Non-human primate model of long-COVID identifies immune associates of hyperglycemia

Post-acute sequelae of SARS-CoV-2 (PASC), or Long COVID, affects 10-30% of individuals infected with SARS-CoV-2, manifesting as various metabolic diseases, including type 2 diabetes (T2D), myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), breathlessness, thrombosis, and neuropsychiatric sequelae. SARS-CoV-2 infection elevates the risk of diabetes, suggesting a link between hyperinflammatory responses and metabolic PASC. Glucose homeostasis involves hormonally regulated glucose uptake by tissues (liver) and the gut microbiota, with hepatic glucose production (gluconeogenesis and glycogenolysis) and β-cell function also playing crucial roles. Inflammation, particularly involving chemokines like CCL25, is implicated in impaired glucose homeostasis. However, the mechanisms linking SARS-CoV-2 infection to prolonged hyperglycemia are poorly understood due to a lack of suitable animal models. This study aimed to develop a non-human primate (NHP) model of metabolic PASC using SARS-CoV-2-infected African green monkeys (AGMs) to analyze blood biochemistry, virology, and immunology parameters longitudinally and investigate potential mechanisms underlying metabolic PASC. The study also explored whether vaccination during acute infection could ameliorate immunometabolic dysregulation.

Existing literature highlights the significant prevalence and diverse manifestations of long COVID, emphasizing the impact on metabolic health, particularly the increased risk of type 2 diabetes. Studies have linked hyperinflammatory responses to the severity of acute COVID-19 and the development of metabolic PASC. The importance of the immune system's role in metabolic reprogramming is well established, particularly the involvement of immune cells. However, the precise mechanisms by which SARS-CoV-2 infection triggers prolonged hyperglycemia remain unclear due to a dearth of appropriate animal models. The existing literature underscores the role of the liver and pancreas in glucose homeostasis and suggests that impaired insulin secretion and increased hepatic glucose production could contribute to hyperglycemia in COVID-19. Inflammation's role is supported by studies demonstrating a correlation between elevated inflammatory molecules and impaired glucose homeostasis in various diseases. The lack of a suitable animal model for studying metabolic PASC has hampered our understanding of the long-term metabolic consequences of SARS-CoV-2 infection.

Fifteen African green monkeys (13 females, 2 males) were intranasally and intratracheally infected with SARS-CoV-2 (strain 2019-nCoV/USA-WA1/2020). Five animals received the BNT162b2 (Pfizer/BioNTech) mRNA vaccine 4 days post-infection, while 10 remained unvaccinated. The animals were monitored weekly for 18 weeks, with virological, clinical, blood chemistry, and immunometabolic assessments. Nasal and pharyngeal swabs were collected for viral RNA quantification by qPCR. Blood samples were collected for blood chemistry, cytokine/chemokine analysis, and immune/antibody response profiling. Tissues (liver, pancreas, duodenum, lung) were collected at necropsy (18 weeks post-infection) for virological, metabolic, and clinical evaluations. The OLINK Proximity Extension Assay (PEA) was used to analyze a panel of inflammatory proteins. Flow cytometry was employed to assess immune cell subsets and cytokine production. RNAscope and immunohistochemistry were performed to detect SARS-CoV-2 RNA and proteins in tissues. Statistical analyses included Wilcoxon matched-pairs signed-rank test, Spearman's rank correlation, Mann-Whitney U test, and PERMANOVA.

SARS-CoV-2-infected AGMs exhibited persistent hyperglycemia up to four months post-infection. A plasma chemokine signature during acute COVID-19 was identified, correlating with the metabolic defect. Hyperglycemia correlated with inflammatory T-cell populations and hepatic glycogen levels. Vaccination on day 4 post-infection had a favorable metabolic effect, resulting in significantly lower blood glucose levels in the vaccinated group compared to the unvaccinated group over time. The study also found no substantial long-term SARS-CoV-2 replication in the liver, pancreas, or duodenum. An unbiased analysis revealed 15 significantly upregulated plasma analytes at week 1 post-infection; 5 (CCL25, CDCP1, Flt3L, CCL8, SCF) maintained significance after Benjamini-Hochberg correction. CCL25 and GDNF were significantly elevated and positively correlated with blood glucose levels. Further analysis showed a significant positive correlation between the percentage of activated CD8+ T cells producing IFNγ and TNF and blood glucose levels. Increased liver glycogen levels were observed in infected animals, correlating with blood glucose levels.

This study provides the first comprehensive analysis of the temporal dynamics of hyperglycemia in a non-human primate model of long COVID. The findings demonstrate that SARS-CoV-2 infection in AGMs results in early-onset, persistent hyperglycemia, mirroring human observations. The identified chemokine signature, particularly CCL25 and GDNF, suggests a potential role for inflammation in mediating this metabolic dysfunction. The positive correlation between activated CD8+ T cells and hyperglycemia suggests a potential contribution of the adaptive immune response. The absence of substantial viral persistence in extrapulmonary tissues suggests that the metabolic sequelae may be driven by the early inflammatory response rather than ongoing viral replication. The favorable effect of vaccination during acute infection highlights the potential for early intervention strategies to mitigate long-term metabolic complications.

This study establishes a valuable non-human primate model for studying metabolic PASC, providing insights into the immunologic and metabolic alterations associated with long COVID. The identification of CCL25, GDNF, and activated CD8+ T cells as key factors involved in hyperglycemia opens new avenues for research. Future studies could explore therapeutic strategies targeting these factors to prevent or treat COVID-19-related metabolic disorders. Further investigation into the precise roles of GDNF and the chemokine signature in the pathogenesis of hyperglycemia is warranted.

The study had a bias towards female AGMs due to limited availability of animals, which could limit the generalizability of the findings. Not all data from the OLINK analysis met quality control standards, potentially affecting the interpretation of some findings. The study did not investigate the role of glucagon in the observed hyperglycemia. The absence of a larger sample size also restricts statistical power and definitive causality assessments. The focus on metabolic aspects of long COVID restricts the applicability of the findings to other manifestations of PASC.