Danuglipron

Danuglipron (PF-06882961) in type 2 diabetes: a randomized, placebo-controlled, multiple ascending-dose phase 1 trial
Aditi R. Saxena 1 ✉, Donal N. Gorman 2, Ryan M. Esquejo1, Arthur Bergman1, Kristin Chidsey1, Clare Buckeridge1, David A. Griffith 1 and Albert M. Kim 1
Agonism of the glucagon-like peptide-1 receptor (GLP-1R) results in glycemic lowering and body weight loss and is a thera- peutic strategy to treat type 2 diabetes (T2D) and obesity. We developed danuglipron (PF-06882961), an oral small-molecule GLP-1R agonist and found it had comparable efficacy to injectable peptidic GLP-1R agonists in a humanized mouse model. We then completed a placebo-controlled, randomized, double-blind, multiple ascending-dose phase 1 study (NCT03538743), in which we enrolled 98 patients with T2D on background metformin and randomized them to receive multiple ascending doses of danuglipron or placebo for 28 d, across eight cohorts. The primary outcomes were assessment of adverse events (AEs), safety laboratory tests, vital signs and 12-lead electrocardiograms. Most AEs were mild, with nausea, dyspepsia and vomiting most commonly reported. There were no clinically meaningful AEs in laboratory values across groups. Heart rate generally increased with danuglipron treatment at day 28, but no heart-rate AEs were reported. Systolic blood pressure was slightly decreased and changes in diastolic blood pressure were similar with danuglipron treatment at day 28, compared with placebo. There were no clinically meaningful electrocardiogram findings. In this study in T2D, danuglipron was generally well tolerated, with a safety profile consistent with the mechanism of action of GLP-1R agonism.

D
iabetes is estimated to affect more than 460 million people worldwide, with the number of people with diabetes pro- jected to reach more than 570 million by 2030. Within this
population, approximately 90% of those affected will have T2D1. Despite available treatments for T2D, only around 50% of patients in the United States achieve glycated hemoglobin (HbA1c) levels below the treatment goal2. Activation of the GLP-1R with exogenous GLP-1 has been shown to stimulate insulin release, inhibit gluca- gon secretion in a glucose-dependent manner, delay gastric empty- ing3,4, increase satiety and suppress food intake5. As such, there have been efforts to identify therapeutic GLP-1R agonists. The effects of GLP-1R agonism stated above are also associated with potent effi- cacy in glycemic lowering and body weight loss and these effects are observed across the GLP-1R agonist drug class6. Additionally, clinical trials have demonstrated a reduction in the risk of cardio- vascular events with different GLP-1R agonists7–9.
While several injectable, peptidic GLP-1R agonists are approved for the treatment of T2D (with the earliest agents approved at least 15 years ago10) these agents continue to account for only a limited proportion of diabetes prescriptions11. Several reasons could explain the slow adoption of injectable GLP-1R agonists, including poor adherence of this patient population to injectable therapies12 and the general patient preference for oral medicines compared with injectable therapies13. In late 2019, the oral formulation of the pep- tidic GLP-1R agonist semaglutide was approved for clinical use in T2D14. However, adequate absorption of oral semaglutide requires co-formulation with a gastric permeation enhancer and fasting to achieve sufficient systemic exposure14,15.
Danuglipron (PF-06882961) is an orally bioavailable, potent (cyclic AMP half-maximal effective concentration (EC50) = 13 nM),

small-molecule GLP-1R agonist (molecular weight of 555.6 Daltons) that selectively activates the GLP-1R in primates and humans16. Given the species-specific activation of GLP-1R by danuglipron, a new humanized GLP-1R (hGLP-1R) mouse model was devel- oped to characterize the nonclinical pharmacodynamic profile of danuglipron. Here we report the results of a randomized, double-blind, placebo-controlled, multiple ascending-dose phase 1 study to assess the safety, tolerability and pharmacokinetic and pharmacodynamic profiles of danuglipron in participants with T2D.
Results
Preclinical study. To study the in vivo activity of danuglipron, which selectively activates the primate and human GLP-1R, we gen- erated a hGLP-1R knock-in mouse model following a previously reported cloning strategy17. We used this knock-in mouse model to determine a response to liraglutide, which activates both rodent and primate GLP-1R. We observed a blunted anorectic response to liraglutide, compared with its wild-type (WT) counterpart (data not shown), suggesting defective GLP-1R signaling in the central nervous system. Therefore, we generated a new hGLP-1R knock-in mouse model using an alternative cloning strategy that enabled proper protein expression and activity of hGLP-1R. Contrary to our findings with the previously published hGLP-1R knock-in mouse model, we found that liraglutide, dosed subcutaneously (0.3 mg kg−1), resulted in similar improvements in glucose tolerance (Extended Data Fig. 1a,b) and reductions in food intake (Extended Data Fig. 1c) in our hGLP-1R knock-in and WT mice.
We then evaluated the glucose tolerance of our hGLP-1R knock-in mice dosed subcutaneously (for enhanced bioavailability of the compound) with vehicle or danuglipron in an intraperitoneal

1Pfizer Worldwide Research and Development, Cambridge, MA, USA. 2Pfizer Worldwide Research and Development, Cambridge, UK.
✉e-mail: [email protected]
NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine 1079

ARTICLES NATURE MEDICINE

292 screened

194 excluded for not meeting protocol eligibility requirements

98 randomized
3:1, PFI-06882961:placebo

10 mg BID
N = 9
15 mg BID
N = 9
50 mg BID
N = 10
70 mg BID
N = 9
120 mg BID
N = 9
120 mg BID ST
N = 9
120 mg QD
N = 8
200 mg QD CR
N = 10
Placebo
N = 25

All 9 completed

8 completed
1 discontinued due to TEAE
8 completed
2 discontinued (1 due to TEAE and 1 withdrawal by participant)

All 9 completed

All 9 completed

All 9 completed

All 8 completed
7 completed
3 discontinued for other reasons

All 25 completed

98 evaluated for primary end point of safety and tolerability
Fig. 1 | Participant disposition. Enrollment diagram (compliant with CONSORT flow diagram).

glucose tolerance test (IPGTT). Unlike WT mice, our hGLP-1R knock-in mice administered danuglipron displayed improved glu- cose tolerance (Extended Data Fig. 2a), reduced blood glucose area under the time–concentration curve from 0 to 120 min (AUC0–120; Extended Data Fig. 2b), increased plasma insulin (Extended Data Fig. 2c) and increased plasma insulin area under the concentra- tion–time curve (AUC) from 0–120 min (AUC0–120; Extended Data Fig. 2d). We also examined the anorectic response of our hGLP-1R knock-in mice after danuglipron administration by tracking ad libitum food intake during the dark cycle. We found that danuglip- ron (30 mg kg−1) significantly reduced food intake in our hGLP-1R knock-in mice, but not in WT mice, at 2.5, 5.5 and 15.5 h after dose (Extended Data Fig. 2e; P < 0.0001).
Clinical trial population and baseline characteristics. We sought to evaluate the safety, tolerability and pharmacodynamic and phar- macokinetic profiles of multiple ascending doses of danuglipron in patients with T2D taking metformin. A total of 292 patients were screened for eligibility and 98 patients taking metformin were ran- domized. The full inclusion and exclusion criteria are provided in Methods. Patients were enrolled into eight cohorts to receive either danuglipron or placebo for 28 d, with 73 participants ran- domized to danuglipron and 25 randomized to placebo. The eight danuglipron doses were as follows: 10 mg twice daily (BID), 15 mg BID, 50 mg BID, 70 mg BID, 120 mg BID, 120 mg BID slow titration (ST), 120 mg once daily (QD) and 200 mg QD controlled-release (CR) (Fig. 1 and Extended Data Fig. 3). Titration schedules are shown in Supplementary Table 1. Baseline demographic character- istics and physical measurements are presented in Table 1.

Safety and tolerability profile. A total of 92 participants completed the study; 2 participants discontinued due to treatment-related treatment-emergent AEs (TEAEs) (1 each from the 15 mg BID and

1080
50 mg BID groups), 1 due to withdrawal by the participant and 3 due to other reasons (Supplementary Table 2). Two participants dis- continued from danuglipron dosing due to TEAEs (one on day 24 and one on day 27 of the dosing period), but continued in the study (Supplementary Table 3). Nine participants had dose reduction or temporary discontinuation of dosing due to TEAEs: one in the pla- cebo group, four in the 50 mg BID group (for whom temporary or permanent down-titration to 15 mg BID was permitted), one in the 120 mg BID group, two in the 120 mg BID ST group and one in the 200 mg QD CR group (Supplementary Table 3).
In total, 319 TEAEs were reported by 83 participants (84.7%) (Table 2). Most TEAEs (294 out of 319 (92.2%)) were mild in sever- ity. Twenty-three TEAEs were of moderate severity, of which 18 were considered treatment-related. Two severe TEAEs were reported in the 120 mg BID ST group, one of which (nausea) was considered treatment-related; the other occurred during the follow-up period, was not considered treatment-related and was also the only serious AE in this study. No deaths occurred in this study.
The most frequently reported all-causalities TEAEs were nausea (49.0%), dyspepsia (32.7%), vomiting (26.5%), diar-
rhea (24.5%), headache (23.5%) and constipation (20.4%). The danuglipron 10-mg BID and 15-mg BID groups had the low- est number of TEAEs with 8 and 16 events, respectively, whereas the 120-mg BID ST group had the highest number of TEAEs with
50 events. The incidence of gastrointestinal TEAEs was lower in the placebo group (44.0%), compared with danuglipron, with the exception of the 10-mg BID group (22.2%). Gastrointestinal TEAEs were most commonly reported in the 120-mg BID ST (100%), 120-mg QD (100%), 200-mg QD CR (90.0%) and 50-mg BID (90.0%) groups (Table 2). The majority of gastrointestinal TEAEs in all groups were considered treatment-related. Nausea, diarrhea and vomiting were among the most commonly reported gastrointestinal TEAEs.
NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine

NATURE MEDICINE

The time course of nausea and vomiting varied according to target dose level and titration scheme (Extended Data Fig. 4). For danuglipron target dose levels up to 70 mg BID, the propor- tion of participants experiencing nausea typically decreased over the 4-week dosing duration. However, with higher target doses of danuglipron (120 mg QD, 120 mg BID and 200 mg QD CR), where titration to achieve the target doses occurred more rapidly, the pro- portion of participants experiencing nausea increased in weeks 3 and 4, compared with weeks 1 and 2. The proportion of partici- pants experiencing vomiting was lower than those experiencing nausea, but the general time course was similar to that observed with nausea.
One participant in the danuglipron 120-mg BID group expe- rienced mild hypoglycemia after missing a meal or snack, with a measured blood glucose of 64 mg dl−1; this TEAE resolved in 15 min on the same day of onset, was self-limited and was considered treatment-related (Supplementary Table 4).
Mean time-matched changes from baseline (CFB) (double dif- ferences) for triplicate measurements of supine heart rate, systolic blood pressure (BP) and diastolic BP were assessed from day 1 to day 28 (Fig. 2). Heart rate generally increased with danuglipron treatment at day 28, with mean differences from baseline ranging from −6.6 to +11.8 beats per minute (b.p.m.) across doses of danug- lipron; mean heart-rate differences from baseline ranged from
ARTICLES

