Early SARS-CoV-2 dynamics and immune responses in unvaccinated participants of an intensely sampled longitudinal surveillance study.

Background: A comprehensive understanding of the SARS-CoV-2 infection dynamics and the ensuing host immune responses is needed to explain the pathogenesis as it relates to viral transmission. Knowledge gaps exist surrounding SARS-CoV-2 in vivo kinetics, particularly in the earliest stages after exposure.
Methods: An ongoing, workplace clinical surveillance study was used to intensely sample a small cohort longitudinally. Nine study participants who developed COVID-19 between November, 2020 and March, 2021 were monitored at high temporal resolution for three months in terms of viral loads as well as associated inflammatory biomarker and antibody responses. CD8 + T cells targeting SARS-CoV-2 in blood samples from study participants were evaluated.
Results: Here we show that the resulting datasets, supported by Bayesian modeling, allowed the underlying kinetic processes to be described, yielding a number of unexpected findings. Early viral replication is rapid (median doubling time, 3.1 h), providing a narrow window between exposure and viral shedding, while the clearance phase is slow and heterogeneous. Host immune responses different widely across participants. Conclusions: Results from our small study give a rare insight into the life-cycle of COVID-19 infection and hold a number of important biological, clinical, and public health implications.

Introduction

A comprehensive understanding of early infection viral dynamics and associated host immune responses is key to describing the underlying disease pathogenesis and is needed to inform effective public health measures and clinical management policies. Characterizing the viral load kinetics in a number of diverse patient populations also can be instrumental in developing new antiviral drugs and therapies. Advances in the management of acute or chronic viral diseases—such as influenza[1], human immunodeficiency virus (HIV)[2,3], and hepatitis C virus (HCV)[4]—were aided by foundational studies on clinical viral dynamics. There remain a number of knowledge gaps surrounding severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) kinetics in coronavirus disease 2019 (COVID-19) patients, and how the viral dynamics interplay with disease progression. Infections with SARS-CoV-2 can be described by two main stages—the viral proliferation and clearance phases—that typically end with a long tail of low-level, persistent viral RNA shedding.

A number of longitudinal clinical studies have examined the SARS-CoV-2 clearance phase[5-9] after the establishment of infection, but little is known about the rapid, exponential proliferation (i.e., viral growth) phase after exposure. Characterization of the early phase of the viral life-cycle is challenging due to its occurrence before symptoms, if any, and its short duration. Prospective, observational clinical studies to investigate this phenomenon would require large participant cohorts committed to frequent serial sampling, which is logistically difficult.

Controlled human challenge studies have been successful at elucidating the viral kinetics of milder diseases than COVID-19, where effective treatment strategies were available, such as human influenza[1]. In this model, volunteers are deliberately exposed to an infectious challenge agent to study the subsequent infection and the potential benefits of experimental interventions (e.g., antiviral agents, vaccine candidates). Human challenge studies using SARS-CoV-2 could overcome some of the practical limitations of observational clinical studies as participants would be closely monitored in a controlled setting. However, given our limited understanding of COVID-19 and the potential for significant morbidity associated with acute disease presentation as well as persistent, long-lasting symptoms (i.e., so-called “long-COVID”), human challenge studies involving SARS-CoV-2 are controversial and face an ethical dilemma that has been the subject of considerable debate[10-14]. Controlled infection models also suffer from a number of scientific limitations borne out of their inherent artificial nature, such as the choice of the viral strain, the size of the viral inoculum, the mode of inoculation, and the age of the participants, as only young and healthy subjects typically can be enrolled. Two SARS-CoV-2 human challenge studies are ongoing in the United Kingdom[15]. One study reported that 18 out of 34 volunteers (aged 18–29 years) became infected following intranasal inoculation with wild-type virus (SARSCoV-2/human/GBR/484861/2020), and no serious safety signals were detected[16].

Deepening our nascent understanding of the SARS-CoV-2 dynamics can hold important implications for managing the pandemic. For example, an effective strategy for curbing the spread of SARS-CoV-2 relies on the rapid, early identification of infected individuals followed by isolation. Test-based screening is playing a critical role in these efforts, as symptom presentation is not a reliable indicator of infectiousness[17]. Since March 23, 2020, we have been conducting a continuous, ongoing workplace clinical study involving the longitudinal and intensive characterization of COVID-19 prevalence and incidence at the Oak Crest Institute of Science (Oak Crest), a nonprofit scientific research organization in Southern California. The intensely sampled observational surveillance study has enabled unvaccinated participants who developed COVID-19 to be identified in the early stages of the viral proliferation phase and allowed them to be followed at high temporal resolution. The kinetics of SARS-CoV-2 production and clearance, along with the concomitant host immune responses, reported here hold a number of important biological, clinical, and public health implications, as discussed in detail below.

