key: cord-0967122-wfpojxof
authors: Gandhi, S.; Klein, J.; Robertson, A.; Pena-Hernandez, M. A.; Lin, M. J.; Roychoudhury, P.; Lu, P.; Fournier, J.; Ferguson, D.; Mohamed Bakhash, S. A.; Muenker, M. C.; Srivathsan, A.; Wunder, E. A.; Kerantzas, N.; Wang, W.; Pyle, A.; Wilen, C. B.; Ogbuagu, O.; Greninger, A. L.; Iwasaki, A.; Schulz, W. L.; Ko, A. I.
title: De novo emergence of a remdesivir resistance mutation during treatment of persistent SARS-CoV-2 infection in an immunocompromised patient: A case report
date: 2021-11-09
journal: medRxiv : the preprint server for health sciences
DOI: 10.1101/2021.11.08.21266069
sha: 7532872d30f16d046d6abe7a2205d05a78ee590e
doc_id: 967122
cord_uid: wfpojxof

SARS-CoV-2 remdesivir resistance mutations have been generated in vitro but have not been reported in patients receiving treatment with the antiviral agent. We present a case of an immunocompromised patient with acquired B-cell deficiency who developed an indolent, protracted course of SARS-CoV-2 infection. Remdesivir therapy alleviated symptoms and produced a transient virologic response, but her course was complicated by recrudescence of high-grade viral shedding. Whole genome sequencing identified a mutation, E802D, in the nsp12 RNA-dependent RNA polymerase which was not present in pre-treatment specimens. In vitro experiments demonstrated that the mutation conferred a ~6-fold increase in remdesivir IC50 but resulted in a fitness cost in the absence of remdesivir. Sustained clinical and virologic response was achieved after treatment with casirivimab-imdevimab. Although the fitness cost observed in vitro may limit the risk posed by E802D, this case illustrates the importance of monitoring for remdesivir resistance and the potential benefit of combinatorial therapies in immunocompromised patients with SARS-CoV-2 infection.

iii with persistent SARS-CoV-2 infection from whom a RDV resistance mutation was identified during recrudescence of viral shedding following treatment with the antiviral agent.

A woman in her 70s received a course of rituximab and bendamustine for the treatment of Stage Figure 1d ), but she did not experience dyspnea, hypoxemia or recurrence of respiratory symptoms during her hospitalizations and subsequent course of illness. The workup for other sources of fever, which included a bone marrow biopsy, was unrevealing. CD19 + B-cells were not identified in the bone marrow biopsy or by flow-cytometry of peripheral blood mononuclear cells (PBMC). The patient was initiated on treatment with filgrastim 1-2 times per week and monthly intravenous immunoglobulin (IVIG) to manage her neutropenia and hypogammaglobulinemia, respectively, and both were continued after hospital discharge on day 51.

The patient continued to have daily fevers, except for a 30-day period of defervescence (days 103-132), refractory neutropenia, anemia, and positive RT-PCR test results with Ct values <25 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint iv when she was hospitalized at our institution on day 145. Her chest CT on day 140 was notable for increased opacities compared with prior examinations (Figure 1d) . The patient initiated a 10day course of RDV (days 148-157), which precipitated a resolution of her fever (day 149, Figure   1a ), normalization of CRP (day 156, Figure 1c ) and improvement in the opacities on chest CT (day 162, Figure 1d ). An initial virologic response to therapy was observed, with an increase in infusion of casirivimab-imdevimab 7,8 on day 163 after expanded access for the use of the monoclonal antibody (mAb) therapy was approved. A rapid and sustained virologic response was observed with undetectable cultured virus, low sgRNA and negative NP RNA in specimens by days 164, 166 and 217, respectively. The patient's anosmia resolved roughly 17 days after administration of casirivimab-imdevimab (day 180. Figure 1a) . During the following five-month period (days 160-292), the patient did not have recurrence of COVID-19 related symptoms and had a chest CT which showed minimal opacities (day 170, Figure 1d ), and had high anti-SARS-CoV-2 S1 protein IgG titers (Figure 2a) . Peripheral blood neutrophil counts, hemoglobin levels and serum inflammatory markers normalized during convalescence (Figure 1c ).

