key: cord-0898706-tiu86ugm authors: Cheemarla, Nagarjuna R.; Watkins, Timothy A.; Mihaylova, Valia T.; Wang, Bao; Zhao, Dejian; Wang, Guilin; Landry, Marie L.; Foxman, Ellen F. title: Magnitude and timing of the antiviral response determine SARS-CoV-2 replication early in infection date: 2021-01-27 journal: medRxiv DOI: 10.1101/2021.01.22.21249812 sha: a4cd57a4078eb2a2ad3a7f7982ae1f66b6029279 doc_id: 898706 cord_uid: tiu86ugm The interferon response is a potent antiviral defense mechanism, but its effectiveness depends on its timing relative to viral replication. Here, we report viral replication and host response kinetics in patients at the start of SARS-CoV-2 infection and explore the impact of these kinetics experimentally. In both longitudinal patient nasopharyngeal samples and airway epithelial organoids, we found that SARS-CoV-2 initially replicated exponentially with a doubling time of ~6hr, and induced interferon stimulated genes (ISGs) with delayed timing relative to viral replication. Prior exposure to rhinovirus increased ISG levels and blocked SARS-CoV-2 replication. Conversely, inhibiting ISG induction abrogated interference by rhinovirus and enhanced SARS-CoV-2 replication rate. These results demonstrate the importance of initial interferon-mediated defenses in determining the extent to which SARS-CoV-2 can replicate at the start of infection and indicate that biological variables that alter the airway interferon response, including heterologous induction of innate immunity by other viruses, could profoundly impact SARS-CoV-2 susceptibility and transmission. The novel coronavirus SARS-CoV-2 emerged in late 2019 and has led to a global pandemic, causing 2 over 98M infections and 2.1M deaths at the time of this writing (Dong, 2020) . This huge impact has 3 motivated efforts to understand the host immune response to this virus, both to better predict patient 4 outcomes and to design interventions. For an emerging viral infection such as SARS-CoV-2, innate 5 immune responses can be particularly important in host protection, as these responses do not require 6 prior exposure to effectively defend against a pathogen. Studies based on patient samples indicate that 7 dysregulation of innate immune responses late in infection drives immunopathology in severe COVID-8 19 (Galani et al., 2021; Lee and Shin, 2020; Lucas et al., 2020) , but there are relatively few reports 9 describing host responses to SARS-CoV-2 in patients at the start of infection, when innate immune 10 mechanisms are most likely to contribute to host defense. 12 SARS-CoV-2 enters the body and first replicates in the upper respiratory tract, achieving the highest 13 viral load in the first few days following infection (Wolfel et al., 2020; Zou et al., 2020) . High viral load in 14 the nasopharynx correlates strongly with transmissibility in contact tracing studies, and significant viral 15 replication following exposure is likely a prerequisite, although certainly not the only factor, for COVID-Recent evidence supports a protective role for the interferon response in COVID-19, although there is 27 also evidence that the virus antagonizes this response. Recombinant interferon blocks SARS-CoV-2 28 replication in vitro, and genetic deficiencies in the Type I interferon response as well as anti-interferon 29 autoantibodies have been linked to greater COVID-19 disease severity (Bastard et al., 2020; 30 Lokugamage et al., 2020; Pairo-Castineira et al., 2020; Vanderheiden et al., 2020; Zhang et al., 2020) . Furthermore, early data from trials of recombinant Type I or Type III interferon for COVID-19 indicate a 32 therapeutic benefit, particularly when patients are treated early in disease (Feld, 2020; Monk, 2020; 33 . However, during initial infection of the upper respiratory tract, the kinetics of ISG 34 induction by SARS-CoV-2 are not clear. ISG expression in SARS-CoV-2 infected epithelia can be 35 observed in vitro and in patients, but there is also strong evidence that SARS-CoV-2 antagonizes the 36 interferon response, which likely affects the magnitude and timing of this response (Banerjee et al., 37 2020; Blanco-Melo et al., 2020; Konno et al., 2020; Martin-Sancho et al., 2020; Ravindra et al., 2020; 38 Xia et al., 2020; Zhou et al., 2020) . Since a major beneficial function of ISGs is preventing viral 39 replication, the kinetics of the interferon response early in infection are likely to determine its protective 40 impact, and host and environmental factors which modulate