Within-host evolution of SARS-CoV-2 in an immunosuppressed COVID-19 patient as a source of immune escape variants

Abstract

The origin of SARS-CoV-2 variants of concern remains unclear. Here, we test whether intra-host virus evolution during persistent infections could be a contributing factor by characterizing the long-term SARS-CoV-2 infection dynamics in an immunosuppressed kidney transplant recipient. Applying RT-qPCR and next-generation sequencing (NGS) of sequential respiratory specimens, we identify several mutations in the viral genome late in infection. We demonstrate that a late viral isolate exhibiting genome mutations similar to those found in variants of concern first identified in UK, South Africa, and Brazil, can escape neutralization by COVID-19 antisera. Moreover, infection of susceptible mice with this patient's escape variant elicits protective immunity against re-infection with either the parental virus and the escape variant, as well as high neutralization titers against the alpha and beta SARS-CoV-2 variants, B.1.1.7 and B.1.351, demonstrating a considerable immune control against such variants of concern. Upon lowering immunosuppressive treatment, the patient generated spike-specific neutralizing antibodies and resolved the infection. Our results suggest that immunocompromised patients could be a source for the emergence of potentially harmful SARS-CoV-2 variants.

Fig. 1. Summary of the clinical course of the SARS-CoV-2-positive kidney transplant patient.

Temporal overview of (a) hospitalization, (b) immunosuppressive treatment (daily dose in mg/day), and (c) antiviral therapy (daily dose in mg/day), Remdesivir was given 200 mg on the first day and 100 mg/day 2 to 10. d Diagnostic SARS-CoV-2 RT-qPCR cycle threshold (Ct) values of oropharyngeal swabs over time. Day 0 indicates the first positive RT-qPCR result in March 2020, 12 days after kidney transplantation. The dotted line indicates the cutoff value (Ct ≥ 40) between positive and negative results. e Attempts of virus isolation from oropharyngeal swabs. f Detection of spike S1-subunit- and nucleoprotein (N) specific antibodies by ELISA. The dotted line indicates the anti-S1 ELISA cutoff at 1.1 arbitrary unit (AU).

Fig. 2. SARS-CoV-2 whole genome sequencing and phylogenetic analysis.

Phylogenetic analysis of the viral sequences obtained from patient swabs between day 0 to day 140, after the first positive RT-qPCR result in March 2020. The sequences were aligned to a set of representative SARS-CoV-2 genome sequences from the Freiburg area (a) and from Germany (b) between February and April 2020 which have been deposited in the GISAID data bank (Supplementary table 2 and 3). The circularized maximum-likelihood phylogenetic tree was constructed with IQ-Tree (GTR + F + I) and rooted on the Wuhan-Hu-1 reference sequence (NC_045512). The sequences obtained from the immunosuppressed patient are indicated as red dots and lineage-defining mutations are indicated at the respective branches. The scale represents nucleotide substitutions per site. c Schematic overview of the viral genome variations from patient swab samples (day 0-140) in comparison to the Wuhan-Hu-1 reference sequence. The heatmap summarizes the positions in the viral genome and the variant frequencies in the different samples (cut off values of 25 and 10% for the S gene, respectively). The days of sampling are indicated at the right and the heatmap color intensity indicates variant frequencies. Stars denote non-synonymous mutations leading to amino acid substitutions in the spike protein (> 50 % of reads). d Schematic overview of the SARS-CoV-2 spike protein including the S1 and S2 cleavage products and functional domains such as the N-terminal domain (NTD), receptor-binding domain (RBD), receptor binding motif (RBM), S1/S2 proteolytic furin cleavage site, fusion peptide (FP), heptad repeat regions (HR1/HR2), transmembrane domain (TM) and C-terminal domain (CT). Selected non-synonymous changes in the spike (S) gene from panel c are indicated. e Summary of mutations found in the spike protein of the patient sequences obtained on d14, d105, and d140 (>50 % of reads) in comparison to circulating new variants of concern: alpha, B.1.1.7, beta, B.1.351, and gamma, P.1.

Fig. 3. Structure of the SARS-CoV-2 spike trimer.

The spike structure (PDB accession number: 7BNM) with the most prominent mutations in the patient viral sequences is shown in the surface presentation. The NTD is colored in blue and the RBD in red. Close-ups of the single NTD and RBD regions defined by boxes are presented as ribbons. The location of the deletions in the NTD and amino acid substitutions in the RBD are indicated by black residues. Furthermore, the deletions in the NTD are displayed as amino acid alignments at the right.

Fig. 4. The late SARS-CoV-2 isolate with mutations in the spike protein does not affect viral fitness.

