A new study by researchers from the University of Maryland School of Medicine, Baltimore-USA. And J. Craig Venter Institute-Maryland-USA warns that accessory gene mutations in emerging SARS-CoV-2 variants are altering pathogenesis.
The study findings found that mutations in accessory proteins may lead to less clinical disease, extended time toward knowing an infection exists in a person and thus increased time for transmission to occur. Furthermore, by inhibiting immune response and exhibiting less clinical symptoms, the new variants could cause far more long-term damage in the human host..
The study findings were published on a preprint server and are currently being peer reviewed. https://www.biorxiv.org/content/10.1101/2022.05.31.494211v1
A unique characteristic of coronaviruses is the inclusion of genes in the 3’ end of the genome that are unique to each coronavirus family.16 These genes are called “accessory genes” because, while unique to each family, they are not essential for replication.17
Many of these genes across the coronavirus family have been shown to alter host pathways including interferon signaling, cell cycle progression, and other assorted anti-viral responses to viral infection that enhance replication and pathogenesis.9,10,18,19
We designed two ways to analyze the functional consequences of the accessory genes in SARS-CoV-2 in vivo. First, we produced deletion viruses that deleted each ORF from our infectious clone of SARS-CoV-2. Second, we produced mutant SARS-CoV-2 viruses that contained the spike of each previously circulating variant.
As variants of SARS-CoV-2 have emerged, the increasing incidence of mutations both within spike and outside of spike were noted. Although, the mutations in spike may enhance entry and kinetics of infection, the mutations observed in the other genes may alter disease severity through interactions with the host immune system.
In our experiments we sought to differentiate between the role of spike mutations in each variant and those in other genes of the variant genome. As many of these non- spike mutations in the genome are in the accessory ORFs, we decided to couple our work with our variant spikes in WA-1 viruses with our work with accessory deletion viruses. This allows us to determine which accessory ORFs contribute significantly to pathogenesis through infection with our deletion viruses, and to link mutations in the corresponding variant ORFs to impacts on ORF function and pathogenesis.
Our work with the accessory deletion viruses has revealed that ORF3a and ORF3b contribute significantly to viral replication in K18-hACE2 mice. Mice infected with a WA- 1ΔORF3a/b deletion virus demonstrated attenuated weight loss and significantly reduced lung titers at both day 2 and day 4.
Our finding that an ORF3a/b deletion virus is attenuated in mice is supported by previously published work as does our finding that WA-1ΔORF6 is not attenuated in mice.20 This paper also found that deletions of ORF7a, ORF7b, and ORF8 show reduced lung titer, which we did not find in our experiments. This may be due to inoculation differences in our
model where mice were infected with a non-lethal dose of 103 pfu instead of 105 pfu as used in the previous study. It is possible that we may see attenuation in weight loss and viral replication of WA-1ΔORF7a/7b, and WA-1ΔORF8 at higher doses or in different mouse backgrounds, which were not studied here.
The lung cytokine and chemokine profiles of mice infected with WA-1ΔORF3a/b demonstrated reduced expression of inflammatory cytokines and chemokines compared to WA-
1. This difference is most likely attributed to the fact the WA-1ΔORF3a/b virus demonstrates attenuated replication in mice. As expected, the attenuated replication of WA-1ΔORF3a/b results in the downregulation of cytokines and chemokines involved in neutrophil recruitment, including Ccl7, Csf3, Cxcl1, Cxcl3, and Cxcl5. There is a small subset of genes that are upregulated in the WA-1ΔORF3a/b lungs. Two of these upregulated genes are Il4 and Il5, which function to drive the Th2 response.
Interestingly, COVID-19 is known to skew to a Th2 response through stimulating the production of Il4 and Il5.21 Given that the WA-1ΔORF3a/b virus is attenuated compared to WA-1, we would expect a downregulation of these cytokines. The downregulation of Il4 might in part be explained by its role in the tissue remodeling process, with the faster clearance of WA-1ΔORF3a/3b allowing for tissue repair to take place.22
Another of these upregulated genes, adipoq, encodes the insulin-sensitizing hormone, adiponectin. Interestingly, reduced adiponectin levels are associated with severe respiratory failure in COVID-19 patients.23 Future work to characterize the role of these cytokines and chemokines SARS-CoV-2 pathogenesis will focus on how these pathways interact with the accessory genes of the virus.
