Cumulative SARS-CoV-2 mutations and corresponding changes in immunity in an immunocompromised patient indicate viral evolution within the host

Different scenarios explaining the emergence of novel variants of concern (VOC) of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have been reported, including their evolution in scarcely monitored populations, in animals as alternative hosts, or in immunocompromised individuals. Here we report SARS-CoV-2 immune escape mutations over a period of seven months in an immunocompromised patient with prolonged viral shedding. Signs of infection, viral shedding and mutation events are periodically analyzed using RT-PCR and next-generation sequencing based on naso-pharyngeal swabs, with the results complemented by immunological diagnostics to determine humoral and T cell immune responses. Throughout the infection course, 17 non-synonymous intra-host mutations are noted, with 15 (88.2%) having been previously described as prominent immune escape mutations (S:E484K, S:D950N, S:P681H, S:N501Y, S:del(9), N:S235F and S:H655Y) in VOCs. The high frequency of these non-synonymous mutations is consistent with multiple events of convergent evolution. Thus, our results suggest that specific mutations in the SARS-CoV-2 genome may represent positions with a fitness advantage, and may serve as targets in future vaccine and therapeutics development for COVID-19.


Point to point response letter concerning the Manuscript (NCOMMS-21-39674A)
Authors: Thank you for considering our manuscript , formerly entitled " First evidence of in-host immune escape mutations of SARS-CoV-2 in an immunocompromised patient: a possible new source of mutations?" for publication. The title is now changed to "The mutational steps of SARS-CoV-2 to become like Omicron within seven months: the story of immune escape in an immunocompromised patient" in view of current developments of Omicron and other variants of concern (VOC).
We appreciate the critical comments of the reviewers. We have carefully considered these comments and suggestions and revised the manuscript accordingly. With these improvements, we hope that the current version can meet the Journal's standards for publication. The following is a point-by-point response to all those comments and a list of changes we have made to the manuscript. The new changes and revisions in the manuscript are highlighted in green.

Reviewer comments and point-to-point reply:
Reviewer #1 (COVID-19, convalescent plasma therapy) (Remarks to the Author): I have read the manuscript named 'First evidence of in-host immune escape mutations of SARS-CoV-2 in an immunocompromised patient: a possible new source of mutations?' with interest. Below are my comments on the manuscript: 1. Did the authors use antiviral therapy for SARS-CoV-2?
Authors: We did not use any antiviral therapy for SARS-CoV-2 and added this information now: "No antiviral therapeutics were administered to the patient at any time of the infection." Page 7, Line 225-6.
2. The authors have given the antibody titres of the patient in a timeline basis. However, the antibody levels did not reflect the antibody of the patient as the patient was given IVIG three times, on the 67th, 77th and 112nd days. We know that the commercial IVIG products include anti-SARS-CoV-2 antibodies.
Authors: Thank you very much for pointing this out. We have taken it into account in the discussion: After identification of the batch used and consultation of the manufacturer, we confirm that pre-pandemic plasma was used for preparation of the IVIG batch administered to our patient. According to the manufacturer's pre-testing of the product, no SARS-CoV-2 specific antibodies were detectable in the batch used at any time. Therefore, we may assume that the antibodies detected by us serologically originate from the patient's own immune defence.
3. The authors may add the RTX therapy date as well as IVIG date in the figure that includes the immunoglobulin levels of the patient.  4. The authors suggested in the discussion (line 501-2) that the case report documents a patient with an adequate humoral but impaired cellular immune response. The patient has very low antibody titre (301mg/dl). So, we cannot suggest that the patient has adequate humoral immune response. Another point is that the authors did not give any absolute lymphocyte count and lymphocyte subset value. So, they cannot indicate accurately that the patient has impaired cellular immune response.
Authors: We fully agree and changed the sentence: "In summary, our case report documents the medical phenomenon of persisting SARS-CoV-2 infection in an immunocompromised patient with impaired humoral and cellular immune response." Page 18, Line 560 -1 Furthermore, we checked throughout the manuscript including abstract to point out the impaired function of presumably both the humoral and cellular immune response.
We agree that the patient presents a very low total IgG concentration, probably due to the B-cell depleting therapy with rituximab and may impair the evaluations of specific antibody testing by low total IgG counts. However, in the avidity assay, the spike-specific IgG concentrations were adjusted to 100 IU/ml to avoid bias by low antibody concentrations as described in methods and references.
However, in the cellular assay the PBMC counts were adjusted to 1 x 10 6 cells/mL to avoid bias by different cell numbers. However, an impairment of cellular reactivity is likely caused by immunosuppressants and the underlying malignant condition of the patient and low CD4+ naïve T cell counts and CD4/CD8 ratio.
5. Discussion: Line 506-507: The authors have written 'With the aid of NGS, we witnessed the directed mutational changes of SARS-CoV-2 probably enforced by exclusively humoral immune defence without destruction of infected cells by cell mediated immunity.' The authors may change this sentence as 'With the aid of NGS, we witnessed the directed mutational changes of SARS-CoV-2, probably enforced by insufficient humoral immune defence.
Authors: Dear Reviewer, thank you for this improvement. We changed the sentence according to your specifications. Page 18, line 563 Minor points:

