230]3d), showed amelioration of
   glucose intolerance (Fig. [231]3e,f) and showed no signs of hepatic
   steatosis (Fig. [232]3g), decreased free fatty acids in plasma (Fig.
   [233]3h) and protection against increased liver enzymes, while no
   changes in other liver or kidney functional parameters were observed in
   both groups (Fig. [234]3i–k). Importantly, SANA (200 mg per kg per day)
   not only promoted weight loss, but also showed increased effectiveness
   compared to salicylate (200 mg per kg per day) and metformin (300 mg
   per kg per day) in glycaemic control during the 3-week period studied
   (Fig. [235]3l–o).

Fig. 3. Treatment of obese mice with SANA promotes weight loss and
amelioration of glucose intolerance and liver damage.

   [236]Fig. 3
   [237]Open in a new tab

   Adult (6-month-old) mice treated with SANA (200 mg per kg per day,
   p.o.) after 5 weeks of a HFD. a, Representative photograph and X-ray
   image at treatment end. b, Weight gain and percentage of initial
   weight. n: HFD = 7; HFD + SANA = 8. c, Cumulative food intake during
   treatment. n = 2 per group. d, Fasting glucose at week 8. n: HFD = 7;
   HFD + SANA = 8. e,f, GTT at week 8. n: HFD = 6; HFD + SANA = 7 (e) and
   GTT AOC n = 6 per group (f). g, Liver macroscopic appearance and H&E
   staining. h, Plasma free fatty acid levels at week 8. n: HFD = 7;
   HFD + SANA = 6. i–k, Liver ALT and AST plus liver and renal function
   markers (TP, total protein; ALB, albumin; GLO, total globulins; TBIL,
   total bilirubin; BUN, blood urea nitrogen) in plasma/serum at week 8.
   Hepatic transaminases: n = 6 per group; liver function: n: HFD = 6;
   HFD + SANA = 7; renal function: n: HFD = 7; HFD + SANA = 5. l,m,
   Percentage of initial weight and representative image of HFD-fed mice
   treated with SANA or SAL (200 mg per kg per day, p.o.) alone or
   combined with metformin (MET, 300 mg per kg per day, gavage). n:
   HFD = 5; HFD + MET = 5; HFD + SANA = 5; HFD + SAL = 4;
   HFD + SAN + MET = 5. n, Cumulative food intake. n = 2 per group. o,
   Evolution of fasting glucose. n = 5 per group. Data are the
   mean ± s.e.m. Statistical analyses used were two-way ANOVA followed by
   Tukey’s multiple-comparisons test (b, c and l), unpaired two-sided
   Student’s t-test (d, f and h–k) and paired two-sided Student’s t-test
   (o).

   [238]Source data

Proteomic analysis in iWAT shows that SANA induces thermogenesis

   The phenotype triggered by SANA pointed to increased energy expenditure
   in adipose tissue as a putative mechanism of action. Based on the
   recent emergence of alternative thermogenic pathways, we used unbiased
   proteomic analysis in iWAT as a discovery approach. Label-free
   quantitative proteomic analysis in iWAT of mice revealed a consistent
   shift in the pattern of protein expression between obese mice and obese
   mice treated with SANA (Fig. [239]4a). Proteins uniquely detected in
   HFD or HFD + SANA, as well as those differentially abundant between
   conditions were pinpointed (Fig. [240]4b,c and Supplementary Table
   [241]1). Combined, 109 proteins were increased in HFD + SANA iWAT
   tissue. Kyoto Encyclopedia of Genes and Genomes pathway enrichment
   analysis showed that SANA stimulated catabolism, oxidative
   phosphorylation and thermogenesis (Fig. [242]4d and Supplementary Table
   [243]1). Next, we extended the proteomic analysis to isolated
   mitochondria from iWAT. Heat map and differential analyses from
   mitochondrial proteomes showed a dramatic change in the protein
   expression profile in SANA-treated compared to placebo-treated animals
   (Fig. [244]4e–g and Supplementary Table [245]1). In this case, 380
   proteins were identified as statistically overrepresented in HFD + SANA
   (Fig. [246]4f,g). Pathway enrichment analysis showed that thermogenic,
   metabolic and tricarboxylic acid cycle, arginine-related,
   glycine-related and proline-related metabolic pathways were upregulated
   by SANA (Fig. [247]4h and Supplementary Table [248]1). These results
   resembled those previously described^[249]19 after cold-induced
   thermogenesis in mice and pointed to an increased activity of the
   creatine-dependent thermogenesis. Interestingly, among the proteins
   detected only after SANA treatment, glycine amidinotransferase (GATM),
   the rate-limiting step in creatine synthesis (Fig. [250]4f), and
   creatine kinase M-type (CKM) were present (Fig. [251]4b). Consistent
   with the metabolic signature of creatine-dependent thermogenesis
   described previously^[252]19 and the upregulation of GATM (Fig.
   [253]4f), we found a significant increase in creatine levels in iWAT
   from SANA-treated mice (Fig. [254]4i).

Fig. 4. Proteomic analysis of whole tissue and isolated mitochondria from
iWAT to SANA.

   [255]Fig. 4
   [256]Open in a new tab

   a–d, Whole iWAT proteomics performed in mice fed with HFD or HFD + SANA
   (400 mg per kg per day, p.o.) for 8 weeks. n = 3 per group. a, Heat map
   displaying unique proteins per row (labels shown every third protein)
   across biological replicates. b, Venn diagram of proteins exclusively
   detected in each condition (P < 0.05), including CKM. c, Volcano plot
   of common proteins with differential abundance (Benjamini–Hochberg
   q < 0.05); each dot represents a protein detected in at least four of
   six replicates; darker dots are statistically differential, and red
   dots are mapped to the enriched biological processes in d. d,
   WebGestalt pathway over-representation analysis for proteins
   overexpressed in HFD + SANA versus HFD. e–h, Same proteomic analysis
   conducted on isolated mitochondria from iWAT, highlighting
   mitochondrial Gatm. n = 3 per group; each sample is pooled from 2–3
   mice. i, Total creatine levels in iWAT were measured by mass
   spectrometry. n = 5 per group. Data are the mean ± s.e.m. Statistical
   analyses used were Bayesian framework with a Poisson distribution to
   identify exclusively detected proteins (b and f), multiple unpaired
   two-sided Student’s t-tests (one per protein) for volcano plots (c and
   g) and an unpaired two-sided Student’s t-test (i).

   [257]Source data

SANA stimulates creatine-dependent thermogenesis

   In line with the proteomic signature, thermal images showed that
   SANA-treated mice had increased surface temperature (Fig. [258]5a).
   Tissue microscopic analysis in iWAT showed an appearance of interspaced
   adipocytes with decreased lipid droplet size, consistent with a
   beigeing phenotype (Fig. [259]5b). Indeed, SANA stimulated
   ^18F-fluoro-2-deoxyglucose (^18FDG) uptake in vivo in iWAT (Fig.
   [260]5c). However, with the exception of the brain, glucose uptake in
   other tissues was not affected by SANA treatment (Extended Data Fig.
   [261]4a). Isolated inguinal white adipocytes from mice treated with
   SANA for 8 weeks demonstrated increased mitochondrial respiration (Fig.
   [262]5d), consistent with the signature from tissue proteomic data.
   SANA also induced increased respiration when human intact
   differentiated adipocytes were treated for 24 h (Extended Data Fig.
   [263]4b). However, direct treatment of adipocytes, isolated
   mitochondria from iWAT or brain with SANA had no direct effect on
   mitochondrial respiration (Extended Data Fig. [264]4c–e), demonstrating
   the effect of SANA on adipocyte respiration is delayed and potentially
   mediated by a genomic mechanism rather than acute regulation of
   mitochondrial respiration. Therefore, next we explored the possibility
   that SANA could induce changes in gene expression profiles in mouse
   iWAT. In fact, we observed that SANA stimulated the expression of Ckmt1
   and Ckm (Fig. [265]5e). Importantly, the increase in mRNA for Ckmt1 was
   also reflected on increased CKMT1 protein levels (Fig. [266]5f). Other
   genes involved in creatine metabolism, transport and synthesis
   including Slc25a4 (Ant1), Gatm, Oat, Prodh and Slc6a8 (CrT) were also
   upregulated by SANA (Fig. [267]5e). In contrast, Ucp1 was downregulated
   in mice treated with SANA when measured by mRNA (Fig. [268]5e), but no
   significant changes were observed in UCP1 protein levels by western
   blot (Fig. [269]5f). The thermogenesis-related genes Cidea, Prdm16 and
   Ppargc1a (Pgc1ɑ) were also upregulated by SANA (Fig. [270]5e).
   Importantly, treatment of differentiated human white adipocytes with
   SANA stimulated the expression of regulators of creatine metabolism,
   including CKMT1 and CKMT2 (sarcomeric mitochondrial CK; Extended Data
   Fig. [271]4b). Surprisingly, we found that Ckmt1-KO mice treated with
   SANA showed a ~50% increase in mortality rate after 1 week of treatment
   (no mortality occurred in wild-type (WT) mice), an effect that was
   rescued when Ckmt1-KO mice were acclimatized at thermoneutrality during
   the treatment with SANA (Fig. [272]5g). Under thermoneutral conditions,
   the glucose-lowering effect of SANA was lost in Ckmt1-KO mice (Fig.
   [273]5h), showing that SANA may require CKMT1 for its effects on
   glucose homeostasis and thermoregulation in vivo. SANA also stimulated
   the expression of genes involved in creatine metabolism and
   thermogenesis in brown adipose tissue (BAT). Unlike in iWAT, where Ckm
   and Ckmt1 expression is upregulated by SANA, in BAT Ckm and Ckmt2
   increased in expression (Fig. [274]5e,i). Other thermogenesis-related
   genes such as Cidea, Pgc1ɑ, Gatm and Prodh were upregulated in BAT,
   whereas Ucp1 was not altered (Fig. [275]5i). These changes in gene
   expression profile in BAT were reflected by increased creatine (Fig.
   [276]5j) and creatine kinase activity in isolated mitochondria (Fig.
   [277]5k). Measurements of respiration in isolated mitochondria of
   animals treated with SANA showed increased state II respiration (Fig.
   [278]5l). However, BAT mitochondrial respiration inhibited by GDP (Fig.
   [279]5l and Extended Data Fig. [280]4f) and stimulated after titration
   with oleate (Fig. [281]5l), attributable to UCP1 activity^[282]20, was
   not altered by SANA. Although not quantitative, these effects support
   the notion that increased state II respiration was not dependent on
   UCP1 regulation. Direct addition of SANA to BAT-isolated mitochondria
   also had no effect on GDP-inhibited respiration (Extended Data Fig.
   [283]4g). We further confirmed these findings by treating obese UCP1-KO
   mice with SANA and found that the compound promoted weight and fat mass
   loss (Fig. [284]5n and Extended Data Fig. [285]4h–j) without affecting
   food consumption (Extended Data Fig. [286]4h). To further support the
   thermogenic potential of SANA, we performed cold challenge experiments.
   Mice treated with SANA (20 mg per kg body weight, s.c., for 5 days)
   exposed to cold showed improved thermogenic response compared to
   controls (Extended Data Fig. [287]4k), which was reflected in
   upregulation of creatine kinases and thermogenic markers (Extended Data
   Fig. [288]4l) and increased creatine levels in iWAT (Extended Data Fig.
   [289]4m). To confirm that the improved response to cold was dependent
   on creatine, mice were treated with the creatine-depleting drug
   β-guanidinopropionic acid (β-GPA)^[290]21, which can only be used
   acutely as chronic exposure can induce weight loss^[291]19,[292]21.
   Mice were treated with SANA (20 mg per kg per day, s.c.) or
   SANA + β-GPA daily (0.4 mg per kg per day, intraperitoneally) for 5
   days. Treatment with β-GPA abolished the effect of SANA on thermogenic
   response (Fig. [293]5m), confirming that creatine is necessary for the
   effect of SANA on thermogenesis. No changes in body weight or food
   consumption were detected among treatments (Extended Data Fig.
   [294]4n,o). Muscle activity was also assessed during cold exposure. As
   creatine is a major muscle metabolite, we tested the potential for SANA
   to affect muscle function. Electromyographic analysis showed that
   shivering thermogenesis was not affected by SANA (Extended Data Fig.
   [295]5a). Creatine kinase expression and mitochondrial function in
   skeletal muscle and cells were not affected by SANA (Extended Data Fig.
   [296]5b–d), although there was a significant increase in muscle
   creatine levels in vivo (Extended Data Fig. [297]5e). Aerobic exercise
   capacity was performed and revealed that SANA did not impact muscle
   physiology (Extended Data Fig. [298]5f). Furthermore, treatment with
   SANA did not generate any observable deleterious effects on cardiac
   mitochondrial respiration (Extended Data Fig. [299]5g) or function
   (Extended Data Fig. [300]5h).

