In this study, we analyzed the plasma proteome of pediatric patients with severe DKA compared to age- and sex-matched insulin-controlled. After correcting for multiple comparisons, we identified 214 differentially expressed proteins with 162 upregulated and 52 downregulated (adj P < 0.05 and a fold change > 2). Using NLP, we mapped the expression patterns of these proteins across various organs (e.g., digestive, endocrine, neurological) and cell types (e.g., leukocytes, epithelial, endothelial). Pathway analysis revealed 38 enriched pathways, which were correlated with clinical variables, biochemistry, and GCS. This is the first comprehensive proteome profiling of pediatric DKA.

Our study specifically targeted patients with severe DKA (Segerer et al. 2021). The mean HbA1C level among these patients was 11.9%, indicating poor glucose control over the 2–3 months preceding the DKA episode (American Diabetes Association 2021). Additionally, these patients exhibited low HCO3 levels, characteristic of metabolic acidosis, along with elevated BUN, likely resulting from dehydration and catabolic processes.

Our analysis of the DKA plasma proteome identified a greater number of DEPs than previous studies focusing on specific inflammatory markers (Karavanaki et al. 2011; Omatsu et al. 2014; Woo et al. 2016a; Woo et al. 2016b). Consistent with our previous DKA research, we found highly elevated levels of PR3 (PRTN3; Woo et al. 2016b) and significant increases in MMP8 and MMP9 levels (Woo et al. 2016a). Many of the DEPs were linked to inflammation and metabolism. For example, interleukin-6 (IL6; upregulated) is a potent inducer of the acute phase response, while growth differentiation factor 15 (GDF15; upregulated) regulates food intake, energy expenditure, and body weight in response to metabolic and toxin-induced stresses, affecting cellular stress and β-cell function (Xu et al. 2022; Mohammad et al. 2023). Additionally, we noted a decrease in members of the TNF superfamily, some of which have been previously implicated in DKA (Rochfort et al. 2016).

The clinical manifestations of DKA are closely tied to organ-specific protein expression patterns. While early signs of organ dysfunction may not be immediately apparent, rapid deterioration can occur without prompt treatment. Common initial symptoms include polyuria, polydipsia, and polyphagia (Dhatariya et al. 2020; Alois and Rizzolo 2017). As DKA progresses, worsening ketosis and acidosis trigger systemic responses across multiple organs. Gastrointestinal symptoms such as nausea, vomiting, abdominal pain, and tenderness often arise, sometimes leading to upper gastrointestinal bleeding (Umpierrez and Freire 2002). Respiratory complications, including Kussmaul breathing, are common due to ketoacidosis and may, in rare cases, lead to spontaneous pneumothorax (Vallabhajosyula et al. 2016). Although less frequent, acute respiratory distress syndrome (ARDS) and pneumomediastinum, often associated with a cytokine storm, are serious complications (Horiya et al. 2022; Bialo et al. 2015).

Free thyroid hormone levels decrease as DKA severity increases, with acidosis contributing to this decline (Xing et al. 2021). The pituitary gland plays a key role in DKA pathogenesis, particularly in glucose metabolism and stress responses (Voss et al. 2019). Counter-regulatory hormones drive the metabolic disturbances that perpetuate the ketoacidosis cycle (Dhatariya et al. 2020; Voss et al. 2019). Our pathway analysis revealed significant enrichment in pathways related to "response to hormone" and "response to endogenous stimulus," indicating cellular and systemic changes in response to hormonal signals. Upregulated pathways involving peptide hormones and nitrogen-containing compounds suggest increased activity of the renin–angiotensin–aldosterone system and elevated levels of glucagon, catecholamines, cortisol, and corticotropin-releasing hormone during DKA (Voss et al. 2019). These hormones promote lipolysis, glycogenolysis, gluconeogenesis, and proteolysis, leading to the accumulation of nitrogen compounds. Combined with hyperglycemia and glycosuria, this contributes to dehydration and hyperosmolarity (Glaser 2005; Svart et al. 2016).

Lipid metabolism is significantly disrupted in DKA (Voss et al. 2019). Our results show marked changes in adipose tissue signaling, with increased signals for food intake and lipid mobilization into adipocytes, supported by elevated markers like GDF15, FABP4, FABP5, LDLR, and ANGPTL4 (Svart et al. 2016; Wang et al. 2021). Pathways related to lipid and fatty acid metabolism, particularly "lipid localization," "lipid transport," and "fatty acid transport," were significantly enriched. These findings align with known DKA pathophysiology, where counter-regulatory hormones drive lipolysis, mobilizing free fatty acids to the liver and worsening ketosis (Alois and Rizzolo 2017; Laffel 1999). Additionally, lipid mobilization and signaling may influence DKA-associated inflammation (Glass and Olefsky 2012).

