A diminished immune response underlies age-related SARS-CoV-2 pathologies
Comparison of young and older golden hamsters infected with SARS-CoV-2 shows that age reduces the magnitude and duration of the innate response and of tissue repair, expands suppressor T cells and IL-17-driven neutrophil recruitment, and lowers germinal centre B cell frequency and neutralizing antibody potency without raising lung virus titres.
- Senior authors
- Benjamin R. tenOever
- Correspondence
- Benjamin R. tenOever
Research areas & themes
Citation
Oishi K, Horiuchi S, Frere J, Schwartz RE, tenOever BR. A diminished immune response underlies age-related SARS-CoV-2 pathologies. Cell Reports. 2022. Volume 39, issue 13, article 111002.
DOI 10.1016/j.celrep.2022.111002. PMID 35714615. PMCID PMC9181267.
Oishi and Horiuchi contributed equally. tenOever is the lead contact.
One-sentence contribution
Comparison of young and older golden hamsters infected with SARS-CoV-2 shows that age reduces the magnitude and duration of the innate response and of tissue repair, expands suppressor T cells and IL-17-driven neutrophil recruitment, and lowers germinal centre B cell frequency and neutralizing antibody potency without raising lung virus titres.
Executive summary
Morbidity and mortality from SARS-CoV-2 rise sharply with age, and the biology behind that pattern was incompletely defined. The authors used the golden hamster, which reproduces many features of COVID-19, to compare animals of six to nine weeks with animals of forty weeks or more after intranasal infection. Lung virus titres were higher in young animals at one and three days and comparable between cohorts at five and seven days, so the aged phenotype is not explained by greater viral replication. Transcriptional profiling of uninfected lung showed that TGF-beta signalling, NF-kappaB signalling and cell division signatures were already lower in older animals at baseline. After infection, chemokine induction was diminished and delayed in older animals, and although the interferon-stimulated gene profile was comparable through three days it fell away rapidly thereafter despite equivalent virus. Ki67 staining showed reduced proliferation in the lungs of older animals, indicating slower repair. Flow cytometry showed that older animals failed to expand CD3 positive T cells or Th1 cells, carried higher frequencies of FoxP3 positive regulatory and CD8 T cells, showed elevated CCR6 positive CD8 T cells and IL-17 staining in both lymphocytes and epithelium, and recruited more CD11b positive neutrophils. Spike-specific B cell numbers and anti-receptor-binding-domain IgG titres were unchanged, but germinal centre B cells were markedly reduced in lung, spleen and draining lymph node, and serum neutralizing potency fell by more than sixty percent. Human COVID-19 cadaver lung transcriptomes showed IL-17 induction correlating with age.
Scientific context
Viruses that evade host defences generally cause most disease in the very young and the very old, but SARS-CoV-2 was noted early on to fall disproportionately on the aged. Ageing was known to alter immune function in several relevant ways described by other groups. Regulatory T cells, which secrete TGF-beta and suppress B and T cell function, accumulate with age. IL-17, the signature cytokine of T helper 17 cells, also increases with age and can activate neutrophils. Neutrophils contribute to early clearance but also to tissue injury, and neutrophil counts are used clinically as a marker related to acute respiratory distress syndrome in COVID-19. Previous hamster studies from other laboratories had established the model as reproducing COVID-19 biology and had reported that greater disease severity in older animals corresponded with lower rather than higher virus replication together with significant tissue damage. What had not been done was a molecular characterisation of the virus-host interaction and of both immune arms across age in that model, which is the gap this study addresses. The authors state that they corroborate the earlier severity and replication observations and extend them.
Central question
What changes in the innate and adaptive immune response to SARS-CoV-2 with age, and can those changes rather than differences in virus replication account for the greater morbidity seen in older hosts?