the timing of this response may be key 41 determinants of whether the virus can amplify to a high viral load following infection. One factor that can potentially modulate antiviral defenses in the airway epithelium is induction of the 44 ISGs by other viruses, and such effects may be particularly important in limiting viruses which 45 successfully block autologous interferon induction. Rhinovirus, the most frequent cause of the common Of reads mapping to the human genome, 1770 RNAs differed significantly between SARS-CoV-2+ 79 patients and control subjects ( Fig 1A) . These included 1567 protein-coding genes, of which 1245 (79.4%) 80 were enriched and 322 reduced in patients relative to controls. The most significantly enriched genes in 81 the nasopharynx of SARS-CoV-2 patients were known interferon stimulated genes (ISGs), including three 82 NP ISG transcripts previously shown by our group to accurately identify patients with viral respiratory Examination of gene expression across patient samples revealed several patterns (Fig 1E-G) . First, the 90 45 most significantly enriched genes were all interferon stimulated genes, according to the Interferome 91 database (Rusinova et al., 2013) . ISGs appeared to be co-regulated within individual patients, i.e. patients with high expression of one ISG tended to have high expression of other ISGs (Fig 1E) . This 93 was also demonstrated by analysis of the correlation between reads for different ISGs across samples 94 ( Fig 1F) . Second, ISG expression appeared to be loosely correlated with viral load, with those patients 95 with the highest viral load (Fig 1E, left) tending to have higher ISG expression than those with the lowest 96 viral loads (Fig 1E, right) . However, while all SARS-CoV-2+ samples showed enrichment of ISG 97 expression compared to controls, direct comparison of DEGs in patient groups with distinct clinical 98 characteristics (sex, age, or outpatient/admitted status) showed no significant differences in ISG 99 expression. Foxman, 2018). Consistently, we observed a significant positive correlation between NP CXCL10 protein 105 level with the NP mRNA level of Cxcl10 in SARS-CoV-2+ patient NP samples (Fig 1E, G) . Together, 106 these results indicated that across subjects with diverse clinical presentations, SARS-CoV-2 induced a 107 robust interferon response in the nasopharynx and that the NP CXCL10 protein level correlated with ISG 108 expression at the RNA level. Relationship between SARS-CoV-2 viral load and NP host response in vivo 111 Next, we sought to examine the host response during SARS-CoV-2 infection in a larger set of samples 112 from patients evaluated in our healthcare system in March and April of 2020 (n=140) (Fig 2 and Fig S2A, 113 B), to gain further insight into the relationship between viral replication, disease status, and host response 114 to infection. Based on our previous studies and our finding that NP CXCL10 protein level correlated with 115 ISG expression in SARS-CoV-2 positive samples (Fig 1) , we used NP CXCL10 as an indicator of the 116 nasopharyngeal antiviral response. First, we examined NP CXCL10 level in SARS-CoV-2+ individuals 117 who were tested as outpatients and not admitted to the hospital, compared to those who were admitted. Notably, we observed significantly higher NP CXCL10 levels in outpatients compared to admitted patients 119 (Fig 2A, p=0 .0019). To understand the reason for this, we first examined patient age, since as a group, 120 admitted patients were significantly older than outpatients (16 years older on average, Fig S2C) . However, there was no correlation between age and CXCL10 level ( Fig S2D) . Next, we examined viral 122 load. We were initially surprised to find that admitted patients had significantly lower viral loads than 123 outpatients ( Fig 2B) . This suggested that the main factor driving CXCL10 level was viral load. Supporting 124 this idea, correlation analysis showed a significant positive correlation between NP CXCL10 level and 125 viral load by RT-qPCR (for all patients, r 2 =0.2030, p<0.0001, Fig 2C) . This correlation was also seen in 126 separate analyses of outpatients and admitted patients, with but no significant difference in the slope of 127 the CXCL10 vs viral load correlation between these groups, although there was a trend towards a higher 128 slope in outpatients ( Fig 2C) . We also observed no significant relationship between sex and NP CXCL10 129 or sex and viral load in this sample set (Fig S2 E,F) . . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Prior work on SARS-CoV-2 has shown that the nasopharyngeal viral load is highest in the first few days 131 of infection, and that the more severe symptoms of COVID-19 requiring hospitalization occur in the 132 second or third week of infection (Cevik, 2020) . Therefore, we hypothesized that admitted patients may 133 have shown lower viral loads at the time of testing than outpatients because they presented later in 134 infection, after peak viral replication in the nasopharynx. Consistently, outpatients tended to report fewer 135 days of symptoms prior to testing compared to admitted patients, although this information was only 136 available for a subset of patients (about one-third, n=44; Fig 2D) . To further evaluate the relationship between viral replication and the innate antiviral response in the 140 nasopharynx, we examined viral load and NP CXCL10 data in longitudinal samples. First, we examined 141 viral load from 29 inpatients from March 12 and April 30, 2020 for whom we had at ≥8 sequential tests 142 results for SARS-CoV-2 with at least the first sample tested using the CDC assay (our clinical laboratory 143 also had other testing platforms) (Fig 3, Table S1 ). At this time, most serial testing was aimed at patient 144 clearance for discharge. Consistently, the majority of patients (15/29) showed low viral loads (Ct N1>21) 145 which remained low throughout the time course. Another common pattern was high viral load in the first 146 sample (Ct N1<20) followed by a decline in viral load over time, similar to patterns reported in the literature 147 for patients who presented close to the start of symptomatic illness (7/29 patients, Fig 3A) . These patients 148 showed high CXCL10 level in the sample with peak viral load and a decline in NP CXCL10 after the viral 149 load had decreased ( Fig 3B) . One patient had a consistently high NP viral loads for 20 days and did not 150 survive (not shown). Finally, a third pattern was seen in a several patients (6/29), in which the first sample 151 had a low viral load which subsequently increased to a high peak level (Ct N1<20), then decreased over 152 time (Fig 3) . This pattern is consistent with patient presentation close to the start of infection. Two of 153 these patients had no symptoms of SARS-CoV-2 and the virus was detected incidentally on screening 8 test N1 Ct 34.6, N2 Ct 35.4), suggesting that this might have been the first day of infection for this patient 157 (L2, Fig 3C) . The other four patients presented with acute symptoms including fever, and in some cases 158 cough and/or shortness of breath. NP CXCL10 in these patients was undetectable or low in the first 159 positive sample with low viral load, then rose with viral RNA, and then subsequently declined as viral load 160 declined. Together, the longitudinal data show a correlation with NP CXCL10 and viral load in individual 161 patients over time, similar to what we observed across 140 patients tested at a single time point (Fig. 2) . Notably, for patient L2, the only patient for whom three samples were available prior to peak viral load, 163 there appeared to be a delay between CXCL10 production relative to viral replication during the first few CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 183 Viral replication in organoid cultures appeared to follow an exponential curve for the first 72hr of infection. Therefore, we used curve-fitting to exponential growth to estimate the doubling time, which was 5.858 hr 185 (95% C.I. of 4.85-7.357 hr., based on 20 y-values, 5 per time point; Fig 4G) . For patient L2, viral load 186 data from first three SARS-CoV-2+ time points also appeared to follow exponential growth, therefore we 187 used the same method to estimate the SARS-CoV-2 doubling time in vivo from this data, which was 188 6.454 hr (95% C.I. 4.261-13.30 hr based on 3 y-values, Fig 4H) . For all other patients from whom viral 189 load increased in serial samples (Fig 3) , we had only one sample prior to peak viral load. We asked what 190 the doubling times for SARS-CoV-2 would be in these samples if we assumed exponential replication 191 between the first and peak viral RNA values. The calculated doubling times across patients ranged from 192 3.048-6.509 hr for samples less than or equal to 5 days apart. For the two patients with a larger sampling 193 interval (L44, L12), calculated doubling times were 9.455 and 12.58 hr, although these calculations would CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 5B). We then evaluated whether rhinovirus infection altered the expression of ACE2, the SARS-CoV-2 209 entry receptor. ACE2 was originally reported to be an ISG, but a subsequent study reported that full-210 length ACE2, which functions as an entry receptor, is not an ISG, and that a truncated form, dACE2, is 211 an ISG but is not a functional SARS-CoV-2 entry receptor (Onabajo et al., 2020; Ziegler et al., 2020) . Consistent with this finding, we observed that dACE-2 was significantly induced by rhinovirus infection 213 (~14-fold) and that, as expected for an ISG, induction was prevented by blocking activation of IRF3, a 214 transcription factor downstream of viral RNA sensors, using the inhibitor BX795 (Clark et al., 2009) . In 215 contrast, full-length ACE2 expression was slightly but significantly increased by rhinovirus infection(~2-216 fold) and this change was not abrogated by BX795, suggesting a different mechanism of induction (Fig 217 S3.) Rhinovirus infection had no effect on expression of TMPRSS2 (not shown). Next, we evaluated SARS-CoV-2 replication and ISG induction following infection of airway epithelial 220 organoids, with or without prior rhinovirus infection. SARS-CoV-2 viral load increased exponentially in 221 infected cultures without prior RV infection, as observed previously (Fig 4) , but showed essentially no 222 increase when cultures had been exposed to rhinovirus 3 days prior ( Fig 5C) . Evaluation of ISG 223 expression over the course of infection showed that at early time points of SARS-CoV-2 infection (24, 48 224 hr, and sometimes 72hr), ISGs were significantly more highly expressed in RV-preinfected cultures that 225 in cultures infected with SARS-CoV-2 without prior RV exposure ( Fig 5D) . This included several ISGs 226 which have been previously reported to limit coronavirus replication or for which polymorphisms are CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10. 1101 /2021 To better understand the timing and breadth of the epithelial host response to rhinovirus that appeared 01A entry receptor LDL-R being ubiquitously highly expressed throughout the culture (Fig S4F,G) . Although rhinovirus was only detected in a small subset of cells in infected cultures at day 5 (1.67%), ISGs were elevated in all cells compared to mock-treated cultures (Fig 5F) , demonstrating that rhinovirus 248 infection induces a robust bystander antiviral response in uninfected cells that lasts at least 5 days. Blocking ISG induction restores SARS-CoV-2 replication following rhinovirus infection 251 Next, to test whether suppression of SARS-CoV-2 replication by rhinovirus was dependent upon the host 252 cell interferon response, we pre-treated cells with the signaling inhibitor BX795 18hr before rhinovirus 253 infection, which prevents interferon and ISG induction by rhinovirus (Fig 6A, B) (Clark et al., 2009; Wu et 254 al., 2020) . There was no significant difference in SARS-CoV-2 viral load in RNA isolated from organoid 255 cultures at 72 hr for cultures with and without drug treatment, (Fig 6B) . The effect of BX795 treatment 256 was much more striking in the setting of sequential rhinovirus-SARS-CoV-2 infection. As seen previously 257 (Fig 5) , prior infection with rhinovirus suppressed SARS-CoV-2 replication by >1000-fold, but replication 258 was restored by BX795 pre-treatment ( Fig 6C) . These results indicate that IRF3 signaling is critical for 259 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. towards an increase (5-10x) was also seen at 96 hrs post-infection, although due to variability among 279 replicates this difference was not statistically significant (Fig 7A, B) . Based on these results, we checked 280 the effect of BX795 on viral shedding into the apical wash in the MOI 0.5 infection (Fig 6) but in this case 281 results were similar to those observed in organoid cultures (not shown). Next, based on the increase in 282 viral RNA from 1hr to 72 hr during the low MOI infection, we estimated the effect of BX795 on the SARS- CoV-2 doubling time in organoid culture. Assuming exponential growth between 1hr and 72 hr, the SARS- CoV-2 doubling time without BX795 was 5.127 hr (95%C.I. 3.889 to 7.518 hrs), and with BX795 was 285 . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. Induction of all of these ISGs was suppressed by BX795 pre-treatment (Fig 7D-F) . For BST2 mRNA, 289 blocking ISG induction with BX795 revealed a decrease in BST2 mRNA below the baseline level during 290 SARS-CoV-2 infection, suggesting that SARS-CoV-2 may antagonize BST2 expression at the mRNA 291 level, in addition to other mechanisms whereby SARS-CoV-2 or SARS-CoV have been reported to 292 antagonize BST2 at the protein level (Martin-Sancho et al., 2020; Taylor et al., 2015) . The effects of 293 BX795 on SARS-CoV-2 during low MOI infection indicate that the epithelial antiviral response induced 294 by SARS-CoV-2 does limit viral replication, albeit to a lesser extent than interference by rhinovirus. Taken together, our results show that SARS-CoV-2 undergoes exponential replication at the start of . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint SARS-CoV-2 replication in the nasopharynx is known to peak during the first week of infection, but the 302 biological variables governing the rate and magnitude of viral amplification are not fully understood 303 (Cevik, 2020; He et al., 2020) . Nasopharyngeal viral load early in infection correlates with likelihood of 304 transmitting the infection, and robust viral amplification in the respiratory tract is also likely a prerequisite, 305 although certainly not the only factor, for progression of COVID-19 disease. Here we present evidence 306 that during the first few days of SARS-CoV-2 infection, the airway interferon response plays a protective 307 role by curtailing viral replication in its initial target tissue: the airway epithelium. Specifically, our data 308 show that the extent of viral replication is determined by the magnitude and timing of the host interferon 309 response during a critical time window at the start of infection. The interferon response is a potent mechanism of antiviral innate defense at mucosal surfaces and 312 effectively curtails replication of many viruses, most of which also antagonize this host response to some Consistently, we observed robust ISG induction in all patients using RNA-Seq of nasopharyngeal RNA, 320 regardless of disease severity or other biological variables (Fig 1) . However, there is also convincing 321 evidence that SARS-CoV-2 antagonizes the interferon response in infected cells through multiple 322 mechanisms (Banerjee et al., 2020; Konno et al., 2020; Martin-Sancho et al., 2020; Xia et al., 2020) . Our 323 observations with organoid culture (Fig 4) , and in the one patient for whom we had multiple longitudinal 324 samples prior to peak viral load (patient L2, Fig 3A) show exponential viral replication at the start of 325 SARS-CoV-2 infection. ISGs are also induced, but it is unclear when this host response becomes 326 . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint functionally effective. We found that in the first 72 hours of infection, prior to the peak host response, 327 blocking ISG induction had no effect in SARS-CoV infection at MOI 0.5 (Fig 6) , but did increase viral 328 replication at ten-fold lower MOI (Fig 7; MOI 0.05). Together, these observations support a model in which 329 antagonism by SARS-CoV-2 attenuates but does not prevent the interferon response and thereby creates CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10. 1101 /2021 Conversely, host factors that enhance interferon-mediated defenses could be impactful in reducing peak 352 viral load or even preventing infection altogether. One factor that is known to alter ISG expression in the 353 nasopharynx is recent viral infection. In this study, we focused on rhinovirus, the most frequent cause of 354 the common cold. Recent epidemiological studies have shown that this virus is much more prevalent in 355 the upper respiratory tract than previously appreciated (Foxman and Iwasaki, 2011; Jartti et al., 2008) . For example, in a recent year-round study, about one-third (34%) of all nasal samples from young children CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10. 1101 /2021 and ISG induction in response to its own replication, including specific targeting of ISGs effective in 378 blocking the coronavirus life cycle such as BST2 (Martin-Sancho et al., 2020; Taylor et al., 2015) . The concept of viral interference based on the host interferon response assumes a host with intact innate 381 immune defenses, which may not always be the case. In our experimental model, there was a profound 382 difference in the outcome of rhinovirus-SARS-CoV-2 co-infection in the presence and absence of an 383 intact host cell interferon response (Fig 6) . With an intact host response, viral loads of both viruses 384 decreased, but when the interferon