Scientists were chasing an incretin hormone, and when GLP-1 was finally discovered, we found that it had a pronounced satiety effect, slowed down gastric emptying, and actually reduced postprandial insulin response. These mechanisms are the basis for the highly efficacious GLP-1 analogues that today offer safe and effective treatment in millions of people living with obesity. Moreover, the combined GLP-1 mechanisms of weight loss and delayed carbohydrate absorption may also be the key drivers of remission of type 2 diabetes and reduced cardiovascular events found by GLP-1 analogues.
With the discovery of GLP-1 in 1986 simultaneously by Jens Juul Holst and Joel Habener a cascade of important physiological discoveries followed [1]. Their identification of the amino acid sequence of the biologically active GLP-1 hormone laid the groundwork for drugs in the management of type 2 diabetes. Gut peptide hormones such as GLP-1 could potentially also have an impact on appetite regulation by effects on gastric emptying and direct CNS effects, but we needed a robust methodology that could be used to measure acute effect of putative hormones on subjective appetite and spontaneous food intake. We* developed a study design using visual analogue scales (VAS) for measurement of subjective appetite recordings during a fixed breakfast meal test, and combined it with a subsequent ad libitum lunch meal test that measured consumed energy. The putative satiety effect of a compound would be seen as an effect on satiety VAS during the breakfast meal and on the spontaneous caloric intake at the lunch meal. VAS had been mainly used to measure subjective sensations (e.g. pain) and studies of appetite lacked validated measures. In an initial pilot study, we demonstrated the need for more comprehensive validation [2].
Time line showing the discovery of GLP-1 i 1986, the satiety hormone effect i 1996, and the synthesis of the long-acting analogues liraglutide and semaglutide, and the key persons involved.
Together with John Blundell, we developed the protocol: on two different test days we examined 55 healthy individuals and recorded their appetite sensations before breakfast and every 30 min during the 4.5 hr postprandial period under the same conditions [3]. We also examined if diet standardization the days prior to the test days influenced the results. The study showed the validity of quantification of hunger, satiety, and prospective food consumption measured by VAS as well as measurement of ad libitum energy consumption and showed that the appetite recordings during the breakfast meal predicted the energy intake during the lunch meal [3]. These results provided us with the essential methodology to design the GLP-1 infusion study and provided information on the estimated number of participants needed for adequate statistical power. This VAS paper is today a classic methods paper in this field [3] and has been cited 2,382 times.
* In this article “we” is used as a reference to the group of scientists, dieticians, and technical staff in my department who contributed to own studies cited, and the key collaborating scientists from other labs including Jens Juul Holst and John Blundell. Contributors from my own department are mentioned in the Acknowledgment.
In 1994 I took up the idea with Holst that GLP-1 could also be a mediator of meal-induced satiety, and we designed a study to infuse GLP-1 into normal human volunteers. Meanwhile, in January 1996 the first rodent study suggested that GLP-1 could be a central regulator of feeding behavior and satiety based on intracerebroventricular injection studies [4], but soon thereafter the results of other rodent studies using peripherally injected GLP-1 argued against an appetite effect [5].
The study was conducted in 1995–1996, and we used commercially available synthetic, human GLP-1 (7–36 amide) [6]. We conducted the trial in 20 young, healthy, normal-weight men in a placebo-controlled, randomized, blinded, crossover design. Infusion of GLP-1 or saline was initiated simultaneously with the consumption of the breakfast test meal. VAS were used to assess appetite sensations, and at lunchtime an ad libitum test meal was offered. We found that, after the energy-fixed breakfast, GLP-1 markedly enhanced satiety and fullness compared with placebo (Fig. 2) [6]. Furthermore, spontaneous energy intake at the ad libitum lunch was reduced 12% by GLP-1 infusion (Fig. 2). We were quite excited by the findings, as this study was the first to demonstrate a physiological role of GLP-1 in appetite control and energy intake in humans.
The pivotal findings of the study in which we infused GLP-1 in human subjects showed that it enhanced satiety reduced hunger during a fixed breakfast meal and reduced spontaneous energy intake by 12% during an ad libitum lunch meal [6].
An invention was submitted for a patent on 12 November 1996 [7], and the scientific article was submitted for publication in the Journal of Clinical Investigation 20 June 1997 and published 1st February 1998 [6]. The experimental set-up and execution of the studio are reproduced in a short video with actors.
In September 1998, Näslund et al. failed to detect an effect of GLP1 on food intake [8]. However, that study was done in only 6 subjects, raising concerns about the lack of statistical power to detect a clinically relevant effect. In 1999, Gutzwiller et al. showed a dose-dependent reduction in food intake with GLP-1 infusion in 16 healthy male subjects [9]. Subsequently, we showed that GLP-1 induces satiety equally well in people with obesity [10]. These effects could be at least partially attributed to the slowing of gastric emptying. By 2000, seven studies had been published on the acute effect of peripheral GLP-1 administration on ad libitum energy intake. Four found energy intake to be significantly reduced, whereas the remaining studies failed to find an effect – again potentially due to limited statistical power and also low GLP-1 infusion rates. We therefore conducted an individual participant data analysis, pooling raw data from the existing studies [11]. Six infusion studies provided us with 147 participants, and we found that GLP-1 reduced energy intake dose-dependently in both lean and overweight participants (a mean of 13%, P< 0.001, Fig. 3) [11]. There was also a dose-response effect on the reduction in emptying rate.
Data obtained from n = 147 volunteers following infusions with GLP-1 or saline. Adopted from [11].
In close collaboration with Jens Juul Holst, we conducted a series of infusion studies on