Table 1 | Baseline demographic characteristics of study population and baseline parameters of interest
Parameter All participants Placebo Danuglipron (PF-06882961) dose
(N = 98) (N = 25)
10 mg BID 15 mg BID 50 mg BID 70 mg BID 120 mg BID 120 mg BID 120 mg QD 200 mg QD
(N = 9) (N = 9) (N = 10) (N = 9) (N = 9) ST (N = 9) (N = 8) CR (N = 10)
Age, years 57.4 (7.22) 57.6 (7.71) 54.0 (6.22) 56.1 (9.53) 60.2 (7.97) 58.3 (5.32) 55.8 (7.68) 58.6 (6.69) 58.3 (3.81) 57.6 (8.30)
Sex
Male 51 (52.0) 12 (48.0) 7 (77.8) 3 (33.3) 5 (50.0) 4 (44.4) 7 (77.8) 4 (44.4) 4 (50.0) 5 (50.0)
Female 47 (48.0) 13 (52.0) 2 (22.2) 6 (66.7) 5 (50.0) 5 (55.6) 2 (22.2) 5 (55.6) 4 (50.0) 5 (50.0)
Race
White 70 (71.4) 18 (72.0) 7 (77.8) 5 (55.6) 8 (80.0) 7 (77.8) 6 (66.7) 4 (44.4) 8 (100.0) 7 (70.0)
Black or African American 27 (27.6) 7 (28.0) 2 (22.2) 3 (33.3) 2 (20.0) 2 (22.2) 3 (33.3) 5 (55.6) 0 3 (30.0)
Native Hawaiian or other Pacific Islander 1 (1.0) 0 0 1 (11.1) 0 0 0 0 0 0
Ethnicity
Hispanic or Latino 61 (62.2) 18 (72.0) 6 (66.7) 4 (44.4) 7 (70.0) 6 (66.7) 4 (44.4) 4 (44.4) 6 (75.0) 6 (60.0)
Not Hispanic or Latino 37 (37.8) 7 (28.0) 3 (33.3) 5 (55.6) 3 (30.0) 3 (33.3) 5 (55.6) 5 (55.6) 2 (25.0) 4 (40.0)
Duration of diabetes, years 9.5 (6.0) 8.5 (6.8) 11.7 (6.1) 8.1 (4.2) 9.1 (5.1) 11.6 (4.6) 10.8 (4.7) 9.2 (5.9) 9.2 (7.4) 8.7 (7.8)
BMI, kg m−2 32.9 (4.3) 33.2 (4.5) 33.9 (3.4) 35.3 (5.1) 32.6 (3.6) 31.9 (4.3) 35.0 (4.7) 30.7 (4.5) 31.3 (3.3) 31.9 (3.6)
Weight, kg 92.4 (17.6) 94.3 (17.7) 99.6 (12.7) 92.9 (22.5) 87.8 (13.8) 86.9 (18.6) 101.6 (17.4) 87.9 (19.7) 84.9 (15.3) 92.4 (18.0)
HbA1c, % 8.3 (0.8) 8.0 (0.8) 8.2 (0.6) 8.6 (0.6) 8.3 (0.9) 8.3 (0.6) 8.5 (1.0) 8.2 (0.7) 8.2 (1.0) 8.6 (0.9)
MDG, mg dl−1 197.9 (39.4) 183.9 (38.3) 189.9 (20.6) 230.5
(39.4) 178.5 (35.8) 218.5 (45.9) 219.0 (32.4) 179.9 (36.7) 188.8
(24.9) 215.9 (39.6)
FPG, mg dl−1 178.7 (34.4) 167.6 (32.5) 158.3 (23.3) 198.4
(33.1) 168.1 (34.9) 186.3 (32.0) 196.9 (26.1) 176.3 (44.0) 179.9
(28.5) 195.3 (37.5)
FPI, μIU l−1 17.2 (15.0)* 15.6 (10.0) 23.4 (27.5) 15.5 (7.0) 13.2 (7.3) 10.3 (3.9)** 27.1 (26.6) 16.9 (9.3) 15.5 (9.9) 19.8 (18.5)
HOMA-IR 7.5 (6.5)* 6.6 (4.3) 8.0 (7.1) 7.2 (2.1) 5.7 (4.1) 4.5 (1.2)** 13.1 (13.4) 7.6 (5.0) 6.7 (3.9) 9.7 (9.2)
Data are n (%) or mean (s.d.). Baseline for MDG was defined as the AUC24/24 h on day −1. Baselines for FPG, FPI, HOMA-IR, HbA1c and body weight were defined as measurement at day −1, 0 h. HOMA-IR = (FPI x FPG)/405; where the FPI concentration is in μIU ml−1 and the FPG concentrations is in mg dl−1. *n = 97; **n = 8. BMI, body mass index.

−6.4 to −0.8 b.p.m. for placebo. No heart rate–related TEAEs were reported and there were no occurrences of heart rates >120 b.p.m. Regarding BP, most danuglipron groups seemed to have greater mean decreases in systolic BP, compared with placebo and most groups had similar mean CFB in diastolic BP compared with pla- cebo by day 28 (Fig. 2).
Treatment with danuglipron did not significantly affect the mean time-matched double differences from baseline in the electrocar- diogram (ECG) parameters of PR interval, QRS interval, QT inter- val or QTcF intervals from day 1 to day 28, compared with placebo. No ECG findings were reported as clinically meaningful by the investigator.
Across all danuglipron and placebo groups, there were no occur- rences of fasting hypoglycemia. Some isolated laboratory values outside of reference ranges were observed, but there were no clini- cally meaningful adverse trends across treatment groups.
Pharmacodynamic end points. All danuglipron dose levels resulted in statistically significant, dose-responsive reductions in mean daily glucose (MDG) at day 28, compared with placebo. Reductions in least squares (LS) mean (s.e.m.) MDG ranged from
−94.4 (8.6) to −53.4 (9.0) mg dl−1 at day 28, with the greatest reduc- tions observed in the 70-mg BID and 120-mg BID dose groups (Fig. 3a). Mean reductions from baseline (LS mean difference (s.e.m.))

NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine 1081

ARTICLES NATURE MEDICINE

Table 2 | Adverse-event profile, all causalities (≥2 occurrences) in any treatment arm
System Organ Class and Placebo Danuglipron (PF-06882961) dose
Preferred Term (N = 25)
10 mg BID 15 mg BIDa 50 mg BIDa 70 mg BID 120 mg 120 mg BID 120 mg QD 200 mg
(N = 9) (N = 9) (N = 10) (N = 9) BID ST (N = 9) (N = 8) CR QD
(N = 9) (N = 10)
Gastrointestinal disorders 11 (44.0) 2 (22.2) 6 (66.7) 9 (90.0) 7 (77.8) 8 (88.9) 9 (100) 8 (100) 9 (90.0)
Diarrhea 5 (20.0) 0 3 (33.3) 2 (20.0) 3 (33.3) 4 (44.4) 1 (11.1) 2 (25.0) 4 (40.0)
Dyspepsia 4 (16.0) 0 0 5 (50.0) 4 (44.4) 4 (44.4) 7 (77.8) 3 (37.5) 5 (50.0)
Nausea 4 (16.0) 0 2 (22.2) 7 (70.0) 3 (33.3) 8 (88.9) 9 (100) 7 (87.5) 8 (80.0)
Constipation 3 (12.0) 2 (22.2) 2 (22.2) 1 (10.0) 2 (22.2) 1 (11.1) 6 (66.7) 1 (12.5) 2 (20.0)
Vomiting 2 (8.0) 0 0 2 (20.0) 2 (22.2) 7 (77.8) 7 (77.8) 3 (37.5) 3 (30.0)
Abdominal discomfort 0 0 0 0 0 2 (22.2) 2 (22.2) 1 (12.5) 0
Abdominal distension 0 0 0 0 3 (33.3) 0 0 2 (25.0) 0
Eructation 0 0 0 2 (20.0) 0 1 (11.1) 0 0 0
Nervous system disorders 9 (36.0) 0 1 (11.1) 6 (60.0) 4 (44.4) 2 (22.2) 2 (22.2) 4 (50.0) 4 (40.0)
Headache 8 (32.0) 0 1 (11.1) 5 (50.0) 2 (22.2) 0 0 4 (50.0) 3 (30.0)
Dizziness 1 (4.0) 0 0 2 (20.0) 2 (22.2) 2 (22.2) 2 (22.2) 0 1 (10.0)
General disorders and administration-site conditions 3 (12.0) 1 (11.1) 0 7 (70.0) 1 (11.1) 0 2 (22.2) 1 (12.5) 2 (20.0)
Early satiety 2 (8.0) 1 (11.1) 0 5 (50.0) 1 (11.1) 0 2 (22.2) 1 (12.5) 2 (20.0)
Fatigue 1 (4.0) 0 0 3 (30.0) 0 0 0 0 0
Skin and subcutaneous tissue disorders 3 (12.0) 0 2 (22.2) 1 (10.0) 0 0 2 (22.2) 1 (12.5) 0
Dermatitis contact 2 (8.0) 0 1 (11.1) 1 (10.0) 0 0 0 1 (12.5) 0
Rash 1 (4.0) 0 1 (11.1) 0 0 0 2 (22.2) 0 0
Infections and infestations 2 (8.0) 1 (11.1) 0 0 0 0 0 0 0
Upper respiratory tract infection 2 (8.0) 1 (11.1) 0 0 0 0 0 0 0
Metabolism and nutrition disorders 1 (4.0) 0 0 1 (10.0) 1 (11.1) 6 (66.7) 2 (22.2) 3 (37.5) 5 (50.0)
Decreased appetite 1 (4.0) 0 0 1 (10.0) 1 (11.1) 6 (66.7) 2 (22.2) 3 (37.5) 5 (50.0)
Eye disorders 0 0 0 4 (40.0) 1 (11.1) 0 0 0 1 (10.0)
Dry eye 0 0 0 2 (20.0) 0 0 0 0 1 (10.0)
Vision blurred 0 0 0 2 (20.0) 1 (11.1) 0 0 0 0
Investigations 0 0 0 1 (10.0) 1 (11.1) 2 (22.2) 2 (22.2) 2 (25.0) 3 (30.0)
Alanine aminotransferase increased 0 0 0 0 0 0 1 (11.1) 0 2 (20.0)
Weight decreased 0 0 0 1 (10.0) 1 (11.1) 2 (22.2) 1 (11.1) 2 (25.0) 1 (10.0)
Musculoskeletal and connective tissue disorders 0 0 2 (22.2) 0 0 0 0 0 0
Back pain 0 0 2 (22.2) 0 0 0 0 0 0
Data are n (%). Participants were counted only once per treatment per event. All data collected since the first dose of study treatment were included. Totals for the number of participants at a higher level were not necessarily the sum of those at the lower levels, as a participant might have reported ≥2 different AEs within the higher level category. Placebo was pooled across all cohorts. Medical Dictionary for Regulatory Activities v.22.0 coding was applied. aThe 15-mg BID and 50-mg BID doses were not titrated.

