SARS-Cov2, is a newly emerged strain of coronavirus responsible for causing the coronavirus disease-2019 pandemic (COVID-19), wreaking havoc worldwide. This virus belongs to the genus Betacoronaviruswhose other members-severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV) are known for two previous outbreaks. Although, SARS and MERS caused deadly pneumonia with mortality rates of 10% and 36% respectively1, much higher as compared to 2.3% for SARS-CoV2 infections2, the overall number of deaths from SARS-CoV2 far outweighs the previous outbreaks. At genome level, SARS-CoV2 is more similar to a bat SARSr-CoV showing 96% genome identity than that of SARS and MERS sharing about 79% and 50% of genome identity respectively3,4. Two of the significant predictors of severity and death in patients affected by SARS, MERS and pandemic influenza are diabetes and uncontrolled glycemia5-7. Interestingly, majority of deaths in the SARS-CoV2 infection have occurred in patients suffering from metabolic co-morbidities and indeed metabolic co-morbidities have been reported to portend worse disease outcomes in multiple studies carried out in different geographical locations, with completely different ethnic populations.8-10 But, the reason for this often-fatal role of metabolic co-morbidities, more apparent in the case of SARS-CoV2 infection, is not yet mechanistically clear. We aimed at understanding the pathogenetic mechanisms underlying this prognostic implication of metabolic disorders in COVID-19 that may differentiate SARS-CoV2 infection from few other respiratory viruses that caused more limited pandemics in the recent past.
In order to do so, we undertook a meta-analysis of three publicly available gene expression studies (RNA-sequencing as well as microarray data) done on human lung epithelial cells (HLECs) (either A549 or Calu-3 cells) infected with different respiratory viruses, viz. SARS-CoV2, Respiratory Syncytial Virus or RSV, H1N1 Influenza, SARS and H3N2 Influenza11-13. The HLEC transcriptome on infection with these viruses revealed transcriptional signatures shared between the different viruses as well as transcriptional regulations exclusive to each of the viruses, as expected (Figure 1A). SARS-CoV2 was found to drive differential regulation of maximum number of genes among the datasets analyzed. Further analysis of the significantly regulated gene expressions revealed the key enriched pathways that are differentially regulated among these viruses (Figure 1B, C; Supplemental information). In Figure 1B and C, we represented pathways, among those enriched, which have role in cellular metabolic regulation and for which the nature of regulation of the pathway could be ascertained and excluded pathways that are related to diseases of no apparent relevance. Amongst these, modulation of a few key metabolic pathways (viz. AMPK signaling pathway, HIF1αsignaling pathway, mitophagy, AGE-RAGE signaling pathway, pentose phosphate pathway etc.) as well as the cellular senescence pathway and lysosome mediated degradation pathway were found to be exclusive to SARS-CoV2 infection (Figure 1B,C). The AGE-RAGE signaling pathway and cellular senescence pathway were upregulated (Figure 1B), whereas the AMPK signaling pathway, HIF1α signaling pathway, mitophagy, pentose phosphate pathway and the lysosome pathway were all downregulated (Figure 1C). Interestingly, dysregulation of most of these metabolic pathways (upregulation of AGE-RAGE signaling and downregulation of AMPK signaling pathway, HIF1α signaling pathway, mitophagy as well as the pentose phosphate pathway) have all been previously implicated in systemic disease pathogenesis observed in insulin resistance, Type 2 diabetes and its associated complications14-18.
Type 2 diabetes mellitus (T2DM) and obesity are two of the major co-morbidities portending worse prognostic outcomes in Covid-19 patients8-10. Interestingly,previous reports have established the increased association of asthma and chronic obstructive pulmonary disease(COPD) with these metabolic disorders19. Infact, these diseases associated with increased systemic metabolic disorder, are known to cause metabolic dysregulations in HLECs, which possibly underlie the predisposition of diabetic individuals towards asthma and COPD20,21. In order to compare the transcriptomic signature of SARS-Cov2 infected lung and diabetic lung, we performed an integrated analysis of the HLEC transcriptome in response to SARS-CoV2 infection as well as a COVID-19 bronchoalveolar lavage22and autopsied lung transcriptome23 with lung transcriptome from two different datasets of preclinical rodent models of diet-induced obesity24,25 (Figure 2A). Interestingly, we found significant overlaps in how some of these key pathways are transcriptionally regulated in lung in these two distinct clinical contexts of SARS-CoV2 infection and metabolic disorder (Figure 2B). In Figure 1D, selection of pathways were based on whether 1) pathway enrichment and the nature of regulation are shared between at least two datasets, 2) are not enriched in any of the other 4 virus datasets (RSV, SARS, H1N1 and H3N2), 3) are related to cellular metabolic regulation, 4) are represented in either Figure 1B or 1C and 5) are not primarily linked to diseases of no apparent relevance. Thus, the similarity in the dysregulation of pathways in HLECs/lung upon SARS-CoV2 infection and metabolic disorders was apparent.
Our meta-analysis led us to hypothesize that with underlying metabolic co-morbidities, e.g. T2DM, this already present metabolic dysregulation of the HLECs is further aggravated upon SARS-CoV2 infection leading to faster and more widespread disruption of epithelial integrity and pulmonary damage, leading to worse disease outcomes. This provides a possible explanation for the distinction observed in the characteristics of the population subgroups worst affected by SARS-CoV2 and the other above-mentioned viruses including viruses belonging to the same family, such as SARS.
We envisage that any therapeutic agent that can target these key featured pathways, as shown in Figure 2B, should be of interest in SARS-CoV2 infection. This also points to a hitherto untapped potential for use of the widely used anti-diabetic drug metformin in COVID-19, which is an established modulator of most of these dysregulated pathways26-30. Indeed, metformin has been reported to inhibit AGEs, thus downregulating the AGE-RAGE signaling pathway associated with diabetes complications26. Metformin also enhances mitophagy in patients with Type 2 diabetes27. This anti-diabetic drug has also been reported to maintain cellular integrity by hindering cellular senescence28. The role of metformin in upregulation of the AMPK signaling pathway is one of its most well-established mechanisms of action involved in treating diabetes, reported in multiple studies29. Metformin has also been reported to modulate fatty acid metabolism30.
In addition to its projected effect on maintenance of lung epithelial integrity through these metabolic pathways in the epithelial cells, metformin has also been shown to boost CD8+ T cell memory though metabolic reprogramming, an effect which will also add to its potential therapeutic action in COVID-1931,32. In a preclinical model on pulmonary effects of air pollution it was also shown that metformin can attenuate the production of the proinflammatory cytokine IL-6 from alveolar macrophages33, which is also a critical component of the hyperimmune response in COVID-1934. Moreover, vascular thromboembolism has been shown to be a unique feature in a number of COVID-19 patients in different studies35,36 and metformin has been shown to have an antithrombotic effect through inhibition of platelet activation and maintenance of endothelial integrity37,38. Thus, metformin, due to its multifaceted pharmacodynamics, is well poised for acting as a therapeutic agent in SARS-CoV2 infections. But for ascertaining the therapeutic potential of metformin, mechanistic studies in the context of SARS-CoV2 infection, as well as comprehensive clinical trials in COVID-19 patients with or without metabolic co-morbidities, are warranted.