Methods

Ethics statement

All human research under OCIS-05, “Longitudinal Characterization of COVID-19 Prevalence and Incidence in a Small Working Institution with Both Public Health and Diagnostic Aims”, was approved by Aspire IRB (Aspire Study # 1281548) and conducted according to the Declaration of Helsinki. All study participants provided written informed consent or assent. There were some minors participating in the study who were capable of understanding and signing an IRB-approved assent form in addition to the IRB informed consent form completed by their parent or legal guardian.

Clinical study design

The workplace SARS-CoV-2 surveillance clinical study was initiated by the Oak Crest Institute of Science (Oak Crest, https://www.oak-crest.org/), a small nonprofit academic science research organization located in Monrovia, CA, on 23 March, 2020, has been running without interruptions and is ongoing at the time of writing. The study design has been described in detail elsewhere[18]. Briefly, all Oak Crest employees, students, and volunteers were asked to participate in the prospective, longitudinal, observational study designed to last 12 weeks, or longer. Those choosing not to participate had no negative employment or finance-related consequences but were asked to work from home exclusively. Household members from the above-described study population also were invited to participate in the study. Swab samples (nasal and oral) were collected between 8:30 and 9:00 AM three times per week—with the exception of periods of low, local SARS-CoV-2 positivity rates where the testing frequency was reduced to twice weekly—from participants needing access to the Oak Crest facilities while they were isolated in their motor vehicles.

Saliva and blood sample collection and processing

Optional saliva samples were self-collected in Falcon tubes (50 ml) at the participant’s home or in their sealed vehicle, and stool swabs were collected at the participant’s home. Specific written instructions were provided to participants opting to provide these specimens. Blood (5–8 ml, ×2) was collected for cytokine and antibody testing by a licensed phlebotomist using Vacutainer (Becton, Dickinson and Company, Franklin Lakes, NJ) tubes for serum (spray-coated silica) and plasma (spray-coated K2EDTA) in the Oak Crest parking lot, while the participant remained comfortably seated in their vehicle.

For the analysis of cellular responses to SARS-CoV-2, blood was collected in vacutainer vials (362753, Becton, Dickinson and Company) by standard venipuncture and centrifuged in a horizontal rotor (i.e., swing-out bucket) for 15 min at 2000 × g, room temperature. The peripheral blood mononuclear cell (PBMC) layer aggregated in a whitish layer just under the plasma layer, which was removed to separate vials and stored at −80 °C for future analysis. The PBMC layer was removed into a sterile 15 ml conical centrifuge tube, taking care not to disrupt the separation. Cell media (RPMI 1640, 11835030, Thermo Fisher Scientific, Waltham, MA) was added up to the volume of 10 mL while resuspending the cells. An aliquot (10 µL) was removed for cell counting, and the remaining sample was centrifuged again at 2000 × g for 15 min. The resulting supernatant was aspirated, and the PBMC pellet was resuspended in freshly prepared, ice-cold freezing media consisting of dimethyl sulfoxide in fetal bovine serum (10% v/v, 26140079, Thermo Fisher Scientific) to a final concentration of 3.0 × 106 cells mL−1. The cellular suspensions were dispensed as 1.0 mL aliquots into pre-chilled, labeled vials. The PBMC samples were stored at −80 °C for 24 h and then transferred to liquid nitrogen storage until use.

Analysis of clinical nasal swab specimens by transmission electron microscopy (TEM)

Nasal swab samples collected from participants who tested positive for SARS-CoV-2 RNA by RT-qPCR and a SARS-CoV-2 RT-qPCR negative control were stored directly either in glutaraldehyde in PBS (5% v/v) or formaldehyde in PBS (8% w/v). These fixative concentrations were chosen two-fold above our standard mixture to maintain concentrations above accepted SARS-CoV-2 inactivation thresholds under all circumstances. The samples were allowed to react at room temperature for 12 h to further ensure complete virus inactivation and then stored at 4 °C until sample preparation. Swab segments were cut with a razor blade, and formaldehyde-fixed samples were fixed further in glutaraldehyde in PBS (1% v/v). The segments were post-fixed in aqueous osmium tetroxide (2% w/v), block-stained in aqueous uranyl acetate (1% w/v), dehydrated in an ethanol series, and embedded in Spurr’s resin. The resulting blocks were sectioned 50–70 nm thin and collected on formvar filmed 2 × 1 mm slot grids, stained with aqueous uranyl acetate (2% w/v) and Reynolds lead citrate, and imaged at 80 kV in a Model EM10 (Carl Zeiss AG, Oberkochen, Germany) TEM equipped with an Gatan Erlangshen ES1000W (Pleasanton, CA) CCD camera. Images were enhanced for brightness/contrast as needed using ImageJ.