in MDG on day 28 relative to placebo were similar for the 120-mg BID ST (−50.30 mg dl−1 (9.9)), 120-mg QD (−47.60 mg dl−1 (10.3))
and 200-mg QD CR (−55.99 mg dl−1 (10.2)) groups. These reduc- tions were larger than those observed in the 10-mg BID, 15-mg BID and 50-mg BID groups, but not as large as those in the 70-mg BID or 120-mg BID groups. On day 28, the modeled mean CFB (90% confidence interval (CI); P value) in MDG relative to placebo for the 10, 15, 50, 70 and 120-mg BID doses were −29.5 mg dl−1 (−45.9,
−13.0; 0.0038), −41.9 mg dl−1 (−59.9, −23.9; 0.0002), −24.6 mg dl−1
(−41.8, −7.4; 0.020), −61.5 mg dl−1 (−78.5, −44.5; <0.0001) and
−65.5 mg dl−1 (−82.6, −48.5; <0.0001), respectively. The modeled
1082
mean CFB (90% CI) in MDG relative to placebo for the 120-mg BID ST, 120-mg QD and 200-mg QD CR doses were −50.3 mg dl−1 (−66.8, −33.9; <0.0001), −47.6 mg dl−1 (−64.8, −30.4; <0.0001) and
−56.0 mg dl−1 (−73.0, −39.0; <0.0001), respectively.
Although baseline MDG levels varied across dosing groups (Table 1), all danuglipron groups achieved mean MDG levels of
≤149.2 mg dl−1 by day 14 and ≤141.5 mg dl−1 by day 28. For the danuglipron dose groups of 70 mg BID, 120 mg BID, 120 mg BID ST and 120 mg QD, mean MDG levels were <120 mg dl−1 by day 28 (Fig. 3b). Glucose profiles on day 28 for all treatment groups are shown in Extended Data Fig. 5.
NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Placebo
PF-06882961 dose 10 mg BID

15 mg BID

50 mg BID

70 mg BID

a
Mean time-matched double difference (b.p.m.)
CFB in heart rate 15

10

5
120 mg BID
120 mg BID ST 120 mg QD
200 mg QD CR

0

–5

–10

b

0 1 2 3

4 5 6 7 8 9 10 11 12
Planned time post-dose (h)

CFB in systolic BP 15

Mean time-matched double difference (mm Hg)
10

5

0

–5

–10

–15

–20

c

0 1 2 3 4 5 6 7 8 9 10 11 12
Planned time post-dose (h)

Mean time-matched double difference (mm Hg)
CFB in diastolic BP 10

5

0

–5

–10
0 1 2 3 4 5 6 7 8 9 10 11 12
Planned time post-dose (h)

Fig. 2 | Change from baseline in vital signs at day 28. Danuglipron (PF-06882961) resulted in mild increases in heart rate, compared with placebo, over 28 d, with declines in systolic and diastolic BPs at most dose levels. a, CFB in heart rate. b, CFB in systolic BP. c, CFB in diastolic BP. All plots are of mean time-matched double differences versus time in the first 12 h after dose administration in the morning. The time-matched double difference was calculated using the following steps: subtract pre-dose value on day 1 from all post-dose values (1); subtract the value at 0 h on day −1 from all other values on day
−1 (2); take the difference between the adjusted post-dose value in (1) and its time-matched value in (2). Unplanned readings were excluded from the presentation. Means of replicates were used in the calculations.
NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine 1083

ARTICLES NATURE MEDICINE

a

0

LS mean CFB in MDG, mg dl–1 (90% CI)
–20

–40

–60

–80

–100

Placebo

PF-06882961 dose
mean CFB (90% CI; P value) in FPG relative to placebo for the 10, 15, 50, 70 and 120 mg BID doses were −16.1 mg dl−1 (−34.9,
2.8; 0.16), −23.7 mg dl−1 (−43.4, −3.9; 0.05), −4.7 mg dl−1 (−23.5,
14.2; 0.68), −42.6 mg dl−1 (−61.7, −23.6; 0.0003) and −44.3 mg dl−1
(−63.7, −25.0; 0.0003), respectively. The modeled mean CFB (90% CI; P value) in FPG relative to placebo for the 120 mg BID ST, 120-mg QD and 200-mg QD CR doses were −44.2 mg dl−1 (−63.0,
−25.4; 0.0002), −14.9 mg dl−1 (−34.6, 4.9; 0.21) and −30.1 mg dl−1
(−49.4, −10.8; 0.011), respectively. There were no occurrences of FPG ≤ 70 mg dl−1 in any randomized participant.
All danuglipron BID dose groups achieved mean FPG levels of
≤137.0 mg dl−1 by day 14, despite mean baseline FPG levels being varied across dosing groups (Table 1). By day 28, all danuglipron BID dose groups achieved mean FPG levels of ≤127.3 mg dl−1, with the exception of the 50-mg BID group, where four partici- pants required down-titration to 15 mg BID. For danuglipron dose groups of 10 mg BID, 70 mg BID, 120 mg BID, 120 mg BID ST and 200 mg QD CR, mean FPG values were <126 mg dl−1 by day 28 (Fig. 4b), with the danuglipron 70-mg BID, 120-mg BID and 120-mg BID ST dose groups achieving mean FPG values <110 mg dl−1 at day 28 (Fig. 4b).
CFB in fasting plasma insulin (FPI; LS mean (s.e.m.)) at day 28 across the danuglipron dose groups ranged from −6.0 μIU ml−1 (2.2)

b
Placebo

10 mg BID
PF-06882961 dose

15 mg BID

50 mg BID

70 mg BID
to 3.6 μIU ml−1 (2.2), compared with −3.2 μIU ml−1 (1.2) for the pla- cebo group. There were no clear dose-related trends in FPI at day 28 (Extended Data Fig. 6a).

120 mg BID
240

220

Absolute mean in MDG (mg dl–1)
200

180

160

140

120

100
0
120 mg BID ST
120 mg QD

14
Day
200 mg QD CR

28
CFB in homeostatic model assessment of insulin resistance (HOMA-IR) (LS mean (s.e.m.)) at day 28 across the danuglipron dose groups ranged from −4.3 (0.9) to −0.2 (0.9), compared with
−2.5 (0.5) for placebo at day 28. Only the 70-mg BID and 120-mg BID dose groups demonstrated statistically significant declines in HOMA-IR relative to placebo (Extended Data Fig. 6b).
Body weight was statistically significantly reduced in the danuglipron 70-mg BID, 120-mg BID, 120-mg BID ST, 120-mg QD and 200-mg QD CR groups at day 28, compared with placebo. Body weight reductions (LS mean (s.e.m.)) ranged from −7.2 kg (0.7) to −2.2 kg (0.7) (Extended Data Fig. 7a), compared with −1.8 kg (0.4) for placebo. Relative to placebo, declines in body weight (s.e.m.; P value) for these dose groups ranged from −5.5 kg (0.8;
<0.0001) to −1.7 kg (0.9; 0.054). Danuglipron 120 mg BID demon- strated the greatest decline from baseline of −7.2 kg (0.7), with a placebo-adjusted decline of −5.5 kg (0.8; <0.0001).
HbA1c, analyzed as an exploratory safety assessment, was signif- icantly reduced for all danuglipron doses at day 28, compared with placebo. CFB (LS means (s.e.m.)) in HbA1c ranged from −1.2% (0.1) to −0.9% (0.1), which were all statistically significantly lower, com- pared with −0.5% (0.1) for the placebo group (Extended Data Fig. 7b). Relative to placebo, declines in HbA1c (SE; P value) ranged from

Fig. 3 | Change from baseline in mean daily glucose at day 28 and over
time. a,b, Danuglipron (PF-06882961) improved MDG in a dose-responsive manner at 28 d. CFB in MDG at day 28 (a). MDG over time (b). Plot in a represents LS means and 90% CIs from a mixed-model repeated measures (MMRM) model. Baseline was defined as the value collected at day −1, 0 h. MDG was defined as AUC24/24 h. Placebo n = 25; 10 mg n = 9, P = 0.0038; 15 mg n = 8, P = 0.0002; 50 mg n = 8, P = 0.0195; 70 mg n = 9, P < 0.0001;
120 mg BID n = 9, P < 0.0001; 120 mg BID ST n = 9, P < 0.0001; 120 mg QD
n = 8, P < 0.0001; 200 mg QD CR n = 7, P < 0.0001. Statistical significance, compared with placebo, was predefined as a two-sided P value <0.1 (indicated by *) based on the t-statistic from the MMRM model.