a Schematic overview of the sequence variations in SARS-CoV-2 genomes detected in early (d14) or late (d105) swab samples and isolated viruses. The heatmap illustrates the positions and the frequency of major variations in the viral genome (cut off 10%). The days of isolation are indicated at the right. The heatmap colors represent the variant frequencies. In ORF7b, L14* indicates a frameshift mutation due to a deletion of two nucleotides. b Immunofluorescence analysis of SARS-CoV-2 infected cell cultures. VeroE6 cells were infected with the virus isolates, d14 or d105, and the prototypic lineage B.1 isolate, Muc-IMB-1, using 0.1 plaque-forming units (pfu)/cell. At 8 h post-infection, the cells were fixed and stained with SARS-CoV-2 N- and S-specific antibodies (red). In addition, F-actin (white) and nuclear DNA (DAPI, blue) were detected. The scale bar represents 10 µm. c Western blot analysis of viral protein expression. Calu-3 cells were infected with 0.001 pfu/cell. Cells were lysed 8 h, 24 h, 48 h, and 72 h post-infection and analyzed using N- and S-specific antibodies. Detection of β-actin was used as a loading control. Panels b and c show representative data of two independent experiments. d, e Growth of the two patient isolates in VeroE6 (d) and Calu-3 (e) cells infected with the d14 or d105 isolates using 0.001 pfu/cell. At different time points post-infection, cell culture supernatants were collected and viral titers were determined. The log-transformed titers are shown as means ± SD of results from three independent experiments. Significance was determined via two-way ANOVA with a Sidak´s multiple comparison test, **p < 0.01, ***p < 0.001, ns=non significant. f and g In vivo infection experiments. Weight loss (f) and survival (g) of 8 to 12 weeks-old K18-hACE2 mice intranasally infected with 200 pfu of d14 (n = 5), 2000 pfu of d14 (n = 4), 200 pfu of d105 (n = 7) or 2000 pfu of d105 virus (n = 7). Signs of disease and body weight loss were monitored daily for 14 days. In panel f, data are presented as mean values ± SEM. Source data are provided as a Source Data file.

Fig. 5. Delayed seroconversion and viral escape from the spike protein-specific antibody response.

a, b Detection of neutralizing activity of immune sera against SARS-CoV-2 variants. 100 pfu of the d14 and d105(del) isolates were incubated for 60 min at room temperature with serial dilutions of the patient sera. Sera obtained from naïve (– ctrl) or convalescent individuals (+ ctrl) served as negative and positive controls. Virus neutralization was determined by plaque assay on VeroE6 cells. Virus titers are indicated as percentages (mean ± SD) of the titer of the untreated virus inoculum. The dotted lines indicate the cutoff value between positive (<50%) and negative (>50%) neutralization. Shown are the means of three biological replicates. a Sera from the immunocompromised patient. The times of blood withdrawal are indicated. b Convalescent sera from COVID-19 patients suffering from mild, moderate, or severe disease or human post-vaccination (BNT162b2 mRNA) sera. c–e Neutralization capacity of SARS-CoV-2 antisera using VSV*∆G(FLuc) vector pseudotyped with the SARS-CoV-2 spike protein and coding for firefly luciferase. The pseudotyped viruses were incubated with serial dilutions of a COVID-19 convalescent serum prior to inoculation of VeroE6 cells. Pseudotyped virus infection was monitored 16 h post-infection by measuring the firefly luciferase activity in the cell lysates. The control without serum was set to 100%. c Neutralization of VSV*∆G(FLuc) pseudotyped with the early and late SARS-CoV-2 spike variants (d14 and d105). d, e Neutralization of VSV*∆G(FLuc) pseudotyped with the d14 spike protein containing the individual or combined mutations found in the late d105 and d140 sequences. Immune sera from two different convalescent COVID-19 patients (c, d) or a vaccinated person (e) were analyzed. The neutralization was determined by calculating the NT₅₀ via a non-linear regression (variable slope, four parameters). Shown are means ± SD (n = 3). Statistics were calculated with a one-way ANOVA (Tukey’s multiple comparison test), ns = non-significant, *p < 0.05, **p < 0.01, ***p < 0.001. The exact p-values are given in the figure. Source data are provided as a Source Data file.

Fig. 6. Late SARS-CoV-2 d105 isolate elicits cross-reactive protective immunity in mice.

Sera were collected from convalescent K18-hACE2 mice at least 21 days post-infection with Muc-IMB-1 (n = 5 blue), d14 (n = 4), d105(del) (n = 13, red) or d105 (n = 2, orange) virus isolates. a Anti-SARS-CoV-2 IgG titers of serially diluted sera (mean ± SD) were determined using virus-infected cells and indirect immunofluorescence analysis (IFA). b, c Neutralization of d14 and d105 virus isolates by convalescent mouse sera obtained after infection with wild-type SARS-CoV-2, d14 and Muc-IMB-1 (anti-wt sera, n = 9) (b), or with variant d105 virus isolate (pooled data using anti-d105 (n = 2, orange) and anti-d105(del) sera (n = 13, black)) (c). Neutralization capacity was determined by incubating 400 pfu of either virus isolate with serial dilutions of the mouse sera. The mixture was then applied to VeroE6 cells and infected cells were stained with N-specific antibodies. d, e Neutralization of B.1.1.7 and B.1.351 variants of concern by mouse convalescent sera was determined as described in panels b and c. Neutralization titers, NT₅₀, are meant as the highest dilution for each individual serum causing 50% reduction of infectivity. Each serum titer (b–e) is shown as the mean out of two independent experiments. Significance was determined via a two-tailed, paired t test with *p < 0.05, **p < 0.01, ***p < 0.001. The exact p-values are given in the figure. f, g Convalescent animals are protected against re-challenge infection. Weight loss (f) and survival (g) of convalescent K18-hACE2 mice (mean ± SEM), challenged one to four months after the prime infection. Animals primarily infected with d14 and Muc-IMB-1 viruses (pooled wt survivors, n = 7), or with d105(del) virus (n = 14) were intranasally challenged with 100,000 pfu of d14 or d105(del) viruses (2–7 mice per group, as indicated). As a control, naïve 8 weeks old K18-hACE2 mice were intranasally infected with 100,000 pfu of d14 or d105(del) isolate viruses (n = 2 per group). Source data are provided as a Source Data file.