As the emergence of variants of SARS-CoV-2 occurred, we noted the appearance of point mutations in the accessory genes of SARS-CoV-2. To elucidate the impact of non-spike mutations on viral pathogenesis, we synthesized variant spikes in the WA-1 backbone. We focused on the synthesis of the B.1.351, B.1.1.7, and P.1 lineages. When comparing the in vitro replication of the parent variants to their paired variant spikes in WA-1, we did not see many significant differences in replication in VeroE6 cells. The only significant difference we saw was a significant decrease in supernatant viral titer at 72 hours of P.1 S in WA-1 when compared to P.1.
In the K18-hACE2 mice, the absence of in vitro replicative differences for B.1.351 S in WA-1 and B.1.1.7 in WA-1 were reflected. Mice infected with B.1.351 in WA-1 exhibited no differences in weight loss, lung titer, and brain titer compared to mice infected with B.1.351.
This was also true for mice infected with B.1.1.7 when compared to mice infected with B.1.1.7 S in WA-1. Results from the BALB/c experiment differed significantly for the B.1.351 S in WA- 1/B.1.351 pairing, with B.1.351 S in WA-1 showing attenuated weight loss and lung titers compared to B.1.351. The mice infected with B.1.1.7 S in WA-1 showed an increase in lung titer on day 2, but no difference in weight loss was seen when compared to B.1.1.7.
The most significant replicative differences in both BALB/c and K18-hACE2 mice was seen with P.1 and P.1 S in WA-1. In both sets of mice, mice infected with P.1 S in WA-1 demonstrated increased weight loss and significantly higher lung titers by plaque assay and qPCR than mice infected with P.1. In the K18-hACE2 mice, the brain RNA and brain titers trended with this data as well, although the values were not significant.
Given the significant differences in P.1 S in WA-1 and P.1 replication in vivo, we analyzed cytokine and chemokine gene expression on both the day 2 and day 4 lungs of these mice. Compared to P.1, the P.1 S in WA-1 cytokine and chemokine profiled differed the most on Day 2. Most of the significant changes seen involved the upregulation of genes involved in neutrophil recruitment. The most upregulated gene was cxcl5, which has been implicated as the major chemoattractant for neutrophils in SARS-CoV-2 and a major cause of inflammation.24 It is therefore not unexpected that this gene would be upregulated due to the increased lung titer of
P.1 S in WA-1 compared to P.1. The downregulated genes of interest include thpok and Il5, which are important for driving the differentiation of CD4+ T-cells. This finding could suggest that there is some inhibition of the CD4+ T-cell recruitment pathway by the non-spike genes in WA-1 that is attenuated or absent in P.1. There is also evidence of the loss of early interferon antagonism in the non-spike genes in P.1, as IFNγ levels in lungs from the P.1 S in WA-1 mice were reduced at day 2 compared to the P.1 mice despite the higher lung titers in the mice infected with P.1 S in WA-1. However, this difference was not seen at Day 4.
Together this work demonstrates that ORF3a/b have substantial roles in pathogenesis and host responses to SARS-CoV-2. It is interesting to note that two of the variant spike in WA-1 viruses studied here, B.1.351 and P.1, possess mutations in ORF3a. This points to the fact the emergence of mutations in the variant accessory ORFs, particularly ORF3a, contribute to viral pathogenesis.
The possible advantageous effects on viral fitness and likelihood of transmission of these mutations is supported by the continued identification of mutations in these ORFs in new variants. We interpret this data to suggest that mutations outside of spike may be driving critical phenotypes of SARS-CoV-2 infection and disease. Although spike mutations may allow for better engagement of ACE2 to facilitate entry into cells as well as evasion of antibodies, the mutations in the accessory proteins may actually be negatively impacting disease leading to less severe clinical phenotypes.
The balance of these two strategies may confer longer courses of virus replication and spread in the lungs while allowing for increased time for virus transmission from one infected person to their contacts. We hypothesize that this balance is critical for further evolution of SARS-CoV-2 and, as more variants emerge, we will identify additional mutations outside of spike that contribute significantly to viral replication, transmission, and pathogenesis.