Fig. 5. SANA stimulates mitochondrial respiration, a gene expression
signature of creatine-dependent thermogenesis and a UCP1-independent loss of
weight.

   [301]Fig. 5
   [302]Open in a new tab

   a–f,i–k, Measurements at 8 weeks of treatment in mice fed with HFD,
   HFD + SANA or HFD + SAL at 400 mg per kg per day, p.o. a, Thermal
   images (left), with inguinal surface temperatures quantification
   (right). n: HFD = 5; HFD + SANA = 8. b, Representative H&E staining of
   iWAT. c, In vivo ^18FDG uptake in iWAT. n = 5 per group. d, Oxygen
   consumption of adipocytes isolated from iWAT. n = 7 per group. e,
   Quantitative PCR with reverse transcription (RT–qPCR) analysis of
   thermogenic markers in iWAT. n: Supplementary Table [303]1. f, Western
   blot of iWAT-isolated mitochondria, n = 7 for CKMT1; n = 4 for CKMT2
   and UCP1. g, Survival curves for WT and Ckmt1-KO mice treated with
   HFD + SANA (20 mg per kg per day, s.c.) at 23 °C or 28 °C. n:
   wt_SANA_23 °C = 6; Ckmt1-KO_SANA_23 °C = 17; wt_SANA_28 °C = 7;
   Ckmt1-KO_SANA_28 °C = 8. h, Fasting glucose in WT and Ckmt1-KO mice at
   28 °C after HFD and SANA or vehicle treatment. n: WT-HFD = 6;
   WT-HFD + SANA = 7; Ckmt1-KO-HFD = 5; Ckmt1-KO-HFD + SANA = 7. i,
   RT–qPCR analysis of thermogenic markers in BAT, n: Supplementary Table
   [304]1. j, Total creatine levels in BAT measured by mass spectrometry.
   n = 8 per group. k, Creatine kinase activity in BAT-isolated
   mitochondria, n: HFD = 15; HFD + SANA = 14. l, Respiration analysis in
   BAT-isolated mitochondria. n = 6 per group. m, β-GPA treatment (0.4 mg
   per kg per day, i.p.) in SANA-treated mice for 5 days (20 mg per kg per
   day, s.c.) before a 3-h cold exposure. n: Control = 14; β-GPA = 9;
   SANA = 13; SANA + β-GPA = 9. n, Weight gain of UCP1-KO mice treated
   with SANA (20 mg per kg per day, s.c.) or vehicle daily for 4 weeks.
   Mice were maintained at thermoneutrality and fed with a HFD 15 weeks
   before treatment. n = 5 per group. Mitochondrial assays utilized
   oligomycin (oligo), FCCP (carbonyl
   cyanide-p-trifluoromethoxyphenylhydrazone), antimycin A (AA), rotenone
   (ROT), pyruvate (P), malate (M) and fatty acid-free BSA. Data are the
   mean ± s.e.m. Statistical analyses used were paired two-sided Student’s
   t-test (a), unpaired two-sided Student’s t-tests (d, f and j–l),
   one-way ANOVA followed by Bonferroni post hoc for multiple comparisons
   (c, e and I), two-way ANOVA followed by Tukey’s multiple-comparisons
   test (h, m and n) and log-rank (Mantel–Cox) test (g). Vertical P values
   indicate multiple comparisons of each experimental condition versus the
   control condition (HFD in c, e and i).

   [305]Source data

Extended Data Fig. 4. SANA stimulates creatine-dependent thermogenesis and
promotes weight loss independently of UCP1 expression and activation.

   [306]Extended Data Fig. 4
   [307]Open in a new tab

   (a) 18FDG uptake in different tissues of mice fed with HFD, HFD + SANA
   or HFD + SAL, 400 mg/kg/day, PO for 8 weeks. n = 5 per group. (b)
   Differentiated human white adipose cells (TERT-hWA) were incubated with
   SANA (100 μM), for 24 hours. Left: mitochondrial respiration. n:
   Control=44; SANA = 48. Right: thermogenic markers measured by qRT-PCR.
   n = 6 per group. (c) Respiration of adipocytes isolated from mice fed
   with HFD for 8 weeks. Cells were incubated with SANA 100 μM or vehicle
   (DMSO) n = 6 per group. (d-e) Respiration of mitochondria isolated from
   adipocytes (D) or brain (E), incubated with SANA (100 μM) or vehicle.
   n = 4 per group. (f) Effect of GDP on respiration, calculated from data
   in Fig. [308]5l. n = 6 per group. (g) Effect GDP on UCP1 activity of
   isolated mitochondria from BAT, incubated with SANA (100 µM) or
   vehicle. n = 3 per group. (h-i) Cumulative food intake (H) and tissue
   mass (I) of WT and UCP1-KO mice treated with SANA (20 mg/kg/day SC),
   measured at the end of treatment. n = 5 per group. (j) UCP1 expression
   in BAT from WT and UCP1-KO. n = 4 per group. (k) Cold challenge. Mice
   were treated with SANA (20 mg/kg/day, SC) or vehicle. n = 14 per group.
   (l–m) Effect of SANA on cold response in mice. (L) Expression of
   thermogenic markers in iWAT measured by qRT-PCR after 6 hours of cold
   exposure. n: Supplementary [309]Tables. (M) Creatine quantification in
   iWAT in mice treated as described in L. n = 5 per group. (n-o) Effect
   of the creatine antagonist β-GPA in mice treated with SANA previous to
   cold. (N) Weight gain. n: Control=15; β-GPA = 9; SANA = 16;
   SANA+β-GPA = 10 and (O) cumulative food intake. n: Control=3;
   β-GPA = 2; SANA = 3; SANA+β-GPA = 2. Data: mean ± SEM. Statistical
   analyses: one-way ANOVA followed by Bonferroni post hoc for multiple
   comparisons (A); two-way ANOVA followed by Tukey’s multiple comparisons
   test (H, I, K); unpaired two-sided Student’s t-test (B -G, L-M).
   Vertical p-values indicate multiple comparisons of each experimental
   condition vs the control condition (HFD in A).

   [310]Source data

Extended Data Fig. 5. Effect of SANA on skeletal and cardiac muscle function.

   [311]Extended Data Fig. 5
   [312]Open in a new tab

   (a) Electromyogram at 15 and 30 min time points during cold exposure of
   mice treated daily with SANA (20 mg/kg/day, SC) or vehicle for 5
   consecutive days before cold exposure. n = 4 per group. (b) qRT-PCR
   analysis of creatine kinases expression in gastrocnemius muscle. n:
   n = 5 for Ckb, Ckm and Ckmt2; n = 4 for Ckmt1 (c) Respiration of
   gastrocnemius permeabilized fibers from mice treated or not with SANA
   (20 mg/Kg/day, SC), and after addition of substrates (PM:
   pyruvate-malate; SUCC: succinate; ADP) and inhibitors (OLIGO:
   oligomycin and ROT: rotenone). n: Control=8; SANA = 7. (d) Assessment
   of mitochondrial respiration in differentiated C2C12 myoblasts n = 3
   per group. (e) Creatine levels measured by NMR of mice treated with HFD
   or HFD + SANA (400 mg/kg/day, PO). n = 10 per group. (f) Aerobic muscle
   capacity measured in the treadmill at the end of treatment. Mice were
   fed during 22 weeks with HFD and HFD + SANA (or SAL) at 400 mg/kg/day,
   PO. n = 9 per group. (g) Respiration of permeabilized fibers from heart
   muscle from mice treated with SANA (20 mg/kg/day, SC). n: Control=5;
   SANA = 3. (h) Echo Cardiac function assessment in mice in HFD or
   HFD + SANA (or SAL) at 400 mg/kg/day, PO after 22 weeks of treatment.
   n = 5 per group. Data: mean ± SEM. Statistical analyses: one-way ANOVA
   followed by Bonferroni post hoc for multiple comparisons (f, h);
   two-way ANOVA followed by Tukey’s multiple comparisons test (a);
   unpaired two-sided Student’s t-test (b, c, d, e, g).

