Mutational analysis of severe acute respiratory syndrome coronavirus 2 in immunocompromised patients with persistent viral detection using whole genome sequencing.

Dear Editor, During the coronavirus disease 2019 (COVID-19) pandemic of more than three years, several variants have evolved from the previously prevalent strains, being categorized as variants of concern (VOCs), variants of interest (VOIs), variants of high consequence and variants being monitored.1 The origin of new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants is unclear, but one possible explanation is that they stem from immunocompromised patients.2 Whole-genome sequencing (WGS) is a useful tool for detecting new mutations and emerging SARS-CoV-2 variants.3 Here, we used WGS to investigate the features of nonsynonymous SARS-CoV-2 mutations that appeared in immunocompromised patients with persistent viral detection during the Omicron-prevalent era. This prospective study was conducted at a 2732-bed tertiary teaching hospital from February to November 2022. We enrolled immunocompromised adults within 12 weeks of initial SARS-CoV-2 diagnosis and gathered nasopharyngeal swabs, saliva and blood samples on a weekly basis. We also performed real-time reverse transcription-polymerase chain reaction tests for SARS-CoV-2, viral cultures, plaque reduction neutralization tests, and WGS on at least two serial samples from each patient. The details of patient enrollment, sample collection, and laboratory procedures are explained in the Supporting Information. A total of 37 WGS results from 13 SARS-CoV-2 patients were included in the final analysis (Figure S1). The baseline features of the study subjects are described in Table S1. The WGS analysis results were obtained from each immunocompromised patient with a median frequency of three times (interquartile range [IQR] 2–3). The median interval between consecutive WGS analyses was 20 days (IQR 15–46 days). The patients acquired a median of two nonsynonymous mutations (IQR 1–7), excluding temporary mutations. The specific mutations compared with the Wuhan Hu-1 reference genome and the acquired mutations in subsequent WGS are presented for each patient in Figure 1 and Table 1. Among the total 87 nonsynonymous mutations, 16 (18.4%) and 13 (14.9%) mutations were classified as persistent and temporary mutations, respectively. More than half of the mutations were detected in the ORF1ab region (Figure S2). There were 29 mutations in the S region, 12 of which were associated with immune evasion (see Supporting Information). Also, 13 mutations in the ORF1ab, S, and M regions were the defining mutations of the major variants, including Omicron BA.1, BA.2.75, BA.4/5, and several XBB subvariants,4 and eleven of these mutations occurred in the S region (Table 2). The proportion of acquired mutations that were defining mutations of other variants was higher in the S region (11/29, 37.9%) than in the whole genomic region (13/87, 14.9%). Table S2 outlines the number of nonsynonymous mutations associated with immune evasion and the defining mutations of the major variants for each patient. The V792I mutation in the nsp12, also known as V5184I in the ORF1ab region, is reported to be associated with viral resistance to remdesivir.5 Patient H acquired this mutation 142 days after SARS-CoV-2 diagnosis. Before the acquisition of this mutation, the patient had prolonged exposures to remdesivir, dexamethasone, and baricitinib for 28, 17 and 15 days, respectively (Figure 2). This patient also received high-dose steroids (≥ equivalent doses of prednisolone 0.3 mg/kg daily) for more than two months. While B-cell depletion is considered the main factor affecting the period of SARS-CoV-2 shedding,6 the presence of high neutralizing antibody titers does not always ensure eradication of SARS-CoV-2 infection.7 Patient J shed the virus persistently from days 72–79 despite maintaining a high titer of neutralizing antibodies since the initial COVID-19 diagnosis (Figure 2). WGS analyses were conducted on days 79 and 98, and the missense mutation S:L452Q, detected on day 79, seemed to have ‘reverted’ by day 98. This mutation has been reported to be associated with immune evasion and a decreased sensitivity to neutralizing antibodies.8 While we could not determine the exact duration of the presence of the S:L452Q mutation or assess the status of T-cell immunity for this patient, persistent viral shedding might be attributed to this mutation and its diminished sensitivity to neutralizing antibodies. This study investigated the dynamics and characteristics of SARS-CoV-2 mutations in immunocompromised patients with persistent viral detection during the Omicron era. Each patient acquired a median of two amino acid substitutions over a median of 51 days, which equals 14.2 substitutions per year. In comparison, other studies from the pre-Omicron era reported nonsynonymous mutation rates of 24.4–52.4 substitutions per year for immunocompromised patients.7, 9, 10 While our study involved a larger cohort of such patients, some exhibited a milder immunocompromised status than those in previous studies, potentially leading to variations in the mutation rates. Several mutations seem to have emerged sporadically, distributed throughout the whole SARS-CoV-2 genome. This distribution pattern mirrors findings from an earlier study that also reported a sporadic distribution of various mutations across the SARS-CoV-2 genome in immunocompromised individuals.2, 7, 9, 10 The ORF1ab region, accounting for up to 21,290 nucleotides (71.2%) of the 29,900 total, housed more than half of the nonsynonymous mutations identified in this study. The S region, consisting of 3,822 nucleotides (12.8%), harboured about one-third of the mutations. The adjusted mutation numbers per kilobase were 2.1 for the ORF1ab region and 7.6 for the S region. Additionally, mutations known to contribute to immune escape, or those defining other variants designated as VOIs or VOCs, primarily arose in the S region. This observation aligns with findings from other studies.7, 9 The immunocompromised patients in this study were predominantly infected with BA.2 or BA.2.3 sub-lineages. We identified mutations typical of BA.4/5, BA.2.75, BQ.1 and various XBB subvariants in the SARS-CoV-2 genomes from these patients. Notably, during the pre-Omicron era, immunosuppressed patients were found to acquire nonsynonymous mutations linked to subsequent SARS-CoV-2 lineages.2, 7, 9 These observations suggest that persistent viral infections in immunocompromised patients could drive the acquisition of new mutations, leading to the adaptive evolution of SARS-CoV-2. Therefore, tracking these mutations might provide insights into viral adaptation and the advent of new SARS-CoV-2 variants. This study has several limitations. Viral evolution within populations can be influenced by infection prevalence and immune landscapes, as well as ethnic genetic predispositions.11 Our study was conducted during the Omicron-prevalent era, and the lack of data from the pre-Omicron period limits the generalization of our findings. Also, the patients in our study might not fully represent the spectrum of immunity statuses in immunocompromised patients. Specifically, eleven out of the thirteen patients in our study had hematologic malignancies, and eight had not received SARS-CoV-2 vaccines. This particular immunity profile could give rise to mutations distinct from those observed in other immunocompromised populations.11 Consequently, there might be some potential for regional or immunological biases in our findings. In conclusion, during the Omicron-prevalent era, SARS-CoV-2 genomes of immunocompromised individuals with persistent viral detection exhibited several mutations. These mutations have been reported to be associated with immune evasion, remdesivir resistance and new variant emergence. Given the rise of new subvariants with mutations associated with immune evasion or remdesivir resistance and the potential for immunocompromised individuals to shed viable viruses, decisions regarding the termination of isolation for immunocompromised patients with SARS-CoV-2 infection should be approached with caution. The authors declare no conflict of interest. Korea National Institute of Health, Grant/Award Numbers: 2022-ER1609-00, 2022-NI-043-00 and 6634-325-210; Ministry of Science and Information & Communications Technology, Republic of Korea, Grant/Award Number: NRF-2022M3A9I2017241; Ministry of Education, Republic of Korea, Grant/Award Number: 2021R1A6C101C570 All data supporting the findings of this study are available within the paper and its supplementary material and from the corresponding authors upon reasonable request. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.