Shedding of culturable virus, seroconversion, and 6-month follow-up antibody responses in the first 14 confirmed cases of COVID-19 in the United States

Abstract We aimed to characterize presence of culturable virus in clinical specimens during acute illness, and antibody kinetics up to six months post-onset, among 14 early US COVID-19 patients. We isolated viable SARS-CoV-2 from rRT-PCR-positive respiratory specimens collected during days 0-8 post-onset, but not after. All 13 patients with two or more serum specimens developed anti-spike antibodies; 12 developed detectable neutralizing antibodies. We did not isolate virus after detection of neutralizing antibodies. Eight participants provided serum at six months post-onset; all retained detectable anti-spike IgG, and half had detectable neutralizing antibodies. Two participants reported not feeling fully recovered at six months.


Introduction
Since identification of the first confirmed COVID-19 cases, numerous studies have described a high percentage of seroconverting patients, with detection of viral RNA in respiratory and non-respiratory specimens even after resolution of symptoms [1][2][3][4]. However, the duration and peak of infectiousness, is still not fully understood, though has been reported to peak around symptom onset [5] Given the short time frame of SARS-CoV-2 circulation, little is known about the longevity of immune response. Recent studies have indicated waning of serum antibodies by 4 months post-onset among mildly symptomatic and asymptomatic convalescent patients [6,7] while others have indicated persistence of pan-immunoglobulin antibodies and neutralizing antibodies at 4-5 months [8,9]. Furthermore, the persistence of symptoms and possible longer-term effects of SARS-CoV-2 infection need further characterization [10]. In this investigation, we sought to characterize the ability to isolate virus from patient samples from the first 14 reported symptomatic U.S. COVID-19 cases in relation to antibody seroconversion. Additionally, we examined the persistence of serum antibody responses at six months, and participant health and healthcare use in the six months since illness.

Specimen collection and human subjects research
Identification, initial interview, and specimen collection for patients A-L were described previously [1]. Patients M and N were identified shortly after these cases through screening of repatriated individuals returning from Wuhan, China. Patients were contacted by their local or state health department to participate in voluntary 6-month follow-up blood collection and interview and were excluded if they were no longer residing in the United States. Participants A c c e p t e d M a n u s c r i p t 5 were interviewed about their health and healthcare use since their COVID-19 illness using structured and open-ended questions. This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy. * Forms for the initial public health investigation were approved under Office of Management and Budget, number 0920-1011.

Cell culture, viral isolation, and microneutralizations
To characterize recovery of viable virus, available rRT-PCR positive and inconclusive specimens collected during days 0-29 post illness onset were used for viral culture (Supplemental Fig 1). Prior to virus culture, swabs in viral transport media, stool, serum and A c c e p t e d M a n u s c r i p t 6 sputum sere stored at -80°C. Patient specimens were thawed, diluted serially in DMEMdilution and Vero CCL-81 cells were innoculated [13]. Cultures showing evidence of cytopathic effect (CPE) were tested by SARS-CoV-2 rRT-PCR and viral recovery was defined as successful if the first viral passage had an N1 C t at least 2 lower than the clinical specimen. After confirmation, virus propagation continued to generate laboratory stocks.
Before microneutralization, sera were heat inactivated for 30 minutes at 56°C. Sera were diluted 2-fold in DMEM and 100 TCID 50 per well SARS-CoV-2 WA_1/USA/2020 were added to diluted sera in triplicate at 1:20-1:640. Freshly trypsinized Vero CCL-81 were added to diluted sera and virus and incubated for 5 days. Cells were fixed and stained with crystal violet fixative. Endpoint titers was defined as the highest dilution to block viral CPE in all 3 wells.

Data analysis
Data were analyzed using GraphPad Prism 7 and R v.3.6.2. Ct values for each of the three targets (N1, N2, and N3) were compared among respiratory specimens from which virus was recovered versus not recovered using a Mann-Whitney test. ELISA titers were calculated by performing Log 2 transformations, performing spline analysis, and extrapolating the titer at the cutoff of 0.4. Pan-Ig and neutralizing antibody titers were then compared among patients receiving and not receiving supplemental oxygen using a Mann-Whitney test.
Comparisons of antibody titers were performed using Log 2 data transformation and linear regression with coefficients of determination (R 2 ) presented as measures of correlation.

