We identified a 17q12 microdeletion in a patient allowing the diagnosis of MODY5. The diagnosis of MODY5 exclusively relies on genetic analyses that are not available in our routine clinical practice, and about half of cases can be explained by 17q12 microdeletions [4, 5] that cannot be detected by sequencing-based approaches. Because of these technical difficulties, MODY5 has been underdiagnosed and we need to accumulate evidence to establish diagnostic algorithms and improve therapeutic approaches. Recent advancements in massively parallel sequencing have enabled us to identify pathogenic mutations responsible for a wide variety of diseases. However, we need to recognize both the promises and pitfalls of this technique. As we have demonstrated in this case, we cannot identify deletions of genomic loci with this approach. Furthermore, the presence of pseudogenes can hamper exome-based diagnosis, as demonstrated in the case of by 21-hydroxylase deficiency [12].

The detailed genotype–phenotype correlation in MODY5 cases awaits further investigation [1, 4]. The clinical manifestations in cases of HNF1B mutations may exhibit differential patterns depending on the functional domains affected by the mutations, considering the complex regulation of HNF1B transcriptional activities as exemplified by interactions with other proteins [13]. In cases of 17q12 microdeletions, genomic loci spanning 1.5 Mb on average are deleted that encode genes such as AATF, ACACA, C17orf78, DDX52, DHRS11, DUSP14, GGNBP2, HNF1B, LHX1, MRM1, MYO19, PIGW, SYNRG, TADA2A and ANHIT3 [14]. The phenotypic features [14, 15] may vary among individuals depending on the influenced regions. It has been reported that insulin secretory capacities and ages of onset in MODY5 cases vary among individuals [1, 2]. However, these observations have been discussed mainly based on the requirement of insulin therapy that reflects both insulin secretion and insulin sensitivity. Intriguingly, our case showed a markedly decreased serum glucagon level that would have significantly contributed to the relatively low daily dose of insulin. While counter regulatory hormones have been often dismissed in the analyses of MODY5, our case suggests that we need to pay more careful attention to these hormones. Notably, it was recently reported that deficiency of acetyl-CoA carboxylase 1 (ACC1) encoded by the ACACA gene blunts glucagon secretion [16]. Since the ACACA gene is encoded in close proximity of the HNF1B gene and is centrally located in the commonly deleted regions in the 17q12 syndrome [14], 17q12 microdeletions would delete the ACACA gene together with HNF1B. Although potential coexistence of other mutations may be responsible for the decreased plasma glucagon concentration in this case, we were not able to find any pathogenic nonsynonymous mutations in proglucagon gene that encodes glucagon, the processing enzyme (i.e., prohormone convertase 2, PC2) or other genes related to alpha-cell functions. She has haploinsufficiency of ACACA gene since our exome-sequencing survey did not reveal any pathogenic mutations in the ACACA gene in the intact allele. Our observation is in sharp contrast to the increased glucagon concentrations in diabetic patients [17], as well as the fact that glucagon secretion is increased in the absence of insulin-mediated suppression [18]. The S148L mutation in HNF1B was also reported to be associated with increased glucagon secretion with compensatory mechanisms [19]. These observations implicate that the 17q12 deletion may specifically decrease plasma glucagon concentrations. Whereas the glucagon level in a case of 17q12 deletion recently reported was not affected, the discrepancy may arise because of the different ELISA systems employed, since the specificity of glucagon detection systems has been a matter of debate [20], or because of differences in the genomic regions affected [21]. Despite this uncertainty, this perspective deserves further scrutiny to better understand glucose homeostasis in MODY5 cases. The therapeutic potential of inhibition of glucagon’s action in type 1 diabetes has been extensively investigated [22, 23], and a phase II trial to examine the effects of the glucagon receptor antagonist volagidemab was reported recently [24]. Our case can also be a piece of human data to facilitate our understanding of this perspective. The potential contributions of ACACA gene deletion to the renal phenotypes in MODY5 have been reported [1], which was recently supported by a rodent study [25] albeit with some debates regarding whether ACC1 alone suffices or not [26]. In other words, MODY5 cases with relatively preserved renal functions may carry a 17q12 microdeletion rather than mutations in the HNF1B gene, and we may be able to choose our genetic approaches based on the clinical course of renal functions of patients although the controversies need to be clarified experimentally and clinically in the future.

It is also important to acknowledge the limitations in this study. Although we surveyed her genome using the exome-sequencing approach as well as technologies to detect microdeletions, we may need to consider potential contributions of mutations in regulatory regions such as promoters and introns that were not covered by our approaches in this study [27]. In addition, we assumed that the diminished glucagon would be at least in part responsible for the enhanced insulin sensitivity. However, MODY5 patients especially with microdeletion tend to be leaner [1], which may be responsible for her insulin sensitivity. Moreover, we were not able to quantify her insulin sensitivity with historically established approaches (i.e., homeostasis model assessment of insulin resistance (HOMA-IR), Matsuda index, insulin sensitivity factor or the glucose clamp technique), which could be another limitation for us to compare our case with others reported in the literature. Lastly, another issue remains to be solved is the question whether or not the haploinsufficiency of ACACA gene is sufficient to cause the diminished glucagon secretion observed in our case. The alpha-cell-specific ACACA knockout mice exhibited some residual mRNA expression of ACACA gene, and neither gene dose effects nor different concentrations of the ACC inhibitor was tested in the preceding report [16]. Therefore, this question needs to be addressed by careful studies in the future. However, we would speculate that there might be additional unidentified defects in our case to reconcile all previous reports including ours to ensure consistency.

In conclusion, we diagnosed a case of MODY5 carrying a 17q12 microdeletion. When we encounter patients with clinical features consistent with MODY5, the potential pitfalls in exome sequencing should be kept in mind. Furthermore, we would like to emphasize that more attention may need to be paid to glucagon, since MODY5 is prejudiced to be a disease with impaired insulin secretion in a simplified view.