Glucagon-like peptide 1 receptor (GLP-1R) agonists fail to reduce weight or improve glucose control in a sizable minority of individuals with type 2 diabetes. We hypothesized that stimulation of the hypothalamic-pituitary-adrenal (HPA) axis by GLP-1R agonists, thus inducing cortisol secretion, could explain this unresponsiveness to GLP-1R agonists. To assess the effects of GLP-1R agonist treatment on the HPA axis, we selected 10 individuals with type 2 diabetes with (n = 5 women and 5 men) and nine without (n = 4 and 5) an adequate response to GLP-1R agonists and used [68Ga]Ga-NODAGA-exendin-4 positron emission tomography/computed tomography to quantify GLP-1R expression in the pituitary. Oral glucose tolerance and 24-h urinary cortisol excretion were measured in all participants. Pituitary tracer uptake was observed in all participants, with no significant difference between responders and nonresponders. Pituitary tracer uptake correlated with the area under the curve for ACTH, urinary cortisol-to-creatinine ratio, and age. Interestingly, men had higher pituitary tracer uptake than women. In conclusion, this study does not indicate a role for pituitary GLP-1R expression or HPA axis stimulation in the difference in treatment response to GLP-1R agonists among individuals with type 2 diabetes. The findings of substantial pituitary GLP-1R expression and the significant sex differences require further research.

Article Highlights

  • It is unclear why some individuals with type 2 diabetes are unresponsive to treatment with glucagon-like peptide 1 receptor (GLP-1R) agonists, but hypothalamic-pituitary-adrenal (HPA) axis activation could play a role.

  • We used [68Ga]Ga-NODAGA-exendin-4 positron emission tomography/computed tomography to compare pituitary GLP-1R expression between responders and nonresponders to treatment with GLP-1R agonists.

  • Pituitary GLP-1R expression and HPA axis activation did not differ between responders and nonresponders to GLP-1R agonist treatment. In addition, pituitary radiolabeled exendin uptake was markedly higher in men than in women.

  • Further study is required to explain treatment differences and understand sex differences in pituitary radiolabeled exendin uptake.

Glucagon-like peptide 1 receptor (GLP-1R) agonists are widely used for the management of type 2 diabetes (1,2). GLP-1R agonists improve glucose homeostasis by stimulating insulin and suppressing glucagon secretion and promote weight loss by delaying gastric emptying and reducing appetite through neuronal pathways (3). GLP-1R agonists are associated with gastrointestinal adverse effects, such as nausea and diarrhea, which are dose dependent and often diminish over time, but do not explain the treatment effects of GLP-1R agonists (4).

Regardless of adverse effects, there are considerable individual differences in the response to GLP-1R agonist treatment, in that up to 30% of individuals with type 2 diabetes are partly or completely unresponsive to GLP-1R agonist treatment (5–7). Treatment with GLP-1R agonists in these so-called nonresponders does not result in (clinically meaningful) improvements in glucose control or weight loss. Some potential mechanisms behind the heterogeneity in treatment response to GLP-1R agonists have been suggested. Genome-wide association studies have highlighted the importance of genetics in the variability in response to treatment with GLP-1R agonists among individuals with type 2 diabetes (8,9). Furthermore, low β-cell function was associated with reduced response to GLP-1R agonist treatment, particularly with respect to glucose control (10), and women seem to have a greater response in terms of weight loss as compared with men (11). However, the pathophysiologic mechanisms explaining the heterogeneity in treatment response to GLP-1R agonists remain largely unknown. Identifying predictors of individual responses to GLP-1R agonists is important in the context of cost effectiveness and allowance for more personalized therapy.

To noninvasively study GLP-1Rs in humans, the positron emission tomography (PET) tracer [68Ga]Ga-NODAGA-exendin-4 (radiolabeled exendin) was developed, which is mainly used to determine β-cell mass (12) or detect insulin-producing neuroendocrine tumors, called insulinomas (13). In ongoing studies, we observed uptake of radiolabeled exendin in the pituitary of obese individuals, with high interindividual variability (14). GLP-1Rs in the pituitary may be involved in the regulation of the hypothalamus-pituitary-adrenal (HPA) axis, leading to cortisol secretion in the adrenal glands. Cortisol, a glucocorticoid hormone, has a pivotal role in the metabolic system, in which it increases glucose levels, decreases insulin sensitivity, stimulates appetite, increases food intake, and induces weight gain (15–17). Indeed, previous research demonstrated increased cortisol levels after administration of the GLP-1R agonist exendin (18).

We therefore hypothesized that GLP-1R agonists can lead to activation of the HPA axis, with increased cortisol release inducing metabolic unresponsiveness to GLP-1R agonists. To test this hypothesis, we compared GLP-1R expression in the pituitary measured by PET/computed tomography (CT) imaging with radiolabeled exendin between responders and nonresponders to treatment with GLP-1R agonists for type 2 diabetes and simultaneously assessed the stimulatory effect of a GLP-1R agonist on the HPA axis.

Study Participants

Individuals with type 2 diabetes were included at the Radboud University Medical Center (Nijmegen, the Netherlands). All study participants were recruited from the diabetes outpatient clinic of the Radboud University Medical Center (Nijmegen, the Netherlands) and Rijnstate Hospital (Arnhem, the Netherlands).

All participants had used a GLP-1R agonist for at least 12 months, although discontinuation because of treatment failure was allowed. An adequate treatment response was defined as a decrease in HbA1c of ≥5 mmol/mol (0.5%) and/or weight loss of ≥5% of initial body weight within 1 year of GLP-1R agonist treatment. Conversely, nonresponsiveness was defined as a decrease in HbA1c of <5 mmol/mol (0.5%) and no weight loss within 1 year of GLP-1R agonist treatment, despite adequate adherence. In addition, the responsible clinician was consulted to confirm potential participants as nonresponders. General exclusion criteria were pregnancy, breastfeeding, age <18 years, pituitary disorders, and renal or liver dysfunction.