DKA triggers systemic inflammation through the activation and recruitment of various leukocyte subtypes, even without infection. The severity of DKA is correlated with elevated levels of WBCs, neutrophils, and monocyte (Nichols et al. 2022). Our study confirms this association, showing leukocytosis in all DKA patients. We identified over thirty proteins linked to altered regulatory activities critical for inflammation, particularly those associated with leukocytes. Specifically, we observed increased levels of pro-inflammatory cytokines (e.g., IL6, IL15), innate immune components (e.g., PTX3, AZU1, CSF3), and markers of adaptive immune cells (e.g., IL1RL1, LEPR). In contrast, inhibitory hematopoietic mechanisms (e.g., RGMA, SMAD5, AREG) and anti-inflammatory pathways (e.g., CLEC4C, OGN) were downregulated. Pathways such as "signaling receptor regulator activity" and "signaling receptor activator activity" were significantly altered, reflecting enhanced pro-inflammatory signaling, leukocyte recruitment, and cytokine production. Enrichment in pathways related to "response to peptide hormone" and "signaling receptor activator activity" suggests increased extracellular and intracellular signaling. This immunological dysregulation can contribute to endothelial activation and increased vascular permeability, potentially leading to complications such as vasogenic edema (Omatsu et al. 2014; Woo et al. 2016b; Rochfort et al. 2016), rhabdomyolysis (Bialo et al. 2015), and acute pancreatitis due to repeated inflammation (Hahn et al. 2010).

In severe DKA, acute neurological dysfunction often manifests as a decreased GCS score, which can be worsened by cerebral edema (Eisenhut 2018; Bialo et al. 2015). Long-term neurological sequelae, such as neuronal damage and memory impairment, are linked to the severity of acidosis and are more pronounced in younger individuals (Bialo et al. 2015; Ghetti et al. 2020). Prolonged hyperglycemia further damages vascular structures, leading to cellular impairment and an increased risk of cerebrovascular events (Glaser 2005; Hamed et al. 2017). Our analysis revealed upregulation of pro-angiogenic factors (e.g., ADAMTS8) and pro-survival pathways (e.g., KITLG), alongside downregulation of pro-apoptotic factors (e.g., TNFSF12, TNFSF10), likely exacerbating inflammatory and vascular complications. We also observed increased levels of coagulation factors (e.g., HGF), platelet aggregation markers (e.g., TIMP4, SELP), and changes in cell-to-cell and cell-to-matrix interactions (e.g., ADAM23, SELP). Elevated levels of serine proteases (e.g., PR3, or PRTN3) and vasodilators (e.g., ADM, CALCA) further contribute to vascular derangements (Woo et al. 2016b; Hoffman et al. 2019).

Our analysis revealed both positive and negative associations between enriched pathways and clinical/biochemical variables. Elevated levels of HCO3 and GCS were negatively associated with several pathways. Specifically, HCO3 negatively correlated with pathways related to hormone response, receptor regulation of endogenous stimuli, and the metabolism of nitrogen, oxygen, and organonitrogen compounds, with weaker correlations in lipid metabolism. This is likely due to HCO3's role in maintaining acid–base balance (Kamel and Halperin 2015). Metabolic acidosis may increase counterregulatory hormones like cortisol, glucagon, growth hormone, and catecholamines to elevate blood pH (Voss et al. 2019), while oxygen compounds support cellular respiration under hypoxia. Hyperventilation compensates for acidemia, and acidosis treatments can activate nitrogen pathways, potentially worsening the condition (Guh et al. 1997). GCS was negatively correlated with pathways linked to activity modulation, likely reflecting prolonged diabetes decompensation and its impact on signaling pathways and clinical outcomes (Dhatariya et al. 2020; Glaser 2005).

Positive correlations were most pronounced with lactate, WBC counts, and PO2. Lactate levels strongly correlated with receptor-ligand modulation, likely due to its role in regulating metabolic processes, such as insulin-mediated anti-lipolysis (Ahmed et al. 2010; Lu et al. 2011). Elevated lactate can result from anaerobic glycolysis in poorly perfused tissues, reduced hepatic clearance, stress-induced aerobic glycolysis, and mitochondrial dysfunction, all of which enhance cellular signaling (Lu et al. 2011; Liu et al. 2021). Elevated WBC counts likely reflect systemic inflammation, with no underlying infections identified (Aon et al. 2024; Hamtzany et al. 2023). Increases in specific subtypes, such as lymphocytes, may indicate enhanced cellular trafficking associated with metabolic inflammation (Dhatariya et al. 2020; Kamel and Halperin 2015). Activated lymphocytes promote inflammation and produce reactive oxygen species (Kitabchi et al. 2004). Additionally, changes in oxygen availability during metabolic acidosis may explain the positive correlation with PO2 (Kraut and Madias 2010).

Our study provides the first comprehensive evaluation of the DKA proteome, though several limitations must be addressed. First, while we included a balanced yet limited number of matched participants, our findings remain statistically significant after correction for multiple comparisons. Second, we focused solely on severe DKA patients to identify metabolic and inflammatory changes, but future studies should include a broader spectrum of DKA severity to improve generalizability. Third, although we identified differentially expressed proteins in plasma prior to insulin therapy, the lack of longitudinal samples prevents tracking changes in protein levels over time. This limits our ability to assess fluctuations during DKA treatment. Fourth, pediatric DKA patients experience osmotic diuresis, which could concentrate plasma proteins; this factor was considered in our analysis. Fifth, differences in T1D duration between cohorts, with some participants presenting with DKA as their first manifestation, could have influenced results. Lastly, despite significant associations between GCS and differentially expressed proteins, the limited GCS range warrants cautious interpretation of the data.

The identification of DEPs in severe DKA could improve patient stratification for future studies and clinical care. In this study, we identified 214 proteins and 38 signaling pathways that distinguish DKA patients from insulin-controlled diabetes. These findings provide valuable insights into the DKA pathophysiology. Future research should include longitudinal data to explore how initial protein profiles relate to short- and long-term outcomes.