   [313]Source data

SANA stimulates energy expenditure during thermoneutrality

   One caveat of envisioning thermogenesis as a suitable approach to treat
   obesity in humans is that they spend most of their time under
   thermoneutral conditions. Interestingly, disruption of creatine
   metabolism in adipose tissue decreases energy expenditure and promotes
   obesity in thermoneutrality^[314]22,[315]23. Administration of SANA
   under thermoneutral conditions (Fig. [316]6a) also protected against
   DIO (Fig. [317]6b), with no effect on accumulated food intake (Fig.
   [318]6c). SANA was also effective in preventing diet-induced
   hyperglycaemia (Fig. [319]6d) and glucose intolerance (Fig. [320]6e),
   further supporting that SANA may be a suitable treatment for obesity
   under thermoneutral conditions. To clearly determine that increased
   energy expenditure due to thermogenesis is a driving force for the
   protection against obesity but not a side effect of differences in body
   weight, we followed a protocol of acute HFD feeding (Fig. [321]6f) as
   previously described^[322]23. During this short treatment there was no
   difference in weight gain among groups (Fig. [323]6g). However, thermal
   imaging showed that SANA promoted an increase in body heat dissipation
   (Fig. [324]6h,i). Consistently, mice treated with SANA during acute HFD
   feeding showed a significant increase in energy expenditure (Fig.
   [325]6j–l). These results are consistent with previous reports showing
   that a HFD is necessary for activation of creatine metabolism and
   energy expenditure in adipose tissue^[326]23. Interestingly, treatment
   of mice with SANA fed with normal chow showed no changes in energy
   expenditure (Extended Data Fig. [327]6). To further confirm our
   results, we administered the β[3]-adrenergic agonist CL316,243 after 3
   days of acute HFD feeding. Single treatment with CL316,243 stimulated a
   transient increase in energy expenditure, which was significantly
   increased in SANA-treated mice, irrespective of body weight, as shown
   by analysis of covariance (ANCOVA) analysis (Fig. [328]6m,n).
   Additionally, after acute HFD feeding at 28 °C, the mice were exposed
   to cold and surface temperature was assessed. Mice treated with SANA
   showed increased heat dissipation measured by thermal imaging (Fig.
   [329]6o,p). Finally, we used the acute HFD model to explore the effect
   of SANA isomers and analogues on creatine-dependent thermogenesis. We
   synthesized two different SANA isomers, 3-(2-nitroethenyl)salicylic
   acid and 4-(2-nitroethenyl)salicylic acid. In these isomers, the
   nitroalkene group was linked to the carbons 3 and 4 of salicylate,
   respectively, and 5-(2-nitroallyl)salicylic acid, with an extended
   carbon chain. We named them 3-SANA, 4-SANA and E-SANA, respectively
   (Extended Data Fig. [330]7a). SANA analogues were tested for their
   capacity to induce thermogenic response under acute HFD treatment
   (Extended Data Fig. [331]7a). Of all the compounds tested, only SANA
   activated thermogenesis, measured by infra-red imaging (Extended Data
   Fig. [332]7b,c) or increased the expression of creatine kinases in BAT
   (Extended Data Fig. [333]7d), showing that not only is the effect
   dependent on the presence of nitroalkene, but also the location is
   crucial for its function.

Fig. 6. SANA stimulates energy expenditure and protects against obesity under
thermoneutral conditions.

   [334]Fig. 6
   [335]Open in a new tab

   a, Diagram of long‐term treatment: Mice were fed HFD, HFD + SAL or
   HFD + SANA (20 mg per kg per day, s.c.) for 8 weeks. Created with
   [336]BioRender.com. b, Weight gain and percentage of initial weight. n:
   HFD = 13; HFD + SAL = 10; HFD + SANA = 11. c, Cumulative food intake.
   n = 2 per group. d, Fasting glycaemia at weeks 0 and 8. n = 8 per
   group. e, GTT (left) and GTT AOC (right) at week 8. n = 8 per group. f,
   Diagram of acute treatment. Mice were given a HFD with or without SANA
   (20 mg per kg per day, s.c.) at 28 °C. Created with [337]BioRender.com.
   g, Body weight. n = 9 per group. h, Representative thermal image of HFD
   and HFD + SANA mice at the end of the acute treatment. i,
   Quantification of surface temperature from thermal images. n = 5 per
   group. j–l, Energy expenditure (EE) evaluations from the acute
   treatment. n: HFD = 14; HFD + SANA = 13; data were pooled from two
   independent experiments. j, EE measured over a 24-h period. k,
   Regression analysis of EE data from j. l, Individual changes in EE
   during the acute treatment. m,n, EE evaluations of CL316,243-treated
   mice at the end of the acute treatment. n: HFD = 5; HFD + SANA = 6 (m)
   EE measurement. n, Regression analysis of EE from m. o,p,
   Thermoneutral-to-cold (4 °C) challenge. Representative thermal image
   after 1 h at 4 °C (o). Surface temperature quantification from images
   (p). n = 5 per group. Data are the mean ± s.e.m. Statistical analyses
   used were one-way ANOVA followed by Bonferroni post hoc for multiple
   comparisons (e), two-way ANOVA followed by Tukey’s multiple-comparisons
   test (b and c), paired two-sided Student’s t-test (d, i and p),
   unpaired two-sided Student’s t-test (g and l) and ordinary least
   squares and ANCOVA for regression analyses (k and n).

   [338]Source data

Extended Data Fig. 6. Supporting additional information to CLAMS data of mice
under acute HFD and SANA treatment under thermoneutral conditions.

   [339]Extended Data Fig. 6
   [340]Open in a new tab

   (a) Surface temperature quantification from thermal images of mice fed
   with ND and treated with SANA under thermoneutral conditions (day 9).
   n = 9 per condition. (b) Left: EE measurements over a 24-hour period at
   the end of the acclimation period at thermoneutrality (day 7). Right:
   regression plot of these data. n: CTL = 14; SANA = 13. (c) Total
   activity over a 24-hour period during acclimation (day 7). n: CTL = 14;
   SANA = 13 (d) Left: EE Measurements over a 24-hour period at the end of
   SANA treatment under ND feeding (day 9). Right: regression plot of
   these data. n: CTL = 6; SANA = 7 (e) Total activity over a 24 h period
   at the end of SANA treatment under HFD. n: CTL = 14; SANA = 13. (f)
   Mice weight at the beginning and end of the acute HFD treatment. n:
   CTL = 14; SANA = 13. (g) Mice weight after acute single treatment with
   CL316,243 (1 mg/kg). n: HFD = 5; HFD + SANA = 6. (h) Respiratory
   exchange ratio (RER) from mice under acute HFD treatment at the end of
   acclimation period (day 7), at the end of ND + SANA (day 9) and at the
   end of HFD + SANA (day 12). n: CTL = 14; SANA = 13. Data: mean ± SEM.
   Statistical analyses: paired two-sided Student’s t-test (A), unpaired
   two-sided Student’s t-test (G, H), two-way ANOVA followed by Tukey’s
   multiple comparisons test (f), ANCOVA (b, d).

   [341]Source data

Extended Data Fig. 7. Effect of SANA isomers (3-SANA, 4-SANA) and analog
(E-SANA) on creatine induced thermogenesis.

   [342]Extended Data Fig. 7
   [343]Open in a new tab

   (a) Chemical structure of synthesized and tested analogs and
   experimental strategy to evaluate their effect on thermogenesis. Mice
   were treated with different compounds (20 mg/kg/day, SC) following the
   same protocol described in Fig. [344]6f (acute HFD). Created with
   [345]BioRender.com. (b) Representative thermal images. (c) Quantitation
   of thermal images corresponding to BAT, iWAT and tail regions at the
   end of the treatment. n = 8 per group. (d) qRT-PCR analysis of creatine
   kinases in BAT at the end of the treatment. n: Supplementary
   [346]Tables. Data: mean ± SEM. Statistical analyses: One way ANOVA
   followed by Bonferroni post hoc for multiple comparisons (c, d).
   Expressed p-values indicate multiple comparisons of each experimental
   condition vs the condition of mice treated with SANA (HFD + SANA in c
   and d).

   [347]Source data

Analysis of safety and pharmacodynamics of SANA in human volunteers

   Based on all the accumulated preclinical evidence supporting that SANA
   might be a suitable first-in-class drug for treatment of obesity and
   comorbidities, we performed a randomized, double-blind,
   placebo-controlled phase 1A/B clinical trial in healthy lean volunteers
   and in volunteers with obesity. A single ascending dose (SAD; total
   cohort size of 17 volunteers) and a multiple ascending dose (MAD; total
   cohort size of 24 volunteers) were evaluated in lean and volunteers
   with obesity, respectively (Extended Data Figs. [348]7a and [349]8a).
   For clinical development purposes, SANA was referred to as MVD1. The
   primary endpoint of the study was to determine the safety of MVD1.
   Secondary endpoints included tolerability and pharmacokinetics.
   Additionally, exploratory endpoints (that is, body weight, glucose and
   insulin) were also analysed during the MAD part of the study. The
   demographic and baseline characteristics of the cohorts are presented
   in Extended Data Table [350]1. Regarding the safety of MVD1, only two
   patients showed adverse events (AEs) that were interpreted to be
   definitively related to the drug (renal reversible tubular damage,
   evidenced by proteinuria and glucosuria), which occurred with the
   highest dose of the SAD. No definitively drug-related AEs were observed
   during the MAD. One patient that was withdrawn from MAD cohort 1 (Fig.
   [351]7a) was later determined to have a pre-existing condition
   ([352]Methods). A summary of the patients with at least one possible or
   definitively drug-related AE is presented in Table [353]1. The most
   frequent possibly drug-related AEs were headache or soft stools. No
   severe AEs were observed during the study (both SAD and MAD). Liver
   function was normal in all patients after 15 days of treatment
   (Supplementary Table [354]1). Pharmacokinetics profiles both in SAD and
   MAD were like the ones determined in mice (Extended Data Fig. [355]8c,d
   and Supplementary Table [356]1). Importantly, a pharmacokinetics
   profile in day 15 of MAD was comparable to day 1, showing that drug
   does not accumulate in the body over time (Extended Data Fig. [357]8d
   and Supplementary Table [358]1).

Extended Data Fig. 8. Extended clinical information.

   [359]Extended Data Fig. 8
   [360]Open in a new tab

   (a-b) Clinical trial design for Single Ascending Dose (SAD) (a) and
   Multiple Ascending Dose (MAD) (b). (c) Pharmacokinetics of MVD1 in SAD
   and MAD, Day 1. The calculation of the pharmacokinetics parameters is
   presented in Supplementary [361]Tables. (d) Pharmacokinetics of the
   last dosage given to volunteers from MAD at day 15 of administration.
   The calculation of the pharmacokinetics parameters is presented in
   Supplementary [362]Tables.

   [363]Source data

Extended Data Table 1.