At day 28, all danuglipron dose groups had dose-responsive reductions from baseline in fasting plasma glucose (FPG; LS mean (s.e.m.)) ranging from −76.9 (9.8) to −37.3 mg dl−1 (9.8), where most reductions from baseline were statistically significantly greater than those observed with placebo (Fig. 4a). On day 28, the modeled

1084
−0.7% (0.1; <0.0001) to −0.4 (0.2; 0.017) across the danuglipron dose groups.
Postprandial CFB in plasma glucose, insulin, glucagon and C-peptide were assessed on day 28 by calculating the AUC from 0 to 4 h after a mixed-meal tolerance test (MMTT). There were statis- tically significant reductions in glucose AUC0–4 in all danuglipron dose groups, relative to placebo (Extended Data Fig. 8a). Reductions from baseline in insulin AUC0-4 were statistically significant for danuglipron 120 mg BID and 120 mg QD, compared with placebo and no consistent dose-related pattern was observed (Extended Data Fig. 8b). Similarly, no consistent pattern was observed for mean CFB in C-peptide AUC0–4. The lowest doses of danuglipron (10 mg BID and 15 mg BID), as well as the 200 mg QD CR group, had sta- tistically significant increases from baseline in C-peptide AUC0–4, compared with placebo. Only the 120 mg QD group demonstrated a statistically significant reduction in C-peptide AUC0–4, compared with placebo (Extended Data Fig. 8c). CFB in glucagon AUC0–4 were not statistically significant from placebo for any danuglipron dose
NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

a

0

LS mean CFB in FPG, mg dl–1 (90% CI)
–20

–40

–60

–80

–100

Placebo

PF-06882961 dose

The dose-normalized exposures (AUC24(dn)) of the 200 mg QD CR and 120 mg QD formulations were similar to the AUC24(dn) of the BID formulations on day 28.
Median maximum plasma concentrations (Cmax) at day 28 were reached between 3 to 10 h post-dose (median Tmax) for most danuglipron dosing arms (Extended Data Fig. 9); absorption was slower in the danuglipron 200 mg CR QD arm where median Tmax was 14 h. Mean half-life (t½) values ranged from 4.7 to 8.1 h at day 28 across all treatments; no apparent trends were observed across treatments, regimens or doses. Inter-participant variability for danuglipron exposure was based on geometric mean AUC24 and Cmax, with the percent coefficient of variation (%CV) values rang- ing from 31% to 87% for AUC24 and 32% to 94% for Cmax on day 28 across all treatments and cohorts. For all danuglipron doses, the amount of drug excreted unchanged in urine was minimal (<0.1%), indicating that the urinary excretion is not a primary clearance pathway of danuglipron.
Discussion
Administration of danuglipron, a selective, oral small-molecule GLP-1R agonist, was shown to be generally safe and well tolerated over 28 d. Most AEs were mild, with nausea, dyspepsia and vom- iting the most commonly reported. While increases in heart rate

b
Placebo
PF-06882961 dose

10 mg BID 15 mg BID

50 mg BID

70 mg BID
were observed with higher doses of danuglipron, analysis of other vital signs, laboratory evaluations and ECGs were generally no dif- ferent, compared with placebo. Analysis of exploratory outcomes

120 mg BID
220

200

Absolute mean in FPG (mg dl–1)
180

160

140

120

100

80
120 mg BID ST
120 mg QD
200 mg QD CR
revealed substantial dose-responsive reductions in glycemic indices (assessed by MDG, FPG and HbA1c) and body weight in patients with T2D over 28 d. The glycemic effects in this study were consis- tent with the effects generated in a new, hGLP-1R knock-in mouse model, in which administration of danuglipron resulted in acute improvements in glucose tolerance. Similarly, declines in body weight observed in this study were consistent with the acute reduc- tions in food intake in the hGLP-1R mouse model.
The safety and tolerability profile of danuglipron seems to be in line with that of peptidic GLP-1R agonists6,10, with the most com- mon TEAEs being nausea, dyspepsia, vomiting, diarrhea, head- ache and constipation. The incidence of gastrointestinal TEAEs in danuglipron dose groups ranged from 22% to 100%, compared with 44% in the placebo group. Overall, the incidence of gastroin- testinal TEAEs increased with higher doses of danuglipron, with the highest incidence (100%) occurring at higher doses (120 mg BID ST and 120 mg QD). As this study administered multiple doses of danuglipron and because the mechanism of action was expected to result in gastrointestinal TEAEs, prospective titration was incorpo-

0 4 8 14 18 21 28
Day

Fig. 4 | Change from baseline in fasting plasma glucose at day 28 and over time. Danuglipron (PF-06882961) improved FPG at all dose levels in a dose-responsive manner in 28 d. a, CFB in FPG. b, FPG over time. Plots in a represent LS means and 90% CIs from an MMRM model. Baseline was defined as the value collected at day −1, 0 h. Placebo n = 25; 10 mg n = 9,
P = 0.1605; 15 mg n = 8, P = 0.0500; 50 mg n = 8, P = 0.6829; 70 mg n = 9,
P = 0.0003; 120 mg BID n = 9, P = 0.0003; 120 mg BID ST n = 9, P = 0.0002;
120 mg QD n = 8, P = 0.2142; 200 mg QD CR n = 7, P = 0.0111. Statistical significance, compared with placebo, was predefined as a two-sided P value
<0.1 (indicated by *) based on the t-statistic from the MMRM model.

and there was no dose-related trend observed at day 28 (Extended Data Fig. 8d).
Pharmacokinetic results. Danuglipron plasma exposure at day 28, as measured by the geometric mean of the AUC over 24 h (AUC24), increased in an approximately dose-proportional man- ner across all danuglipron BID doses (Supplementary Table 5).
rated into dosing paradigms after clinical data were obtained from patients treated with danuglipron at the 15-mg BID and 50-mg BID dose levels. The 70-mg BID dose, which incorporated prospective titration, had a lower incidence of nausea compared with the 50-mg BID dose group. Although higher incidences of nausea and vomit- ing occurred with the highest doses of danuglipron (such as 120 mg BID), the titration schemes required to reach these doses involved rapid titration (due to the short overall duration of study and the desire to maintain the top dose for ≥2 weeks), with increases in dose every 2–4 d. In phase 1 and phase 2 dose-ranging studies with pep- tidic GLP-1R agonists, similar degrees of nausea and vomiting have been observed at the highest doses administered18,19. Slower titra- tion schemes have been shown to reduce rates of nausea and vomit- ing with GLP-1R agonists10, so we anticipated that danuglipron will have lower rates of nausea and vomiting when administered with more gradual titration schemes.
Increases in heart rate were observed with higher doses of danuglipron, which have also been observed with peptidic GLP-1R agonists20. The magnitude of increase in heart rate in the current study was in line with short-term clinical data published with mar- keted GLP-1R agonists, including dulaglutide and liraglutide18,21.

NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine 1085

ARTICLES NATURE MEDICINE

Although mean CFB in diastolic BP were not clearly observed, mean CFB in systolic BP suggested a trend for decline at higher doses of danuglipron by the end of the dosing period. Similar patterns have been observed with administration of peptidic GLP-1R agonists19,22. Given the frequent coexistence of hypertension in patients with T2D23, declines in systolic BP could offer additional therapeutic benefits in this patient population.
The pharmacodynamic profile of danuglipron showed robust declines in the glycemic biomarkers of MDG and FPG, in addition to significant declines in HbA1c and body weight after only 28 d. Compared with placebo, significant declines in MDG were observed with all doses of danuglipron, with greater declines observed with higher doses of danuglipron. The dosing and titration schemes for danuglipron doses of 10 mg BID, 15 mg BID, 50 mg BID, 70 mg BID and 120 mg BID were designed to incorporate administration of a stable dose level of danuglipron through at least days 15 to 28 of the dosing period to permit steady-state assessment of MDG on day 28. Based on the translation of MDG to predicted effects on HbA1c in longer duration studies24,25, the placebo-adjusted declines in MDG observed with danuglipron would be expected to provide placebo-adjusted declines in HbA1c of at least 1% long term (such as 12 weeks). The placebo-adjusted declines in HbA1c that were observed on day 28 were >0.8% at several danuglipron dose lev- els, so with longer term dosing over several months, even greater declines in HbA1c would be expected.
In addition to the substantial declines in MDG observed through- out the dosing period, the mean levels of MDG observed for all danuglipron dose groups reached levels of ≤149.2 mg dl−1 by day 14 and ≤141.5 mg dl−1 by day 28, with multiple danuglipron dose levels achieving mean MDG levels of <120 mg dl−1. These levels of MDG, which were achieved rapidly within this 28-d study, are expected to correspond to HbA1c levels of <7%, based on the relationship of average glucose to HbA1c demonstrated in the A1c-Derived Average Glucose study26. These data indicate that all dose levels of danuglipron administered in this study could result, on average, in optimal glycemic control with respect to current treatment guide- lines for the management of T2D27.
Declines in FPG were anticipated in the placebo group dur- ing the study conduct due to the controlled design of this phase 1 study; participants were administered all meals and all concomitant medications according to their prescribed schedule in the clinical research unit. However, declines in FPG associated with all doses of danuglipron were greater in magnitude than with placebo, with most doses having significantly greater declines than placebo. These declines occurred without causing fasting hypoglycemia. In fact, only one TEAE of mild, nonfasting hypoglycemia was reported in the study. The greater magnitude in reduction of MDG, com- pared with FPG, in participants receiving danuglipron is consistent with the observation that GLP-1R agonism attenuates postprandial increases in glucose, partially by slowing gastric emptying28.
Large declines in body weight were observed by day 28 in sev- eral danuglipron dosing groups, including 70 mg BID and 120 mg BID. The mean baseline body weight for the study population was within the range of other studies with participants with T2D. However, the degree of weight loss reported in this study was larger than has been noted in other phase 1 and 2 studies with inject- able GLP-1R agonists or dual GLP-1R-glucose-dependent insuli- notropic polypeptide agonists, but direct comparisons are limited by differences in titration schemes and study durations18,19. In the current study, a reduction in body weight (LS mean decline) of −4.4 kg at day 28 was reported with danuglipron 70 mg BID, a dose level with a lower incidence of nausea and vomiting compared with other dose levels. The higher dose of 120 mg BID had sub- stantial weight loss of −7.2 kg (LS mean decline) by day 28, but also had a higher incidence of nausea and vomiting than other doses, with a majority of participants experiencing nausea (eight of nine

1086
participants) and/or vomiting (seven of nine participants). Published data with marketed GLP-1R agonists indicate that higher incidence of nausea and vomiting are typically experienced within the first 4 weeks of dosing and, if experienced, are asso- ciated with greater weight loss over time29. Although there were no clear changes in FPI levels with any danuglipron dose, com- pared with placebo, HOMA-IR declined with danuglipron doses of 70 mg BID and 120 mg BID; these doses also resulted in substan- tial weight loss. While the limitations of HOMA-IR assessment in the setting of the 4-week administration of a GLP-1R agonist are acknowledged, an association between reduced insulin resistance and reductions in body weight has been demonstrated in longer term studies with peptidic GLP-1R agonists30.
The pharmacokinetic profile of danuglipron demonstrated dose-proportional increases in plasma exposure. Doses were administered with the morning and evening meals to standardize administration in this inpatient study. However, danuglipron has been shown to have similar pharmacokinetic exposure, as measured by AUC24, in both the fed and fasted states16. This stands in contrast to the recently approved oral formulation of the peptidic GLP-1R agonist semaglutide, which has fasting requirements because mini- mal to no measurable systemic exposure is observed when semaglu- tide is administered in the fed state15.
Strengths of the study included the controlled, inpatient, phase 1 study design in the target patient population. Study end points were assessed frequently and there was control of dietary influences to minimize potential confounders of pharmacodynamic assess- ments. MDG was assessed by sampling plasma glucose at 14 time points over a 24-h period to provide a more complete picture of average glucose on the selected study days. Furthermore, the phar- macodynamic biomarkers measured in this study had direct clinical relevance to the study population and indicated the potential for therapeutic benefits with longer term dosing.
Limitations of the study included the 28-d dosing duration, which required rapid titration to reach higher doses and limited the assessment of the tolerability of these doses with longer, slower titration schemes. In addition, no placebo run-in period was incor- porated due to the prolonged inpatient stay involved in the study, possibly causing the mild declines in pharmacodynamic biomarkers observed in the placebo group.
In summary, we present the first reported multidose clinical data with danuglipron, a selective, potent, oral small-molecule GLP-1R agonist administered in participants with T2D. Danuglipron had a favorable safety and tolerability profile, which was generally con- sistent with the peptidic GLP-1R agonist class. The results of this study support further clinical development of danuglipron for the treatment of T2D and obesity.
Online content
Any methods, additional references, Nature Research report- ing summaries, source data, extended data, supplementary infor- mation, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41591-021-01391-w.

