Computational Model Explores Molecular Pathways Linking Malnutrition to Type 5 Diabetes

A scientific study published in Results in Chemistry by researcher Kevin-Armel Fotso Pechi combines a structured literature review with in silico reverse molecular docking to investigate potential molecular mechanisms connecting chronic malnutrition with pancreatic beta-cell dysfunction in malnutrition-associated diabetes, recently termed Type 5 Diabetes Mellitus. Rather than claiming established clinical causal mechanics, the study presents an exploratory, hypothesis-generating model examining how severe nutritional deprivation disrupts metabolic homeostasis. The computational docking simulations reveal that three nutritionally relevant compounds—vitamin D, glucose, and palmitic acid—exhibit distinct, ligand-specific binding affinities and residue interaction profiles with critical regulatory enzymes: glucokinase, AMP-activated protein kinase, dipeptidyl peptidase-4, and superoxide dismutase 1. These computational observations propose biologically plausible pathways through which chronic undernutrition may impair glucose sensing, alter cellular energy balance, accelerate oxidative stress, and compromise insulin secretion in nutritionally vulnerable populations without the autoimmune destruction or primary insulin resistance typical of other diabetes forms.

The condition historically identified as malnutrition-related diabetes mellitus has long been observed in populations experiencing prolonged undernutrition across low- and middle-income regions, particularly in parts of Africa and Asia. Between the 1950s and 1980s, clinical researchers identified young, lean individuals presenting with severe diabetic symptoms diverging from classical autoimmune type 1 diabetes and obesity-linked type 2 diabetes. Affected individuals exhibited profound insulin deficiency without detectable islet autoimmunity or marked insulin resistance. During the 1980s, the World Health Organization formally recognized malnutrition-related diabetes as a distinct category, subdividing it into fibrocalculous pancreatic diabetes, featuring pancreatic calcifications, and tropical pancreatic diabetes, associated with nutritional deprivation. In its 1999 classification update, the World Health Organization withdrew this separate category, reclassifying such presentations under secondary diabetes forms, specifically pancreatic diabetes or type 3c. Recently, the International Diabetes Federation reintroduced this metabolic entity under the designation Type 5 Diabetes Mellitus, establishing an analytical framework while acknowledging diagnostic criteria remain unstandardized.

Epidemiological documentation of malnutrition-associated diabetes remains limited and fragmented due to the historical absence of universal diagnostic guidelines, impeding precise surveillance across affected populations. Available evidence indicates that this condition predominantly impacts young, impoverished individuals in low- and middle-income regions, notably sub-Saharan Africa, South and Southeast Asia, and rural Latin America. Global estimates compiled in the review suggest that approximately 20 to 25 million individuals may be affected worldwide, though these numbers require cautious interpretation due to diagnostic heterogeneity and scarce population data. In rural South Asian regions, including parts of India, malnutrition-associated diabetes is estimated to account for roughly 5 to 10 percent of all diagnosed diabetes cases. Specific regional data cited in the review report an estimated prevalence of 0.019 percent in India’s general population and 0.36 percent among Indian diabetic cohorts. In Latin America, regional estimates report a prevalence below 5 percent, while sub-Saharan Africa presents clinical clusters among young, lean subjects without reliable epidemiological totals.

Unlike type 1 diabetes, which involves autoimmune beta-cell destruction, and type 2 diabetes, which is primarily driven by peripheral insulin resistance, malnutrition-associated diabetes is defined predominantly by insulin deficiency arising from impaired pancreatic functional capacity and reduced beta-cell mass. The developmental origins of health and disease framework, supported by Hales and Barker’s 1992 thrifty phenotype hypothesis, indicates that inadequate fetal and early neonatal nutrition induces permanent alterations in glucose-insulin metabolism that heighten adult diabetes risk. Experimental studies demonstrate that maternal protein restriction reduces beta-cell mass and suppresses insulin secretion in offspring. In addition, chronic pancreatic stress represents an essential pathological contributor, combining oxidative damage, low-grade inflammation, and deficiencies in micronutrients such as zinc, vitamin B12, and vitamin D. In several tropical regions, chronic dietary exposure to cyanogenic compounds from insufficiently processed cassava serves as an established environmental risk factor, promoting pancreatic fibrosis, islet hypoplasia, tissue atrophy, and gut dysbiosis that triggers systemic metabolic endotoxemia.

