Delayed engagement of host defenses enables SARS-CoV-2 viremia and productive infection of distal organs in the hamster model of COVID-19

lab-led

In golden hamsters, productive SARS-CoV-2 replication in the airways generates circulating type I and III interferon that primes every organ against infection, and blunting or bypassing that airway response permits viremia and productive infection of liver, kidney, spleen and brain.

Lucia Carrau; Justin J. Frere; Ilona Golynker; Alvaro Fajardo; Cristobal F. Rivera; Shu Horiuchi; Tyler Roonprapunt; Judith M. Minkoff; Daniel Blanco-Melo; Benjamin TenOever

2023 · Science Signaling · primary research

Senior authors

Benjamin TenOever

Correspondence

Benjamin TenOever

Research areas & themes

Citation

Carrau L, Frere JJ, Golynker I, Fajardo A, Rivera CF, Horiuchi S, Roonprapunt T, Minkoff JM, Blanco-Melo D, TenOever B. Delayed engagement of host defenses enables SARS-CoV-2 viremia and productive infection of distal organs in the hamster model of COVID-19. Science Signaling. 2023. Volume 16, issue 789, article eadg5470.

DOI 10.1126/scisignal.adg5470. PMID 37311033.

One-sentence contribution

In golden hamsters, productive SARS-CoV-2 replication in the airways generates circulating type I and III interferon that primes every organ against infection, and blunting or bypassing that airway response permits viremia and productive infection of liver, kidney, spleen and brain.

Executive summary

COVID-19 presents very differently between individuals, and complications outside the lung are common in severe cases, which has raised the question of whether extrapulmonary replication contributes to severity. Working in the golden hamster, which reproduces many features of human infection, the authors asked where the virus actually replicates early after airway exposure, where the host antiviral response appears, and what connects the two. Bulk RNA sequencing of nine organs three days after intranasal challenge showed an interferon signature in every organ examined, while infectious virus by plaque assay was largely confined to lung and olfactory bulb, with less consistent recovery from heart and gastrointestinal tract. Whole blood carried an interferon stimulated gene signature without interferon transcripts of its own, and a bioassay using hamster fibroblasts detected roughly sixty units per millilitre of circulating interferon at one day after infection. Dexamethasone treatment delayed airway interferon stimulated gene induction and reduced lung phagocyte infiltration without changing early lung titers, and under that treatment infectious virus appeared in liver, spleen, olfactory bulb and gastrointestinal tract, with low level viremia detectable after amplification. Delivering virus intravenously bypassed the airway entirely and produced productive infection of kidney, liver, spleen, heart and gastrointestinal tract. Priming animals by intranasal infection before an intravenous challenge reduced distal viral loads. Circulating interferon was detected only in animals with lung titers.

Scientific context

SARS-CoV-2 delays the host interferon response by compartmentalizing double-stranded RNA in double membrane vesicles and by degrading host mRNA, while NF kappa B driven chemokine and cytokine production proceeds, producing an unbalanced response in the lungs characterized by high chemokines with delayed type I and III interferon and heavy neutrophil recruitment. The interferon response that does appear is thought to arise once infected cells die and their contents, including viral double-stranded RNA and host DNA, are sampled by phagocytes and other cells bearing Toll-like receptors. Clinically, COVID-19 shows substantial heterogeneity, with severe disease concentrated among older and comorbid individuals and frequently accompanied by cardiac, cognitive and gastrointestinal complications, which has led several groups to propose that distal replication contributes to severity. Deficits in innate immune signaling, including inborn errors and autoantibodies against type I interferon, are associated with critical outcomes, and genetic ablation of type I but not type III interferon signaling in hamsters had been shown to raise viral RNA. The golden hamster had been established by this laboratory and by others as a model reproducing transmission, clinical presentation and comorbidity effects. What had not been resolved was whether the interferon response seen in organs away from the airway reflects local replication or a signal arriving from elsewhere, and whether that response does anything.

Central question

Does the antiviral state observed throughout the body after airway SARS-CoV-2 infection arise from productive replication in those organs or from interferon generated in the lung and carried in the circulation, and does that systemic priming restrict where the virus can establish infection?