Participants who still used a GLP-1R agonist at the time of inclusion, usually responders, were asked to temporarily discontinue this medication during their participation, for at least three times the dose interval of the particular GLP-1R agonist. All procedures were performed at the Radboud University Medical Center within a period of 1 week. All procedures were approved by the local institutional ethics review committee (Nijmegen, the Netherlands), and all participants provided written informed consent in accordance with the provisions of the Declaration of Helsinki (ClinicalTrials.gov identifier NCT03923114).

Oral Glucose Tolerance Test

Participants underwent a 75-g oral glucose tolerance test (OGTT) to assess β-cell function. They were asked to temporarily halt all diabetes medication 24 h before the test, except for short-acting insulin, which was terminated from midnight. The OGTT was performed in the morning, preceded by an overnight fast. Blood was drawn to determine baseline fasting glucose, C-peptide, insulin, and HbA1c levels, as well as creatinine and liver enzymes (ALT and AST). Subsequently, participants were asked to drink 75 g anhydrous glucose in water within a timeframe of 5 min. Blood samples were collected at 0, 10, 20, 30, 45, 60, 90, 120 and 180 min after the glucose intake to determine glucose, C-peptide and insulin levels.

The areas under curve (AUCs) for glucose, C-peptide, and insulin over the whole course were computed. The AUCC-peptide, the ratio of AUCC-peptide to AUCglucose, and HOMA2-%β were used to assess β-cell function. The ratio of AUCC-peptide to AUCglucose in the first 30 min of the OGTT was calculated to determine first-phase insulin secretion. In addition, insulin sensitivity and insulin resistance were assessed by HOMA2-%S and HOMA2-IR, respectively. The HOMA indices were calculated based on the updated HOMA model (19,20).

HPA Axis Stimulation Test

To assess the stimulatory effect of GLP-1R agonist treatment on the HPA axis, all participants underwent an HPA axis stimulation test. The stimulation test was performed in the morning, preceded by a 4-h fasting period. Before the test, blood was drawn to determine baseline ACTH and cortisol levels. Participants then received a subcutaneous injection of 10 μg exenatide (Byetta; AstraZeneca). Blood samples were collected at 15, 30, 45, 60, 75, 90, 105, 120, 135, 150, 165, 180, 210, and 240 min after the injection to determine ACTH and cortisol levels.

Because we were interested in the effect of both subcutaneously and intravenously injected exendin on the HPA axis, additional blood samples were collected during the PET/CT session at 5, 15, 30, 45, 60, 90, and 120 min after intravenous [68Ga]Ga-NODAGA-exendin-4 injection to determine cortisol and ACTH levels. The AUCs for both subcutaneously and intravenously stimulated ACTH and cortisol levels were computed. In addition, percentage changes in both subcutaneously and intravenously stimulated ACTH and cortisol were calculated after 180 and 60 min, respectively.

To examine a potential difference in baseline cortisol levels, 24-h urine was collected from participants twice, from which urinary-free cortisol and creatinine levels were determined. The ratio of cortisol to creatinine was calculated to correct for the effect of urine concentration.

PET/CT Acquisition

Participants were asked to fast for 4 h and when applicable to halt the use of short-acting insulin 6 h before the scan to prevent hypoglycemia. Blood was drawn to determine baseline cortisol, ACTH, and glucose levels. Additionally, blood pressure was measured before and after the [68Ga]Ga-NODAGA-exendin-4 injection. Dynamic PET images of the head were acquired directly after a slow intravenous bolus injection of 101 ± 7.2 MBq (range 88–122 MBq) [68Ga]Ga-NODAGA-exendin-4 (peptide dose 4–7 μg) over a course of 1 min. The radiolabeled exendin tracer was prepared as previously described (21). The PET/CT scan was performed using the Siemens Biograph 40 mCT time-of-flight PET/CT scanner. During the dynamic PET, venous blood sampling was performed every 20 s between 1 and 3 min and at 5, 15, 30, 45, and 60 min after injection to be used for kinetic analysis. A low-dose CT scan without contrast agent of the head was used for attenuation correction and anatomic identification (transaxial matrix 512 × 512 [0.98 mm voxel size]; CT slide width 3 mm).

The 60-min dynamic PET scan of the head was followed by an additional static PET/CT scan (one bed position; 10 min/bed position) of the upper abdomen to quantify pancreatic tracer uptake as a measure for β-cell mass. Both dynamic and static PET images were reconstructed according to the following settings: three iterations, 21 subsets, and postreconstruction Gaussian filter of 3 mm in full width at half maximum. The PET image size was 200 × 200 (4.07 mm pixel size).

Magnetic Resonance Imaging Acquisition

Five magnetic resonance imaging (MRI)–compatible participants with the highest maximum standardized uptake value (SUVmax) of radiolabeled exendin uptake in the pituitary were invited to undergo an MRI scan of the head to gain more detailed information on the anatomic location of radiolabeled exendin uptake. The MRI scan was performed using a Siemens MAGNETOM Prisma 3T MRI system using a head-neck coil. A T1-weighted image was obtained with three-dimensional magnetization-prepared rapid gradient-echo sequence. For the analysis, the last frame of the dynamic PET scan was fused with the MRI scan. The PET/MRI fusion of the fifth participant was excluded from analysis due to artifacts.