   Demographic Characteristics of Volunteers in SAD and MAD Studies and
   Baseline-to-End (Day 1 to Day 15) Treatment Profile of MAD Cohorts
   Demographic and clinical characteristics of the single ascending dose
   (SAD) participants at baseline
   SAD GROUP Number of patients 17 Race or ethnic group Number (%) Average
   body weight/weight range (kg) 74 /58-102
   Average patient age (years) 36 White 15 (88%) Average Body Mass Index /
   range 25 / 20-30
   Age range (years) 18-58 Asian 1 (6%) BMI Distribution Number (%)
   Number (percent) of female patients 8 (47%) Black 0 <20 0 (0%)
   Other 1 (6%) ≥20 - <25 7 (41%)
   Number (percent) of male patients 9 (53%) Hispanic or Latino 4 (24%)
   ≥25 - ≤30 10 (59%)
   Doses by Cohort mg/dose Doses (mg/dose/Kg) by Cohort Average (range)
   SAD 1 200 SAD 1 2.8 (2.7 - 3.3)
   SAD 2 400 SAD 2 5.3 (3.3 - 6.9)
   SAD 3 800 SAD 3 11.4 (8.3 - 13.9)
   Demographic and clinical characteristics of the multiple ascending dose
   (MAD) participants at baseline
   MAD GROUP Number of patients 24 Race or Ethnic Group Number (%) Average
   body weight / weight range (kg) 93 / 77-109
   Average patient age (years) 35 White 13 (57%) Average body mass index
   31 / 28-34
   Age range (years) 26-55 Asian 7 (30%) BMI distribution Number (%)
   Number (percentage) of female patients 5 (22%) Black 1 (4%) ≥28 - <30
   10 (43%)
   Other 2 (9%) ≥30 - ≤35 13 (57%)
   Number (percent) of male patients 18 (78%) Hispanic or Latino 4 (17%)
   >35 0
   Doses by cohort mg/dose Doses (mg/dose/Kg) by Cohort Average (range)
   MAD 1 100 MAD 1 1.0 (0.9 - 1.2)
   MAD 2 150 MAD 2 1.7 (1.4 - 1.9)
   MAD 3 200 MAD 3 2.1 (1.9- 2.2)
   Characteristics of MAD participants in each cohort at the baseline (day
   1) and at the end of the treatment (day 15)
   PLACEBO
   Weight (kg) BMI Fasting Glucose (g/L) Fasting Insulin (mU/L) HOMA-IR
   Fructasamine (umol/L)
   day 1 day 15 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day 15 day 1
   day 15
   Placebos p = 0,21 (n = 3) p = 0,24 (n = 3) p = 0,17 (n = 5) p = 0,19 (n
   = 5) p = 0,13 (n = 5) p = 0,48 (n = 5)
   P1 108.3 ND 33.1 ND 0.92 0.86 9 7 2.0 1.5 < 30 324
   P2 82.4 ND 28.1 ND 0.86 0.83 8 9 1.7 1.8 295 303
   P3 83.7 81.7 30.8 30.1 0.85 0.76 13 11 2.7 2.1 256 242
   P4 103.6 103.5 34.4 34.4 0.74 0.77 7 5 1.3 1.0 263 198
   P5 89.1 88.2 29.6 29.3 0.74 0.72 6 6 1.1 1.1 215 238
   ACTIVE
   Weight (kg) BMI Fasting Glucose (g/L) Fasting Insulin (mU/L) HOMA-IR
   Fructasamine (umol/L)
   MAD Coh.1 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day
   15 day 1 day 15
   200 mg/day p = ND (n = 1) p = ND (n = 1) p = 0,01 (n = 5) p = 0,10 (n =
   5) p = 0,05 (n = 5) p = 0,22 (n = 5)
   P6 97.7 ND 30.7 ND 1.24 1.01 18 20 5.5 5.0 < 30 356
   P7 101.5 ND 29.2 ND 0.92 0.76 4 4 0.9 0.8 < 30 326
   P8 100.4 ND 32.5 ND 0.94 0.85 14 2 3.2 0.4 323 300
   P9 90.3 ND 32.9 ND 0.90 0.83 25 14 5.6 2.9 286 274
   P10 104.1 101 33.8 32.8 0.94 0.81 21 7 4.9 1.4 212 ND
   MAD Coh.2 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day
   15 day 1 day 15
   300 mg/day p = 0,01 (n = 6) p = 0,02 (n = 6) p = 0,47 (n = 6) p = 0,58
   (n = 6) p = 0,65 (n = 6) p = 0,40 (n = 6)
   P11 81.1 78.9 28.3 27.5 0.81 0.81 5 6 1.0 1.2 313 322
   P12 95.7 95.1 29.4 29.2 0.85 0.83 2 14 0.4 2.9 292 264
   P13 78.1 76.9 30.3 29.8 0.92 0.86 13 11 3.0 2.3 266 250
   P14 77.1 75 30.4 29.6 0.85 0.79 8 9 1.7 1.8 251 249
   P15 83.7 83.7 29.0 29.0 0.79 0.83 6 6 1.2 1.2 312 326
   P16 109.1 107.5 34.3 33.8 0.77 0.79 8 4 1.5 0.8 184 171
   MAD Coh.3 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day 15 day 1 day
   15 day 1 day 15
   400 mg/day p = 0,01 (n = 6) p = 0,007 (n = 6) p = 0,31 (n = 6) p = 0,04
   (n = 6) p = 0,03 (n = 6) p = 0,004 (n = 6)
   P17 92.2 90.3 29.8 29.2 0.85 1.01 2 1 0.4 0.2 340 274
   P18 91 88.2 30.4 29.5 0.9 0.83 7 6 1.6 1.2 280 221
   P19 94.5 89.4 28.4 26.9 0.79 0.77 5 4 1.0 0.8 259 191
   P20 102.4 101.9 32.4 32.2 1.01 0.88 23 18 5.7 3.9 279 213
   P21 94.1 90.6 29.1 28.0 0.92 0.85 13 6 3.0 1.3 245 225
   P22 91.1 89.5 32.7 32.1 1.1 0.88 8 6 2.2 1.3 394 274
   [364]Open in a new tab

   P1 to P22- ID multiple ascending dose (MAD) participants ; cohort
   (Coh.) of participants; ND, not determinated; P value (P) obtained
   performing paired t-test statistical analyses compairing day 1 vs day
   15

   Demographic characteristics of volunteers in SAD and MAD studies and
   baseline-to-end (Day 1 to Day 15) treatment profile of MAD cohorts.

Fig. 7. Study design and main metabolic parameters induced by SANA (under the
name of MVD1 during clinical studies) in a randomized, double-blind,
placebo-controlled study phase 1A/B, first-in-human clinical trial in healthy
volunteers with overweight or obesity.

   [365]Fig. 7
   [366]Open in a new tab

   a, Clinical scheme of trial design for MAD. The complete design for SAD
   and MAD is described in Extended Data Fig. [367]8. Created with
   [368]BioRender.com. b, Analysis of weight evolution in the MAD cohort 3
   between day 1 and day 15. n = 6 per group. All remaining data for all
   cohorts can be found in Extended Data Table [369]1. c, Analysis of
   percentage of body weight loss between day 1 and day 15 in placebos and
   treated volunteers in MAD (cohort 2 and cohort 3). n: Placebo = 3; MVD1
   150 mg/12 h = 6; MVD1 200 mg/12 h = 6). Data are presented as the
   mean ± s.e.m. d–g, Analysis of plasma glucose (d), insulin levels (e),
   HOMA-IR test (f) and fructosamine levels (g), comparing day 1 and day
   15 of MAD cohort 3 volunteers treated with MVD1 200 mg/12 h. n = 6. All
   remaining data for all cohorts can be found in Extended Data Table
   [370]1. Statistical analyses used were paired two-sided Student’s
   t-test (b and d–g) and one-way ANOVA followed by Bonferroni post hoc
   for multiple comparisons (c).

   [371]Source data

Table 1.

   Summary of the AEs reported during SAD and MAD
   Treated participants Participants with possibly drug-related AEs
   Participants with definitively drug-related AEs
   SAD cohort (dose) n n n
   1 (200 mg) 4 0 0
   2 (400 mg) 4 3^a 0
   3 (800 mg) 3 0 2^b
   MAD cohort (dose) n n n
   1 (200 mg) 6 5^c 0
   2 (300 mg) 6 4^d 0
   3 (400 mg) 6 5^e 0
   [372]Open in a new tab

   ^aOne participant with mild, self-resolving diarrhoea, one with
   headache and one with tiredness. Mild, self-resolved AEs.

   ^bTwo with renal tubular damage, one of whom also had nausea and mild,
   self-resolving diarrhoea. Mild, self-resolved AEs.

   ^cTwo with headache, two with loose stools and one with
   lightheadedness. Mild, self-resolved AEs.

   ^dOne with headache, one with mild, self-resolving diarrhoea, one with
   abdominal pain and one with dysphagia. Mild, self-resolved AEs.

   ^eOne with constipation, one with soft stools, two with headaches and
   one with increased thirst. Mild, self-resolved AEs.

   Finally, we analysed the exploratory endpoints of the study during the
   MAD. The MAD study was carried as an inpatient study, and all meals
   containing a high-carbohydrate diet (250–300 g per day) were provided
   ad libitum to the volunteers (see meal plan in the [373]Methods). The
   body weight and glucose metabolism marker evolution for all cohorts and
   placebos is presented in Extended Data Table [374]1. Although the study
   was designed and powered to determine safety and tolerability, when we
   analysed volunteers in cohort 3 (200 mg/12 h) we found that volunteers
   with overweight or obesity showed a significant decrease in body weight
   in 15 days of treatment (Fig. [375]7b). Indeed, volunteers showed
   weight loss of about 3% at maximum dose (Fig. [376]7c), which is
   comparable to the weight loss found for semaglutide after the same
   period of treatment^[377]2. Furthermore, the volunteers not only lost
   weight, but also showed improvement in glucose metabolism. Except for
   one, all volunteers at 200 mg/12 h showed amelioration of plasma
   glucose (Fig. [378]7d). Insulin levels (Fig. [379]7e) and insulin
   resistance (HOMA-IR) were significantly improved during the treatment
   with 200 mg/12 h (Fig. [380]7f). Importantly, fructosamine levels,
   which is a read-out of average blood glucose levels over the past 2–3
   weeks, was significantly improved in all MAD 3 cohort (200 mg/12 h)
   volunteers during the period of treatment (Fig. [381]7g). Volunteers
   treated with placebo did not show any effect in these parameters
   (Extended Data Table [382]1). One volunteer started the study with a
   body weight of 102.40 kg and ended with 101.90 kg. Interestingly, the
   initial glucose and insulin were 1.01 g l^−1 and 23 mU l^−1, ending in
   0.88 g l^−1 and 18 mU l^−1, respectively. In consequence, the HOMA-IR
   improved from 5.7 to 3.9. Although preliminary, these results suggest
   that SANA may promote improvement of the glucose–insulin axis
   independently of body weight. This topic warrants further
   investigation. Overall, even though the clinical study was designed to
   determine safety and tolerability, the clinical data show that SANA not
   only is safe and well tolerated but also appears to have promising
   salutary clinical results in patients with obesity.