To identify SARS-CoV-2 mutations that arose during the course of illness, we longitudinally sampled multiple sources of the patient's tissues and secretions (Supplemental Table 1 ) and performed WGS on an Illumina NextSeq platform using the Swift SARS-CoV-2 multiplex All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint v amplicon sequencing panel 9 . Viral N1 or N2 RNA was detected in 19 of 21 nasopharyngeal, 10 of 10 saliva, 10 of 10 stool and 0 of 10 whole blood specimens. Among the 39 samples with detectable N1 or N2 RNA, we assembled 27 whole SARS-CoV-2 genomes from nasopharyngeal (12) , saliva (9) and stool (6) specimens. One of the nasopharyngeal specimens was obtained from the initial phase of illness (day 36, Figure 2b ). Phylogenetic analysis found that sequenced genomes belonged to a single, non-intermixed, lineage (Pango B1) within Nextstrain clade 20C ( Figure S1 ), indicating that viral genomes identified during the course of illness were derived from intra-host diversification following infection with a single strain.

Analysis of assembled viral genomes identified a mutation, E802D, in nsp12, whose detection in patient specimens was temporally associated with RDV therapy. E802D was not identified at an allele frequency above 1% in either the specimen obtained during the initial phase of illness (day 36) or in specimens collected during the first 5 days (days 148-152) of RDV therapy. The mutation was first detected 7 days after initiation of RDV therapy (day 155) and accounted for 22 .6% and 96% of the allele frequency in nasopharyngeal and saliva specimens, respectively, by day 160. The E802 residue of nsp12 resides in the palm sub-domain which contains a portion of the residues that comprise the active site of the SARS-CoV-2 RNA polymerase 1, 10 . The E802 residue participates in an electrostatic network with D804 and K807 which stabilizes the loop involved with binding to the nascent RNA (Figure 2c , S2a). E802D has been identified in an in vitro RDV resistance selection experiment and was found to confer a ~2.5-fold increase in IC50 to the drug 6 . Together, the temporality of E802D emergence in the patient, its location in nsp12 and the in vitro identification of the same mutation with a RDV resistance phenotype support the plausibility that that RDV treatment of the patient selected for variants with the E802D mutation, which in turn contributed to the observed rebound viral shedding. All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint vi To validate the resistance phenotype of E802D, we engineered this mutation and an E802A mutation, which had been also shown to confer RDV resistance 6 , into an infectious molecular clone of SARS-CoV-2/WA01 (icSARS-CoV-2-mNG), which expresses the mNeon Green reporter and is ORF7a depleted 11 . Replication kinetics of the E802D and E802A mutants in Vero-E6 cells revealed decreased viral replication relative to parental icSARS-CoV-2-mNG ( Figure 2d ), suggesting that these mutations may impart a fitness cost. In the presence of high concentrations of RDV (5 µM), the E802D mutant replicated to higher titers than the parental virus ( Figure S2b ), confirming that this mutation confers resistance to RDV. RDV dose-response curves demonstrated that E802D and E802A mutants had significant increases in IC50 values (4.2 μM and 2.7 μM, respectively) relative to parental icSARS-CoV-2-mNG (0.7 μM) ( Figure 2e ).

Assessment of RDV cytotoxicity demonstrated limited impact on cell death rates below 10 μM, indicating that the RDV resistance phenotype conferred by mutations at E802 may occur at physiologically relevant concentrations (Fig S2c) 6 .