Experimental strategy
The design holds the virus constant and varies host age, using two golden hamster cohorts defined by published criteria for the species, six to nine weeks against forty weeks or more, both given the same intranasal dose. Because earlier work had already indicated that older animals do not carry more virus, the strategy is built to look past titre. Lung transcriptomes are taken from uninfected animals of both ages first, so any post-infection difference can be read against a documented baseline difference, and then across a time course of one, three, five and seven days, with the days five and seven comparisons singled out because virus levels are equivalent there and any host difference therefore cannot be attributed to viral load. Transcriptional inference is then checked against protein and cell-level measurement, with Ki67 immunohistochemistry validating the proliferation signature and flow cytometry resolving the immune populations that bulk RNA can only imply. The adaptive arm is approached through methods the group had previously developed for the hamster, which matters because commercial reagents for this species are limited. Antibody quantity and antibody quality are deliberately separated, with ELISA against the receptor binding domain measuring the former and a plaque reduction neutralization test measuring the latter, so that a defect in affinity maturation can be distinguished from a defect in antibody production. Finally, the IL-17 finding is tested for relevance beyond the model by examining RNA sequencing of human COVID-19 cadaver lung tissue stratified by age.
Key findings
-
Lung virus titres were significantly higher in young hamsters at one and three days post-infection and comparable between cohorts at five and seven days (Figure 1B). Greater disease in older animals is therefore not accompanied by greater lung replication.
-
In uninfected animals, gene set enrichment showed TGF-beta signalling, NF-kappaB signalling and proliferation signatures significantly diminished in older lungs (Figure 1C, Table S1). The authors note that NF-kappaB signalling has been reported as required for SARS-CoV-2 replication and suggest this may contribute to the lower early virus in older animals, which is interpretation drawn from prior work rather than tested here.
-
Chemokine induction after infection was diminished in older animals, with lower Ccl2, Ccl4, Ccl8, Ccl28 and Il1b. Older animals showed a rebound at three days but never reached the peak levels seen in young animals at five and seven days (Figure 1D).
-
Interferon-stimulated gene profiles were comparable between cohorts through the first three days but declined rapidly in older animals thereafter despite equivalent virus at five days (Figure 1E). The authors state that the reason for this decline remains unclear and later attribute it, as an interpretation, to diminished immune cell recruitment.
-
At five days, older animals showed a marked increase in neutrophil-associated transcripts S100a8 and S100a9 with log2 fold change above five, and significant downregulation of macrophage and dendritic cell markers Cd163, Clec4a, Mafb and Slamf8 (Figure S1C, Table S4). At seven days, histone-modification-associated transcripts including Tet2, Tet3, Ogt and Suz12 and the activated T cell marker Nfat5 were lost in older animals (Table S5).
-
Ki67 transcript was reduced roughly fourfold in older animals, and immunohistochemistry confirmed markedly fewer Ki67 positive cells in older lungs at five and seven days (Figure 1F, Table S3). The authors read this as a defect in the kinetics or the biology of lung repair.
-
Total immune-enriched cell recruitment to the lung was comparable between cohorts, but older animals had significantly higher CD45 positive cells at baseline and this reversed by seven days (Figures 2A and 2B). Young animals expanded CD3 positive T cells while older animals did not (Figure 2C), CD4 positive T cells plateaued in older animals from an already high baseline (Figure 2D), and Th1 cells defined as CXCR3 positive CD4 cells rose in young and fell in older animals (Figure 2E). CXCR3 positive CD8 cells expanded significantly in both cohorts (Figure 2F).
-
Older animals showed a pronounced increase in FoxP3 positive CD4 regulatory T cells and FoxP3 positive CD8 T cells at seven days (Figure 3A), accompanied by a fivefold rise in Tgfb transcript. RNA sequencing showed loss of TGF-beta receptor subunit expression approaching an order of magnitude (Figure S4A). The authors interpret this combination as elevated TGF-beta production in an environment depleted of its receptor, invoking published negative feedback on receptor expression, and note that it sits awkwardly with the reduced repair signalled by Ki67.
-
CCR6 positive CD8 T cells were elevated in older animals while CCR6 positive CD4 cells were not (Figure 3B), and immunohistochemistry showed markedly more IL-17 positive cells in older lung parenchyma, in epithelium as well as in lymphocytes (Figure 3C). CD11b positive presumed neutrophils were significantly elevated in older animals after infection despite a lower baseline (Figure 3D), corroborated by CD11b and myeloperoxidase staining (Figure 3E, Figure S3).