Received: 16 October 2020; Accepted: 10 May 2021;
Published online: 14 June 2021

References
⦁ International Diabetes Federation. IDF Diabetes Atlas 9th edn. (International Diabetes Federation, 2019).
⦁ Ali, M. K. et al. Achievement of goals in US diabetes care, 1999-2010.
N. Engl. J. Med. 368, 1613–1624 (2013).
⦁ Kreymann, B., Williams, G., Ghatei, M. A. & Bloom, S. R. Glucagon- like peptide-1 7-36: a physiological incretin in man. Lancet 2, 1300–1304 (1987).

NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine

NATURE MEDICINE

⦁ Nauck, M. A. et al. Glucagon-like peptide 1 inhibition of gastric emptying outweighs its insulinotropic effects in healthy humans. Am. J. Physiol. 273, E981–E988 (1997).
⦁ Flint, A., Raben, A., Astrup, A. & Holst, J. J. Glucagon-like peptide 1 promotes satiety and suppresses energy intake in humans. J. Clin. Invest. 101, 515–520 (1998).
⦁ Aroda, V. R. A review of GLP-1 receptor agonists: evolution and advancement, through the lens of randomised controlled trials. Diabetes Obes. Metab. 20, 22–33 (2018).
⦁ Gerstein, H. C. et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet 394, 121–130 (2019).
⦁ Marso, S. P. et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N. Engl. J. Med. 375, 1834–1844 (2016).
⦁ Marso, S. P. et al. Liraglutide and cardiovascular outcomes in type 2 diabetes.
N. Engl. J. Med. 375, 311–322 (2016).
⦁ Drucker, D. J., Habener, J. F. & Holst, J. J. Discovery, characterization, and clinical development of the glucagon-like peptides. J. Clin. Invest. 127, 4217–4227 (2017).
⦁ Montvida, O., Shaw, J., Atherton, J. J., Stringer, F. & Paul, S. K. Long-term trends in antidiabetes drug usage in the US: real-world evidence in patients newly diagnosed with type 2 diabetes. Diabetes Care 41, 69–78 (2018).
⦁ Cooke, C. E., Lee, H. Y., Tong, Y. P. & Haines, S. T. Persistence with injectable antidiabetic agents in members with type 2 diabetes in a commercial managed care organization. Curr. Med. Res. Opin. 26, 231–238 (2010).
⦁ Holko, P., Kawalec, P. & Mossakowska, M. Quality of life related to oral, subcutaneous, and intravenous biologic treatment of inflammatory bowel disease: a time trade-off study. Eur. J. Gastroenterol. Hepatol. 30, 174–180 (2018).
⦁ Novo Nordisk A/S. RYBELSUS® (semaglutide) package insert (2020).
⦁ Buckley, S.T. et al. Transcellular stomach absorption of a derivatized glucagon-like peptide-1 receptor agonist. Sci. Transl. Med. ⦁ https://doi. ⦁ o⦁ rg/10.1126/scitranslmed.aar7047⦁ (2018).
⦁ Griffith, D.A. et al. A small-molecule oral agonist of the human glucagon-like peptide-1 receptor. Preprint at bioRxiv ⦁ https://doi. ⦁ o⦁ rg/10.1101/2020.09.29.319483⦁ (2020).
⦁ Jun, L. S. et al. A novel humanized GLP-1 receptor model enables both affinity purification and Cre-LoxP deletion of the receptor. PLoS ONE 9, e93746 (2014).
⦁ Coskun, T. et al. LY3298176, a novel dual GIP and GLP-1 receptor agonist for the treatment of type 2 diabetes mellitus: from discovery to clinical proof of concept. Mol. Metab. 18, 3–14 (2018).
ARTICLES
⦁ Nauck, M. A. et al. A phase 2, randomized, dose-finding study of the novel once-weekly human GLP-1 analog, semaglutide, compared with placebo and open-label liraglutide in patients with type 2 diabetes. Diabetes Care 39, 231–241 (2016).
⦁ Drucker, D. J. The cardiovascular biology of glucagon-like peptide-1. Cell Metab. 24, 15–30 (2016).
⦁ Lovshin, J. A. et al. Liraglutide promotes natriuresis but does not increase circulating levels of atrial natriuretic peptide in hypertensive subjects with type 2 diabetes. Diabetes Care 38, 132–139 (2015).
⦁ Liakos, A. et al. Effect of liraglutide on ambulatory blood pressure in patients with hypertension and type 2 diabetes: a randomized, double-blind,
placebo-controlled trial. Diabetes Obes. Metab. 21, 517–524 (2019).
⦁ Lastra, G., Syed, S., Kurukulasuriya, L. R., Manrique, C. & Sowers, J. R. Type 2 diabetes mellitus and hypertension: an update. Endocrinol. Metab. Clin. North Am. 43, 103–122 (2014).
⦁ Kjellsson, M. C., Cosson, V. F., Mazer, N. A., Frey, N. & Karlsson, M. O. A model-based approach to predict longitudinal HbA1c, using early phase glucose data from type 2 diabetes mellitus patients after anti-diabetic treatment. J. Clin. Pharmacol. 53, 589–600 (2013).
⦁ Lledó-García, R., Mazer, N. A. & Karlsson, M. O. A semi-mechanistic model of the relationship between average glucose and HbA1c in healthy and diabetic subjects. J. Pharmacokinet. Pharmacodyn. 40, 129–142 (2013).
⦁ Nathan, D. M. et al. Translating the A1C assay into estimated average glucose values. Diabetes Care 31, 1473–1478 (2008).
⦁ American Diabetes Association. 6. Glycemic targets: standards of medical care in diabetes—2020. Diabetes Care 43, S66–S76 (2020).
⦁ Deane, A. M. et al. Endogenous glucagon-like peptide-1 slows gastric emptying in healthy subjects, attenuating postprandial glycemia. J. Clin. Endocrinol. Metab. 95, 215–221 (2010).
⦁ Lean, M. E. J. et al. Tolerability of nausea and vomiting and associations with weight loss in a randomized trial of liraglutide in obese, non-diabetic adults. Int J. Obes. 38, 689–697 (2014).
⦁ Fonseca, V.A. et al. Reductions in insulin resistance are mediated primarily via weight loss in subjects with type 2 diabetes on semaglutide. J. Clin. Endocrinol. Metab. https://doi.org/10.1210/jc.2018-02685⦁ (2019).

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature America, Inc. 2021

NATURE MEDICINE | VOL 27 | JUNE 2021 | 1079–1087 | www.nature.com/naturemedicine 1087

ARTICLES NATURE MEDICINE

Methods
Preclinical evaluations. All animal experiments performed at Pfizer were in accordance with federal, state, local and institutional guidelines governing the use of laboratory animals in research. The procedures used in this study were reviewed and approved by Pfizer’s Institutional Animal Care and Use Committee. The facilities that supported all animal studies are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.
Mice expressing hGLP-1R were generated by genOway. The hGLP-1R gene was inserted by homologous recombination in embryonic stem cells derived from C57Bl/6N mice using a vector containing the hGLP-1R cDNA in frame with mouse Glp1r exon 2. This design preserved the whole mouse intron 1 and kept the endogenous structure of the mouse gene intact. Embryonic stem cells were injected in blastocysts and implanted in pseudo-pregnant females to develop to term. Heterozygous mice carrying the hGLP-1R gene were generated by breeding
chimeric mice with C57BL/6N Cre deleter mice. Insertion of hGLP-1R cDNA into the mouse Glp1r locus was validated by genotyping using PCR and further analyzed by DNA sequencing. Littermate male WT and hGLP-1R knock-in mice on a C57Bl/6N background were generated from heterozygous mating. Mice were housed individually in Innovive cages (Innorack IVC Mouse 3.5) in a temperature- and humidity-controlled environment (22 ± 1 °C), maintained on a 12-h light, 12-h dark cycle with ad libitum access to water and food (LabDiet, PicoLab Rodent Diet 20).
Intraperitoneal glucose tolerance test. Five-month-old male WT (n = 13) and hGLP-1R knock-in (n = 13) mice were acclimated to the laboratory environment for a minimum of 7 d before study. Mice were acclimated to handling and to subcutaneous and intraperitoneal saline injections, for a minimum of 3 d before initiation of the IPGTT. Overnight fasted (14 h) mice were dosed subcutaneously with vehicle (1 MEq NaOH in 2% (v/v) Tween 80 (in 5% (v/v) polyethylene glycol 400 (PEG 400) in 5% (w/v) dextrose) in deionized water) or danuglipron (PF- 06882961) 3 mg kg−1. Fifteen minutes after vehicle or danuglipron injection, mice were administered 40% dextrose at 2 g kg−1 via intraperitoneal injection. Blood glucose was measured via tail nick using a handheld glucometer (AlphaTRAK
2, Zoetis) at 16 and 1 min before and 15, 30, 45, 60, 90 and 120 min after administration of dextrose. To measure plasma insulin, blood was collected in EDTA-coated tubes at 1 min before and 15, 30, 60 and 120 min after dextrose administration. Blood samples were centrifuged (Eppendorf Centrifuge 5417R) to collect plasma. Plasma samples were stored at −80 °C until analysis of plasma insulin. Plasma insulin was measured using ALPCO Mouse Ultrasensitive Insulin ELISA (ALPCO).

Food intake study. Five-month-old male WT (n = 13) and hGLP-1R knock-in
(n = 13) mice were acclimated to handling and subcutaneous saline injections for 4 d before study. Mice were randomized according to body weight and average daily food intake. Body weights were measured using Mettler Toledo XS2002S and food weights were measured using Sartorius Secura 2101-1S. Ninety minutes before the start of the dark cycle (zeitgeber time (ZT) = 10.5), WT and hGLP-1R knock-in mice were dosed subcutaneously with vehicle (1 MEq NaOH in 2% (v/v) Tween 80 (in 5% (v/v) PEG 400 in 5% (w/v) dextrose) in deionized water) or danuglipron (30 mg kg−1). Food weights were measured 2.5, 5.5 and 15.5 h after treatment (ZT = 13, ZT = 16 and ZT = 2, respectively).

