Using UK Biobank data encompassing multiple omics data, MRI imaging and hospital registers, we here demonstrated that increased plasma levels of proglucagon are independently associated with obesity, MASLD and type 2 diabetes. In addition, proglucagon levels were significant predictors of the risk of type 2 diabetes development over a 14-year follow-up period. Although causality cannot be established, our data support the idea that increased proglucagon directly contributes to development of type 2 diabetes [24].

One of the two antibodies in the Olink proglucagon assay binds within the first 100 amino acids of proglucagon, whereas the binding site of the other antibody is proprietary information and has not been disclosed. It is therefore unknown which section(s) of the proglucagon-derived peptide is measured by the assay, and to what extent. We previously performed a pilot in vitro analysis of human plasma spiked with 100 pmol/l of glucagon, GLP-1 7–36, GLP-1 9–36, GLP-2 and oxyntomodulin. The proglucagon assay from Olink measured both forms of GLP-1, whereas neither glucagon, GLP-2 or oxyntomodulin were detected (https://github.com/nicwin98/Olink). Antibodies raised against GLP-1 have the (dis)advantage of measuring all molecular forms containing the amino acid sequence of GLP-1. Importantly, this includes MPGF and GLP-1 1–36, both of which derive from alpha cell processing of proglucagon [25, 26]. In fact, an older study revealed that elevated levels of fasting and arginine-induced immunoreactive GLP-1 in individuals with type 2 diabetes turned out to be the pancreatic peptides MPGF and GLP-1 1–36 [27]. Thus, to interpret analyses of GLP-1 concentrations, it is essential to know if the antibodies used are N-terminal wrapping (i.e. specific for the full-length [1–36], active [7–36], or total [9–36] GLP-1), C-terminal wrapping, or binding anywhere in the middle of the peptide, in which case the assay would measure all forms as well as MPGF.

We observed that proglucagon was elevated in individuals with type 2 diabetes (Fig. 1b). Numerous studies have described that GLP-1 may be reduced, and glucagon increased, in type 2 diabetes [28, 29]. This suggest that the proglucagon assay is likely measuring a pancreatic peptide that includes the GLP-1 amino acid sequence (possibly MPGF). In line with this hypothesis, we observed an increase in plasma proglucagon in carriers of cAMP LoF variants of the glucagon receptor (Fig. 4d). Since glucagon resistance in MASLD results in a compensatory increase in glucagon levels [30], similar mechanisms may underlie the finding of increased proglucagon (alpha cell-derived) in individuals with reduced glucagon receptor cAMP signalling.

Plasma proglucagon showed a positive association with BCAA. BCAA catabolism is not regulated by glucagon, and BCAA do not directly stimulate glucagon secretion [31, 32]. However, BCAA stimulate GLP-1 secretion [33] and play an important role in the pathogenesis of insulin resistance [34, 35]. A limitation of the current study is the lack of markers of insulin resistance, such as plasma insulin levels. We previously observed that glucagon resistance and insulin resistance may coexist but importantly also occur independently of each other, highlighting the differential pathophysiological mechanisms underlying glucagon and insulin resistance [36].

In agreement with previous studies, we found a link between MASLD and glucagon. We used MRI–PDFF to diagnose MASLD, but we lacked data on the severity of fibrosis, which is the most critical predictor of clinical outcomes in MASLD. Given the probable close relationship between glucagon and the metabolic alterations in MASLD, the degree of liver fat is likely to indicate the risk of developing cardiometabolic conditions, including diabetes. Our findings therefore support speculations that there is a two-way connection between MASLD and diabetes and suggest that glucagon could be a main factor in the pathogenesis.

Consistent with prior research, we identified an association between plasma proglucagon and prevalent type 2 diabetes. Metformin treatment is known to increase GLP-1 and reduce glucagon secretion, thus, the increase in proglucagon by metformin in individuals with type 2 diabetes may represent increased plasma GLP-1 levels. Conversely, insulin is an effective inhibitor of glucagon secretion from alpha cells, and the markedly lower levels of proglucagon in individuals on insulin is most likely a consequence of lower glucagon levels. Collectively, these data support the hypothesis that the proglucagon assay measures a combination of intestinal and pancreatic proglucagon products (probably GLP-1 and MPGF, respectively).