Introduction
Line 44: EDCD? Please write before abbreviation (European Center for Disease prevention and control).

Authors: We fully agree and now added the correct name before abbreviation. Line 18, page 2
Line 57: reconvalescent plasma? What is the difference from convalescent plasma?
Authors: Thank you for this advice. We changed it to the correct term "convalescent plasma" throughout the manuscript.

Authors
This manuscript describes an interesting case of a single immunocompromised individual who experienced prolonged SARS-CoV-2 infection with shedding of infectious virus. The authors demonstrate time-dependent intra-host evolution of SARS-CoV-2 including the emergence of multiple mutations associated with immune escape that have been found in VOI/VOCs. Additionally, the emergence of escape mutations primarily occurs later in disease in the setting of an improved humoral immune response (based on serology testing) that is sufficient to decrease viral titres without eradication. The main conclusion is that prolonged viral replication in immunocompromised hosts can be a source of viral evolution, including the selection of specific mutations associated with immune escape.
The mechanisms driving SARS-CoV-2 evolution and the selection of immune escape mutations is of broad interest. However, there are now many reports of SARS-CoV-2 viral evolution in immunosuppressed hosts (For example -PMID 33176080, 33831372, 33915337, 34278371, 34737266, 34431691, 34319130, 33248470, 33545711). Many of these reports have documented the emergence of viral variants harboring mutations associated with VOI/VOCs and/or humoral immune escape, including many mutations described in this manuscript. Given that this phenomenon has now been reported multiple times and many of these reports include detailed immunologic characterizations and have identified the emergence of escape mutations, this case adds support to what is already known rather than describing a new finding.
Authors: We agree that this phenomenon has been reported several times. We added the references as well as the paragraph: " Meanwhile, there are more studies that shed light on the evolution of immune escape variants in immunocompromised patients and support the results of our study (12,(62)(63)(64)(65)(66)(67)(68). Nonetheless, our study not only shows the accumulation of an unusually high number of immune escape mutations in a single patient, which to a strikingly high degree evolved in parallel in various variants of concern. The chronology of mutation events during 7 months of infection shows a rapid accumulation of non-synonymous mutations which in part were persistent, in part temporary or even repeatedly acquired and lost." Page 18, Line 549-53.
Other recommendations: -As there are other reports of in-host evolution of escape mutations during infections in immunosuppressed individuals the title "first evidence" is incorrect Authors: Thank you for giving our work your valuable time, attention and expertise. We agree that in the meantime, to some extent also due to the long handling process, it is not a "first evidence" anymore. However, in view of the dramatic dynamic of the pandemic with occurrence of VOC, we reviewed our previously submitted manuscript in the light of the appearance of the omicron variant and found an astonishing similarity to this VOC. Therefore, we reanalysed our case and revised our manuscript in view of new VOCs and propose to change the title to "The mutational steps of SARS-CoV-2 to become like Omicron within seven months: the story of immune escape in an immunocompromised patient". We have updated the manuscript to include the comparison with the Omicron variant throughout the manuscript. We think that this comparison is indeed a novelty to describe the mutational steps in an individual patient. 50% of the non-synonymous mutations acquired by the investigated strain in our patient are also described for Omicron, and 88.2% of all acquired mutations have already been described in a VOC/VOI. In fact, we found that the chronology of immune escape in our individual patient independently evolved many steps that are now present of the overall SARS-CoV-2 population, to become like Omicron and highly similar to the sum of all variants of concern within 7 months.