Measurement of SARS-CoV-2 viral loads in clinical samples by RT-qPCR

The samples (nasal, stool, saliva) were analyzed for SARS-CoV-2 RNA copy numbers by reverse transcription (RT) and quantitative PCR (qPCR) using primer sequences targeting the SARS-CoV-2 nucleocapsid protein (N) gene transcript fragments (N1 and N2) and one human RNase P (RP) gene transcript fragment (RP). Complete methods have been reported elsewhere[18]. Test results typically were available at 1 PM on the same day as when they were collected.

Calculation of SARS-CoV-2 doubling time

The in vivo SARS-CoV-2 doubling time (T) during the exponential growth phase (i.e., proliferation phase) was calculated from the corresponding rate constant (k) according to Eq. 1 and Eq. 2:where, y is the SARS-CoV-2 RNA copy number per swab; y is the initial SARS-CoV-2 RNA copy number per swab; and t is time.

Model fitting of the temporal SARS-CoV-2 concentration trajectories

The employed model generally was based on the framework described by Kissler et al.[8] (available at: https://github.com/gradlab/CtTrajectories). The model used viral load concentration-time data, using the cycle threshold (C) values measured by RT-qPCR. The C value represents the number of thermal cycles needed to amplify the viral RNA, following transcription into complementary DNA (cDNA), to a detectable signal. Since we only had one group in our analysis pipeline, we did not use the hierarchical structure component (i.e., Variant versus NonVariant). We removed sequences of three, or more, consecutive negative test results (C = 40) to avoid overfitting to these trivial values. For the main analysis, prior information was used from a previous analysis[8]. We also conducted a sensitivity analysis using vague priors as well as a strongly biased set of priors to assess robustness to the choice of prior. The settings for these priors are presented above in the Results section.

Plasma cytokine concentration analysis

The concentration of 21 inflammatory markers in cryopreserved plasma samples was measured using the MILLIPLEX® human cytokine, chemokine, and growth factor panel (HCYTA-60K-21C, EMD Millipore, Burlington, MA) bead-based multiplex assay on a MAGPIX® instrument (EMD Millipore) according to the manufacturer’s instructions. The analytes were: soluble CD40L (sCD40L), granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon alpha-2 (IFN-α2), interferon gamma (IFN-γ), interleukin-1 alpha (IL-1α), interleukin-1 beta (IL-1β), interleukin-1 receptor antagonist (IL-1Ra), interleukin-2, −4, −6, −8, −10, −12 p70, −13, −15, −17A (IL-2, IL-4, IL-6, IL-8, IL-10, IL-12 p70, IL-13, IL-15, IL-17A), interferon γ-induced protein-10 (IP-10), monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1 alpha (MIP-1α), macrophage inflammatory protein-1 beta (MIP-1β), and tumor necrosis factor alpha (TNF-α). Measurements below the lower limit of quantification were not reported.

Quantification of serum IgG, IgM, and IgA against SARS-CoV-2

Measurement of serum anti-receptor-binding domain (RBD) IgG, IgM, and IgA concentrations was carried out using an enzyme-linked immunoassay (ELISA) using methods described in detail elsewhere[18,19]. Briefly, 96- well microtiter plates were coated with 2 µg mL−1 recombinant RBD protein in calcium- and magnesium-free phosphate-buffered saline (PBS), followed by triple-washing with PBS containing 0.1% Tween-20 (TPBS), and incubation with PBS containing 3% dried milk (Bioworld, Dublin, OH) 1 h at room temperature (RT). Participant serum was added in duplicate serial dilutions, incubated for 2 h at RT, and washed three times with TPBS. Bound antibodies were detected using goat anti-human IgG, IgM, or IgA conjugated with horseradish peroxidase (Bethyl Laboratories, Montgomery, TX), added in PBS at 1:50,000 at RT for 1 h. After three washes with TPBS, tetramethylbenzidine substrate solution (100 µL, ThermoFisher Scientific, Waltham, MA) was added for 10 min at RT followed by sulfuric acid stop solution (100 µL, ThermoFisher Scientific, Waltham, MA) for light absorption measurements at 450 and 650 nm (Spark 10 M, Tecan, Baldwin Park, CA). Each plate contained a control titration of the anti-RBD monoclonal antibody CR3022 in IgG, IgM, or IgA format (Creative Biolabs, Shirley, NY) to provide a standard curve. Serum anti-RBD IgG binding activity was expressed as an equivalent to a concentration of CR3022. The lower detection limit was ca. 3 ng mL−1 control antibody (ca. 100 ng mL−1 for diluted serum).