To explore potential molecular pathways connecting nutritional deprivation to beta-cell failure, the investigator implemented an in silico reverse molecular docking protocol using AutoDock Vina. Four key protein targets governing glucose homeostasis, cellular energy regulation, and oxidative defense were analyzed: glucokinase (PDB ID 1Q18), the pancreatic glucose sensor; AMP-activated protein kinase (PDB ID 4YEE), an energy homeostasis regulator; dipeptidyl peptidase-4 (PDB ID 5KBY), an incretin-degrading enzyme; and superoxide dismutase 1 (PDB ID 5J07), an essential antioxidant enzyme. Target protein structures were retrieved from the Protein Data Bank or generated through SWISS-MODEL homology modeling when experimental structures were absent. The study docked three nutritionally representative ligands: glucose (PubChem CID 5793), reflecting glycemic sensing; vitamin D (PubChem CID 5280795), reflecting micronutrient deficiency; and palmitic acid (PubChem CID 985), representing saturated fatty acid lipotoxicity. Docking grids encompassed active or predicted binding pockets identified by CurPocket, operating with an exhaustiveness parameter of 8 to calculate binding affinities in kilocalories per mole.

Molecular docking simulations demonstrated that vitamin D achieved the strongest binding affinities among evaluated ligands across all four metabolic protein structures, primarily mediated by extensive hydrophobic contacts. Within dipeptidyl peptidase-4, vitamin D produced its most favorable binding score in cavity C1 at −8.5 kcal/mol across a volume of 1967 cubic angstroms, interacting with Chain D residues including Phe357, Tyr631, Trp659, Val656, Tyr662, and Arg358. In glucokinase, vitamin D recorded an optimal binding affinity of −8.2 kcal/mol in cavity C2 (619 cubic angstroms), engaging Chain A residues Cys65, Pro66, Ile67, Thr68, and Glu96 through Thr102, and Chain B residues Glu123, Val125, Gly127, Lys128, Pro129, Phe255, and Tyr290. In superoxide dismutase 1, vitamin D bound cavity C3 with −8.2 kcal/mol (441 cubic angstroms), contacting Gln4, Gly5, Ile6, Asn8, Trp21, Ala98, Val108, and Val110. For AMP-activated protein kinase, vitamin D showed highest affinities in cavity C4 at −8.1 kcal/mol (756 cubic angstroms) and cavity C2 at −7.9 kcal/mol (1311 cubic angstroms), forming a hydrogen bond at Ser100 and pi-sigma interactions at Trp133.

In contrast to the hydrophobic binding modes observed with vitamin D, glucose exhibited moderate binding affinities across all four target enzymes, characterized predominantly by polar conventional hydrogen bonding. When docked against glucokinase, glucose exhibited preferential binding within cavity C2 at −6.2 kcal/mol (1494 cubic angstroms) and cavity C3 at −6.1 kcal/mol (285 cubic angstroms). Contact analysis revealed conventional hydrogen bonds with catalytic residues Asn100 and Glu157 in Chain A, alongside interactions involving Cys20, Ser24, Gly25, Glu26, and Ile27 in Chain B. In AMP-activated protein kinase, glucose displayed highest binding affinities in cavity C3 at −5.9 kcal/mol (1676 cubic angstroms), cavity C4 at −5.5 kcal/mol (1126 cubic angstroms), and cavity C5 at −5.4 kcal/mol (707 cubic angstroms), establishing hydrogen bonds with Asn150 and Thr148. In dipeptidyl peptidase-4, glucose bound cavity C1 at −5.8 kcal/mol (1967 cubic angstroms) via hydrogen bonds with Arg669, Asn595, Gly352, Asp678, and Glu668. For superoxide dismutase 1, glucose attained −4.8 kcal/mol in cavity C1 (628 cubic angstroms) via bonds with Ser23 and Asp55.