These findings suggest that substitutions at residue 802 in nsp12 are novel mediators of RDV resistance. We speculate that these mutations distort the active site in a way that either enables the enzyme to exclude RDV or alleviates the steric clash mediated by S861, thereby allowing the enzyme to escape RDV-mediated chain termination 1, 10 . Further biochemical and structural characterization will be needed to delineate the mechanism by which substitutions at residue E802 confer resistance to RDV.

The therapeutic impact of RDV resistance mutations at E802 may be attenuated by a fitness cost imparted by these mutations which we observed in vitro. Although RDV therapy has been widely administered to patients during the pandemic, E802D and all substitutions at residue 802 have been found in 122 (0.003%) and 270 (0.007%), respectively, of the 4.1M genome All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint vii sequences obtained from patient isolates in the GISAID database (www.gisaid.org; accessed 10/14/21). Consistent with the observation of a fitness cost, the allele frequency of E802D in nasopharyngeal specimens from the patient decreased from 22.6% to <1% between the period (day 160-162) after completion of RDV therapy and prior to casirivimab-imdevimab administration ( Figure 2b ). However, E802 allele frequencies were significantly higher (50-96% at days 158-167) in saliva specimens during this period and did not decline until casirivimabimdevimab was administered. An explanation for these discordant findings is unclear and may relate to the differential compartmentalization of variants within the respiratory tract or the stochastic nature of sampling the nasopharyngeal compartment as opposed to saliva where virus from a wider geographic distribution may pool. Given that saliva is infrequently sampled, RDV resistance mediated by E802 mutations may be under-detected. Continued surveillance will be needed to elaborate whether the occurrence of E802D, as well as other substitutions at this residue, will be limited because of an underlying fitness cost or alternatively, emerge to pose a broader risk for RDV resistance.

We identified two additional mutations, A504V in nsp14 (exonuclease) and I115L in nsp15 (endoRNAse), whose allele frequencies increased during and after RDV treatment ( Figure S3 ). However, A504V in nsp14 was present (allele frequency 85.1%) in a specimen obtained during her initial phase of illness (day 36). Although I115L in nsp15 was not detected in the early-phase specimen, the plausibility of a mutation in the endoRNAse gene which confers RDV resistance is less clear.

The case highlights additional issues for the management of SARS-CoV-2 infection in immunocompromised patients. Firstly, the course of infection was notable for being indolent despite a six-month period of sustained high-grade viral shedding and underlying lung All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint viii involvement. We identified six mutations, including a frameshift mutation in ORF3a (N257fs), which were present throughout the course of illness prior to RDV treatment ( Figure S4 ). Yet week after casirivimab-imdevimab was administered (day 170), we identified a new spike protein mutation, A348S in the RBD domain, which persisted to day 177 ( Figure S6c ). WGS did All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint ix not detect the mutation in specimens collected after day 177 because of low amounts or undetectable viral RNA. A348S is not proximal to the mAb binding sites ( Figure S6d) 16 and was not identified in a comprehensive analysis of resistance mutations 7 . In the setting of a negligible cellular immune response, mAb therapy cleared viral shedding, abrogated the residual complaint of anosmia, and resolved the blood dyscrasias.

In summary, we identified the de novo emergence of a RDV-resistance mutation, E802D, following initiation of RDV in an immunocompromised patient with persistent SARS-CoV-2 infection. While the finding is limited to a single case and requires confirmation of its generalizability in larger patient populations, it suggests that RDV can impart selective pressure in vivo to drive evolution of the virus. E802D is associated with a fitness cost in vitro which may limit the broader impact of this mutation on the development of secondary resistance during treatment and the risk for primary resistance through transmission of resistant variants. Yet, our findings underscore the importance of immunocompromised hosts with uncontrolled viral replication as a source of genetic diversification 4, 17 and selection of mutations that may potentially impart adverse consequences for antiviral therapy. Enhanced genomic surveillance of immunocompromised patients may thus be warranted. As observed in this case, initiation of anti-SARS-CoV-2 mAb may serve as a therapeutic option to achieve rapid and sustained virologic responses and improved clinical outcomes in immunocompromised patients.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint

x Online Methods

Regulatory approval to administer an 8g dose of casirivimab/imdevimab was obtained from the (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xi Biosciences, Ann Arbor, MI, USA) as described previously 9 . Libraries were quantified using fluorometric methods (Quant-iT dsDNA high sensitivity kit, Life Technologies, Carlsbad, CA, USA), normalized, and sequenced on the Illumina NextSeq 500 or 2000 instruments (Illumina, San Diego, CA, USA) using 2x150 reads.

Reads were processed using a custom bioinformatic pipeline (https://github.com/greningerlab/covid_swift_pipeline), as described 9 . Briefly, raw FASTQs were trimmed of adapter content and low-quality reads, then aligned to the NCBI reference Wuhan-1 (NC_045512). PCR primers were soft-clipped with Primerclip (https://github.com/swiftbiosciences/primerclip) and a consensus genome was generated from this alignment using bcftools v1.9 19 . Variants were examined longitudinally with a modified Python script adapted from LAVA (https://github.com/greninger-lab/lava).

A k-mer based approach was used to identify reads corresponding to subgenomic RNAs. After removing adapters and low-quality regions, junction-spanning reads corresponding to individual sgRNAs were filtered using BBDuk (https://sourceforge.net/projects/bbmap/) against a custom FASTA file composed of 30 nucleotides of the 5' leader sequence and 30 nucleotides of the respective 3' junctions of downstream ORFs 20 , allowing an edit distance of 1. In order to avoid calling of genomic reads, we constrained matches to those that had 10 consecutive matches of 31mers.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xii As described previously, patient whole blood was collected in sodium heparin-coated vacutainers and kept gently agitating at room temperature until sample pick-up by IMPACT team members.

All blood was processed on the day of collection. Plasma samples were collected after centrifugation of whole blood at 400g for 10 minutes at room temperature (RT) without brake.

The undiluted serum was then transferred to 15-ml polypropylene conical tubes, and aliquoted and stored at −80 °C for subsequent analysis. PBMCs were isolated using Histopaque (Sigma-Aldrich, #10771-500ML) density gradient centrifugation in a biosafety level 2+ facility. After isolation of undiluted serum, blood was diluted 1:1 in room temperature PBS, layered over Histopaque in a SepMate tube (StemCell Technologies; #85460) and centrifuged for 10 minutes at 1,200g. The PBMC layer was isolated according to the manufacturer's instructions. Cells were washed twice with PBS before counting. Pelleted cells were briefly treated with an ACK lysis buffer for 2 minutes and then counted. Percentage viability was estimated using standard Trypan blue staining and an automated cell counter (Thermo-Fisher, #AMQAX1000). (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. 

ELISAs were performed as previously described 22 . Briefly, Triton X-100 and RNase A were added to serum samples at final concentrations of 0.5% and 0.5mg/ml respectively and incubated at room temperature (RT) for 30 minutes before use to reduce risk from any potential virus in serum. 96-well MaxiSorp plates (Thermo Scientific #442404) were coated with 50 μl/well of recombinant SARS Cov-2 S1 protein (ACROBiosystems #S1N-C52H3-100ug) at a concentration of 2 μg/ml in PBS and were incubated overnight at 4 °C. The coating buffer was removed, and plates were incubated for 1h at RT with 200 μl of blocking solution (PBS with 0.1% Tween-20, 3% milk powder). Serum was diluted 1:50 in dilution solution (PBS with 0.1% Tween-20, 1% milk powder) and 100 μl of diluted serum was added for two hours at RT. Plates were washed three times with PBS-T (PBS with 0.1% Tween-20) and 50 μl of HRP anti-Human All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xiv IgG Antibody (GenScript #A00166, 1:5,000) diluted in dilution solution added to each well.