Quantitative Image Analysis

The dynamic PET/CT images were analyzed using PMOD software (version 4.4; PMOD Technologies, Zurich, Switzerland). A spherical volume of interest (VOI) was drawn manually in the pituitary (radius 5 mm). Another VOI was drawn in the confluence of sinuses (radius 5 mm). For the dynamic analysis, a time-activity curve was generated for each VOI to determine the activity concentration (kBq/mL) at each time point. A reconstruction of the last frame of the dynamic PET was generated and fused to the static CT to determine the SUVmax (unitless) in the pituitary by correcting for the injected activity and body weight.

To determine β-cell mass, VOIs were manually drawn in the pancreas, kidneys, and proximal part of the duodenum. VOIs of the pancreas and proximal duodenum were drawn to include all (possible) uptake on PET images of the upper abdomen using the CT as reference in case of misalignment. VOIs of the kidneys were based on the CT images. Because clearance of radiolabeled exendin via the kidneys results in high renal tracer uptake, VOIs of the kidneys were dilated by 8 mm to account for spillover activity. Subsequently, VOIs of the proximal duodenum and kidneys were subtracted from the VOI of the pancreas to correct for a possible overlap in anatomic location, thereby preventing overestimation of the pancreatic tracer uptake. The mean SUV (SUVmean [unitless]) in the spillover-corrected pancreas, representing β-cell mass, was calculated from these VOIs on static PET images.

Statistical Analysis

Data were analyzed using Prism software (version 10.0.0; GraphPad Software, San Diego, CA). Normality tests were performed followed by either an independent t test including Welch correction or Mann-Whitney U test to assess group differences. Mean differences within a group were assessed by one-way ANOVA with post hoc Tukey testing to correct for multiple comparisons. Linearity between variables was checked using Pearson correlation (two tailed). Acquired variables are expressed as mean ± SD, median (interquartile), or n (%). P values <0.05 were considered statistically significant.

The HPA axis stimulation test data of one nonresponder were excluded from analysis because the test revealed very low ACTH levels (<0.3 pmol/L) after stimulation with both subcutaneous and intravenous injections of exendin. In addition, samples for intravenously stimulated ACTH and cortisol levels could not be obtained in all participants because of a lack of space in the PET/CT scanner; therefore, these data were limited.

Data and Resource Availability

The data sets generated and analyzed during the current study are available from the corresponding author upon reasonable request. No applicable resources were generated or analyzed during the current study.

We included 19 individuals with type 2 diabetes, 10 of whom had responded well to GLP1-R agonist treatment (responders) and nine of whom had not had an adequate treatment response (nonresponders) according to the inclusion criteria. One nonresponder was excluded from analysis because of high mucosal tracer uptake in the nose, disturbing the quantification of tracer uptake in the pituitary region. Baseline participant characteristics are listed in Table 1, and information regarding responsiveness to GLP-1R agonist treatment can be found in Supplementary Table 1. The two groups did not differ with respect to sex, age, BMI, or diabetes duration. β-Cell function as reflected by baseline fasting C-peptide levels was significantly higher in responders compared with nonresponders (P = 0.045), whereas β-cell function calculated by HOMA2 was similar between both groups (Table 1).

Table 1

Participant characteristics

VariableResponders
(n = 10)
Nonresponders
(n = 8)
P
GLP-1R agonist type   — 
 Exenatide — 
 Liraglutide — 
 Semaglutide — 
Effects of GLP-1R agonist treatment   — 
 Decrease in HbA1c, mmol/mol (%) 25 ± 10 (2.3 ± 0.9) 0 ± 6.3 (0 ± 0.6) <0.0001 
 Weight loss, % 7.8 (2.1–9.3) 0 (0–0.3) 0.0054 
Female sex 5 (50) 4 (50) — 
Age, years 57.2 ± 9.8 62.5 ± 9.8 0.27 
BMI, kg/m² 32 (29–35) 33 (29–39) 0.78 
Weight, kg 104 ± 18 103 ± 19 0.90 
Height, cm 176 ± 8 176 ± 11 0.89 
Diabetes duration, years 11.1 ± 7.6 13.6 ± 5.7 0.43 
Fasting glucose, mmol/L 11.5 ± 3.2 14 ± 4.0 0.27 
HbA1c, mmol/mol (%) 57 ± 16 (7.4 ± 1.4) 66 ± 8 (8.2 ± 0.7) 0.14 
Fasting C-peptide, nmol/L 1.40 ± 0.49 0.86 ± 0.51 0.045 
Insulin, units/L 27 (25–53) 154 (18–449) 0.23 
HOMA2-%S, %* 26 (21–34) 32 (27–79) 0.21 
HOMA2-IR* 3.8 ± 1.1 2.9 ± 1.7 0.23 
HOMA2-%β, %* 58 ± 35 32 ± 17 0.06 
Creatinine, μmol/L 74 ± 21 74 ± 21 0.99 
ALT, units/L 33 ± 19 28 ± 6 0.41 
AST, units/L 25 ± 9 23 ± 6 0.45 
VariableResponders
(n = 10)
Nonresponders
(n = 8)
P
GLP-1R agonist type   — 
 Exenatide — 
 Liraglutide — 
 Semaglutide — 
Effects of GLP-1R agonist treatment   — 
 Decrease in HbA1c, mmol/mol (%) 25 ± 10 (2.3 ± 0.9) 0 ± 6.3 (0 ± 0.6) <0.0001 
 Weight loss, % 7.8 (2.1–9.3) 0 (0–0.3) 0.0054 
Female sex 5 (50) 4 (50) — 
Age, years 57.2 ± 9.8 62.5 ± 9.8 0.27 
BMI, kg/m² 32 (29–35) 33 (29–39) 0.78 
Weight, kg 104 ± 18 103 ± 19 0.90 
Height, cm 176 ± 8 176 ± 11 0.89 
Diabetes duration, years 11.1 ± 7.6 13.6 ± 5.7 0.43 
Fasting glucose, mmol/L 11.5 ± 3.2 14 ± 4.0 0.27 
HbA1c, mmol/mol (%) 57 ± 16 (7.4 ± 1.4) 66 ± 8 (8.2 ± 0.7) 0.14 
Fasting C-peptide, nmol/L 1.40 ± 0.49 0.86 ± 0.51 0.045 
Insulin, units/L 27 (25–53) 154 (18–449) 0.23 
HOMA2-%S, %* 26 (21–34) 32 (27–79) 0.21 
HOMA2-IR* 3.8 ± 1.1 2.9 ± 1.7 0.23 
HOMA2-%β, %* 58 ± 35 32 ± 17 0.06 
Creatinine, μmol/L 74 ± 21 74 ± 21 0.99 
ALT, units/L 33 ± 19 28 ± 6 0.41 
AST, units/L 25 ± 9 23 ± 6 0.45 