Experimental strategy

The design separates the question of where virus is from the question of where the response is, and then removes the response to see what changes. First, the two are measured in parallel in the same animals, with bulk RNA sequencing across nine organs set against plaque assay and viral RNA quantification in the same tissues at one and three days, plus immunohistochemistry for the interferon stimulated protein Mx alongside viral nucleocapsid in lung and kidney. Second, the blood is examined as the candidate conduit, using RNA sequencing of whole blood to show an interferon stimulated gene signature and using the absence of Ifnb and Ifnl reads in blood to argue that those cells are responding to interferon rather than sensing virus, since interferon transcripts are induced only on direct detection of viral double-stranded RNA. Because no hamster interferon reagents were available, circulating interferon was quantified indirectly with a bioassay in which serum is applied to BHK-21 hamster fibroblasts, which cannot themselves make type I or III interferon, and Mx induction is read against a universal interferon standard curve. Third, the priming is removed in two independent ways that fail differently. Dexamethasone suppresses de novo transcription and therefore delays the response while leaving the airway infection in place, and intravenous inoculation places virus in the circulation without an airway infection to generate the signal. Fourth, priming is restored deliberately by infecting intranasally and then challenging intravenously three days later, with control groups receiving buffer by one route or the other, and organs harvested at four days so that the readout precedes an adaptive response. Detection of very low viremia used an amplification step on permissive Vero E6 cells before plaque assay.

Key findings

1. Three days after intranasal challenge, transcriptional profiling of olfactory bulb, brain, liver, kidney, gastrointestinal tract, pancreas, spleen, heart and lung showed organ specific clustering and shared responses, with type I and III interferon signatures the most prominent enriched annotation in every organ, corroborated by Isg15 quantitative RT-PCR (Figure 1A and Figure S1).

2. Infectious virus was not present in most of those organs. Plaque assay recovered high titers in lung, consistent recovery from olfactory bulb, less consistent recovery from heart, and a single gastrointestinal tract sample the authors attribute to possible ingestion of inoculum (Figure 1B). Viral RNA was elevated only in lung and gastrointestinal tract, and the same confinement was seen at one day after infection (Figure 1C and Figure S1D). Mx protein was detected in kidney as well as lung by immunohistochemistry (Figure 1D).

3. Whole blood showed an interferon stimulated gene signature at both one and three days, including Ifit3, Mx, Isg15, Oasl and Irf7, while no replication competent virus was recovered from blood and viral N RNA was detected only at one day at low level (Figures 2A and 2B, Figures S2A and S2B).

4. No Ifnb or Ifnl reads were found in blood, whereas Ifnl appeared in lung at one day and rose by three days and Ifnb appeared in lung at three days (Figures 2C and 2D). The authors read the absence of interferon transcripts in blood as indicating that blood cells are responding to interferon rather than directly sensing virus. Ifna transcripts were not detected in any sample, which the authors attribute to incomplete or incorrect annotation of that gene family in the hamster genome.

5. The serum bioassay detected roughly sixty units per millilitre of interferon equivalents at one day after infection, stable at three days (Figure 2E). Taken with the preceding points, the authors support a model in which lung derived interferon enters the circulation and primes distal organs.

6. Dexamethasone delayed rather than abolished the airway response. Lung titers were unchanged over the first three days, but treated animals failed to control replication thereafter, with titers rising after three days, replication competent virus present at seven days and clearance only by nine days (Figure 3A). Mx positive cell counts were reduced at three days (Figure 3B), lung interferon stimulated gene induction was delayed by RNA sequencing without other aspects of host biology being affected (Figure 3C), and IBA1 positive phagocyte infiltration was reduced (Figure 3D). Flow cytometry showed reduced B cell populations and reduced T cell frequency after challenge (Figures S3B and S3C).

7. Dexamethasone reduced distal Isg15 induction at one day across kidney, liver, gastrointestinal tract, brain and heart, with levels similar to vehicle by three days (Figure 3E).

8. Under dexamethasone, infectious virus appeared in liver, spleen, olfactory bulb and gastrointestinal tract in addition to lung, although not with complete penetrance (Figures 3F and 3G). Direct plaque assay of blood and sera was negative, but amplification on Vero E6 cells revealed low level viremia in roughly fifty to seventy five percent of dexamethasone treated animals at one day only (Figure 3H). Animals with no detectable blood virus at one day were also negative in kidney, liver, spleen, heart and gastrointestinal tract.