After 1 h of incubation at RT, plates were washed three times with PBS-T. Plates were developed with 100 μl of TMB Substrate Reagent Set (BD Biosciences #555214) and the reaction was stopped after 15 min by the addition of 2 N sulfuric acid. Plates were then read at a wavelength of 450 nm and 570nm.

As previously described 21 , Vero E6 kidney epithelial cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 1% sodium pyruvate (NEAA) and 5% fetal bovine serum (FBS) at 37°C and 5% CO2. The cell line was obtained from the ATCC and has been tested negative for contamination with mycoplasma. SARS-CoV-2, strain USA-WA1/2020, was obtained from BEI Resources (#NR-52281) and was amplified in Vero E6 cells. Cells were infected at a MOI 0.01 for four three days to generate a working stock and after incubation the supernatant was clarified by centrifugation (450g × 5min) and filtered through a 0.45-micron filter. The pelleted virus was then resuspended in PBS then aliquoted for storage at − 80°C. Viral titers were measured by standard plaque assay using Vero E6 cells. All experiments were performed in a biosafety level 3 with the Yale Environmental Health and Safety office approval.

SARS-CoV-2 titers from longitudinal nasopharyngeal swabs were assessed by conventional plaque assay. Nasopharyngeal swabs were collected by trained clinical staff and swabs stored and frozen in universal viral transport media (VTM) prior to plaquing. Each nasopharyngeal sample was ten-fold serially diluted (dilution range from 1:10 to 1:1000000). Diluted NP All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xv samples were added on a monolayer of Vero E6 cells. After 1 hour of infection, DMEM (2% FBS), and 1.2% of Avicel (Dupont, 9004346) were overlaid. 72 hours post inoculation, media was removed and 4% Paraformaldehyde added for 1 hour. Cells were stained with 0.1% of crystal violet to visualize SARS-CoV-2 plaques. Plaques were counted by single operator. Each sample was tested in duplicate.

production.

The reverse genetics system used to generate E802D and E802A SARS-CoV-2 mutants is described at length elsewhere 11 . The icSARS-CoV-2 mNG (mNeonGreen) backbone previously Steps to produce full-length infectious virus were performed as described previously 11 , with the following minor modifications: Plasmid propagation and digestion were performed as described and excision of DNA bands from the gel was done by using a Blue-Light transilluminator (Accuris). Next, equal molar amounts of the seven plasmids were reduced by half relative to the prior reports to perform the full-length genomic ligation: F1 (.305 μg), F2 (.325 μg), F3 (.375 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xvii with 500 uL of each of the clarified supernatants. Supernatants were collected for each mutant virus at 48 hours post infection and tittered as described previously.

RNA was extracted from SARS-CoV-2 mutants using the MagMAX viral/pathogen nucleic acid isolation kit, and tested with the multiplexed RT-qPCR variant assay 23 . Using a standard dilution series, Ct values were converted to RNA copies per µL. Extracted RNA was diluted to 1 million copies per µL and used as input into the COVIDSeq Test RUO version for library preparation, in duplicate. Pooled libraries were sequenced on the NovaSeq (paired-end 150) at the Yale Center for Genome Analysis. Single nucleotide variants were called at a minimum frequency threshold of 0.02 and minimum read depth of 400X using iVar (version 1.3.1) 24 , and filtered between duplicate replicates. Introduction of the E802D and E802A was confirmed at a minimum frequency of 0.82-0.9 in initial stocks.

Isolates from initial stocks were expanded and used to infect Vero-E6 cells with 0.01 MOI of each E802D and E802A mutant virus to assess their replication kinetics. Supernatants were collected at 8, 24 and 48 hours post-infection. Titers for each time point were assessed by conventional plaque assay as described above. Each virus was assessed in triplicate.