-
In human COVID-19 cadaver lung RNA sequencing, IL-17 induction correlated statistically with age, with IL-17 target genes including Il6, Cxcl2 and Cxcl1 trending in the same direction (Figures S4B and S4C).
-
Total B cell frequency, class switching, and spike-specific B cell frequency in lung and spleen showed no significant age-related difference (Figure 4A, Figures S5A and S5B), and anti-receptor-binding-domain IgG and IgG2 titres were also unchanged (Figure 4C). Spike-specific germinal centre B cells defined as Bcl6 positive were markedly reduced in older animals in lung, spleen and mediastinal lymph node (Figure 4B).
-
Serum from older animals neutralized SARS-CoV-2 less effectively, with 50 percent plaque reduction titres of 1,549 for young and 514 for older animals, a reduction of more than sixty percent (Figure 4D). Because antibody quantity was unchanged, the authors read this as a deficit in affinity maturation rather than in production.
Mechanistic model
The study does not establish a definitive causal mechanism, and the authors describe the adaptive phenotype as most certainly multifactorial. The account the data support is correlational and layered. Older animals begin with lower baseline NF-kappaB and TGF-beta signalling and lower proliferative capacity in the lung. On infection they mount chemokine and interferon-stimulated gene responses that are delayed, lower in magnitude, and not sustained, which the authors attribute to reduced immune cell recruitment rather than to differences in virus. In parallel the suppressor compartment expands, with more regulatory T cells and FoxP3 positive CD8 T cells producing more TGF-beta, which the authors propose creates an environment favouring IL-17-producing cells, whose product then recruits neutrophils through NF-kappaB-driven chemokine induction. Neutrophil influx is offered as the link to tissue damage, citing prior work on neutrophils in COVID-19 pathology. On the humoral side, the proposed chain runs from expanded regulatory T cells to suppression of germinal centre formation to impaired affinity maturation and therefore to weaker neutralizing antibody at unchanged antibody titre. Each link in these chains is supported by an association measured in the same animals rather than by intervention. No depletion, blockade, or transfer experiment is performed, so causality between regulatory T cell expansion and germinal centre loss, or between IL-17 and neutrophil recruitment, is proposed rather than demonstrated. The authors also flag an internal tension they do not resolve, namely that TGF-beta is associated with tissue repair yet repair markers are reduced, and offer loss of receptor expression as a candidate reconciliation. The suggestion that IL-17 antagonists might be therapeutically useful is stated explicitly as requiring further testing and validation.
Conceptual or technical advance
The study separates three things that age-related severity might otherwise be assumed to share a cause. Virus burden is not elevated in older animals, antibody quantity is not reduced, and immune cell recruitment in bulk is not reduced, yet the quality and duration of the response are all diminished. That decomposition is what makes the phenotype tractable, because it points at affinity maturation, at response duration, and at the suppressor and IL-17 axis rather than at replication control. The work also applies flow cytometry and antigen-specific B cell methods developed by the group for the hamster, which extends what can be measured in a model that had mostly been characterised transcriptionally and histologically, and it supplies matched innate and adaptive profiling across age in one experiment. The IL-17 correlation with age in human post-mortem lung provides a point of contact between the model and patients.
Relationship to the broader research program
The paper depends directly on hamster immune profiling methods the laboratory published previously, and it follows the group's earlier work establishing the imbalanced host response to SARS-CoV-2 and the use of the interferon system as a line of defence against SARS-CoV-2 pathogenicity. It shares first and second authors with the coinfection study published by the same group in the same year, and both draw on the same hamster infrastructure and the same framing that the host response rather than viral replication sets the outcome. The observation that NF-kappaB signalling supports SARS-CoV-2 replication is taken from the group's own work with Nilsson-Payant. Marked as category 3 synthesis, the recurring commitment across these papers is that the interesting variable is the shape of the host response over time rather than the peak virus titre, and that the same model can be interrogated by varying the host, whether that host difference is age here or a prior infection in the coinfection study.
Related publications
- Horiuchi et al. 2021 (Science Immunology), methodological foundation. Same laboratory, shares an author, and the source of the hamster flow cytometry and adaptive immune profiling methods used throughout, alongside its own finding on immune memory from SARS-CoV-2 infection in hamsters.