Results
All 14 patients were symptomatic, with symptom resolution among 13 patients reported (based on initial public health interview) at five-36 days after onset (median 13 days). Four patients received supplemental oxygen; one of whom required intensive care A c c e p t e d M a n u s c r i p t 8 We tested serum specimens from 14 patients collected 0-42 days post symptom onset for antibodies; 13 had at least two acute sera available to assess antibody kinetics. All 13 demonstrated seroconversion (i.e. detectable pan-Ig antibodies) during days three-21 (median 8 days) (Figure 2A). One patient did not develop detectable IgM and two did not develop detectable IgA. Twelve patients had neutralizing antibody titers above the level of detection (titers ≥80). The patient who did not have detectable neutralizing antibodies reported only a cough, with viable virus detected at days zero and five post-onset. The median time to neutralizing antibody detection was 10 days post-onset (range: five-28 days) and did not appear to differ by illness severity ( Figure 2B). When comparing maximum titers detected during day 7-42 post-onset in patients who did not vs. did receive supplemental oxygen, the median maximum pan-Ig titers were 5977 and 6400, respectively (Mann-Whitney test, p= 0.33), and median maximum neutralizing antibody titers were 160 and 240, respectively  Table 1). Two patients reported not feeling fully recovered; one reported intermittent headaches and insomnia since their illness or soon after, and another patient with underlying COPD reported feeling mostly well but with some ongoing chest congestion and cough felt to be slightly worse than before their COVID-19 illness.

Discussion
In summary, live virus was not detected more than 8 days post-onset, binding IgG antibodies were still detectable at six months post-onset, and neutralizing antibodies were detected in some patients at six months. These findings are consistent with those shown previously [5,8,9,14]. Two of eight participants reported still not feeling fully recovered at six months post-onset. While these symptoms were self-reported, this supports the growing understanding that long term effects of SARS-CoV-2 infection warrant further investigation.
We recovered live virus from rRT-PCR-positive respiratory specimens with C t values ranging from 16.5-32.6 using a non-quantitative assay, but did not recover live virus from A c c e p t e d M a n u s c r i p t 10 serum or stool specimens, from inconclusive respiratory specimens, or after symptom resolution, despite continued detection of viral RNA. C t values should not be used as direct marker of whether an individual is infectious because specimen handing can affect C t and the assay is not quantitative. Virus recovery indicates that a patient is potentially infectious and capable of transmission, but the opposite may not be true.
Anti-S serum antibodies were detected in all patients by day 21 post-onset, and as early as day three, in line with previous reports [2,3]. Detection of serum antibodies correlated with an inability to culture virus, also consistent with previous reports [15]. This  M a n u s c r i p t 16 Figure 1. Antibody responses, rRT-PCR data, and viral culture data, by patient, in the first 14 patients with COVID-19 in the United States. Patient numbers highlighted with an * indicate that the patient received supplemental oxygen. Antibody responses are shown as blue triangles (pan-Ig ELISA titers) and black triangles (neutralizing antibody titers). The dashed horizontal line shows the limit of detection for the antibody assays, with ELISA titers <100 and neutralizing titers <80 considered not detectable. rRT-PCR data for respiratory samples on which viral culture was attempted are shown as circles for positive specimens (all three gene targets with a Ct <40) and squares for inconclusive specimens (one or two gene targets with a Ct<40); black fill or outline indicates culture negative and red denotes culture positive. A number of key dates are also denoted on each applicable graph: a green X indicates the last reported rRT-PCR test for each patient (inclusive of positive samples for which culture was not attempted); a purple arrow indicates the first date of Remsdesivir administration (if applicable); and each black arrow indicates that last date of reported symptoms.
Note. For Patient I, the last rRT-PCR positive result was 36 days post-onset. For patient F, the last rRT-PCR positive result was 32 days post-onset. Patient H reported symptom resolution 36 days after symptom onset. For Patient M, the last day of symptoms was not available. Figure 2. Kinetics of early antibody responses are shown within the first 45 days; six-month follow-up timepoints are shown after the broken X axis. A. Anti-spike pan-Ig titers vs. days post illness, for patients who did and did not received supplemental oxygen (O 2 ). The dotted line at 100 represents the limit of detection. B. Neutralizing antibody titers for patients who did and did not received supplemental O 2 (same timepoints). The dotted line indicates the limit of detection and the dotted line at the top of the graphs represents the maximum dilution tested.
A c c e p t e d M a n u s c r i p t