9. Intravenous administration bypassed airway control and produced infection of lung, kidney, liver, spleen, heart and gastrointestinal tract by three days, with kidney titers near ten to the fourth plaque forming units per millilitre at both one and three days (Figures 4A and 4B). Kidney infection was confirmed as productive by nucleocapsid staining in parenchymal cells and in endothelial cells of the peritubular capillary network and renal medulla, with elevated Mx in surrounding tissue (Figures 4C to 4F).

10. Intranasal infection generated more circulating interferon than intravenous infection at one day, with comparable levels by three days (Figures 4G and 4H). Across all animals, circulating interferon was detected only in those with positive lung titers regardless of inoculation route, while kidney viral load did not affect circulating interferon (Figures 4I and 4J). The same pattern held using an RNA sequencing derived interferon score (Figures 4K and 4L). A single intravenously infected animal with lung infectious virus at one day also had the highest circulating interferon.

11. Kidney interferon stimulated gene profiles were similar whether infection was intranasal or intravenous for the most highly expressed canonical genes, which the authors interpret as the airway response producing distal protection comparable to what a locally infected organ generates (Figure S4C).

12. Prior airway infection protected distal organs against an intravenous challenge. Animals infected intranasally and then challenged intravenously three days later had significantly less virus in blood and sera, kidney, liver and gastrointestinal tract at four days than animals given buffer intranasally before the same intravenous challenge (Figures 4M and 4N), while Isg15 induction was comparable across groups (Figure S4E).

Mechanistic model

The study does not establish a definitive molecular mechanism, and the central claim is a causal relationship established by intervention rather than a pathway. What the data support is that productive replication in the airway is the source of type I and type III interferon that enters the circulation, that this circulating interferon rather than local replication accounts for the antiviral transcriptional state in organs where no virus is recovered, and that this state restricts where the virus can subsequently establish infection. The strongest support comes from three independent manipulations pointing the same way, namely that dexamethasone delays the airway response and permits distal infection with transient viremia, that intravenous delivery bypasses the airway response and permits productive distal infection, and that deliberately priming by airway infection before intravenous challenge reduces distal viral loads. The inference that blood cells respond to interferon rather than sense virus rests on the absence of Ifnb and Ifnl reads in blood together with their presence in lung, which is indirect. The circulating interferon measurement is a bioassay reading Mx induction in hamster fibroblasts rather than a direct measurement of interferon protein, since the authors state that hamster reagents were unavailable. The data do not identify which cells produce the circulating interferon, do not distinguish the relative contributions of type I and type III interferon, and do not establish the compartment through which virus disseminates, since infectious material was recoverable from blood only after an amplification step. The authors also do not exclude that dexamethasone acts on the distal organs directly in addition to its effect on the airway. The extension of this model to severe human COVID-19 or to Long Covid is offered explicitly as speculation requiring further study.

Conceptual or technical advance

Separating where a respiratory virus replicates from where the host responds, and then showing by intervention that the response in the second place depends on replication in the first, reframes extrapulmonary antiviral signaling as a protective output of the lung rather than as evidence of distal infection. It follows that the heterogeneity of COVID-19 presentation can be considered as a function of the speed and strength of early airway immune engagement, and that individuals whose interferon response is blunted by age, immunosuppression, inborn error or autoantibody may permit the transient viremia that seeds distal organs. Practically, the work supplies a serum bioassay that makes circulating interferon quantifiable in hamsters despite the absence of species specific reagents, and it provides two complementary ways to remove airway priming, one pharmacological and one by route of inoculation, that can be applied to other respiratory pathogens.

Relationship to the broader research program

This extends the laboratory's prior hamster work on SARS-CoV-2, which described a wave of inflammation reaching tissues with little productive replication and proposed that disseminated viral material might account for it. The present study tests the alternative explanation and finds for circulating interferon, and then shows that the systemic response is protective rather than merely present. The recurring position across these studies is that the host response, measured transcriptionally and across whole animals, is the primary object of study and that the model must be chosen so that response can be seen. Category 3 synthesis, visible only when several corpus papers are read together, is that the laboratory repeatedly treats the site of replication and the site of response as separable variables and designs interventions that decouple them. That statement rests on more than this paper alone.