Normally distributed values are given as mean ± SD, nonnormally distributed values as median (interquartile range), and dichotomous values as n (%). Bold font indicates significant association.

*Measured by HOMA2 calculator.

β-Cell function derived from the OGTT, reflected by the AUCC-peptide and the ratio of AUCC-peptide to AUCglucose, did not differ between responders and nonresponders (Supplementary Fig. 1). In addition, the β-cell mass, as determined by quantitative analysis of the radiolabeled exendin PET/CT scan of the pancreas, was similar between responders and nonresponders (SUVmean 4.3 [3.5–6.2] vs. 3.5 [2.8–4.8]; Mann-Whitney U test P = 0.22) (Supplementary Fig. 2).

Two of the eight nonresponders did not undergo the PET/CT scan, one because of serious claustrophobia during the PET/CT scan leading to termination of the scan and the other because of personal reasons unrelated to the study.

The uptake of radiolabeled exendin in the pituitary, expressed as the SUVmax, did not significantly differ between responders and nonresponders (5.2 ± 1.9 vs. 5.5 ± 3.5, unpaired t test P = 0.85) (Fig. 1A), although there was substantial interindividual variation in pituitary tracer uptake (SUVmax range 2.7–12) (Fig. 1B and C and Supplementary Fig. 3). PET/MRI analyses revealed that the tracer uptake in the pituitary was clearly located in the anterior part of the pituitary in two of the four participants undergoing an MRI scan for anatomic correction (Fig. 2A and B). In one participant, differentiation between the posterior and anterior parts of the pituitary was not possible (Fig. 2C), whereas the tracer uptake overlapped in both lobes of the pituitary in another participant (Fig. 2D).

Figure 1

A: SUVmax of the tracer uptake in pituitary between responders and nonresponders (unpaired t test P = 0.85). B and C: [68Ga]Ga-NODAGA-exendin-4 PET/CT image of head of participant with high (B) and low (C) radiolabeled exendin tracer uptake in pituitary, indicated by white arrow.

Figure 1

A: SUVmax of the tracer uptake in pituitary between responders and nonresponders (unpaired t test P = 0.85). B and C: [68Ga]Ga-NODAGA-exendin-4 PET/CT image of head of participant with high (B) and low (C) radiolabeled exendin tracer uptake in pituitary, indicated by white arrow.

Close modal
Figure 2

PET fused to MRI for four participants (sagittal view). Blue and green arrows depict anterior and posterior parts of pituitary shown by hyperintense area, respectively. Uptake of radiolabeled exendin, shown in bright color, is located toward the anterior part of pituitary in two of four participants (A and B).

Figure 2

PET fused to MRI for four participants (sagittal view). Blue and green arrows depict anterior and posterior parts of pituitary shown by hyperintense area, respectively. Uptake of radiolabeled exendin, shown in bright color, is located toward the anterior part of pituitary in two of four participants (A and B).

Close modal

The HPA axis stimulation test did not reveal differences in ACTH or cortisol level between responders and nonresponders on stimulation by subcutaneously injected exenatide (AUCACTH s.c. 1,141 ± 486 vs. 1,366 ± 626; P = 0.45; AUCcortisol s.c. 69 ± 18 vs. 66 ± 11; P = 0.72) (Fig. 3A–D) or intravenously injected radiolabeled exendin (AUCACTH i.v. 530 [360–816] vs. 745 [415–966]; P = 0.71; AUCcortisol i.v. 35 ± 14 vs. 35 ± 10; P = 0.93) (Fig. 3E–H). Furthermore, there were no differences in ACTH or cortisol between groups at specific time points of the stimulation test. In addition, percentage changes in subcutaneously and intravenously stimulated ACTH and cortisol after 180 and 60 min, respectively, were similar between responders and nonresponders (Supplementary Fig. 4), and the urinary cortisol-to-creatinine ratio did not differ between the groups (P = 0.76) (Supplementary Fig. 5).

Figure 3

HPA axis stimulation test revealed similar levels in AUCs for ACTH and cortisol between responders and nonresponders on stimulation with subcutaneously injected exenatide (unpaired t test P = 0.45 and P = 0.72, respectively) (AD) or intravenously injected exendin (Mann-Whitney U test P = 0.71 and unpaired t test P = 0.93, respectively) (EH).

Figure 3

HPA axis stimulation test revealed similar levels in AUCs for ACTH and cortisol between responders and nonresponders on stimulation with subcutaneously injected exenatide (unpaired t test P = 0.45 and P = 0.72, respectively) (AD) or intravenously injected exendin (Mann-Whitney U test P = 0.71 and unpaired t test P = 0.93, respectively) (EH).