Each icSARS-CoV-2 mutant virus (E802D and E802A) was cultured in the presence of increasing concentrations of RDV, ranging from 0.39 to 40 μM. Infections were performed using All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xviii Vero-E6 in 384-well plates containing the corresponding RDV dilutions in phenol-red free DMEM media with 5% FBS and incubated at 37°C for 4 hours. Next, cells were infected with 0.01 MOI of each mutant virus and infected cell frequencies were measured at 48 hours postinfection by mNeonGreen expression by high content imaging (Cytation 5, BioTek) 12 . Cell culture supernatants were also collected for plaque assay. Cell viability in uninfected cells was assessed at 72 h post-infection using the CellTiter-Glo kit (Promega) according to the manufacturer's instructions.

As described previously 25 , patient sera was isolated from whole blood and heat treated for 30 minutes at 56 °C. Sixfold serially diluted plasma, from 1:10 to 1:2430 were incubated with SARS-CoV-2 for 1 h at 37 °C. The mixture was subsequently incubated with Vero E6 cells in a 6-well plate for 1 hour for infection. Cells were overlayed with DMEM supplemented NaHCO3, 2% FBS, and 0.6% Avicel mixture 12 . Plaques were resolved at 40h post infection by fixing in 4% formaldehyde for 1h followed by staining in 0.5% crystal violet. All experiments were performed in parallel with healthy, uninfected control sera to establish degree of protection as reflected in difference in plaque counts.

Phylogenic trees were generated using the Nextstrain software package 26 . 326 background sequences from CT, USA collected April 1 st -September 30 th 2020 were downloaded from GISAID (Supplemental Table 3 ). The tree was rooted using the Wuhan reference strain.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xix

Sequencing data are available under NCBI BioProject no. PRJNA774781 (Supplementary Table   2 ).

Grubaugh at the Yale School of Public Health for sequencing infectious clones and providing advice on the analysis and interpretation of the genomic data. Casirivimab and imdevimab was provided by Regeneron Pharmaceuticals, Inc. for the purpose of emergency compassionate use treatment. We also thank the health care staff at Yale New Haven Hospital, which include Dr.

Jensa Morris, Dr. Anne Spichler, Dr. Frederick Altice and Dr. Nikhil Seval, who cared for the patient during her hospital course and provided key input, and most of all, the patient for her equanimity and contributions to this study. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xx Contributions J.K., A.R., and M.P.H contributed to data collection, analysis, and writing. M.L., P.R., J.F., D.

F., C.V., M.C.M., A.S., E.W, N.K., P.L, W.W., A.P. contributed to data collection and analysis.

O.O., A.G, C.B.W, A.I., and W.L.S. conceived of the study and contributed to study design, interpretation, and editing of the manuscript. S.G. and A.I.K conceived of the study and contributed to study design, data interpretation, writing, and regulatory approvals. 

Mechanism of SARS-CoV-2 polymerase stalling by remdesivir

A living WHO guideline on drugs for covid-19

No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 9

IDSA Guidelines on the Treatment and Management of Patients with COVID-19

CoV-2 evolution during treatment of chronic infection

Persistence and Evolution of SARS-CoV-2 in an Immunocompromised Host

In vitro selection of Remdesivir resistance suggests evolutionary predictability of SARS-CoV-2

Antibody cocktail to SARS-CoV-2 spike protein prevents rapid mutational escape seen with individual antibodies

REGN-COV2 antibodies prevent and treat SARS-CoV-2 infection in rhesus macaques and hamsters

Sensitive Recovery of Complete SARS-CoV-2 Genomes from Clinical Samples by Use of Swift Biosciences' SARS-CoV-2 Multiplex Amplicon Sequencing Panel

Structural basis for inhibition of the RNA-dependent RNA polymerase from SARS-CoV-2 by remdesivir

Engineering SARS-CoV-2 using a reverse genetic system

Genome-wide CRISPR Screens Reveal Host Factors Critical for SARS

Diverse Functional Autoantibodies in Patients with COVID-19

A single-cell atlas of the peripheral immune response in patients with severe COVID-19