Type 2 diabetes is often diagnosed by general practitioners, and only ~41% of diabetes diagnoses are registered in hospital records [17]. Primary care data are linked to the UK Biobank for ~45% of the participants up until 2016 (England) and 2017 (Scotland and Wales), so to get a longer follow-up period, incident type 2 diabetes was defined from secondary care ICD-10 diagnostic codes. This is a limitation of this study and may explain the flattening of the Kaplan–Meier curve (Fig. 2a) in the later years of the follow-up period, as hospital diagnoses may be registered at a later point than the primary care diagnosis. Furthermore, UK Biobank is not representative of the UK population with evidence of a selection bias towards more healthy volunteers regarding obesity, smoking, and alcohol consumption [37]. There are more female than male participants in the UK Biobank, and to address potential confounding effects related to sex differences, we have included sex as a covariate in all our analyses. Another limitation is that quantification of liver fat is obtained from the MRI scan performed ~10.5 years after the baseline data was obtained.

Traditionally, the primary focus on glucagon receptor signalling has centred on the adenylate cyclase/cAMP/protein kinase A pathway as the predominant mediator of the hepatic glucose-mobilising actions of glucagon. However, compelling evidence suggests that the PLC/IP3 pathway may be even more important for physiological levels of glucagon, compared with pathologically or pharmacologically elevated levels [38]. Further investigation into the impact of rare missense variants on this pathway is warranted.

The most common missense variant in the glucagon receptor, G40S, has normal to mildly reduced cAMP signalling [39,40,41], and substantially decreased β-arrestin signalling [5]. G40S has previously been linked to non-insulin-dependent diabetes and central adiposity in France and Sardinia [21, 22], but interestingly not in Japan or Finland [42, 43]. We did not find an increased prevalence of type 2 diabetes or obesity in heterozygotes or homozygotes of the G40S variant in the UK Biobank. However, there appeared to be an interaction between plasma proglucagon and G40S, with proglucagon having a stronger association with BMI in carriers of G40S compared with carriers of the wild-type variant (Table 1). In contrast, the association between proglucagon and type 2 diabetes was weaker in individuals with G40S compared with the wild-type variant (p=0.107, Table 1). Together with the divergent literature on the effects of G40S on metabolic disorders, our results suggest that G40S may play a role in the development of type 2 diabetes and obesity, but that this may require parallel metabolic disruptions leading to altered proglucagon levels.

Previous research has indicated a potential association between cAMP LoF variants and an increased risk of obesity [5]. Since then, exome sequencing data was increased to cover the whole UK Biobank. After an approximate doubling of the sample size of carriers of a cAMP LoF variant, this tendency disappeared, both when BMI was treated as a continuous and dichotomised variable. However, the increased plasma levels of proglucagon in this group of individuals suggest that cAMP signalling is involved in the regulation of a factor, that in turn regulates proglucagon levels in a feedback manner. This has previously been suggested to be amino acids, with alanine being particularly important for this feedback system, termed the liver–alpha cell axis [44,45,46]. In line with this, there was a tendency to increased alanine (p=0.1) in individuals with cAMP variants adjusting for age, sex, BMI, fasting time and creatinine. Because the cAMP LoF variants are very rare, we aimed at creating the most homogenous study population by excluding participants who were not categorised as white, and participants with sex chromosome aneuploidy. This lack of diversity may restrict the generalisability of our findings to other ethnic or racial groups.

The Frameshift category of glucagon receptors was defined rather broadly. Yet, carriers of one of the Frameshift variants were associated with higher levels of liver fat compared with carriers of the wild-type variant. This was not observed in G40S or the cAMP LoF variant groups, suggesting that a signalling pathway other than cAMP and β-arrestin is likely involved. Further functional subdivision and in silico predictions may help select and group variants that are pharmacologically characterised by more stratified LoF phenotypes. Other research groups have found ubiquitination and β-arrestin to be essential for the signalling and internalisation of the glucagon receptor [4, 47], whereas others report only a minor internalisation of the glucagon receptor [48], suggesting that this pathway may not be a major regulator of glucagon receptor signalling. The multiple factors impacting signalling pathways are complex and warrant further exploration.

In conclusion, our study supports the involvement of glucagon signalling in metabolic disorders such as type 2 diabetes and MASLD and that increased proglucagon levels may predispose to type 2 diabetes. Furthermore, the presence of hyperglucagonaemia in obesity, MASLD and type 2 diabetes indicates that distinct mechanisms may drive increased alpha cell secretion. Signalling pathways other than cAMP and β-arrestin recruitment may be important for the metabolic effects of glucagon such as the regulation of liver fat. Identification of the specific molecular characteristics responsible for the beneficial effects of glucagon on hepatic lipid turnover may be crucial in developing improved glucagon co-agonists for the treatment of MASLD and obesity.