Association between prognostic factors and the outcomes of patients infected with SARS-CoV-2 harboring multiple spike protein mutations

The outcome of SARS-CoV-2 infection is determined by multiple factors, including the viral, host genetics, age, and comorbidities. This study investigated the association between prognostic factors and disease outcomes of patients infected by SARS-CoV-2 with multiple S protein mutations. Fifty-one COVID-19 patients were recruited in this study. Whole-genome sequencing of 170 full-genomes of SARS-CoV-2 was conducted with the Illumina MiSeq sequencer. Most patients (47%) had mild symptoms of COVID-19 followed by moderate (19.6%), no symptoms (13.7%), severe (4%), and critical (2%). Mortality was found in 13.7% of the COVID-19 patients. There was a significant difference between the age of hospitalized patients (53.4 ± 18 years) and the age of non-hospitalized patients (34.6 ± 19) (p = 0.001). The patients’ hospitalization was strongly associated with hypertension, diabetes, and anticoagulant and were strongly significant with the OR of 17 (95% CI 2–144; p = 0.001), 4.47 (95% CI 1.07–18.58; p = 0.039), and 27.97 (95% CI 1.54–507.13; p = 0.02), respectively; while the patients’ mortality was significantly correlated with patients’ age, anticoagulant, steroid, and diabetes, with OR of 8.44 (95% CI 1.5–47.49; p = 0.016), 46.8 (95% CI 4.63–472.77; p = 0.001), 15.75 (95% CI 2–123.86; p = 0.009), and 8.5 (95% CI 1.43–50.66; p = 0.019), respectively. This study found the clade: L (2%), GH (84.3%), GR (11.7%), and O (2%). Besides the D614G mutation, we found L5F (18.8%), V213A (18.8%), and S689R (8.3%). No significant association between multiple S protein mutations and the patients’ hospitalization or mortality. Multivariate analysis revealed that hypertension and anticoagulant were the significant factors influencing the hospitalization and mortality of patients with COVID-19 with an OR of 17.06 (95% CI 2.02–144.36; p = 0.009) and 46.8 (95% CI 4.63–472.77; p = 0.001), respectively. Moreover, the multiple S protein mutations almost reached a strong association with patients’ hospitalization (p = 0.07). We concluded that hypertension and anticoagulant therapy have a significant impact on COVID-19 outcomes. This study also suggests that multiple S protein mutations may impact the COVID-19 outcomes. This further emphasized the significance of monitoring SARS-CoV-2 variants through genomic surveillance, particularly those that may impact the COVID-19 outcomes.


Discussion
Our first study in Indonesia investigates the association between several prognostic factors and disease outcomes of COVID-19 patients infected with SARS-CoV-2 harboring multiple S protein mutations. Indeed, our findings showed the effect of hypertension and anticoagulants on the severity of COVID-19 patients from the Indonesian population. Based on our study, the patients with hypertension have a ~ 17-fold higher risk of hospitalization than Table 1. Association between prognostic factors and hospitalization of patients with COVID-19. *Significant (p < 0.05). ACEI angiotensin-converting enzyme inhibitors, ARB angiotensin receptor blocker, CI confidence interval, OR odds ratio.  Table 2. Association between prognostic factors and mortality of patients with COVID-19. *Significant (p < 0.05). ACEI angiotensin-converting enzyme inhibitors, ARB angiotensin receptor blocker, CI confidence interval, OR odds ratio.