IFN-γ ELISpot assay for CD8+ T-cell responses

These assays were performed as previously described[20]. Thawed cryopreserved PBMC were plated at 1–2 million cells/well in RPMI with IL-2 at 50 U mL−1 (NIH AIDS Reagent Repository Program) with a CD3:CD4 bi-specific monoclonal antibody (gift of Dr. J Wong) and cultured for ca. 14 days to yield purified polyclonal CD8+ T cells. These cells were viably cryopreserved until the day of ELISpot assay.

Cells were added to a 96-well filter plate that had been precoated with an anti-IFN-γ antibody (Mabtech, Nacka Strand, Sweden) with the addition of a peptide pool, medium alone (three wells), or medium with PHA (Sigma Aldrich, St. Louis, MO) at 25 μg mL−1. After overnight incubation in a humidified CO2 incubator, the plate was washed and stained with biotinylated anti-IFN-γ antibody (Mabtech, Nacka Strand, Sweden) for visualization using a streptavidin-peroxidase reagent and counting on an automated ELISpot reader (AID, Autoimmun Diagnostika GMBH, Strassberg, Germany) against synthetic overlapping peptide pools spanning SARS-CoV-2 spike, nucleocapsid, matrix, and envelope proteins (NR-52402, NR-52404, NR-52403, NR-52405, BEI Resources, Manassas, VA). The response against each peptide pool was expressed as the raw count minus the mean of the triplicate negative control wells. The two following criteria needed to be met for positivity: ≥50 SFC/106 CD8+ T lymphocytes and ≥ mean and two standard deviations of negative control wells (no peptide).

Statistics and reproducibility

Datasets were analyzed using GraphPad Prism (version 9.3.1; GraphPad Software, Inc., La Jolla, CA). Serum IgA and IgM concentration half-lives (t) were compared using a Wilcoxon matched-pairs signed rank test (paired t-test; nonparametric) and a Mann–Whitney test (unpaired rank test; nonparametric).

Reproducibility of the experiments, including sample sizes as well as the number and nature of replicates, were as follows. Clinical swab specimens collected from study participants were analyzed individually for SARS-CoV-2 RNA by RT-qPCR. Three target sequences (two viral, one host) were amplified simultaneously in a 96-well format, and each plate also included one human specimen control as well as one positive control. Plasma cytokine concentration measurements were carried out using two replicates consisting of paired aliquots from the same clinical sample. Each 96-well plate also contained two quality controls and seven standards, all in duplicate. Serum antibody measurements by ELISA were carried out on individual samples, run in four serial four-fold dilutions, and OD values were compared against a standard curve with the control antibody CR3022 in serial three-fold dilutions on the same 96-well plate. Cryopreserved PBMC samples were analyzed individually for CD8+ T-cell responses using an ELISpot assay. Each sample was run in a single well on a 96-well plate that also included triplicate negative control wells and duplicate positive control wells. The mean of the negative control wells was subtracted from the value of each sample well.

Table 1

Demographics of study participants testing positive for SARS-CoV-2 by RT-qPCR who were followed longitudinally.

Characteristic	Value
No. of participants	9
Female, no. (%)	6 (67)
Male, no. (%)	3 (33)
Age (yrs), median (range)	25 (19–53)
Race and ethnicity, no. (%)
Black or African-American	0
White	9 (100)
Hispanic	6 (67)
Non-Hispanic	3 (33)
Asian	0
Other	0

All participants were unvaccinated against SARS-CoV-2 at the time of the first positive test result.

Table 2

Estimated viral trajectories for different priors presented as means with the 95% confidence interval (CI) in brackets.

	Informative priors	Uninformative priors	Biased priors
Clearance time (d)	15.64 (13.09, 18.81)	17.16 (14.54, 21.29)	15.25 (12.73, 16.00)
Peak Ct value	10.49 (6.93, 14.80)	10.43 (6.79, 15.12)	10.52 (7.02, 14.93)
Proliferation time (d)	3.07 (0.68, 7.08)	5.95 (1.42, 9.91)	2.42 (0.54, 6.91)