Preclinical statistical analyses. For longitudinal data, a linear mixed model was used to fit a model, with glucose and insulin as the response variables, animal as the random effect and treatment, genotype, time and all interactions as fixed effects. A restricted maximum likelihood was used to fit the model with AR(1)
correlation. For nonlongitudinal data, a two-way analysis of variance was used to compare glucose and insulin AUC between genotype and treatment groups. The insulin AUC value for one of the animals was removed from the analysis due to being identified as an influential outlier. The response variables were evaluated for meeting the normality assumption with the Shapiro–Wilk test and Q–Q plots, meeting the homogeneity assumption with the Levene’s test and meeting the sphericity assumption based on the comparison of model Bayesian information criterion with and without inclusion of variance weights. Multiple comparisons
P values were estimated and false discovery rate was used to control the experiment-wise error rate for group comparisons. All analyses were performed with R v.3.5.1.

Clinical study. Oversight. The study was managed by Pfizer Inc. (the sponsor) and conducted by investigators contracted by and under the direction of the sponsor.
This study was conducted in compliance with the ethical principles originating in or derived from the Declaration of Helsinki and in compliance with all
International Conference on Harmonisation Good Clinical Practice Guidelines. A signed and dated informed consent was required from all participants before any

adhering to the study procedures described in the protocol, for keeping records of study treatment and for accurately completing and signing the case report forms supplied by the sponsor. The final protocol and informed consent documentation were reviewed and approved by the central IRB, which was used by each of the investigational centers participating in the study. Investigators were required
to inform the IRB of the study’s progress and occurrence of any serious and/or unexpected AEs.
Study design. This was a randomized, double-blind, sponsor open,
placebo-controlled, multiple dose-escalating study of danuglipron in participants with T2D on a background of metformin monotherapy. As this study administered multiple doses of danuglipron to human participants, the primary objective of the study was to evaluate the safety and tolerability of ascending, multiple, oral doses of danuglipron, with assessment of AEs, safety laboratory tests, vital signs and
12-lead ECGs. Secondary and exploratory objectives included the characterization of the pharmacokinetics and the pharmacodynamic profile of danuglipron.
Pharmacokinetic end points included the plasma pharmacokinetic parameters AUC24, Cmax, Tmax and t½ and urine pharmacokinetic parameters. Pharmacodynamic end points included CFB in response to an MMTT at all post-dose time points for AUC24/24 h glucose (MDG); AUC0–4 for glucose, insulin, glucagon and C-peptide; FPG, FPI and HOMA-IR. These objectives and end points were all prespecified before study initiation. HbA1c was measured prospectively and analyzed as an exploratory safety assessment.
The participants and site staff (except those involved in the preparation of doses) were blinded to administration of active versus placebo in each cohort to permit an unbiased assessment of safety. However, a limited number of sponsor team members were unblinded to permit a real-time review of safety and tolerability to assess potential for treatment induced changes.
Participants received oral doses of danuglipron or placebo in this study.
A total of approximately 12 participants were enrolled in each cohort, with a randomization ratio of 3:1 (9 active and 3 placebo) within each cohort. The study included eight cohorts (Extended Data Fig. 3). danuglipron was administered via the immediate-release (IR) tablet formulation for most of the cohorts. However, a CR formulation was also administered in one cohort (Extended Data Fig. 3).
Participants were admitted to the clinical research unit (CRU) on or before day −2 and were discharged following completion of all assessments on day 30, at investigator discretion. A follow-up visit and a follow-up contact (the latter
typically conducted by phone call) occurred 35–42 d and 56–63 d following the first dose of study treatment on day 1, respectively. Participants were remunerated for their participation in this study.

Study participants. Inclusion criteria for the study population were:
⦁ Concomitant metformin monotherapy as the only anti-hyperglycemic treat- ment. Metformin dose was ≥500 mg d−1 and was stable, defined as no change in the treatment, including dose, for at least 2 months before the screening visit.
⦁ HbA1c ≥ 7.0% and ≤10.5% at screening.
⦁ BMI of 24.5 to 45.4 kg m−2; and a total body weight >50 kg (110 lb).
⦁ Men and women of nonchildbearing potential between the ages of 18 and 70 years, inclusive of age at the time of the screening visit.
⦁ Women of nonchildbearing potential with at least one of the following criteria:
⦁ Achieved postmenopausal status, defined as follows: cessation of regular men- ses for at least 12 consecutive months with no alternative pathological or physi- ological cause and with a serum follicle-stimulating hormone level confirming the postmenopausal state;
⦁ Have undergone a documented hysterectomy and/or bilateral oophorectomy;
⦁ Have medically confirmed ovarian failure.
⦁ All other women (including those with tubal ligations) were considered to be of childbearing potential.
⦁ Evidence of a personally signed and dated informed consent document indicating that the participant had been informed of all pertinent aspects of the study.
⦁ Participants willing and able to comply with all scheduled visits, treatment plan, laboratory tests and other study procedures.
Use of other medications for glycemic control (except metformin) was not permitted in this study, but certain concomitant medications to treat concurrent diseases such as hyperlipidemia and hypertension were permitted.
Exclusion criteria were:
⦁ Evidence or history of clinically significant hematological, renal, endocrine, pulmonary, gastrointestinal, cardiovascular, hepatic, psychiatric, neurological or allergic disease (including drug allergies, but excluding untreated, asympto-

study-specific activity was performed. The Institutional Review Board (IRB) that reviewed and approved this study was Quorum Review Inc.
Three clinical study sites recruited, randomized and dosed participants (Anaheim Clinical Trials, QPS-MRA and Vince and Associates). Recruitment for the study took place between June 2018 and March 2019. The study was completed in June 2019. The investigators at these sites were responsible for
matic, seasonal allergies at the time of dosing).
⦁ Patients who have chronic conditions other than T2D (for example, hypercho- lesterolemia or hypertension) but are controlled by either diet or stable doses of medications could be included (for example, a patient with hypercholester- olemia on appropriate treatment was eligible).
⦁ Participants with any of the following medical conditions:

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE

⦁ Any condition possibly affecting drug absorption (such as previous bariatric surgery, gastrectomy or any area of intestinal resection, active inflammatory bowel disease or pancreatic insufficiency);
⦁ Diagnosis of type 1 diabetes mellitus or secondary forms of diabetes;
⦁ History of myocardial infarction, unstable angina, arterial revascularization, stroke, New York Heart Association Functional Class II–IV heart failure or transient ischemic attack within 6 months of screening;
⦁ Any malignancy not considered cured (except basal cell carcinoma and squa- mous cell carcinoma of the skin); a participant is considered cured if there has been no evidence of cancer recurrence in the previous 5 years;
⦁ History of HIV, hepatitis B or hepatitis C; positive testing for HIV, hepatitis B surface antigen, hepatitis B core antibody or hepatitis C antibody;
⦁ Personal or family history of medullary thyroid carcinoma or multiple endo- crine neoplasia syndrome type 2 or participants with suspected medullary thy- roid carcinoma as per the Principal Investigator’s judgment;
⦁ Acute pancreatitis or history of chronic pancreatitis;
⦁ Acute gallbladder disease.
⦁ At screening, people with an estimated glomerular filtration rate
<60 ml min−1 1.73 m−2 as calculated by the modification of diet in renal disease equation and confirmed via a single repeat, if deemed necessary.
⦁ A positive urine drug test. Participants that have been medically prescribed benzodiazepines and report the use of these drugs to the investigator at the screening visit may be allowed to participate if approved by the sponsor.
⦁ History of regular alcohol consumption exceeding 7 drinks per week for women or 14 drinks per week for men (1 drink = 5 ounces (150 ml) of wine or 12 ounces (360 ml) of beer or 1.5 ounces (45 ml) of hard liquor) within 6 months before screening.
⦁ Administration of an investigational drug within 30 d (or as determined by the local requirement) or five half-lives preceding the first dose of investigational product (whichever is longer).
⦁ Screening supine BP ≥ 160 mm Hg (systolic) or ≥100 mm Hg (diastolic), following at least 5 min of supine rest. If BP is ≥160 mm Hg (systolic) or
≥100 mm Hg (diastolic), the BP should be repeated two more times and the average of the three values used to determine the participant’s eligibility.
⦁ Screening supine 12-lead ECG demonstrating a corrected QT (QTcF) interval
>450 ms or a QRS interval >120 ms. If QTcF exceeds 450 ms or QRS exceeds 120 ms, the ECG should be repeated two more times and the average of the
ARTICLES
interactive response technology system. All cohorts in this study were to complete a total of 28 d of dosing. Cohort 1 was dosed with 15 mg BID or placebo BID for all 28 d with no titration. For cohort 2, there was no scheduled titration and all participants were started on 50 mg BID (or matching placebo) on day 1. However,
if a participant did not tolerate the dose as determined by the principal investigator, down-titration to 15 mg BID (or matching placebo) was permitted, with notification to the sponsor. For the planned cohorts 3, 4 and 5, up to 2 weeks of titration was utilized for dose titration, in addition to at least 14 d of dosing at the target dose level (Supplementary Table 1). For cohorts 6, 7 and 8, scheduled titration schemes were utilized across the full 28-d duration of dosing (Supplementary Table 1).

Study assessments. BP and heart rate were measured in triplicate before dosing and at 1, 2, 4, 8 and 12 h after doing on days −1, 1, 14, 21 and 28. Additional collection times, or changes to collection times, of BP and heart rate were permitted,
as necessary, to ensure appropriate collection of safety data. For body weight measurements, the same scale was used for a particular participant for all body weight measurements obtained at the CRU. All scheduled ECGs were performed after the participant had rested quietly for at least 10 min in a supine position.
When a meal or snack was scheduled at the same time as an ECG, the ECG was performed before the meal or snack. Triplicate 12-lead ECGs were obtained approximately 2–4 min apart; the average of the triplicate ECG measurements collected at each nominal time point on day −1 served as each participant’s
time-controlled baseline QTcF value. ECG data were submitted to a central ECG laboratory for overread measurement of ECG intervals and overall interpretation. The central ECG laboratory was blinded to treatment allocation. The final ECG report from the central laboratory was maintained in the participant’s source documentation and was the final interpretation of the ECG recording. Investigators monitored fasting fingerstick blood glucose using daily glucometer measurements taken before breakfast while the participant was confined to the CRU.
Fasted clinical laboratory tests were performed before the first treatment on days −1, 1, 14, 21 and 28. HbA1c was measured at screening, day −1 and day 28. Unscheduled clinical laboratory measurements were obtained at any time during the study to assess any perceived safety concerns. PPD Central Laboratories was used for the management and analysis of safety laboratory samples.
Pharmacodynamics. Plasma glucose was assessed at 14 time points on days −1, 14 and 28 for calculation of MDG. An MMTT was performed on days −1, 14 and 28, with blood sampling for glucose, insulin, C-peptide and glucagon collected at eight