Close modal

The uptake of radiolabeled exendin in the pituitary correlated significantly with the AUCACTH after subcutaneous stimulation with exenatide (Pearson correlation r = 0.61; P = 0.015), but not after intravenously injected radiolabeled exendin (Fig. 4A and D). Neither the subcutaneously nor the intravenously stimulated AUCcortisol correlated with the radiolabeled exendin uptake in the pituitary (Fig. 4B and E). However, the urinary cortisol-to-creatinine ratio inversely correlated with the tracer uptake in the pituitary (r = −0.55; P = 0.026) (Fig. 4C).

Figure 4

AC: Correlations of SUVmax in pituitary with subcutaneously stimulated AUCACTH s.c. (Pearson correlation r = 0.61; P = 0.015) (A), AUCcortisol s.c. (r = −0.14; P = 0.62) derived from HPA axis stimulation test (B), and urinary cortisol-to-creatine ratio (r = 0.55; P = 0.026) (C). D and E: SUVmax did not correlate with intravenously stimulated AUCACTH i.v. (r = 0.44; P = 0.14) (D) or AUCcortisol i.v. (r = −0.11; P = 0.72) (E). Green and red dots represent responders and nonresponders, respectively.

Figure 4

AC: Correlations of SUVmax in pituitary with subcutaneously stimulated AUCACTH s.c. (Pearson correlation r = 0.61; P = 0.015) (A), AUCcortisol s.c. (r = −0.14; P = 0.62) derived from HPA axis stimulation test (B), and urinary cortisol-to-creatine ratio (r = 0.55; P = 0.026) (C). D and E: SUVmax did not correlate with intravenously stimulated AUCACTH i.v. (r = 0.44; P = 0.14) (D) or AUCcortisol i.v. (r = −0.11; P = 0.72) (E). Green and red dots represent responders and nonresponders, respectively.

Close modal

Radiolabeled exendin uptake in the pituitary positively correlated with age (r = 0.58; P = 0.019) (Fig. 5A), but not with BMI or diabetes duration (Fig. 5B and C). In addition, we did not observe an association between tracer uptake in the pituitary and β-cell function derived from the OGTT, as reflected by the AUCC-peptide, the ratio of AUCC-peptide to AUCglucose, and HOMA2-%β (Supplementary Fig. 6).

Figure 5

Correlations of SUVmax in pituitary with age (Pearson correlation r = 0.58; P = 0.019) (A), BMI (r = 0.38; P = 0.15) (B), and duration of type 2 diabetes (r = −0.069; P = 0.80) (C). Green and red dots represent responders and nonresponders, respectively.

Figure 5

Correlations of SUVmax in pituitary with age (Pearson correlation r = 0.58; P = 0.019) (A), BMI (r = 0.38; P = 0.15) (B), and duration of type 2 diabetes (r = −0.069; P = 0.80) (C). Green and red dots represent responders and nonresponders, respectively.

Close modal

Although the pituitary tracer uptake did not differ between responders and nonresponders, there was a clear sex difference, with men having significantly higher uptake of radiolabeled exendin in the pituitary than women (SUVmax 6.7 ± 2.7 vs. 3.9 ± 1.2; unpaired t test P = 0.024) (Fig. 6A). Furthermore, we observed that men showed significantly higher AUCACTH s.c. compared with women (1,694 ± 712 vs. 878 ± 400; P = 0.013), whereas the AUCcortisol both after subcutaneously and intravenously injected exendin tended to be lower for men (P = 0.17 and P = 0.070, respectively) (Fig. 6B–E). The urinary cortisol-to-creatinine ratio was significantly lower in men compared with women (2.4 ± 0.8 vs. 5.9 ± 1.9; P = 0.00070) (Fig. 6F).

Figure 6

Sex differences in radiolabeled exendin uptake and HPA axis hormones. A: Men showed higher SUVmax of tracer uptake in pituitary compared with women (unpaired t test P = 0.024). BE: HPA axis stimulation test revealed higher AUC for subcutaneously stimulated ACTH in men compared with women (unpaired t test P = 0.013) (B) but similar AUC for subcutaneously stimulated cortisol (unpaired t test P = 0.17) (C), intravenously stimulated ACTH (unpaired t test P = 0.24) (D), and intravenously stimulated cortisol (unpaired t test P = 0.070) (E). F: Urinary cortisol-to-creatine ratio was higher in women compared with men (unpaired t test P = 0.00070).

Figure 6

Sex differences in radiolabeled exendin uptake and HPA axis hormones. A: Men showed higher SUVmax of tracer uptake in pituitary compared with women (unpaired t test P = 0.024). BE: HPA axis stimulation test revealed higher AUC for subcutaneously stimulated ACTH in men compared with women (unpaired t test P = 0.013) (B) but similar AUC for subcutaneously stimulated cortisol (unpaired t test P = 0.17) (C), intravenously stimulated ACTH (unpaired t test P = 0.24) (D), and intravenously stimulated cortisol (unpaired t test P = 0.070) (E). F: Urinary cortisol-to-creatine ratio was higher in women compared with men (unpaired t test P = 0.00070).

Close modal

The main findings of this study are that the uptake of radiolabeled exendin in the pituitary of individuals with type 2 diabetes did not differ between responders and nonresponders to GLP-1R agonist treatment. In agreement, administration of the GLP-1R agonist exenatide did not differently affect the release of the two HPA hormones ACTH and cortisol between the two groups. These data therefore argue against activation of the HPA axis to explain the heterogeneity in the treatment response to GLP-1R agonists.