Discussion

   Obesity and its metabolic complications are a challenging problem for
   pharmaceutical interventions. The multifactorial causes of obesity, and
   the need for long-term interventions with associated side effects
   highlight the necessity for constant pharmacological innovation. The
   irruption of the GLP-1 mimetics provides hope for effective and safe
   treatments for obesity^[383]24. However, the massive use of GLP-1
   agonists is under pharmacological surveillance, and evidence about
   long-term safety, efficacy and compliance is still being collected. As
   an example, gastrointestinal effects, such as gastroparesis, nausea and
   diarrhoea, are among the side effects of GLP-1 treatment that call for
   attention^[384]25. Also, some concerns about sarcopenia exist, although
   more definitive data are still missing^[385]26. In that sense,
   developing alternative and even complementary approaches is of need.
   The use of hybrid drugs based on well-designed modifications of safe
   and inexpensive compounds may offer a suitable way to provide massive,
   safe and affordable therapies for this pandemic disease.

   The current study addressed the effects of a derivative of the ancient
   and long-time-used drug salicylate, 5-(2-nitroethenyl) salicylic acid,
   on the prevention and treatment of DIO and its related metabolic
   abnormalities. The data demonstrate that: (1) SANA is effective for
   weight loss and weight gain prevention during DIO, but not under normal
   diet; (2) the primary metabolic abnormalities associated with obesity
   are improved by SANA; (3) the beneficial effects of SANA are
   substantially greater than those obtained with salicylate; (4) the
   effects of SANA are mediated by stimulation of mitochondrial
   respiration in intact adipocytes in vivo, but not in isolated
   mitochondria in vitro; (5) SANA protects against DIO at thermoneutral
   conditions, a situation closely related to normal human environmental
   conditions; and (6) SANA is safe and well tolerated in healthy lean
   volunteers and also in individuals with overweight or obesity, showing
   promising effects in weight and glucose management in a 15-day-long
   trial.

   Our analysis of adipose tissue indicated that the salutary effects of
   SANA are associated with stimulation of non-shivering thermogenesis, a
   homeostatic mechanism present in all mammals, which constitutes an
   attractive approach for treating obesity^[386]27. Several reports show
   that adipose tissue-dependent thermogenesis is stimulated in humans
   after cold exposures^[387]28,[388]29. Importantly, the effects mediated
   by SANA occur through an increase in creatine-related energy
   expenditure, reminiscent of creatine-dependent thermogenesis, or futile
   creatine cycle^[389]30,[390]31, a heat-production pathway that can
   maintain body temperature in the complete absence of UCP1
   expression^[391]32. Recent studies have demonstrated that the futile
   creatine cycle plays a significant role in thermogenesis, functioning
   alongside UCP1 in both brown and beige adipocytes^[392]32–[393]35. SANA
   stimulates thermogenesis independently of UCP1 expression or activation
   (Fig. [394]5) and does not directly uncouple mitochondria. Salicylate,
   the precursor of SANA, promotes thermogenic activity in murine BAT by
   UCP1-dependent and UCP1-independent mechanisms^[395]12–[396]14,[397]36.
   The latter involves direct uncoupling of mitochondria, something that
   did not happen with SANA. Even more, neither the saturated form of SANA
   (M1), nor different isomers and analogues (3-SANA, 4-SANA, E-SANA) had
   any effect on thermogenesis and expression of creatine markers,
   supporting that the thermogenic effect of SANA is exclusive of this
   form of the molecule.

   SANA activated several key regulatory genes that have been identified
   as requisite to boost creatine metabolism in adipose tissue, including
   different creatine kinases and key enzymes for creatine synthesis.
   Disruption of creatine metabolism in vivo with the creatine antagonist
   β-GPA completely abolished the pro-thermogenic effect of the drug (Fig.
   [398]5m). Originally, CKMT1 was proposed to drive creatine
   phosphorylation in beige adipose tissue^[399]19. More recently, the
   same researchers proposed that CKB controls creatine-dependent
   thermogenesis in BAT^[400]37. However, the role of creatine kinase B
   (CKB) in mitochondrial creatine metabolism^[401]38, and how the
   creatine-dependent thermogenic cycle works has been
   questioned^[402]39,[403]40. In human adipose cells CKB is exclusively
   located in the cytosol^[404]38, different from mouse adipocytes, where
   it has been definitively located into the mitochondria^[405]33. This
   suggests that the pathway is complex, may have species-specific
   regulation, and needs further investigation. On the other hand, several
   reports suggest that CKMT2 is an important player in adipose tissue
   thermogenesis, that is, downregulation of Ckmt2 in adipose cells
   decreases phosphocreatine levels^[406]38, and silencing of
   β[3]-adrenergic receptors in beige/brown adipose cells decreases both
   Ckmt1 and Ckmt2 expression^[407]41. Also, bile acids increase
   mitochondrial function in white adipocytes, stimulating a thermogenic
   programme that is partially dependent on CKMT2 expression^[408]42.
   There is also evidence suggesting that downregulation of Ckmt2 in
   skeletal muscle in hibernating bears is linked to metabolic
   reprogramming involving an increase in glycolysis^[409]43.
   Nevertheless, to date, CKB is the only creatine kinase isoform that has
   been genetically demonstrated to mediate creatine kinase activity in
   mouse adipocytes and in mitochondria of thermogenic fat
   cells^[410]33,[411]37.

   We found that in iWAT, SANA stimulated both Ckmt1 and Ckm during DIO.
   However, in BAT, Ckmt2 and Ckm were the main kinases upregulated by
   SANA. While there is no evidence supporting that Ckm participates in
   creatine-dependent thermogenesis, the differential response among
   treatments and tissues suggests that there may be some level of
   redundancy among contributory kinases, as well as differential
   responses between acute and chronic thermogenic activity. In fact,
   RNA-sequencing analysis of human adipocyte cells^[412]44 shows that
   activation by forskolin promotes an increase in expression of more than
   one creatine kinase, which is also seen in vivo by cold exposure
   (Extended Data Fig. [413]9). Interestingly, an alternative theoretical
   model for creatine-dependent thermogenesis has been proposed, which
   resembles its canonical role in skeletal muscle and other tissues of
   high energy expenditure^[414]45, as a fast energy provider for Ca^2+
   reuptake into the endoplasmic reticulum^[415]39. Under this proposal,
   two kinases—one mitochondrial and one cytosolic—are necessary. This
   model better reconciles with our results using SANA, where Ckm is
   systematically upregulated, and does not necessarily antagonize the
   experimental evidence already published^[416]19,[417]23,[418]37.
   Interestingly, cold exposure promotes an increase in mitochondrial and
   cytosolic creatine kinases both in iWAT and BAT (Extended Data Fig.
   [419]9). Indeed, CKMT1/CKB-double KO mice show defective
   thermoregulation^[420]46, further supporting this idea. Nevertheless,
   recent work shows that TNAP/ALPL is indeed necessary for creatine
   futile cycle and body temperature maintenance in the absence of UCP1,
   demonstrating that CKB and TNAP/ALPL are key for the creatine futile
   cycle in mice during cold exposure^[421]33. These apparent
   discrepancies could imply that, while promoting creatine-dependent
   adipose tissue activation and thermogenesis, SANA does not completely
   phenocopy the pathway activation as during cold exposure. In fact, when
   we studied cold response in BAT and iWAT in untreated mice, we also
   found that Ckb is upregulated (Extended Data Fig. [422]9), but despite
   this creatine kinase is not affected by SANA. It is plausible then,
   that the pharmacological pathway activated by the drug phenocopies the
   physiological thermogenic, cold-elicited creatine-dependent
   thermogenesis, but through activation of different kinases. Our results
   show that SANA-induced thermogenesis relies, at least partially, on
   CKMT1. CKMT1-KO mice had compromised thermoregulatory capacity, but
   only when treated with SANA, not in basal conditions^[423]47.
   Furthermore, when maintained at thermoneutrality, CKMT1-KO mice showed
   no response to SANA in glucose management. Some of us recently proposed
   that CKMT1 is dispensable for mitochondrial function in white adipose
   cells, and CKMT1-KO mice have no alterations in viability, normal
   development or metabolic parameters, including thermoregulation in
   normal or HFD feeding^[424]47. The fact that Ckmt1-KO mice have
   diminished thermoregulatory capacity only when treated with SANA,
   suggests that CKMT1 plays a role in thermoregulation and that Ckmt1-KO
   mice might develop compensatory mechanisms to cope with this
   deficiency. It also suggests that treatment with SANA somehow commits
   mice to creatine-dependent metabolism for thermoregulation, to the
   detriment of other thermogenic pathways, compromising the flexibility
   of the thermogenic response. In fact, depletion of creatine in vivo not
   only abolished the effect of SANA on thermoregulation, but also in some
   cases seemed to worsen it (Fig. [425]5m). Experimental evidence
   suggests the existence of a dynamic cross-talk between UCP1 and
   creatine-dependent thermogenic pathways where they seem to go in
   opposite directions^[426]19. It is plausible then, that hyperactivation
   of creatine metabolism in adipose tissue by SANA will impede the
   activation of UCP1 in case of need. Activation of thermogenesis and
   beigeing of white adipose tissue has been linked to increased
   mitochondrial biogenesis. Our gene expression, including upregulation
   of PGC1α, suggest SANA is promoting mitochondrial biogenesis.
   Nevertheless, when we measured mitochondrial number in iWAT in response
   to SANA, we failed to detect any variations. The cold-elicited increase
   in mitochondrial number in iWAT can be detected after prolonged and
   continuous challenge^[427]48, so it is plausible that failure to detect
   an increase in mitochondrial number is mainly due to technical
   limitations. These questions remain to be addressed and will deserve
   future investigation. Also, the molecular target of SANA remains to be
   determined. Nevertheless, shedding light into this goes beyond the
   scope of this work and will certainly be part of future research.
   Importantly, the complete target of established pharmaceutical
   interventions to diabetes and obesity such as metformin^[428]49 or to
   endogenous ligands such as lac-phe^[429]50 have not been completely
   elucidated, but do not diminish their important clinical and
   physiological roles. It is also possible that, by affecting creatine
   metabolism, other tissues where creatine metabolism is of key
   importance, such as skeletal muscle, are being altered. Indeed,
   although we performed several in vivo studies of muscle function and
   did not see any affectation, we did see increased creatine levels in
   SANA-treated mice. Whether there are alterations in skeletal muscle
   function that are also contributing to the metabolic effects of SANA
   cannot be completely ruled out and deserves future investigation.
   Finally, we evaluated the safety, tolerability, pharmacokinetics and
   pharmacodynamics of SANA in a randomized, double-blind,
   placebo-controlled study phase 1A/B, first-in-human clinical trial. In
   the SAD cohort, SANA was well tolerated up to 400 mg, showing adverse
   side effects at the highest dose (800 mg). In the MAD cohort
   (100–200 mg/12 h), SANA showed a favorable safety and tolerability
   profile for 15 days of treatment, with minor AEs. No moderate or severe
   gastrointestinal adverse effects were reported (Table [430]1). Although
   the aim of the trial was to evaluate the safety of the drug, very
   importantly, we found a significant decrease in body weight in the
   highest dose of the MAD cohorts 2 and 3 (100 mg and 200 mg/12 h), while
   no significant weight loss was measured in the placebo (Fig. [431]7 and
   Extended Data Table [432]1). It is important to highlight that the
   weight loss of ~3% measured in cohort 3 (200 mg/12 h) is very similar
   to the results obtained with semaglutide in the same time frame^[433]2.
   Even more, when metabolic parameters were measured, in cohort 3
   (200 mg/12 h) there was a significant improvement in glucose metabolism
   (HOMA-IR, fructosamine) in volunteers with overweight/obesity, at doses
   averaging ~2 mg per kg per day (Fig. [434]7). It is worth noting that
   the improvement in glucose management in the MAD cohort 3 may be
   greater than expected to the body weight change observed. The
   possibility that SANA has effects on insulin sensitivity that go beyond
   the weight-lowering effect will be the subject of further clinical
   investigation. Also, it is worth mentioning that one of the
   participants in cohort 3 showed increased basal glycaemia despite
   showing decreased body weight and insulin levels. All these aspects
   remain to be studied and warrant a larger clinical trial focused on
   efficacy. Overall, SANA appears to be safe in humans, showing promising
   signs of efficacy, supporting the view that it is a good candidate to
   improve the current therapies for obesity and associated comorbidities.