three QTcF or QRS values used to determine the participant’s eligibility.
⦁ Participants with any of the following abnormalities in clinical laboratory tests
time points up to 4 h following the MMTT. The day −1 MMTT assessments served as the baseline assessment for the pharmacodynamic biomarkers which were

at screening, as assessed by the study-specific laboratory and confirmed by a
reported as AUC over the 4 h (AUC ) following the meal. In addition, FPG and

single repeat test, if deemed necessary:
Aspartate aminotransferase or alanine aminotransferase level ≥1.5 × upper limit of normal (ULN);
Total bilirubin level ≥1.5 × ULN; Fasting C-peptide <0.8 ng ml−1; Thyroid-stimulating hormone > ULN; Serum calcitonin > ULN;
Amylase > ULN or lipase ULN; Blood glucose >270 mg dl−1.
⦁ Fasting fingerstick blood glucose on day −2 of >270 mg dl−1.
⦁ Fertile men who were unwilling or unable to use a highly effective method of contraception as outlined in this protocol for the duration of the study and for the duration of the study and for at least 28 d after the last dose of investiga- tional product.
⦁ Blood donation (excluding plasma donations) of approximately 1 pint (500 ml) or more within 60 d before history of sensitivity to heparin or heparin-induced thrombocytopenia only if heparin is used to flush intravenous catheters.
⦁ Unwilling or unable to comply with the criteria in the lifestyle requirements section of the study protocol.
⦁ Participants who were investigator site staff members directly involved in the conduct of the study and their family members, site staff members otherwise supervised by the investigator or who were Pfizer employees, including their
0–4
FPI levels were assessed and used to calculate HOMA-IR31.
The initial caloric intake/menu assigned to each participant was based on the Harris–Benedict formula (sedentary lifestyle) using the participant’s body weight measured at screening32. On days with MMTT assessments, participants were required to consume all provided food. For breakfast on days with MMTT assessments, a liquid meal consisting of approximately 700 kcal (16 fluid ounces) of Ensure Plus (vanilla) was administered to participants. The timing of blood
samples for assessment of AUC0–4 as part of the MMTT was based on the start time of the liquid meal, such that the 15-min time point was collected approximately
5 min after completion of the liquid meal. In addition, the approximate percentage of food consumed was recorded in the case report form. Participants were not required to consume all provided food during standard meals on other study days.
Venous blood samples were obtained for assessment of plasma glucose, plasma insulin, C-peptide and glucagon and analyzed by a central laboratory (PPD Laboratories). Plasma glucose was measured by enzymatic hexokinase method using a Roche Cobas 8000 c702 analyzer (Roche Diagnostics; CV < 2%). Plasma insulin and C-peptide were measured by ElectroChemiluminescence Immunoassay using a Roche Cobas 8000 e602 instrument (CVs < 4% and <5%, respectively).
Glucagon was measured using an ELISA (Mercodia, Inc.; CV < 9.4%).
Pharmacokinetic assessments. Blood samples were collected at intervals over 24 h after the morning dose on day 1, 14 or 21 and over 48 h following the morning

family members, directly involved in the conduct of the study.
⦁ Other acute or chronic medical or psychiatric condition including recent (within the past year) or active suicidal ideation or behavior or laboratory abnormality that may increase the risk associated with study participation or investigational product administration or may interfere with the interpreta- tion of study results and, in the judgment of the investigator, would make the participant inappropriate for entry into this study.

Administration of study treatment. Danuglipron was supplied as IR or CR tablets for oral administration. Matching placebo tablets were also provided. Study treatment was administered in a blinded fashion for all dosing regimens throughout the study. Dosing occurred with food to standardize administration across participants and cohorts. Dose titration was incorporated to enhance tolerability to higher doses
of danuglipron (Supplementary Table 1). Doses and dose titration schemes were provided to investigators in writing before initiation of dosing in each cohort. Allocation of participants to treatment groups proceeded through the use of an
dose on day 28. Urine was collected over a 24-h period following the morning dose on day 28.
The analyte, danuglipron and internal standard (PF-06974801), were extracted from 100 μl of K2-EDTA human plasma or urine by a liquid/liquid extraction procedure. The compounds were detected and quantified by tandem mass spectrometry in positive ion mode on an AB Sciex API 5000TM equipped with a Turbo Ionspray interface (Syneos Health). Calibration was linear from
0.100 to 100 ng ml−1 using l concentration−2 linear regression. The lower limit of quantification for danuglipron was 0.100 ng ml−1 in plasma and urine.
Pharmacokinetic parameters were calculated using sponsor‐validated electronic noncompartmental analysis software (eNCA, v.2.2.4).
For plasma and urine, interday (between‐day) accuracy was expressed as percent relative error for quality control concentrations and ranged from −2.00% to −0.667% and −23.3% to 4.00%, respectively. Assay precision was expressed as the interday %CV of the mean estimated concentrations of quality control samples and was ≤4.77% and ≤9.13% for plasma and urine, respectively.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE

Statistical analyses. A sample size of 12 participants per multiple ascending-dose cohort (9 active, 3 placebo) was selected to minimize the number of participants exposed to a new chemical entity while permitting adequate characterization of safety, tolerability, pharmacokinetics and pharmacodynamics at each dose level in this phase 1 study.
No formal statistical analyses were planned or conducted for the primary safety analyses. All participants who received at least one dose of study medication were included in the safety analyses. Data were reported in accordance with the sponsor reporting standards, which included summarizing by treatment group, AEs, safety laboratory abnormalities, vital signs and ECG data. The CFB to day 28 for HbA1c was analyzed as an exploratory safety parameter, where it was included in an analysis of covariance model with treatment group as a fixed effect and baseline HbA1c as a covariate. The LS means and LS mean differences from placebo for each treatment group (with 90% CIs) were extracted for tables and plotting.
MDG at each time point was calculated based on an AUC24/24 h comprising all time points that plasma glucose was measured (0, 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 10, 12, 14 and 24 h). The AUC0–4 and AUC24 were calculated using the trapezoidal rule. For baseline glucose AUC24 calculations, if a participant had a missing
baseline (day −1) 24-h glucose value, this value was imputed using the baseline 0-h
glucose value from the same participant. For baseline glucose AUC0–4 and AUC24 calculations, if a participant had a missing baseline 0-h glucose value (day −1), this value was imputed using the baseline 24-h glucose value from the same participant. No imputation was made for any other missing glucose values in the glucose AUC calculations. For other pharmacodynamic end points, AUC0–4 was only calculated in any participant at a particular time point if at least the first, last and 75% of the total number of samples within the given interval was available. No imputation for missing values was conducted for these other pharmacodynamic end points.
MMRM was applied to the CFB on day 14 and day 28 MDG, which included all treatment groups in the model across all randomized participants. The model included treatment, baseline MDG, day, the baseline by day interaction and the day by treatment interaction, with day fitted as a repeated effect and participant as a random effect. An unstructured covariance matrix was used and the Kenward– Roger approximation was used for estimating degrees of freedom for the model parameters. The LS means and LS mean differences from placebo for each day (with 90% CIs) were extracted for tables and plotting.
A similar MMRM model to that described above was applied separately to the CFB to post-dose time points up to and including day 28 for FPG, body
weight, FPI, HOMA-IR, glucose AUC0–4 and insulin AUC0–4. A similar analysis of covariance model to the above was applied separately to the CFB at day 28 for glucagon AUC0–4 and C-peptide AUC0–4.
Statistical significance was predefined as a two-sided P value <0.1. No
adjustment for multiplicity was conducted for any of the statistical analyses as these were only applied to exploratory end points/analyses.
Analysis of pharmacokinetic parameters included all randomized participants who received at least one dose of danuglipron and who had at least one of the pharmacokinetic parameters of interest for danuglipron
calculated. Pharmacokinetic parameters were summarized by matrix (urine or plasma), treatment (differentiating doses, dosing frequencies and formulations were required), study day (day −1, 14, 21 or 28 as applicable) and cohort.
Pharmacokinetic parameters from participants who had down-titrated during the study were not included in the summary statistics presented.
Reporting Summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Upon request and subject to certain criteria, conditions and exceptions (https:// www.pfizer.com/science/clinical-trials/trial-data-and-results), Pfizer will provide
access to individual de-identified participant data from Pfizer-sponsored global interventional clinical studies conducted for medicines, vaccines and medical devices (1) for indications that have been approved in the United States and/or European Union or (2) in programs that have been terminated (development for all indications has been discontinued). Pfizer will also consider requests for the protocol, data dictionary and statistical analysis plan. Data may be requested from Pfizer trials 24 months after study completion. De-identified participant data will be made available to researchers whose proposals meet the research criteria and other conditions and for which an exception does not apply, via a secure portal. To gain access, data requestors must enter into a data access agreement with Pfizer. Source data are provided with this paper.

References
⦁ Shankar, S. S. et al. Standardized mixed-meal tolerance and arginine stimulation tests provide reproducible and complementary measures of β-cell function: results from the Foundation for the National Institutes of Health Biomarkers Consortium Investigative Series. Diabetes Care 39, 1602–1613 (2016).
⦁ Lin, P.-H. et al. Estimation of energy requirements in a controlled feeding trial. Am. J. Clin. Nutr. 77, 639–645 (2003).

Acknowledgements
This study was sponsored by Pfizer Inc. We thank the clinical study participants, the investigators and site coordinators. We thank J. Garren for providing support for nonclinical statistical analysis. Medical writing support, under the direction of the authors, was provided by E. Comeau, CMC Connect, McCann Health Medical
Communications and was funded by Pfizer Inc. in accordance with Good Publication Practice guidelines.

Author contributions
A.R.S. led the design, conduct and analysis of the clinical study, as well as the development of the manuscript. D.A.G. and R.M.E. led the design, conduct and analysis of the mouse studies. D.N.G. and C.B. contributed to the design, conduct and analysis of the clinical study. A.B. contributed to the analysis of the clinical study. A.M.K. contributed to the conceptualization and supervision of the clinical study presented, as well as the writing, review and editing of the manuscript. All authors critically reviewed the manuscript and approved the final draft for submission.

Competing interests
All authors were employees of the sponsor at the time the study was conducted and are also Pfizer shareholders.