The current study aimed to determine the role of the HPA axis in the variability of GLP-1R agonist treatment effects. Previous studies conducted in both animals and humans have suggested a stimulatory role of the GLP-1 hormone and GLP-1R agonists regarding the HPA axis (18,22–24). Our results contrast with these findings but are in agreement with a recent study showing that long-term GLP-1R agonist (dulaglutide) exposure in healthy individuals did not activate the HPA axis (25). Although the HPA hormones and GLP-1R expression were similar between responders and nonresponders, radiolabeled exendin uptake in the pituitary was associated with the AUCACTH s.c., which implies a link between ACTH and GLP-1Rs in the pituitary. Additionally, analyses of the acquired PET/MRI images revealed that the tracer uptake was localized toward the anterior lobe of the pituitary in two of the four participants who underwent MRI for anatomic correlation, which requires replication in a larger sample size. Gil-Lozano et al. (18) suggested that the increase in ACTH after exendin-4 and GLP-1 administration may be mediated directly via the pituitary or by activation via the central nervous system. The first theory is in line with our results showing GLP-1R expression in the pituitary with a possible localization toward the anterior lobe (18). Conversely, the urinary cortisol-to-creatinine ratio was inversely associated with the tracer uptake in the pituitary. Apart from cortisol secretion, ACTH is also involved in aldosterone and androgen secretion from the adrenal glands. However, a previous study reported that exendin-4 desensitized GLP-1Rs in the hypothalamus and pituitary and lead to GLP-1R–mediated aldosterone and corticosterone secretion independent of ACTH (26). More research is required to understand the possible link between pituitary GLP-1Rs and ACTH.

Many studies have been conducted to determine factors predicting the response to GLP-1R agonists, focusing on glycemic control markers, liver enzymes, gut microbiota, and genetics (5,6,10,27–29). Recent genome-wide association studies identified loci in the GLP-1R and ARRB1 genes associated with HbA1c reduction after GLP1-R agonist treatment (9) and demonstrated the importance of coding variation in the GLP-1R gene in the differences in treatment response (8). In addition to genetic susceptibility, Jones et al. (10) highlighted in their analysis with 546 participants that diabetes duration and fasting C-peptide are other factors that may influence the effectiveness of GLP-1R agonists. In contrast, almost all markers of glycemic control and diabetes duration were similar between responders and nonresponders in our current study. The baseline fasting C-peptide, HOMA2-%β, and the ratio of AUCC-peptide to AUCglucose in the first 30 min of the OGTT might suggest a trend toward an impaired β-cell function in nonresponders compared with responders. However, the OGTT-derived measure of β-cell function revealed no significant differences between the groups. Altogether, it remains unclear how to predict treatment response to GLP-1R agonists.

We found a clear difference in the uptake of radiolabeled exendin in the pituitary between men and women. We also observed a higher AUCACTH s.c. in men compared with women, associated with the higher radiolabeled exendin uptake in the pituitary. Preclinical studies have suggested differential GLP-1R expression between sexes, comparable to our study (30,31). However, there are limited data about GLP-1R expression in the human brain and pituitary. Although not observed in our study, women reportedly have a better treatment response to GLP-1R agonists with respect to body weight (11). The estrogen hormone has a prominent role in energy expenditure and feeding behavior (32,33). Estrogen levels were more pronounced after GLP-1R agonist administration in female mice, providing a possible explanation for sex differences in weight management after GLP-1R agonist use (34). However, a GLP-1–estrogen conjugate modulates obesity, hyperglycemia, and dyslipidemia in mice sex independently (33). Although the reason for the higher uptake of radiolabeled exendin and higher ACTH levels in male participants remains unknown, our results provide further support for the interaction between GLP-1Rs, ACTH, and possibly gonadal hormones. Sex-specific dynamics of GLP-1R expression seem to exist in the brain, although these studies are mainly based on animal models. Our study is the first prospective in vivo human study assessing GLP-1R expression in the pituitary. More research is required to explain the sex differences in GLP-1R expression in the pituitary.

Our study has some strengths and limitations. This study used [68Ga]Ga-NODAGA-exendin-4 PET/CT scanning, an advanced technique to determine β-cell mass and GLP-1R expression. We applied this technique to measure pituitary GLP-1R expression in humans and combined it with measurements of HPA axis activation. Another strength of this study is that responders and nonresponders were well matched for important clinical parameters, including sex, BMI, diabetes duration, β-cell function, and β-cell mass. Inherent in the use of a costly imaging technique, a limitation of the study is the small sample size, translating to limited statistical power. For example, we found a positive correlation between radiolabeled exendin uptake in the pituitary and age, which should be interpreted with caution. Despite the relatively small sample size, we were able to create a unique homogeneous group in terms of diabetes management. In most participants, nonresponder status was determined based on liraglutide treatment. It may be that newer generations of GLP-1R agonists show lower nonresponse rates. Nevertheless, the two groups in this study represent the more extreme ends of response rate; therefore, it is likely that these groups will also respond differentially to newer GLP-1R agonists. Another important point includes the acquisition time of the PET scan. This was the first prospective trial studying radiolabeled exendin uptake in the pituitary dynamically. The pituitary is a complex structure as a result of its adjacent arteries and veins and its small size. We observed that the tracer was not completely cleared from blood at 60 min after radiolabeled exendin tracer injection, which could lead to partial volume effects in the VOI of the pituitary, especially in individuals with low tracer uptake. Therefore, allowing a longer incubation time before obtaining static images may lead to improved quantification of pituitary uptake resulting from lower blood activity (higher target-to-background ratio) with less spillover activity.

In summary, this study shows that GLP-1R expression–dependent activation of the HPA axis does not explain the heterogeneity in treatment response to GLP-1R agonists in individuals with type 2 diabetes. We did not observe significant differences in ACTH and cortisol levels or radiolabeled exendin uptake in the pituitary between responders and nonresponders. The findings of substantial pituitary GLP-1R expression and the significant sex differences require further research.