- Blanco-Melo et al. 2020 (Cell), predecessor. Same laboratory. Established the imbalanced host response to SARS-CoV-2 and is cited here for neutrophil recruitment as a hallmark of COVID-19.
- Hoagland et al. 2021 (Immunity), predecessor. Same laboratory, shares authors, and cited for neutrophil recruitment in the hamster model and for leveraging the type I interferon system against SARS-CoV-2 pathogenicity.
- Nilsson-Payant et al. 2021, conceptual extension. Same laboratory. Cited as showing NF-kappaB signalling to be required for SARS-CoV-2 replication, which the authors use to interpret the lower early virus titres in older animals.
- Oishi et al. 2022 (Journal of Virology), companion. Same laboratory, same first two authors, same hamster model and SARS-CoV-2 strain, addressing coinfection rather than age.
Limitations and boundaries
The authors provide their own limitations section, and it is substantial. Commercial anti-hamster reagents are limited, so B cells were defined by MHC class II positive and CD3 negative staining rather than by CD19 or CD20, and neutrophils could not be identified with anti-Ly6G, leaving CD11b and myeloperoxidase as markers that also cover monocytes, macrophages and dendritic cells. Because older hamsters grow large and must be housed individually, and biosafety level 3 caging is limited, cohort sizes were only three to five animals, with four per time point for the virus and transcriptome course. Beyond the authors' own list, the study rests on one virus isolate at one dose delivered intranasally, on lung as the principal tissue, and on a time course ending at seven days, so nothing is established about later convalescence, reinfection, or transmission, and the authors say explicitly that how age affects protection from secondary infection requires further analysis. The two age cohorts are compared as blocks with no intermediate ages, so no dose relationship with age is established. Every mechanistic link is correlational, with no depletion, blockade, adoptive transfer, or receptor knockout, so the causal role of regulatory T cells, TGF-beta, IL-17, or neutrophils in the observed pathology is proposed rather than shown. The human data are a transcriptional correlation in post-mortem lung, which supports relevance but not mechanism. The therapeutic suggestion regarding IL-17 antagonists is explicitly extrapolation.
Audience summaries
25 words
Older hamsters infected with SARS-CoV-2 do not carry more virus. They mount a weaker, shorter immune response, repair their lungs more slowly, and make less potent antibodies.
75 words
Golden hamsters aged forty weeks or more were compared with young animals after SARS-CoV-2 infection. Lung virus was not higher in older animals, yet their chemokine and interferon responses were lower and shorter-lived, lung repair markers were reduced, T cell expansion failed, and suppressor T cells, IL-17 and neutrophils increased. Spike-specific B cells and antibody titres were normal, but germinal centre B cells were scarce and neutralizing potency fell by more than sixty percent.
150 words
Oishi and colleagues compared golden hamsters of six to nine weeks with animals of forty weeks or more after intranasal SARS-CoV-2 infection. Older animals had lower lung titres at one and three days and matched titres thereafter, so the aged phenotype does not follow from greater replication. Uninfected older lungs already showed reduced TGF-beta and NF-kappaB signalling and reduced proliferation signatures. After infection, chemokine induction was delayed and blunted, interferon-stimulated genes were not sustained past three days, and Ki67 staining indicated impaired repair. Flow cytometry showed failed T cell and Th1 expansion alongside expanded FoxP3 positive regulatory and CD8 T cells, elevated CCR6 positive CD8 cells, increased IL-17 in lymphocytes and epithelium, and more CD11b positive neutrophils. Spike-specific B cell frequency and anti-receptor-binding-domain IgG were unchanged, yet germinal centre B cells were markedly reduced and neutralizing titres fell from 1,549 to 514. Human post-mortem lung showed IL-17 correlating with age.
Documented publication relationships
- Immune memory from SARS-CoV-2 infection in hamsters provides variant-independent protection but still allows virus transmission — methodological foundation.
- Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19 — predecessor.
- Leveraging the antiviral type I interferon system as a first line of defense against SARS-CoV-2 pathogenicity — predecessor.