Docking simulations with palmitic acid generated moderate binding energies characterized by a mixture of hydrophobic alkyl contacts and specific polar interactions across the investigated metabolic targets. In glucokinase, palmitic acid showed its strongest binding affinity within cavity C4, scoring −6.9 kcal/mol across a cavity volume of 434 cubic angstroms, forming a conventional hydrogen bond with Ala157, a pi-sigma contact with Phe244, and alkyl interactions with Leu248, Val161, and Ala164. In AMP-activated protein kinase, the saturated fatty acid showed optimal binding in cavity C4 at −5.8 kcal/mol across 756 cubic angstroms, establishing a hydrogen bond with Asn150, a pi-sigma bond with Trp150, and multiple alkyl interactions with Lys102, Trp99, Lys107, and Ile106. In dipeptidyl peptidase-4, palmitic acid reached −5.2 kcal/mol in cavity C5 (832 cubic angstroms), forming a hydrogen bond with Thr565 and hydrophobic contacts with Lys512, Leu477, Phe559, Met509, and Leu504. In superoxide dismutase 1, palmitic acid achieved −4.8 kcal/mol in cavity C2 (628 cubic angstroms), engaging Asn8 through a hydrogen bond and contacting Ile6 and Val110.

By synthesizing these docking observations with existing biochemical literature, the study models how chronic undernutrition may trigger interconnected cellular defects leading to persistent insulinopenia. In pancreatic beta-cells, glucokinase serves as the primary glucose sensor catalyzing the initial step of glycolysis, coupling extracellular glucose availability to intracellular ATP production. Under sustained nutrient deprivation, impaired glucokinase activity reduces glycolytic flux, decreasing the intracellular ATP/ADP ratio. This reduction prevents regular closure of ATP-sensitive potassium channels, suppressing membrane depolarization, reducing calcium influx, and blunting glucose-stimulated insulin secretion. Concurrently, chronic energy stress dysregulates AMPK signaling, while gut dysbiosis and low-grade inflammation increase DPP-4 activity, accelerating enzymatic degradation of incretin hormones such as GLP-1. In parallel, decreased activity of antioxidant enzymes like SOD1 intensifies oxidative stress, promoting beta-cell apoptosis. Strong predicted binding of vitamin D across GCK, AMPK, and DPP-4 suggests that chronic vitamin D deficiency may exacerbate these disrupted pathways, promoting beta-cell dysfunction without invoking classical autoimmune or obesity-related insulin-resistance mechanisms.

The author explicitly notes critical methodological limitations, emphasizing that these computational findings represent exploratory, hypothesis-generating propositions rather than conclusive biological proofs. Because findings rely on in silico reverse molecular docking, they characterize static binding energetics and relative affinities without proving direct functional inhibition, activation, or causality in living systems. Furthermore, the investigation did not include in vitro cell culture experiments, isolated islet assays, or in vivo animal validation to verify whether these predicted interactions occur under physiological conditions. The docking workflow omitted redocking controls with co-crystallized reference ligands and did not include stereochemical Ramachandran plot assessments of modeled proteins, relying on standard template selections and default AutoDock Vina settings. Future research must incorporate dynamic molecular dynamics simulations to evaluate complex stability over time. Translating these preliminary computational insights will require multidisciplinary investigations combining molecular biology, metabolomic profiling, and prospective clinical trials in malnourished populations to establish standardized diagnostic criteria and evaluate targeted nutritional interventions for Type 5 Diabetes Mellitus.

Subject of Research: Computational modeling of molecular mechanisms underlying malnutrition-associated diabetes

Article Title: Nutritional stress and β-cell dysfunction in malnutrition-associated diabetes (Type 5): A hypothesis-driven model integrating literature evidence and in silico analysis

Article References: Pechi, K.-A. F. (2026). Nutritional stress and β-cell dysfunction in malnutrition-associated diabetes (Type 5): A hypothesis-driven model integrating literature evidence and in silico analysis. Results in Chemistry, 32, Article 104044. https://doi.org/10.1016/j.rechem.2026.104044

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.104044

Keywords: Type 5 Diabetes, Malnutrition-Related Diabetes, Beta-Cell Dysfunction, Molecular Docking, Glucokinase, AMPK, Dipeptidyl Peptidase-4, Superoxide Dismutase 1, Nutritional, stress, cell, dysfunction