Clinical trial reg. no. NCT03923114, clinicaltrials.gov

This article contains supplementary material online at https://doi.org/10.2337/figshare.27764883.

Acknowledgments. The authors thank the technicians from the Department of Medical Imaging (Radboud University Medical Center) for their assistance during the PET/CT imaging; the laboratory technicians from the Department of Medical Imaging for their efforts in the radiolabeling process; and all the diabetes consultants from the Department of Internal Medicine (Radboud University Medical Center and Rijnstate Hospital) because they proactively contributed to the inclusion of the participants. The authors also express their greatest gratitude to all the study participants for their valuable contribution.

Funding. This research was funded by ZonMw and Diabetes Fonds under project 459001019 (Imaging Pituitary Activation by Exendin [iPAVE]).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.T. was responsible for the data analysis, generation of all figures/tables, writing of the original draft, and editing of the manuscript. S.T., M.B., and T.J.P.J. contributed to the inclusion of all study participants and performed the study procedures. M.B., T.J.P.J., R.M., C.F., A.C.v.B., C.J.T., B.E.d.G., and M.G. reviewed and edited the manuscript and contributed to data discussion. M.B., B.E.d.G., and M.G. were responsible for the study conceptualization and methodology. M.B. and M.G. contributed to the investigation, supervision, and validation. S.T., M.B., and M.G. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. This work was presented as an oral presentation at the Annual Dutch Diabetes Research Meeting, Wageningen, the Netherlands, 2–3 November 2023, and the Annual Meeting for North Europe Young Diabetologists, Stratford-upon-Avon, U.K., 22–24 May 2024; as a short oral discussion at the Annual Meeting of the European Association for the Study of Diabetes, Hamburg, Germany, 2–6 October 2023; and on an e-poster at the Annual Congress of the European Association of Nuclear Medicine, Vienna, Austria, 9–13 September 2023.