response was blocked, viral loads of both viruses were equal to or 385 higher than in single infections (Fig 6. ) This result illustrates that the expected outcome of a viral co-386 infection is not one-size-fits-all: it is likely to be profoundly dependent upon host innate immune status. There are several important caveats to our study. First, our data indicate a relatively limited effect of the 400 SARS-CoV-2-induced interferon response on initial viral replication when it is the only virus present, at 401 least at high MOI (Fig 6) . However, the recruitment of cells of the immune system to the respiratory tract, 402 as indicated by our RNAseq data (Fig 1) , could considerably amplify ISG induction in vivo. Recent work 403 . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10. 1101 /2021 showing that interferon deficiencies are linked to severe COVID-19 indicates the importance of interferon- . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. We also thank Bryan Pasqualucci and Christopher Castaldi at the Yale Center for Genomic Analysis. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. Upon thawing, RNA was isolated from 140μl of transport medium using the Qiagen Viral RNA isolation 491 . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Statistical significance of differences between conditions was determined by t tests (two-tailed). Linear 565 regression analysis was used to determine association between clinical parameters, such as viral load 566 and NP CXCL10 in clinical samples and to test the null hypothesis that the slope of the association was 567 significantly different from zero. Non-linear regression analysis was used to fit viral growth to an 568 . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint exponential curve (exponential growth with log(population)) to determine virus doubling times and to test 569 the null hypothesis that one curve fit both data sets for SARS-CoV-2 growth curve with and without 570 rhinovirus pre-infection. p<0.05 was considered statistically significant. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 He, J., Feng, D., de Vlas, S.J., Wang, H., Fontanet, A., Zhang, P., Plancoulaine, S., Tang, F., Zhan, 663 L., Yang, H., et al. (2006) . Association of SARS susceptibility with single nucleic acid polymorphisms of 664 OAS1 and MxA genes: a case-control study. BMC Infect Dis 6, 106. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Kudo, E., Song, E., Yockey, L.J., Rakib, T., Wong, P.W., Homer, R.J., and Iwasaki, A. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101 https://doi.org/10. /2021 Mick, E., Kamm, J., Pisco, A.O., Ratnasiri, K., Babik, J.M., Castaneda, G., DeRisi, J.L., Detweiler, 715 A.M., Hao, S.L., Kangelaris, K.N., et al. (2020) . Upper airway gene expression reveals suppressed 716 immune responses to SARS-CoV-2 compared with other respiratory viruses. Nat Commun 11, 5854. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint Figures Fig 1. Transcriptome analysis of RNA isolated from SARS-CoV-2+ nasopharyngeal swabs D IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 TRIM24 IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IL1B IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IRF3 IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A IL1A OAS3 TYMP IFI44 OAS2 IRF7 MX1 OASL IFI44L IFIT3 STAT1 IFITM1 EPSTI1 CMPK2 PARP9 ISG15 RSAD2 DDX60 HERC6 XAF1 IFIT2 SIGLEC1 HLA-A IFIT1 IFITM3 SAMD9 HELZ2 MX2 CXCL10 PARP12 ZBP1 SECTM1 DHX58 IFIH1 OAS1 CXCL11 TRIM22 CYBA LAP3 HERC5 PARP10 LY6E DDX58 SLC15A3 PARP14 USP18 min to max . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; Fig 1. Transcriptome analysis of RNA isolated from SARS-CoV-2+ nasopharyngeal swabs (related to Fig S1) . (A) Volcano plot showing significantly differentially expressed protein-coding genes based on RNASeq of NP swab RNA from SARS-CoV-2 patients (n=30) compared to control SARS-CoV-2 negative subjects (n=8). Transcripts with fold change>2, adjusted p-value<0.05 are highlighted in red. (B) Top 20 ingenuity pathways enriched in SARS-CoV-2+ compared to controls, based on 1770 differentially expressed RNAs. P-value and Z-score for each pathway is indicated on the x-axis. Pathways related to interferon and interferon regulatory factor (IRF) signaling are highlighted in red. (F) Correlation between reads mapping to CXCL10 and reads mapping to other ISGs (Ifit2, OasL, Isg15). (G) Correlation between reads mapping to Cxcl10 and CXCL10 protein measured by ELISA in NP swabassociated viral transport medium. . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. 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 January 27, 2021. ; Table S1 ). Nasopharyngeal viral load and CXCL10 in patients diagnosed prior to peak viral load. (A) Viral load over time in seven longitudinal samples from patients with high viral load in first sample (Ct N1>20). (B) Paired viral RNA and NP CXCL10 measurements at the peak viral load and at the end viral load, defined as the first sample with Ct N1>30, for 6 patients shown in (G) (data not available for one sample). CXCL10 level was significantly different in peak and end samples by paired t-test. (C) -(H) Viral load and NP CXCL10 level in longitudinal samples from SARS-CoV-2+ patients who presented with a low viral load (Ct N1>28) that increased to a high viral load (Ct N1<20). Viral load is expressed as fold change from the limit of detection for the SARS-CoV-2 N1 gene (black circles) and CXCL10 is expressed as pg/ml in the NP-swab associated viral transport medium (red squares). Samples with low levels of RnaseP, an indicator of sample quality, are shown with open symbols. Patient characteristics are described in Table S5 . . CC-BY-NC-ND 4.0 International license It is made available under a 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 January 27, 2021. ; . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint Figure 6 . Effect of pretreatment with BX795 during sequential rhinovirus, SARS-CoV-2 infection Organoid cultures were pretreated with or without BX795 for 18 hr, then mock-infected or infected with HRV01A, incubated for 3 days, then infected with SARS-CoV-2. (A) Effect of BX-795 pre-treatment on ISG induction, 3 days post rhinovirus infection. Bars show fold change in ISG mRNA level in RV infected cultures compared to mock without (left) or with (right) BX-795 pre-treatment. P values indicate significant differences between ISG levels in cultures with or without BX795 pretreatment by t-test. (B) SARS-CoV-2 viral RNA level relative to the limit of detection in organoid cultures, 72 hr post SARS-CoV-2 infection, with and without BX-795 and/or RV pre-treatment. P values indicate significant differences in viral RNA levels, n.s. = not significant. (C) HRV01A viral RNA level relative to the limit of detection in organoid cultures, 72 hr post SARS-CoV-2 infection, with and without BX-795 and/or RV pre-treatment. This graph also includes cultures infected with RV but not subsequently infected with SARS-CoV-2. P values indicate significant differences in viral RNA levels. For all graphs, bars show mean and S.E.M. of 4-6 biological replicates per condition. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10.1101/2021.01.22.21249812 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 27, 2021. ; https://doi.org/10. 1101 /2021 Genetic mechanisms of critical illness in 742 StringTie enables improved reconstruction of a transcriptome from RNA-seq reads LY6E impairs coronavirus fusion and confers immune control of viral disease A single-cell atlas of the airway epithelium reveals the CFTR-rich pulmonary ionocyte Physical distancing in schools for 753 SARS-CoV-2 and the resurgence of rhinovirus CDC 2019-Novel Coronavirus (2019-nCoV) Real-Time RT-PCR Emerging SARS-CoV-2 Variants Single-cell longitudinal analysis of SARS-CoV-2 infection 759 in human bronchial epithelial cells Interferome v2.0: an updated database of annotated interferon-regulated genes Viral Interference: The Case of Influenza Viruses. The Journal of infectious 766 diseases 212 Duration of infectiousness and correlation with RT-PCR cycle threshold values 769 in cases of COVID-19 Comprehensive Integration of Single-Cell Data Severe Acute Respiratory Syndrome Coronavirus ORF7a Inhibits Bone 775 Marrow Stromal Antigen 2 Virion Tethering through a Novel Mechanism of Glycosylation Interference Type I and Type III Interferons Restrict CoV-2 Infection of Human Airway Epithelial Cultures Analytical sensitivity and 782 efficiency comparisons of SARS-CoV-2 RT-qPCR primer-probe sets Retrospective Multicenter Cohort Study Shows Early Interferon Therapy Is Associated with Favorable Clinical Responses Virological assessment of hospitalized patients with COVID-788 2019 CC-BY-NC-ND 4.0 International license It is made available under a We would like to thank Craig Wilen and Wilen lab members for valuable help and advice, and for providing 422 SARS-CoV-2 virus stock. We thank Maureen Owen, Robin Garner, Greta Edelman, the entire staff of