Metabolic Reprogramming of the Host Cell by Human Adenovirus Infection

Abstract

Viruses are obligate intracellular parasites that alter many cellular processes to create an environment optimal for viral replication. Reprogramming of cellular metabolism is an important, yet underappreciated feature of many viral infections, as this ensures that the energy and substrates required for viral replication are available in abundance. Human adenovirus (HAdV), which is the focus of this review, is a small DNA tumor virus that reprograms cellular metabolism in a variety of ways. It is well known that HAdV infection increases glucose uptake and fermentation to lactate in a manner resembling the Warburg effect observed in many cancer cells. However, HAdV infection induces many other metabolic changes. In this review, we integrate the findings from a variety of proteomic and transcriptomic studies to understand the subtleties of metabolite and metabolic pathway control during HAdV infection. We review how the E4ORF1 protein of HAdV enacts some of these changes and summarize evidence for reprogramming of cellular metabolism by the viral E1A protein. Therapies targeting altered metabolism are emerging as cancer treatments, and similar targeting of aberrant components of virally reprogrammed metabolism could have clinical antiviral applications.

Keywords: E1A; E4ORF1; HAdV36; HAdV5; MYC; Warburg effect; glutaminolysis; glycolysis; human adenovirus; metabolism.

Figure 1

Viruses co-opt many cellular metabolic pathways to satisfy their metabolic requirements. These pathways include those used for energy production, primarily glycolysis and oxidative phosphorylation, and macromolecule production, such as for the synthesis of nucleotides or fatty acids. Created with BioRender.

Figure 2

Both cancer cells (A) and virus infected cells (B) often exhibit a characteristic metabolic phenotype known as the Warburg effect. This phenotype is associated with an increase in cellular glycolysis and a concurrent decrease, albeit not a complete reduction, of cellular respiration despite the availability of ample oxygen. In contrast, healthy, uninfected cells (C) preferentially utilize cellular respiration over glycolysis as the main ATP generating pathway. Glutaminolysis is also less active in uninfected non-transformed cells. However, there are uninfected cells (D) that preferentially utilize the Warburg effect. For example, endothelial cells consistently have a Warburg effect-like metabolic phenotype [19]. Activated immune cells, such as effector T cells, activated macrophages, and activated dendritic cells, also shift to a Warburg effect-like metabolic phenotype [20]. Created with BioRender.

Figure 3

Key relevant advances in metabolism research, adenovirus research and technology that allowed for contemporary high-throughput studies on the effect of HAdV infection on cellular metabolism. The early 20th century featured many insights into the basics of cellular metabolism. The discovery of HAdV and early studies on the effect of HAdV on host-cell metabolism were performed in the 1950s. Little further research on the influence of HAdV on cellular metabolism was performed until the 21st century, when advances in metabolomic, proteomic and genomic technology allowed for thorough study of host-cell metabolic changes.

Figure 4

Factors influencing host-cell metabolism with HAdV infection. (A) HAdV induced changes in cellular metabolism are less drastic in cells infected at a high cellular density in comparison to cells infected at a low cellular density [31]. (B) Cell type can influence the metabolic reprogramming enacted by HAdV. Primary cells are usually slower growing than immortalized cells, which is reflected in a lower metabolic rate. Although metabolism is changed across various cell types upon HAdV infection [31], the rate of that change is likely faster in immortalized cells and contributes to rapid viral replication in immortalized cells [42,43,44]. However, even among immortalized cells, those with a phenotype more closely resembling the Warburg effect appear primed for HAdV replication and experience more drastic metabolic changes than immortalized cells with a metabolic phenotype reliant on oxidative phosphorylation [45]. (C) Growing and dividing cells infected with HAdV show more drastic metabolic changes than infected quiescent cells [46]. (D) The metabolic profile of HAdV infected cells changes throughout the course of infection [47]. Initially, HAdV infected cells typically exhibit upregulated glycolysis, amino acid metabolism and nucleotide biosynthesis pathways [47]. Towards the later stages of infection, HAdV infected cells still perform glycolysis, but the majority of metabolic activity is directed towards nucleotide biosynthesis and an upregulation of the pentose phosphate pathway (PPP) occurs [47]. (E) Different HAdV types regulate metabolism through mechanisms related to the functions of HAdV E4ORF1 proteins. Some HAdV types (e.g., HAdVF-40) do not have E4ORF1 and clearly rely on other HAdV proteins to regulate metabolism [48]. E1A, which also varies among HAdV types, is another potential regulator of cell metabolism during infection [49,50,51]. Created with BioRender.

Figure 5

Schematic of how E4ORF1 contributes to MYC-regulated transcription of genes involved in glycolysis according to Thai et al. [26]. E4ORF1 binds to MYC, enhancing the transcriptional activity of MYC, leading to increased transcription of metabolic genes such as HK2 and PFKM1. E4ORF1 can also bind glycolytic genes, which may be how E4ORF1 brings MYC into proximity of these target genes. E4ORF6 appears to play a scaffolding role and enhances E4ORF1 binding to MYC, although E4ORF6 does not appear to bind MYC or glycolytic genes itself [26]. Created with BioRender.

Figure 6

Putative mechanisms by which HAdV E1A regulates transcription of host-cell metabolic genes based on models derived from the literature [49,50,51]. (A) E1A can regulate metabolic gene expression through an interaction with the transcription factor MYC. E1A binds TRRAP, part of the NuA4 complex, which in turn is bound to MYC leading to increased transcription of metabolic genes [49]. PFKM and LDHB are two examples of transcripts that may be regulated due to this interaction based on supplementary data from Zhao et al. [49]. (B) The same paper indicated that E1A may be bound to p300 in addition to TRRAP and MYC leading to the expression of other E1A-regulated genes [49]. Again, CYP11A1 and ALG6 are two examples of metabolic genes potentially regulated by this interaction based on supplementary data from Zhao et al. [49]. (C) E1A can bind to pRB and release the inhibition of E2F-mediated gene transcription by pRB [51]. PRPS2 and PLPP3 are examples of two metabolic genes whose expression are decreased in a HAdVC-5 infection with a non-pRB binding E1A mutant compared to wild type infected cells and therefore could rely on the pRB-binding of E1A for expression during infection [51]. (D) E1A may also mediate the expression of E2F regulated genes through an interaction with DP1, which itself can bind to E2F and activate transcription [50]. No specific transcripts are shown, as this study by Pelka et al. did not include an RNA-seq component [50]. (E) Finally, an interaction between E1A, p300 and pRB may inhibit transcription of metabolism related genes through histone deacetylation [51]. GK and AKR1C3 are two genes that may be regulated by E1A binding to p300 [51]. Image created with BioRender.