1.
Davies
MJ
,
Aroda
VR
,
Collins
BS
, et al
.
Management of hyperglycaemia in type 2 diabetes, 2022. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD)
.
Diabetologia
2022
;
65
:
1925
1966
2.
American Diabetes Association
.
Standards of Care in Diabetes—2023. Abridged for primary care providers
[
published correction appears in Clin Diabetes 2023;41:328
].
Clin Diabetes
2022
;
41
:
4
31
3.
Nauck
MA
,
Quast
DR
,
Wefers
J
,
Meier
JJ.
GLP-1 receptor agonists in the treatment of type 2 diabetes: state-of-the-art
.
Mol Metab
2021
;
46
:
101102
4.
Bettge
K
,
Kahle
M
,
Abd El Aziz
MS
,
Meier
JJ
,
Nauck
MA.
Occurrence of nausea, vomiting and diarrhoea reported as adverse events in clinical trials studying glucagon-like peptide-1 receptor agonists: a systematic analysis of published clinical trials
.
Diabetes Obes Metab
2017
;
19
:
336
347
5.
Tsai
C-Y
,
Lu
H-C
,
Chou
Y-H
, et al
.
Gut microbial signatures for glycemic responses of GLP-1 receptor agonists in type 2 diabetic patients: a pilot study
.
Front Endocrinol (Lausanne)
2021
;
12
:
814770
6.
Kyriakidou
A
,
Kyriazou
AV
,
Koufakis
T
, et al
.
Clinical and genetic predictors of glycemic control and weight loss response to liraglutide in patients with type 2 diabetes
.
J Pers Med
2022
;
12
:
424
7.
Khan
M
,
Ouyang
J
,
Perkins
K
,
Nair
S
,
Joseph
F.
Determining predictors of early response to exenatide in patients with type 2 diabetes mellitus
.
J Diabetes Res
2015
;
2015
:
162718
8.
Lagou
V
,
Jiang
L
,
Ulrich
A
, et al.;
Meta-Analysis of Glucose and Insulin-Related Traits Consortium (MAGIC)
.
GWAS of random glucose in 476,326 individuals provide insights into diabetes pathophysiology, complications and treatment stratification
.
Nat Genet
2023
;
55
:
1448
1461
9.
Dawed
AY
,
Mari
A
,
Brown
A
, et al.;
DIRECT Consortium
.
Pharmacogenomics of GLP-1 receptor agonists: a genome-wide analysis of observational data and large randomised controlled trials
.
Lancet Diabetes Endocrinol
2023
;
11
:
33
41
10.
Jones
AG
,
McDonald
TJ
,
Shields
BM
, et al.;
PRIBA Study Group
.
Markers of β-cell failure predict poor glycemic response to GLP-1 receptor agonist therapy in type 2 diabetes
.
Diabetes Care
2016
;
39
:
250
257
11.
Rentzeperi
E
,
Pegiou
S
,
Koufakis
T
,
Grammatiki
M
,
Kotsa
K.
Sex differences in response to treatment with glucagon-like peptide 1 receptor agonists: opportunities for a tailored approach to diabetes and obesity care
.
J Pers Med
2022
;
12
:
454
12.
Boss
M
,
Buitinga
M
,
Jansen
TJP
,
Brom
M
,
Visser
EP
,
Gotthardt
M.
PET-based human dosimetry of 68Ga-NODAGA-exendin-4, a tracer for β-cell imaging
.
J Nucl Med
2020
;
61
:
112
116
13.
Boss
M
,
Rottenburger
C
,
Brenner
W
, et al
.
68Ga-NODAGA-exendin-4 PET/CT improves the detection of focal congenital hyperinsulinism
.
J Nucl Med
2022
;
63
:
310
315
14.
Deden
LN
,
Booij
J
,
Grandjean
J
, et al
.
Brain imaging of the GLP-1 receptor in obesity using 68Ga-NODAGA-exendin-4 PET
.
Brain Sci
2021
;
11
:
1647
15.
Merabet
N
,
Lucassen
PJ
,
Crielaard
L
, et al
.
How exposure to chronic stress contributes to the development of type 2 diabetes: a complexity science approach
.
Front Neuroendocrinol
2022
;
65
:
100972
16.
George
SA
,
Khan
S
,
Briggs
H
,
Abelson
JL.
CRH-stimulated cortisol release and food intake in healthy, non-obese adults
.
Psychoneuroendocrinology
2010
;
35
:
607
612
17.
Joseph
JJ
,
Golden
SH.
Cortisol dysregulation: the bidirectional link between stress, depression, and type 2 diabetes mellitus
.
Ann N Y Acad Sci
2017
;
1391
:
20
34
18.
Gil-Lozano
M
,
Pérez-Tilve
D
,
Alvarez-Crespo
M
, et al
.
GLP-1(7-36)-amide and exendin-4 stimulate the HPA axis in rodents and humans
.
Endocrinology
2010
;
151
:
2629
2640
19.
Levy
JC
,
Matthews
DR
,
Hermans
MP.
Correct homeostasis model assessment (HOMA) evaluation uses the computer program
.
Diabetes Care
1998
;
21
:
2191
2192
20.
University of Oxford
.
HOMA calculator
. Accessed 2 January 2024. Available from https://www.dtu.ox.ac.uk/homacalculator/
21.
Tokgöz
S
,
Boss
M
,
Prasad
S
, et al
.
Protocol for clinical GLP-1 receptor PET/CT imaging with [68Ga]Ga-NODAGA-exendin-4
.
Methods Mol Biol
2023
;
2592
:
143
153
22.
Heinla
K
,
Vasar
E
,
Sedman
T
,
Volke
V.
A GLP-1 receptor agonist inhibits aldosterone release in healthy volunteers
.
Horm Metab Res
2021
;
53
:
402
407
23.
Gil-Lozano
M
,
Romaní-Pérez
M
,
Outeiriño-Iglesias
V
, et al
.
Effects of prolonged exendin-4 administration on hypothalamic-pituitary-adrenal axis activity and water balance
.
Am J Physiol Endocrinol Metab
2013
;
304
:
E1105
E1117
24.
Hsu
C-C
,
Cheng
J-T
,
Hsu
PH
,
Li
Y
,
Cheng
K-C.
Bolus injection of liraglutide raises plasma glucose in normal rats by activating glucagon-like peptide 1 receptor in the brain
.
Pharmaceuticals (Basel)
2022
;
15
:
904
25.
Winzeler
B
,
da Conceição
I
,
Refardt
J
,
Sailer
CO
,
Dutilh
G
,
Christ-Crain
M.
Effects of glucagon-like peptide-1 receptor agonists on hypothalamic-pituitary-adrenal axis in healthy volunteers
.
J Clin Endocrinol Metab
2019
;
104
:
202
208
26.
Malendowicz
LK
,
Neri
G
,
Nussdorfer
GG
,
Nowak
KW
,
Zyterska
A
,
Ziolkowska
A.
Prolonged exendin-4 administration stimulates pituitary-adrenocortical axis of normal and streptozotocin-induced diabetic rats
.
Int J Mol Med
2003
;
12
:
593
596
27.
Gimeno-Orna
JA
,
Verdes-Sanz
G
,
Borau-Maorad
L
,
Campos-Fernández
J
,
Lardiés-Sánchez
B
,
Monreal-Villanueva
M.
Baseline ALT levels as a marker of glycemic response to treatment with GLP-1 receptor agonists
.
Endocrinología y Nutrición (English Edition)
2016
;
63
:
164
170
28.
Imai
K
,
Tsujimoto
T
,
Goto
A
, et al
.
Prediction of response to GLP-1 receptor agonist therapy in Japanese patients with type 2 diabetes
.
Diabetol Metab Syndr
2014
;
6
:
110
29.
Formichi
C
,
Fignani
D
,
Nigi
L
, et al
.
Circulating microRNAs signature for predicting response to GLP1-RA therapy in type 2 diabetic patients: a pilot study
.
Int J Mol Sci
2021
;
22
:
9454
30.
Díaz-Megido
C
,
Thomsen
M.
Sex-dependent divergence in the effects of GLP-1 agonist exendin-4 on alcohol reinforcement and reinstatement in C57BL/6J mice
.
Psychopharmacology (Berl)
2023
;
240
:
1287
1298
31.
Hofmann
SM
,
Zhang
S
,
Tom
R
,
Müller
TD
,
Tschöp
MH.
Sex-specific dynamics of GLP-1R and GIPR expression in the hypothalamus and the amygdala of mice
.
Diabetes
2023
;
72
(
Suppl. 1
):
69-LB
32.
Xu
Y
,
Nedungadi
TP
,
Zhu
L
, et al
.
Distinct hypothalamic neurons mediate estrogenic effects on energy homeostasis and reproduction
.
Cell Metab
2011
;
14
:
453
465
33.
Finan
B
,
Yang
B
,
Ottaway
N
, et al
.
Targeted estrogen delivery reverses the metabolic syndrome
.
Nat Med
2012
;
18
:
1847
1856
34.
Richard
JE
,
Anderberg
RH
,
López-Ferreras
L
,
Olandersson
K
,
Skibicka
KP.
Sex and estrogens alter the action of glucagon-like peptide-1 on reward
.
Biol Sex Differ
2016
;
7
:
6
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.