Extended Data Fig. 9. Gene expression of creatine kinases during different
thermogenic stimuli.

   [435]Extended Data Fig. 9
   [436]Open in a new tab

   (a) RNAseq expression analysis of all creatine kinases in two clusters
   of human adipocytes progenitor cells (10.1073/pnas.1906512116).
   Respective statistical comparisons were set between vehicle-treated
   differentiated adipocytes (M), and forskolin-treated differentiated
   adipocytes (F). C corresponds to adipocytes maintained in a
   non-differentiated state. n = 5 per group in cluster 1-lipogenic
   adipocytes (Left) and n = 14 per group in cluster 2-thermogenic
   adipocytes (right,). CKM was not present in the public dataset
   analyzed. (b) Analysis of creatine kinases and the Alpl phosphatase
   expression by qRT-PCR in iWAT and BAT from mice fed with ND and
   maintained at room temperature (22 °C) or after a prolonged cold
   challenge (48 hours at 4 °C). iWAT: 22 °C n = 5 for Ckm, Ckmt1 and Ckb;
   n = 4 for Ckmt2 and Alpl; 4 °C n = 7 for Ckm, Ckmt1, Ckmt2 and Ckb;
   n = 8 for Alpl. BAT: 22 °C n = 4 for Ckm and Ckmt2; n = 5 for Ckmt1,
   Ckb and Alpl; 4 °C n = 8 for Ckm, Ckmt1, Ckb and Alpl; n = 7 for Ckmt2.
   Data: mean ± SEM. Statistical analyses: unpaired two-sided Student’s
   t-test (a,b).

   [437]Source data

Limitations of the study

   Although the human results obtained during the phase 1A/B clinical
   trial supports the notion that SANA may be effective in humans, it is
   important to take into consideration that the clinical trial was
   designed with the main objective of determining safety and
   tolerability, with a limited number of volunteers and clinical
   determinations. In that sense, the efficacy of SANA in humans needs to
   be further studied in a phase 2 clinical trial, with longer treatment,
   increased sample number, and extended measurements to validate its
   efficacy and mechanism of action. As such, the efficacy results
   presented here, although promising, are still preliminary and subject
   to further validation.

Conclusion

   Non-shivering thermogenesis is a homeostatic response present in mice
   and humans that can be subject to pharmacological stimulation. Boosting
   of creatine metabolism and thermogenesis in adipose tissue reveals a
   promising therapeutic opportunity for the treatment of metabolic
   diseases. The current study presents compelling preclinical and
   clinical evidence that SANA is a potent drug for the treatment of DIO,
   with a primary mechanism of action being the activation of
   creatine-dependent energy expenditure and thermogenesis in adipose
   tissue. Importantly, this activation also occurs at thermoneutrality, a
   condition that reflects everyday human life. This study provides the
   evidence that boosting creatine metabolism in adipose tissue is an
   effective pathway to prevent and treat DIO. Importantly, although our
   human clinical study was designed as a phase 1A/B safety and
   tolerability study, patients in the MAD cohort receiving the highest
   dose of SANA not only tolerated it well but also showed significant
   decrease in body weight and amelioration of glucose management over a
   15-day treatment. In summary, we present a comprehensive study of a
   molecule as a putative treatment for obesity, from the design,
   characterization, preclinical validation, mechanism of action to a
   phase 1A/B human clinical study. SANA appears to be a promising,
   first-in-class small molecule that could have a critical role in the
   treatment of obesity and metabolic syndrome.

Methods

Reagents and resources

   The key reagents and resources are listed in Supplementary Table
   [438]1.

Cell lines

   TERT-hWA cells were kindly provided by J. B. Hansen^[439]51 and
   cultured in 10% FBS/Advanced DMEM/F-12 media. Cells were induced to
   differentiate with insulin (5 μg ml^−1), dexamethasone (1 μM), IBMX
   (0.5 mM), rosiglitazone (1 μM), human cortisol (1 μM) and T3 (1 nM). At
   day 12 after stimuli, the adipocytes were considered mature. Cell lines
   were purchased from the American Type Culture Collection (ATCC). THP-1
   cells (ATCC, TIB-202) were grown with 10% FBS/RPMI media and
   differentiated to macrophages with 200 nM PMA for 48 h at 37 °C in 5%
   CO[2]. C2C12 cells (ATCC, CRL-1772) were grown with 10% FBS/DMEM media.

Animals

   C57BL/6J (Jackson Labs) mice and zebrafish (Danio rerio) were
   maintained at the Institut Pasteur de Montevideo Animal facility
   (UBAL). The experimental protocol was approved by the Institutional
   Animal Care and Use Committee (IACUC) of the Institut Pasteur de
   Montevideo (CEUA, protocol nos. 003-19 and 006-19). Studies were
   performed according to the methods approved in the protocol.

   For the CLAMS studies, C57BL/6J mice were purchased from the Jackson
   Laboratory, and the experiments were approved by the IACUCs of West
   Virginia University and Mayo Clinic (Jacksonville, USA; IACUC no.
   A00003888-18-R24).

   AMPK, Ckmt1 and UCP1-KO mice were also used. WT and AMPK-KO mice (α1)
   on a C57BL/6J/129 mixed background were bred and maintained at the
   Henry Ford Health animal facility as previously described^[440]52. All
   experiments were approved by the Animal Care Committee (IACUC, 1530).

   Ckmt1-KO and WT mice on a C57BL/6N background were maintained at the
   University of Guelph as previously described^[441]47. All experiments
   were approved by the Animal Care Committee (AUP, 5071) at the
   University of Guelph and met the guidelines of the Canadian Council on
   Animal Care.

   WT and littermate homozygous UCP1-KO mice on a pure C57BL/6J genetic
   background were bred and maintained at University of Sao Paulo animal
   facility. UCP1-KO and WT mice were generated by crossing heterozygous
   UCP1-KO mice (B6.129-Ucp1tm1Kz/J, Jackson Laboratories) on a C57BL/6J
   background. All experiments were approved by the Animal Care Committee
   (CEUA, 6071181120, ICB/USP)

   Except where indicated, all the experiments were performed on young
   male mice (10–12 weeks of age). All animals were housed under a 12-h
   light–dark cycle. Relative humidity was maintained between 30% and 70%.
   Food and water were provided ad libitum.

   Experiments were conducted at either 22 °C or 28 °C (thermoneutrality)
   with free access to food and water. Unless specified, mice were housed
   at 22 °C in groups of five. For thermoneutrality experiments, mice were
   housed individually at 28 °C (30 °C for UCP1-KO and WT littermate mice)
   and acclimatized for one week before the start of the experiment.

HFD feeding

   Male C57BL/6J, AMPK-KO and CKMT1-KO mice were fed a HFD (42% fat and
   0.25% cholesterol, AIN93G, LabDiet), starting at 12 weeks of age and
   for a duration of 8 weeks for metabolic evaluations. For the reversal
   of obesity experiments, adult male mice (6 months of age) were fed for
   5 weeks with a HFD before the onset of the treatments. UCP1-KO mice
   were fed for 15 weeks with a high-fat/high-sucrose diet before
   treatment. Male C57BL/6J mice were also fed a normal diet (46% fat, no
   cholesterol added, D113, SAFE ([442]https://safe-lab.com/safe_en/,
   Germany)

Drug delivery

   SANA or salicylate (salicylic acid) were administered orally (p.o.) by
   mixing the drugs as a powder, with the food. Metformin was prepared in
   a 100 mM phosphate buffer, pH 6.5 and was administered by oral gavage.
   For s.c. SANA, salicylic acid and M1: 5-(2-nitroethyl)salicylic acid
   administration, drugs were prepared in a 100 mM phosphate buffer, pH
   6.5/PEG 400 (50:50; vol/vol). In all cases the vehicle was used as
   control.

Cold challenge experiments

   C57BL/6J male mice previously acclimatized at 22 °C were treated or not
   (control) with SANA (or salicylic acid, 20 mg per kg per day, s.c.), in
   combination or not with β-GPA (0.4 g per kg body weight in PBS pH 7.5,
   i.p.) for 5 days and fed with normal chow diet. At that point, mice
   were housed individually at 4 °C from 1–6 h, with free access to food
   and water. Body temperature was measured before and during the
   challenge (once every hour) with a mouse rectal probe (ThermoWorks).

Thermal imaging

   Thermal images were collected from non-anaesthetized, awake mice with a
   FLIR E6 thermal imaging camera. Regions of interest from thermal images
   were selected and quantified using the FLIR Tools software.