Survived (n = 44) (n, %; mean ± SD) p-value OR (95% CI)
RT-PCR Ct value 18.7 ± 5.0 19.9 ± 3.9 0.57 www.nature.com/scientificreports/ those without hypertension, in line with previous reports 11,19 . In a small retrospective study in China examining 191 patients of the early pandemic, the percentage of patients with hypertension was significantly higher in the non-survivor group than the survivor group (48% vs. 23%, respectively) 20 . The association between hypertension and increased risk of severe COVID-19 was confirmed by a meta-analysis study with a total of 2,893 patients. The study found that hypertension was associated with about a 2.5-fold increase of severe and fatal COVID-19 cases 21 .
The pathogenesis of hypertension affecting the COVID-19 severity is complex. Thus, the effect of hypertension on COVID-19 severity is controversial 22 . Indeed, a more extensive population study in England involving more than 17 million health records showed no association between hypertension and COVID-19 mortality after total adjustment 23 . Noteworthy, the impact of hypertension on the severity of COVID-19 is significant when accompanied by cardiovascular diseases, including myocardial injury 24 . However, our study did not show an association between the use of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (ACEI/ARB) and COVID-19 severity. Similar to hypertension, the effect of ACEI/ARB on COVID-19 severity is still inconclusive 23 . The S protein of SARS-CoV-2 binds to the ACE2 receptor to enter the human cells, suggesting that the use of ACEI/ARB might worsen the prognosis of COVID-19 25 . The downregulation of ACE2 resulted in the upregulation of interleukin 6, one of the pivotal mediators of cytokines storm in severe COVID-19 patients 26 . However, current reports showed that ACEI/ARB was not associated with the poorer outcomes of COVID-19 patients 27,28 .
Our study also demonstrated the association between anticoagulant therapy and COVID-19 mortality with an increased risk of approximately ~ 47-fold. SARS-CoV-2 often induces a pro-coagulative state due to several mechanisms, including endothelial dysfunction, cytokine storm, and complement hyperactivation 29 . While a recent study showed that anticoagulant therapies decreased the mortality of patients with COVID-19, it was not the case with our findings. These differences might be because we grouped the hospitalized and non-hospitalized into one group, classified into anticoagulant versus non-anticoagulant groups. Of note, we have only a limited sample size. These limitations should be considered during the interpretation of our findings. Further study with larger sample size is necessary to clarify and confirm our study.
Most previous reports focused on the impact of VOC on the COVID-19 outcomes, including B.1.1.7 (alpha), B.1.351 (beta), P1 (gamma), and the most recent VOC, B.1.617.2 (delta) [30][31][32][33][34] . Indonesia has reported identifying alpha, beta, and delta variants since January 2021 35 . In this present study, we have not found the VOC and VOI strains in our samples collected from June to October 2020 or before the first detection of VOC (B.1.1.7 lineage) in Indonesia in January 2021. Currently, the delta variant is identified as the most frequent VOC 35 . However, the actual frequency of the circulating VOCs in Indonesia might be biased due to our limited whole-genome sequencing capacities.
Interestingly, we revealed that patients with multiple S protein mutations might have a ~ fivefold higher possibility of being hospitalized than those with none or a single S protein mutation. However, the association between mutation and clinical outcome of COVID-19 is inconclusive. A study in Uruguay found that mutation in structural and non-structural protein was not associated with COVID-19 fatalities 36 . Another recent study analyzed the association between viral genomic variants and COVID-19 outcomes. They showed that 17 variants had a two-fold higher risk of severe COVID-19, while 67 variants were associated with less severe COVID-19 37 . This is in line with another study from France and the US, suggesting that different viral variants may result in different infection severity and risk of hospitalization 38,39 . Since SARS-CoV-2 is an RNA virus, its dynamic evolution is expected to influence its biological characteristic 40 , including its virulence and pathogenicity 37 . Interestingly, as an RNA virus, the critical aspect of the SARS-CoV-2 life cycle is not implied by the protein sequence 41 . Indeed, one study showed the importance of synonymous substitutions on the selection of SARS-CoV-2 42 .
All our samples, except one, contained the D614G variant. Indeed, almost all viruses circulating globally consist of the D614G mutation 35 . It has been shown that the D614G mutation was not associated with the COVID-19 illness 43,44 . A large-scale analysis of the COVID-19 Genomics UK consortium demonstrated that although D614G mutation is associated with higher viral loads, it is not associated with clinical severity and fatality of COVID-19 patients 44 . Another UK study also found no association between VOC-defining mutations with the severity of COVID-19 disease 45 .