Related publications

• Hoagland et al. 2021, leveraging the antiviral type I interferon system as a first line of defense against SARS-CoV-2 pathogenicity, predecessor. The same laboratory's hamster study describing systemic inflammation despite minimal distal replication, which frames the question addressed here.
• Blanco-Melo et al. 2020, imbalanced host response to SARS-CoV-2 drives development of COVID-19, predecessor. Source of the low interferon and high chemokine framing used in the introduction, and Blanco-Melo is a co-author here.
• Boudewijns and colleagues, cited by the authors as showing that ablation of STAT2, and of type I but not type III interferon signaling, increases viral RNA and dissemination in the hamster model, companion. Provides the genetic counterpart to the pharmacological suppression used here.

Limitations and boundaries

All in vivo work is in young male golden hamsters at five to seven weeks, so age and sex effects are not addressed even though advanced age is the comorbidity the authors invoke in interpreting their results. Group sizes are small, generally three or four animals per condition. The authors themselves note inconsistency in recovering infectious material from nonpulmonary tissue and attribute it to genetic diversity in outbred hamsters together with stochasticity, which limits how firmly penetrance can be stated for the dexamethasone results. Dexamethasone is a broad suppressor of de novo transcription and is not specific to interferon signaling, so effects beyond delayed priming cannot be excluded, including direct effects on distal organs. Intravenous administration uses a thousandfold higher dose than the intranasal route, so route and dose are not independent in that comparison. Circulating interferon is measured by bioassay rather than directly, and the assay does not distinguish type I from type III interferon. Ifna could not be assessed at all because of hamster genome annotation. Viremia was detectable only after amplification, so its magnitude and duration are not quantified, and the route of dissemination is not established. The priming experiment was read at four days to precede adaptive immunity, so it speaks to innate priming only, and one animal in the control group was excluded for failing to become infected. Finally, the extension to severe human COVID-19 and to Long Covid is presented by the authors as speculation and is not tested here.

Audience summaries

25 words

Airway infection with SARS-CoV-2 sends interferon into the bloodstream that protects distant organs, and hamsters whose airway response is blunted or bypassed develop infection in those organs.

75 words

SARS-CoV-2 provokes an antiviral response in organs it barely reaches. In hamsters, that response comes from interferon produced in the infected lung and carried in the blood, not from local replication. Suppressing the airway response with a steroid, or injecting virus directly into the bloodstream to skip the lung, allowed productive infection of liver, kidney, spleen and brain. Infecting through the airway first protected those organs against a subsequent bloodstream challenge.

150 words

Golden hamsters infected intranasally with SARS-CoV-2 showed an interferon signature in all nine organs profiled, while infectious virus was largely confined to lung and olfactory bulb. Whole blood carried an interferon stimulated gene signature without interferon transcripts of its own, and a fibroblast bioassay detected circulating interferon by one day after infection. Dexamethasone delayed airway interferon stimulated gene induction and phagocyte infiltration without changing early lung titers, and under that treatment virus appeared in liver, spleen, olfactory bulb and gastrointestinal tract, with low level viremia detectable only after amplification on permissive cells. Intravenous inoculation bypassed the airway and produced productive infection of kidney, liver, spleen, heart and gastrointestinal tract, confirmed in kidney by nucleocapsid staining of parenchymal and endothelial cells. Circulating interferon was present only in animals with lung titers, and prior airway infection reduced distal viral loads after an intravenous challenge, indicating that airway derived interferon restricts tropism.

Discoveries supported by this paper

Documented publication relationships

• Leveraging the antiviral type I interferon system as a first line of defense against SARS-CoV-2 pathogenicity — predecessor. Direction: this paper → linked paper. stated in the Related publications section of 2023-carrau-delayed-engagement-of-host-defense
• Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19 — predecessor. Direction: this paper → linked paper. stated in the Related publications section of 2023-carrau-delayed-engagement-of-host-defense

Pathogens

Technologies