Comprehensive Lab Animal Monitoring System

   Mice were maintained in the Comprehensive Lab Animal Monitoring System
   (Columbus Instruments) for 13 days at 28 °C. They were acclimatized for
   7 days and fed with normal chow. At that point, SANA administration
   (20 mg per kg per day, s.c.) was started and repeated for 6 days. On
   the second day of administration, the normal chow diet was changed to a
   HFD for the subsequent 4 days. On the last day mice were injected with
   CL316,243 (1 mg per kg body weight, i.p. in saline solution). In all
   cases the vehicle was used as control. Oxygen consumption (VO[2]),
   carbon dioxide production (VCO[2]) and activity of individual mice were
   monitored over the total period. VO[2] and VCO[2] values were used to
   calculate respiratory exchange ratio, and VO[2] and respiratory
   exchange ratio values were used to determine energy expenditure (kcal
   h^−1) as described previously^[443]53.

Body composition

   Lean mass and fat mass of individual mice were measured by quantitative
   nuclear magnetic resonance (NMR) using an EchoMRI analyser.

iWAT mitochondrial enrichment

   For each biological replicate, s.c. iWAT from two to three animals
   (according to body weight) were dissected, pooled and washed in
   ice-cold homogenization buffer (250 mM sucrose, 10 mM HEPES, 0.1 mM
   EGTA, pH 7.2). Homogenization buffer was supplemented with 2% BSA
   during the isolation of fat mitochondria and the procedure was
   conducted at 4 °C. Tissues were minced on ice and homogenized in
   homogenization buffer (5 ml per gram of tissue) using a motorized
   Potter–Elvehjem Teflon pestle. Homogenates were filtered through clean
   gauze to remove fat particles and centrifuged at 1,000g for 5 min to
   pellet nuclei and cell debris. The supernatant was transferred to a
   clean tube and centrifuged at 10,000g for 15 min to pellet crude
   mitochondria. The mitochondrial pellet was washed three times with 2 ml
   of homogenization buffer (without BSA) by centrifugation at 10,000g for
   10 min and stored at −80 °C until used.

Proteomics analysis

Sample preparation for LC–MS/MS analysis

   For proteomics analyses, three biological replicates were used per
   condition. Protein quantification was performed by densitometry
   analysis of whole iWAT or iWAT samples enriched in mitochondria
   separated on 12.5% SDS–PAGE gels and using a low-molecular-weight
   protein calibration kit (Amersham, GE Healthcare). For mass
   spectrometry analysis, equal amounts of each protein sample were
   separated only up to 1 cm into the resolving gel, fixed and stained
   with Coomassie Blue R-250. In-gel protein digestion and peptide
   extraction were performed as previously described^[444]52 with minimal
   variations. Briefly, the 1-cm bands were excised, and cysteine residues
   were reduced and alkylated by sequential incubation with 10 mM
   dithiothreitol and 55 mM iodoacetamide. Tryptic in-gel digestion was
   performed by overnight incubation at 37 °C with sequencing-grade
   trypsin (Promega) in a protease:protein ratio of 1:50 (wt/wt). Tryptic
   peptides were extracted from the gel by adding 60% acetonitrile/0.1%
   trifluoroacetic acid in two steps of a 1-h incubation at 30 °C. Samples
   were dried under vacuum and peptides were desalted using ZipTips C18
   (Merck Millipore). Eluted peptides were vacuum dried and dissolved with
   0.1% formic acid.

LC–MS/MS analysis

   Liquid chromatography–tandem mass spectrometry (LC–MS/MS) analysis was
   performed with an UltiMate 3000 nanoHPLC system (Thermo Fisher
   Scientific) coupled to a Q Exactive Plus mass spectrometer (Thermo
   Fisher Scientific). Samples were loaded onto a precolumn (Acclaim
   PepMap 100, C18, 75 µm × 2 cm, 3-µm particle size) and separated with
   an Easy-Spray analytical column (PepMap RSLC, C18, 75 µm × 50 cm, 2 µm
   particle size) at 40 °C using a two-solvent system: (A) 0.1% formic
   acid in water, (B) 0.1% formic acid in acetonitrile. Separation was
   achieved through an elution gradient as follows: 1% to 35% B over
   150 min and 35% to 99% B over 20 min, at a flow rate of 200 nl min^−1.
   The mass spectrometer was operated in positive mode using a top 12
   data-dependent method. Ion spray voltage was set at 2.5 kV and
   capillary temperature at 250 °C. The survey scans were acquired in the
   range of 200–2,000 m/z with a resolution of 70,000 at 200 m/z, an AGC
   target value of 1 × 10^6 and a maximum ion injection time of 100 ms.
   Precursor fragmentation occurred in an HCD cell with a resolution of
   17,500 at 200 m/z, an AGC target value of 1 × 10^4 and a maximum ion
   injection time of 50 ms. Normalized collision energy was used in steps
   of 25, 30 and 35. Dynamic exclusion time was set to 30 s. Each sample
   was injected twice.

Mass spectrometry data analysis

   PatternLab for Proteomics v4.1.1.25 (PatternLab)^[445]54 was used to
   perform peptide-spectrum matching against a Mus musculus proteome from
   UniProt ([446]UP000000589, downloaded on 26 August 2020) and label-free
   quantitative analysis. Search results were filtered by the PatternLab
   Search Engine Processor (SEPro) algorithm with a maximum false
   discovery rate value ≤ 1% at the protein level and 10 ppm tolerance for
   precursor ions. PatternLab’s Venn diagram statistical module was
   performed according to a PatternLab Bayesian model to determine
   proteins uniquely detected in each biological condition using a
   probability value less than 0.05 (refs. ^[447]54,[448]55). PatternLab’s
   TFold module was used to relatively quantify proteins present in both
   biological conditions by a spectrum count-based label-free
   quantification method. Proteins present in at least four biological
   replicates from the total of six were considered for TFold analysis.
   This module uses the Benjamini–Hochberg’s theoretical estimator to deal
   with multiple t-tests and it maximizes the number of identifications
   satisfying a fold change cut-off that varies with the P values
   (Benjamini–Hochberg q < 0.05) while restricting false differential
   proteins mainly due to low abundance^[449]56. The list of statistically
   overrepresented proteins in the condition HFD + SANA, obtained from
   Venn diagram and TFold PatternLab’s modules, was subjected to pathway
   enrichment analysis. The Mus musculus total proteome or the
   mitochondrial proteome obtained in our experiment were used as
   background proteomes for the over-representation analysis of the total
   or mitochondrial protein sets, respectively, using the functional
   enrichment analysis web tool WebGestalt 2019 (ref. ^[450]57). The mass
   spectrometry proteomics data have been deposited in the ProteomeXchange
   Consortium via the PRIDE^[451]58 partner repository under dataset
   identifier [452]PXD030485.

Metabolite analysis by LC–MS/MS and NMR

   All NMR experiments were performed on a Bruker AVANCE III 500 NMR
   spectrometer equipped with a room temperature z-gradient TXI probe and
   operating at ^1H and ^13C frequencies of 500.13 MHz and 125.76 MHz,
   respectively. Water-suppressed ^1H NMR spectra were recorded using a
   one-dimensional nuclear Overhauser effect spectroscopy (1D‑NOESY) pulse
   sequence with pre-saturation at 25 °C. A spectral width of 10 KHz, a
   data size of 32,000, and a total of 128 scans with a relaxation delay
   of 4 s between scans were used to record each experiment. All free
   induction decays were zero-filled to 64,000 points and apodized with a
   0.3-Hz exponential window function before Fourier transformation.
   Spectra were referenced to the a-glucose anomeric proton resonance at
   5.22 ppm present in all samples. Metabolites were identified by
   comparison to spectra from public databases, assisted with data
   obtained from standard gradient‑enhanced HSQC spectra when necessary.
   Metabolite concentrations in the NMR sample were estimated using the
   PULCON method as implemented in the spectrometer and normalized to wet
   tissue mass to yield concentrations in nanomoles per milligram of
   tissue. For LC–MS/MS analysis, the extracts were analysed by a Shimadzu
   LC Nexera X2 UHPLC coupled with a QTRAP 5500 LC–MS/MS (AB Sciex). An
   ACQUITY UPLC BEH Amide analytic column (2.1 × 50 mm, 1.7 μm, Waters)
   was used for chromatographic separation. The extracted multiple
   reaction monitoring peaks were integrated using MultiQuant 3.0.2
   software (AB Sciex).

Respirometry of adipocytes and cells

   Oxygen consumption of cultured differentiated adipocytes and iWAT
   adipocytes was measured with a high-resolution respirometer
   (Oxygraph-2k, OROBOROS INSTRUMENTS) and Seahorse (Agilent
   Technologies). Adipocyte suspension was pipetted into 2.2-ml
   mitochondrial respiration medium (MIR05). For oxygen consumption
   measurements of intact adipocytes, pyruvate (5 mM) was injected, and
   basal cellular respiration was recorded^[453]59. Proton leak
   respiration was assessed by addition of oligomycin (2 μg ml^−1).
   Titration of FCCP (0.5-μM steps) was performed for measuring maximal
   cellular respiration rates. Non-mitochondrial oxygen consumption was
   determined in the presence of antimycin A (0.5 μM) and rotenone
   (0.5 μM). Adipocyte respiration measurements were normalized by the μg
   ml^−1 of DNA. Basal respiration was determined before the addition of
   inhibitors and the uncoupler. Oligomycin-resistant and
   oligomycin-sensitive respiration rates were the ATP-independent (leak)
   and ATP-dependent respiration rates, respectively. Maximum respiration
   rate was obtained after titration with FCCP, and the spare respiratory
   capacity was the difference between the maximum and basal
   respiration^[454]60. Mito-Stress Test assays were performed using the
   Seahorse XFe96 Bioanalyzer and Seahorse 96-well XF cell culture
   microplates according to manufacturer’s instructions.

RNA isolation and real-time RT–qPCR

   Total RNA was isolated from cells or tissues using TRIzol Reagent and
   the Direct-zol RNA MiniPrep kit (Zymo Research) according to the
   manufacturer’s instructions. All primers were synthesized by Integrated
   DNA Technology. The complete list of primers and the number of samples
   used for each gene analysed in mice are shown in Supplementary Table
   [455]1.