Additionally, among 123 chronically shedding immunocompromised patients, no B.1.1.7 VOC-defining mutations were detected 45 . We also observed other S protein mutations in our samples, including L5F, V213A, and S68SR. None of the mutations lies on the receptor-binding domain (RBD) of the S protein.
Interestingly, a previous report showed that one variant in non-RBD S protein, V1176F, might lead to RBD-ACE2 binding changes and was associated with a high mortality rate of COVID-19 46 . Moreover, a recent study revealed that variants within the different proteins of SARS-CoV-2 had been associated with different patients' outcomes 47 . Further in vitro experiments and population studies are essential to clarify whether multiple non-RBD S protein variants associate with the severity of COVID-19 patients. Altogether, determining the effects of single or multiple mutations within or outside the S protein on COVID-19 severity and fatality requires caution. It cannot be inferred from in vitro laboratory experiments.
There are several limitations of our study. First, we have only a limited sample size that may result in bias in our analysis. Second, the S protein continuously evolves, resulting in new mutation(s) that may significantly affect virulence and disease pathogenesis. Third, we only analyzed mutations located within the S protein. Mutations in other structural and non-structural proteins may significantly influence the COVID-19; however, they are not investigated in our study. Fourth, our study only investigated some prognostic factors affecting the COVID-19 outcomes by overall means without considering other factors, including the vaccination status. In addition, the last clinical sample in this study was collected on December 27, 2020, while the COVID-19 vaccination program was started in our country on January 13, 2021. Thus, we suggest that the mutations on the S protein of SARS-CoV-2 were more likely due to natural selection during multiple replications among hosts.  VIRUS ID  3  14  44  73  81  91  93  95  117  129  144  23  59  60  128  129 167   VIRUS ID  196  228  231  299  325  469  663  679  822  945  1022  1115  1179  1197  1198  1230 1244  VIRUS ID  1311  1354  1396  1419  1495  1596  1642  1665  1680  1770  117  143  231  361  383  386 12    VIRUS ID  883  897  9  127  153  169  236  469  507  568  576  588  10  125  153  203  225   NC_045512.2  WUHAN  H  M  N  T  T  V  S  A  R  A  M  T  D  I  M  P  A   EPI_ISL_576130  Y  M  N  T  T  V  S  A  R  A  M  T  D  I  M  P  A   _EPI_ISL_902749  H  M  N  T  T  V  S  A  R  A  M  I  D  I  M  P  A   _EPI_ISL_911709  H  M  N  T  T  V  S  A  R  A  M  T  D  I  M  P  A   EPI_ISL_885142  H  M  N  T  T VIRUS ID  954  958  959  960  962  1259  21  54  57  66  67  71  99  104  106 144 151  VIRUS ID  171  202  207  222  223  224  260  262  275  2  7  8  35  37  210  212  47   NC_045512.2  WUHAN  S  V  F  D  T  G  M  P  N  A  T  I  L  F  H  S  H   EPI_ISL_576130  S  V  F  D  T  G  M  P  N  A  T  I VIRUS ID  194  195  199  203  204  205  234  235  10  29  11  30  32  61  64  65  66   NC_045512.2  WUHAN  S  R  P  R  G  T/G  M  S  P  Q  G  Q  T  F  E  *  Q   EPI_ISL_576130  S  R  P  R  G  T  M  S  P  Q  G  Q  T  www.nature.com/scientificreports/ Figure 1. The evolutionary history was inferred using the Neighbor-Joining method 11 . The optimal tree is shown. The percentage of replicate trees in which associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches 12 . The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Kimura 2-parameter method 13  www.nature.com/scientificreports/ Table 4. Association between multiple S protein mutations with outcomes of patients with COVID-19 and prognostic factors. *Significant. ACEI angiotensin-converting enzyme inhibitors, ARB angiotensin receptor blocker, CI confidence interval, OR odds ratio.  Table 5. Multivariate analysis of the association between prognostic factors and outcomes of patients with COVID-19. *Significant (p < 0.05). ACEI angiotensin-converting enzyme inhibitors, ARB angiotensin receptor blocker, CI confidence interval, OR odds ratio.

Conclusions
Here, our study shows that hypertension and anticoagulant therapy have a substantial impact on the COVID-19 outcomes. Moreover, we suggest the possible association between SARS-CoV-2 mutations within the S protein besides the VOC with COVID-19 outcomes. Our study further suggests the importance of genomic surveillance to monitor SARS-CoV-2 variants, particularly those that might influence the outcomes of COVID-19 patients.

Data availability
All data generated or analyzed during this study are included in the submission. The sequence and metadata are shared through GISAID (www. gisaid. org).