Additional information
Extended data is available for this paper at https://doi.org/10.1038/s41591-021-01391-w.
Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41591-021-01391-w.
Correspondence and requests for materials should be addressed to A.R.S.
Peer review information Nature Medicine thanks Michael Nauck, Stephen Bain, Victor Volovici and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Jennifer Sargent was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.
Reprints and permissions information is available at www.nature.com/reprints.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Extended Data Fig. 1 | Liraglutide improved glucose tolerance and reduced food intake in wt and humanized GLP-1R mice. (a, b) IPGTT study in wt and hGLP-1R mice after subcutaneous dosing with vehicle or liraglutide (0.3 mg/kg). Mice were dosed subcutaneously with vehicle or liraglutide 30 minutes prior to intraperitoneal administration of glucose (2 g/kg): (A) blood glucose; and (B) blood glucose AUC0-120. A total of 5 to 9 mice were used in each dose group. (C) Food intake in wt and hGLP-1R mice 15 hours after subcutaneous dosing with vehicle or liraglutide (0.3 mg/kg). A total of 12 to 17 mice were used in each dose group. Data are plotted as mean ± standard error. Plot (A) was analyzed using a mixed-effect model with AR(1) correlation
and variance adjustments, and plots (B) and (C) were analyzed using one-way ANOVA, where **denotes p < 0.01, ***p < 0.001, and ****p < 0.0001. ANOVA, analysis of variance; AR(1), autoregressive model with lag=1; AUC0-120, area under the concentration–time curve from 0 to 120 minutes; hGLP-1R, humanized GLP-1R; IPGTT, intraperitoneal glucose tolerance test; Lira, liraglutide; ns, not significant; wt, wild type.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE

Extended Data Fig. 2 | Danuglipron (PF-06882961) improved glucose tolerance and reduced food intake in an hGLP-1R mouse model. (a–d), IPGTT study in wt and humanized hGLP-1R mice after subcutaneous dosing with vehicle or PF-06882961 (danuglipron) (3 mg/kg): (a) blood glucose; (b) blood glucose AUC0-120; (c) plasma insulin; and (d) plasma insulin AUC0-120. (e) Food intake in wt and hGLP-1R mice 2.5 hours (ZT = 13), 5.5 hours (ZT = 16), and 15.5 hours (ZT = 2) after subcutaneous dosing with vehicle or PF-06882961 (danuglipron) 30 mg/kg. A total of 6 to 7 mice were used in each dose group. Data are plotted as mean ± standard error. Plots (A), (B), and (E) were analyzed using a mixed-effect model with AR(1) correlation, and plots (C) and (D) were analyzed using two-way ANOVA, where *denotes P < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 (p values adjusted for multiple
comparisons). (A) 0 min p = 0.0053; 15 min = <0.0001; 30 min p < 0.001; 45 min p < 0.0001; 60 min p < 0.0001; 90 min p < 0.0001; 120 min p = 0.0003.
(B) ***p = 0.0001. (C) ***p = 0.0005, **p = 0.006. (D) *p = 0.032. (E) 2.5 hours p < 0.0001; 5.5 hours p < 0.0001; 15.5 hours p < 0.0001. ANOVA, analysis of variance; AR(1), autoregressive model with lag=1; AUC0-120, area under the concentration–time curve from 0 to 120 minutes; hGLP-1R, humanized
GLP-1R; IPGTT, intraperitoneal glucose tolerance test; PF-‘2961, PF-06882961 (danuglipron); wt, wild type; ZT, zeitgeber time.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Extended Data Fig. 3 | Study design. *Represents danuglipron (PF-06882961) dose or matching placebo. 12 planned participants per cohort: 9 active and 3 placebo. 28 days of inpatient dosing for all cohorts. Titration occurred over the duration of the 28 days in the 120 mg BID ST, 200 mg QD CR, and 120 mg QD cohorts. BID, twice daily; CR, controlled-release; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE

Extended Data Fig. 4 | Time course of gastrointestinal TEAEs. The plot presents the proportion of participants who experienced (A) diarrhea, (B) nausea, and (C) vomiting at Weeks 1, 2, 3, and 4. BID, twice daily; CR, controlled-release; QD, once daily; ST, slow titration; TEAE, treatment-emergent adverse event.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Extended Data Fig. 5 | Absolute mean values in plasma glucose over 24 h on Day 28 by treatment arm. The plot presents observed means for each treatment arm at each timepoint. BID, twice daily; CR, controlled-release; MDG, mean daily glucose; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE

Extended Data Fig. 6 | See next page for caption.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Extended Data Fig. 6 | Effect of danuglipron (PF-06882961) on change from baseline in fasting plasma insulin and HOMA-IR at Day 28. Danuglipron (PF-06882961) treatment was associated with a variable plasma insulin response, and HOMA-IR improved with the 70 mg BID and 120 mg BID doses.
(a) CFB in FPI. (b) CFB in HOMA-IR. Plots represent LS means and 90% CIs from an MMRM model. Baseline was defined as the value collected at Day -1, 0 hours. HOMA-IR = (FPI x FPG)/405, where the FPI concentration is in μIU/mL and the FPG concentration is in mg/dL. (A) Placebo n = 25; 10 mg n = 9, p = 0.6115; 15 mg n = 8, p = 0.4340; 50 mg n = 8, p = 0.0080; 70 mg n = 9, p = 0.2777; 120 mg BID n = 9, p = 0.9292; 120 mg BID ST n = 9, p = 0.9663; 120 mg QD n = 8, p = 0.2792; 200 mg QD CR n = 7, p = 0.1075. (B) Placebo n = 25; 10 mg n = 9, p = 0.6947; 15 mg n = 8, p = 0.8518; 50 mg
n = 8, p = 0.0287; 70 mg n = 9, p = 0.0787; 120 mg BID n = 9, p = 0.0788; 120 mg BID ST n = 9, p = 0.1089; 120 mg QD n = 8, p = 0.2535; 200 mg QD CR
n = 7, p = 0.5326. *Statistical significance, compared with placebo, was pre-defined as a two-sided p value of <0.1 based on the t-statistic from the MMRM models. BID, twice daily; CFB, change from baseline; CI confidence interval; CR, controlled-release; FPI, fasting plasma insulin; HOMA-IR, homeostatic model assessment for insulin resistance; LS, least squares; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE

Extended Data Fig. 7 | Effect of danuglipron (PF-06882961) on mean body weight and HbA1c at Day 28. a, b, Dosing with danuglipron (PF-06882961) resulted in dose-responsive CFB in body weight and HbA1c after 28 days of PF-06882961 administration. (a) CFB in body weight. (b) CFB in HbA1c.
Data represent LS means and 90% CIs (a) based on an MMRM and (b) based on an ANCOVA model. Baseline was defined as the value collected at Day -1, 0 hours. (a) Placebo n = 25; 10 mg n = 9, p = 0.5769; 15 mg n = 8, p = 0.3949; 50 mg n = 8, p = 0.2491; 70 mg n = 9, p = 0.0023; 120 mg BID n = 9,
p < 0.0001; 120 mg BID ST n = 9, p < 0.0001; 120 mg QD n = 8, p = 0.0008; 200 mg QD CR n = 7, p = 0.0542. (B) Placebo n = 25; 10 mg n = 9, p = 0.0167;
15 mg n = 8, p = 0.0237; 50 mg n = 8, p = 0.0004; 70 mg n = 9, p < 0.0001; 120 mg BID n = 9, p < 0.0001; 120 mg BID ST n = 9, p < 0.0001; 120 mg QD
n = 8, p = 0.0010; 200 mg QD CR n = 7, p = 0.0006. *Statistical significance, compared with placebo, was pre-defined as a two-sided P value of <0.1 based on the t-statistic from the (a) MMRM model or (b) ANCOVA model. ANCOVA, analysis of covariance; BID, twice daily; CFB, change from baseline; CI, confidence interval; CR, controlled-release; HbA1c, glycated hemoglobin; LS, least squares; MMRM, mixed model repeated measures; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE
ARTICLES

Extended Data Fig. 8 | See next page for caption.

NATURE MEDICINE | www.nature.com/naturemedicine

ARTICLES NATURE MEDICINE
Extended Data Fig. 8 | Effect of danuglipron (PF-06882961) on change from baseline at Day 28 in AUC0-4 for glucose, insulin, glucagon, and C-peptide. Danuglipron (PF-06882961) improved glucose AUC0-4 at all dose levels in a dose-responsive manner in 28 days, while insulin AUC0-4 response was variable, with statistically significant lowering at the 120 mg BID and 120 mg QD dose levels. (a) CFB in glucose AUC0-4. (b) CFB in insulin AUC0-4. (c) CFB in C-peptide AUC0-4. (d) CFB in glucagon AUC0-4. All plots show LS means and 90% CIs. Baseline was defined as the value collected at Day -1, 0 hours. (a) Placebo n = 25; 10 mg n = 9, p = 0.0234; 15 mg n = 8, p = 0.0034; 50 mg n = 8, p = 0.0320; 70 mg n = 9, p < 0.0001; 120 mg BID n = 9, p < 0.0001;
120 mg BID ST n = 9, p < 0.0001; 120 mg QD n = 8, p = 0.0004; 200 mg QD CR n = 7, p = 0.0002. (B) Placebo n = 24; 10 mg n = 8, p = 0.1674; 15 mg n = 8,
p = 0.2985; 50 mg n = 8, p = 0.3293; 70 mg n = 8, p = 0.3562; 120 mg BID n = 9, p = 0.0030; 120 mg BID ST n = 9, p = 0.1025; 120 mg QD n = 8, p = 0.0074;
200 mg QD CR n = 7, p = 0.5884. (C) Placebo n = 25; 10 mg n = 9, p = 0.0056; 15 mg n = 8, p = 0.0479; 50 mg n = 8, p = 0.1520; 70 mg n = 8, p = 0.8454;
120 mg BID n = 9, p = 0.2467; 120 mg BID ST n = 9, p = 0.6303; 120 mg QD n = 8, p = 0.0255; 200 mg QD CR n = 7, p = 0.0856. (D) Placebo n = 25; 10 mg
n = 9, p = 0.9681; 15 mg n = 8, p = 0.2274; 50 mg n = 8, p = 0.8698; 70 mg n = 8, p = 0.7941; 120 mg BID n = 9, p = 0.1005; 120 mg BID ST n = 9, p = 0.8800;
120 mg QD n = 8, p = 0.3025; 200 mg QD CR n = 7, p = 0.7519. *Statistical significance, compared with placebo, was pre-defined as a two-sided p value of <0.1 based on the t-statistic from the (a, b) MMRM models or (c, d) ANCOVA models. AUC0-4, area under the concentration–time curve from 0 to 4 hours; BID, twice daily; CI, confidence interval; CR, controlled-release; LS, least squares; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine

NATURE MEDICINE ARTICLES

Extended Data Fig. 9 | Danuglipron (PF-06882961) pharmacokinetic concentration–time profiles at Day 28 (semi-log scale). Danuglipron (PF- 06882961) exposures increased in an approximately dose-proportional manner across BID and QD dosing regimens. The LLOQ was 0.100 ng/mL. Summary statistics were calculated by setting concentration values below the LLOQ to zero. If a participant received a dose that was not assigned based on the randomized titration scheme, the data from that day were not included. BID, twice daily; CR, controlled-release; LLOQ, lower limit of quantification; QD, once daily; ST, slow titration.

NATURE MEDICINE | www.nature.com/naturemedicine