Clinical study

Participants

   This was a phase 1, first-in-human, single- and multiple-dose
   escalation clinical study, conducted in normal weight or healthy adult
   volunteers with overweight or obesity, performed at CMAX Clinical
   Research Facility, Adelaide, Australia. The safety, tolerability,
   pharmacokinetics and pharmacodynamics of MVD1 following single- and
   multiple-dose administrations was evaluated using a randomized,
   double-blind, placebo-controlled study design. Forty-one adult
   volunteers (aged 18–60 years, body mass index BMI 20–30 kg/m^−2 for SAD
   and aged 25–60 years, BMI 28–35 kg/m^−2) were recruited from the
   general population in the vicinity of Adelaide, Australia, through
   advertisements. The clinical study was performed in accordance with the
   World Medical Association Declaration of Helsinki, the International
   Council for Harmonisation of Technical Requirements for Pharmaceuticals
   for Human Use (ICH) Guideline for Good Clinical Practice (GCP) E6 (R2),
   the applicable regulatory requirements, and the Australian National
   Health and Medical Research Council National Statement on Ethical
   Conduct in Human Research incorporating all updates. Trial registration
   and details can be found at the Australian New Zealand Clinical Trials
   Registry ([456]https://www.anzctr.org.au/, registration number:
   [457]ACTRN12622001519741, date submitted: 21 November 2022, date
   registered: 7 December 2022, last update: 11 August 2024, secondary ID:
   MVD1-AU-001). The Bellberry Human Research Ethics Committee reviewed
   this study in accordance with the National Health and Medical Research
   Council’s National Statement on Ethical Conduct in Human Research
   (2007, incorporating all updates) on the above meeting date. The
   principal investigator and co-investigators were not members of the
   Bellberry Human Research Ethics Committee that reviewed this study. All
   participants provided written informed consent.

Primary, secondary and exploratory endpoints

Primary endpoints

   Safety endpoints included the evaluation of the change from baseline
   and, where appropriate, over time and versus placebo, including: (1)
   Incidence, severity, and relationship of AEs or serious AEs (including
   withdrawals due to AEs). (2) Change in vital signs. (3) Change in
   electrocardiogram parameters. (4) Change in clinical laboratory
   parameters (haematology, serum chemistry, coagulation and urinalysis).
   (5) Stool diary. (6) Nutritional appetite questionnaire (part B only).

Secondary endpoints

   Plasma MVD1 and metabolite M1 pharmacokinetic endpoints included: (1)
   Maximum observed plasma concentration (C[max]). (2) Time to C[max]
   (T[max]). (3) Area under the plasma concentration–time curve from time
   0 to time of last quantifiable concentration (AUC[0–last]). (4) Area
   under the plasma concentration–time curve from time 0 extrapolated to
   infinity (AUC[0–inf]) following a single dose only. (5) Apparent
   terminal elimination half-life (t[1/2]). (6) Terminal elimination rate
   constant (λz). (7) Apparent clearance (CL/F). (8) Apparent volume of
   distribution (Vz/F; part A only). (9) Minimum observed concentration
   C[min] (day 15; part B only). (10) Area under the plasma
   concentration–time curve from time 0 to 12 h (AUC[0–12]). Urine MVD1
   pharmacokinetic endpoints included (but were not limited to): (1)
   Cumulative amount of unchanged drug excreted (Ae) in urine. (2) Renal
   clearance (CL[r])

Exploratory endpoints (MAD only)

   Exploratory endpoints included: (1) Body weight (day −1, day 15 and day
   22); (2) glucose; (3) insulin; (4) fructosamine; (5) the Patient Global
   Impression of Change (PGIC) questionnaire completed by study
   participants at the end of the study treatment period.

   Part A (SAD) consisted of three cohorts (S1–S3) with six participants
   planned per cohort, randomized at a ratio of 2:1 (MVD1:placebo) within
   each cohort. Participants were dosed once a day via the oral route of
   administration. Dose escalation was planned for part A (SAD; 200, 400
   or 800 mg per dose). Participants eligible for inclusion in the study
   were admitted to the investigational site on day −1. Following
   confirmation of eligibility on day 1, participants were enrolled and
   randomized to receive either MVD1 or placebo. Dosing occurred on day 1
   with MVD1 or placebo in a double-blind manner. Participants were
   discharged from the clinical facility on day 4 following completion of
   study assessments and were required to attend the clinic for a final
   follow-up visit on day 8 ± 1 day. For each part A cohort, dosing was
   initiated in two sentinel participants with one participant randomized
   to receive MVD1 and the other randomized to receive placebo. The
   remainder of the participants within each cohort only commenced dosing
   following satisfactory review of sentinel day 2 data by the principal
   investigator/independent medical monitor. Escalation to the next dose
   level occurred following satisfactory Safety Review Committee review of
   available safety data up to day 4 and available pharmacokinetic data
   from all participants at that dose level.

   Part B (MAD) consisted of three cohorts (M1–M3) with eight participants
   planned per cohort, randomized at a ratio of 3:1 (MVD1:placebo) within
   each cohort. Healthy adult volunteers with overweight or obesity were
   screened between day −35 and day −2, inclusive. Participants eligible
   for inclusion in the study were admitted to the investigational site on
   day −1. Following confirmation of eligibility on day 1, participants
   were enrolled and randomized to receive either MVD1 or placebo. Dosing
   occurred on days 1 to 15 with MVD1 or placebo (two daily oral doses of
   study drug taken 12 h apart (morning and evening) on days 1–14 with a
   final single-dose administration on the morning of day 15) in a
   double-blind manner. Participants were discharged from the clinical
   facility on day 18 following completion of study assessments and were
   required to attend the clinic for a final follow-up visit on day 22 ± 1
   day. Body weight was measured on days 1, 15 and 22. Blood samples for
   pharmacokinetic and pharmacodynamic analysis were to be collected
   before the dose and at several time points after the dose. MAD was
   conducted as an inpatient study. Participants were dosed twice a day
   (administered approximately 12 h apart, one in the morning and one in
   the evening) for 14 days, with a single dose administered on day 15 via
   the oral route. All the meals were provided at the facility and
   including at least 250–300 g of carbohydrates per day. At the end of
   the 15-day period, patients were asked to fulfil a quality
   questionnaire (PGIC) regarding appetite and satiety. There were no
   reports of affectation of appetite or satiety by the tested volunteers.
   Participants were required to fast overnight for at least 8 h before
   the screening visit and before admission to the clinical facility on
   day −1. On all dosing days (day 1 to day 15), participants were
   required to fast for at least 2 h before administration of each dose of
   study drug and until at least 2 h after dose. A final overnight fast of
   at least 8 h was required on day 15 before final pharmacodynamic
   measurements on day 16. During all fasting periods in part A (SAD) and
   part B (MAD), no food or beverages were to be consumed by the study
   participants except for water.

Alterations of the original protocol

   In the SAD cohort 3, the last patient was not dosed. The Safety Review
   Committee decided the dose at the no-observed-adverse-effect level was
   already determined and recommended to end the SAD arm of the study. In
   the MAD cohort, one volunteer was excluded from the trial due to a
   protocol deviation in the inclusion criteria (age mismatch with MAD
   inclusion criteria). It was later determined that the volunteer was
   randomized as a placebo. In MAD cohort 1 (100 mg/12 h), one participant
   was withdrawn from the study, following the advice of the Safety Review
   Committee, due to an event of proteinuria deemed mild and unrelated to
   study treatment, and most probably related to a pre-existing renal
   disease that was not detected during the screening period. Due to an
   error in the original protocol, which was later amended, body weight
   could not be determined for two placebo and four treated volunteers in
   cohort 1.

HOMA-IR calculations

   Fasting glucose and fasting insulin were measured from the same blood
   sample and the values obtained were used to calculate HOMA-IR using the
   following formula: HOMA-IR = (insulin (mU l^−1) × glucose (mg
   dl^−1))/405.

Statistical analysis

   No statistical methods were used to predetermine sample sizes, but our
   sample sizes are similar to those reported in previous
   publications^[458]19,[459]23,[460]61,[461]62. For in vivo studies, mice
   were randomly assigned to treatment groups. For clinical studies,
   double blinding and randomization were performed by a third party.
   Samples were processed in random order. Data collection and analysis
   were not performed blind to the conditions of the experiments. Results
   are presented as the mean ± s.e.m. of biologically independent samples.
   In cases where independent data were pooled, the approach is disclosed
   in the corresponding figure legend. Data were tested for normality
   using D’Agostino & Pearson or Shapiro–Wilk tests, and equal variances
   were analysed using the Levene test, meeting the assumptions of the
   statistical tests used. Paired and unpaired two-sided Student’s t-test
   for pairwise comparison, one-way ANOVA followed by Bonferroni post hoc
   for multiple comparisons, two-way ANOVA followed by Tukey’s
   multiple-comparisons test, linear regression and ANCOVA for in vivo
   metabolic analyses and log-rank (Mantel–Cox) test for Kaplan–Meier
   curve comparisons were used to calculate P values. Bayesian framework
   with a Poisson distribution and multiple unpaired Student’s t-tests
   (one per protein) with Benjamini–Hochberg correction for multiple
   comparisons in proteomics analyses were used. In all cases P < 0.05 was
   considered statistically significant. No animals or data points were
   excluded from the analyses. Calculations were done using GraphPad Prism
   and Jasp (JASP Team, 2024, version 0.19.2).

Reporting summary

   Further information on research design is available in the [462]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [463]Supplementary Information^ (4.1MB, pdf)

   MVD1-AU-clinical study report and MVD1-AU-statistical analysis plan.
   [464]Reporting Summary^ (3.8MB, pdf)
   [465]Supplementary Table 1^ (27.9MB, xlsx)

   Proteomic data. Clinical data and criteria. List of number of samples
   for RT–qPCR. Primer, reagent and resource information.

Source data

   [466]Source Data Fig. 1^ (66.9KB, xlsx)

   Statistical source data.
   [467]Source Data Fig. 2^ (60.2KB, xlsx)

   Statistical source data.
   [468]Source Data Fig. 3^ (122.7KB, xlsx)

   Statistical source data.
   [469]Source Data Fig. 4^ (272.1KB, xlsx)

   Statistical source data.
   [470]Source Data Fig. 5^ (98.5KB, xlsx)

   Statistical source data.
   [471]Source Data Fig. 5 and Source Data Extended Data Figs. 3 and 4^
   (844.2KB, pdf)

   Unprocessed western blots.
   [472]Source Data Fig. 6^ (184.5KB, xlsx)

   Statistical source data.
   [473]Source Data Fig. 7^ (14.2KB, xlsx)

   Statistical source data.
   [474]Source Data Extended Data Fig. 1^ (103.6KB, xlsx)

   Statistical source data.
   [475]Source Data Extended Data Fig. 2^ (43.1KB, xlsx)

   Statistical source data.
   [476]Source Data Extended Data Fig. 3^ (450.3KB, xlsx)

   Statistical source data.
   [477]Source Data Extended Data Fig. 4^ (79KB, xlsx)

   Statistical source data.
   [478]Source Data Extended Data Fig. 5^ (81.8KB, xlsx)

   Statistical source data.
   [479]Source Data Extended Data Fig. 6^ (76.3KB, xlsx)

   Statistical source data.
   [480]Source Data Extended Data Fig. 7^ (23.3KB, xlsx)

   Statistical source data.
   [481]Source Data Extended Data Fig. 8^ (18.8KB, xlsx)

   Statistical source data.
   [482]Source Data Extended Data Fig. 9^ (47.8KB, xlsx)

   Statistical source data.

Acknowledgements