First case of persistent pancytopenia associated with SARS-CoV-2 bone marrow infiltration in an immunocompromised patient

Studies in humanized mice and convalescent humans yield a SARS-CoV-2 antibody cocktail

Successful Clearance of 300 Day SARS-CoV-2 Infection in a Subject with B-Cell Depletion Associated Prolonged (B-DEAP) COVID by REGEN-COV Anti-Spike Monoclonal Antibody Cocktail

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR

A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data

No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 9

SARS-CoV-2 genomic and subgenomic RNAs in diagnostic samples are not an indicator of active replication

Longitudinal analyses reveal immunological misfiring in severe COVID-19

A serological assay to detect SARS-CoV-2 seroconversion in humans

Multiplex qPCR discriminates variants of concern to enhance global surveillance of SARS-CoV-2

An amplicon-based sequencing framework for accurately measuring intrahost virus diversity using PrimalSeq and iVar

Delayed production of neutralizing antibodies correlates with fatal COVID

Nextstrain: real-time tracking of pathogen evolution

The authors thank Dr. Michael Jacobs, Dr. Sanjay Bhagani and Dr. David Lowe at the Royal 

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxii All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. scans of chest at indicated days after time of initial diagnosis. The timing of remdesivir and All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxv casirivimab-imdevimab are shown as grey and light blue shading, respectively. RT-PCR results that were positive but performed on assays that did not generate a Ct value are denoted by the open circle in panel b. The timing of filgrastim treatments are denoted by green diamonds in panel b. lsCRP values were converted to hsCRP values using a correction factor of 9.2. Groundglass opacities marked by white arrows in panel d. Abbreviations: Real-time polymerase chain reaction (RT-PCR); Cycle threshold (Ct) Absolute neutrophil count (ANC); Absolute lymphocyte count (ALC); Hemoglobin (Hgb); high-sensitivity C-reactive protein (hsCRP).All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxvi Fig. 2 : de novo emergence of remdesivir resistance mutation during and following treatment with the antiviral agent All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. (red). Shading represents 95% confidence intervals of the IC50 estimate.All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Table 3 ). Samples are annotated by type (nasopharyngeal (NP), saliva (SL), or stool (ST)) and day from diagnosis. Divergence from root reference genome (Wuhan-Hu-1) by nucleic acid changes.All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. 

All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxi was assessed by two-way ANOVA corrected for multiple comparisons (**** p < 0.0001) with 3 biological replicates.All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxiv (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxv CD4/CD8 subset: CD4/CD8+CD45RA+CD127+CCR7+PD-1-; activated CD4/CD8: CD4/CD8+CD38+HLA-DR+;effector memory CD4/CD8:CD4/CD8CD45RA-CD127+CCR7-; exhaustion CD4/CD8: CD4/CD8CD45RA-Tim-3+PD-1+; effTreg: CD4+CD45RA-CD127-CD25hiHLA-DR+. Gating strategy depicted in Fig S7. All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxvii light rain: red) and casirivimab (heavy chain: green; light chain: blue) binding sites.Abbreviations: nasopharyngeal (NP); saliva (SL); receptor binding domain (RBD).All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxviii Figure S7 : Gating strategy for flow cytometry Gating strategies are shown for the key T cell populations described in Figure S5 . The T cell surface staining gating strategy to identify CD8 & CD4 T cells, naïve T cells, TCR-activated T All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xxxix cells, exhaustion T cells, effector memory T cells (Tem), and effector regulatory T cells (eff Treg) are depicted.All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xl Supplemental Table 2 : Accession numbers for submitted samples All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint xli Supplemental Table 3 : GISAID acknowledgements for phylogenetic analysis All rights reserved. No reuse allowed without permission.(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint this version posted November 9, 2021. ; https://doi.org/10.1101/2021.11.08.21266069 doi: medRxiv preprint