-SARS-CoV-2 T cell studies should include additional individuals with or without previous SARS-CoV-2 infection as positive and negative controls to ensure that the SARS-CoV-2 peptide pools / assay is functioning as expected.
Authors: Dear Reviewer, we thank you for this important comment.
For the ELISpot assay a commercially available assay was used and performed according to the manufacturer's protocol.
We clarified the assessment criteria and added a paragraph to describe our standardized laboratory procedures to avoid bias by inter-assay-variations. Pre-pandemic PBMC samples were used to determine the lower cut-off The assessment criteria for the ELISpots were a minimum of 50 SFU in the positive control and a maximum of 10 SFU in the negative control according to the manufacturer's definitions [Pantaleo G, Harari A. Functional signatures in antiviral T-cell immunity for monitoring virus-associated diseases. Nat Rev Immunol. 2006 May;6(5):417-23. doi: 10.1038. When those criteria were fulfilled, the stimulation index (SI) was calculated by dividing the mean SFU numbers in the antigen-specific wells with the mean SFU numbers of the negative control. The test was assessed negative with an SI < 2 according to previous determination of the cut-off by well-defined pre-pandemic PBMC samples and by PBMCs from SARS-CoV-2-naive individuals. The test was suggested to be poorly reactive with an SI between 2 and 7 and reactive with an SI≥7 as defined by the manufacturer [Pantaleo G, Harari A. Functional signatures in antiviral T-cell immunity for monitoring virus-associated diseases. Nat Rev Immunol. 2006 May;6(5):417-23. doi: 10.1038 and samples of patients with acute infection. According to standardized laboratory procedures, in each assay, a standard laboratory control sample of a high-reactive and a non-reactive PBMC sample, respectively, was run to determine inter-assayvariations. Only assays with less than two standard deviations of the high-reactive and the non-reactive PBMC control sample, respectively, were defined valid. Page 19, line 609 ff.
-Editing could significantly improve the presentation and clarity of the manuscript Authors: We have updated the presentation of the manuscript to include the Omicron variant and made several language and terminology improvements. Furthermore, we have been able to include all the reviewers' valuable suggestions for corrections, which, we hope, greatly improves the manuscript. The title has been changed and all figures have been updated.
Page 8, Line 225: The authors better write number instead of percentages. Authors: We agree and changed the entire paragraph: "Overall, 17 of the 22 mutations (77.3%) acquired by the investigated strain convergently evolved in other VOC, mainly in the Alpha and Omicron variants. In the spike-coding region, the proportion of acquired mutations identical to mutations of VOC is even higher -15 of the 17 mutations acquired there (88.2%) are found in other VOC. Overall, SARS-CoV-2 developed eleven persistent mutations during the study period of 140 days as well as eleven temporary mutational events. The chronology of intra-host mutational events is displayed in Fig. 3 and 4. An overview of the total number as well as a characterization of mutations accumulated by the investigated strain during the 7-month study period are shown in Table 2." Page 8, line 186 ff.
Page 5, Line 147-50: Are there are 26, 22 substitutions, or 17. Authors: The sentences 147-150 seem to cause confusion. We have deleted one sentence and shortened the list of all acquired substitutions. A precise, correct and -in our opinion -clear listing of all acquired mutations of the strain during the observation period is depicted in Table 2: Table 2: Listing of all persistent and temporary non-synonymous mutations that the strain has accumulated over the 7-months study period. n total VOC/VOI [%] n in spike VOC/VOI %