Immune resilience despite inflammatory stress promotes longevity and favorable health outcomes including resistance to infection - PubMed (original) (raw)

. 2023 Jun 13;14(1):3286.

doi: 10.1038/s41467-023-38238-6.

Sunil K Ahuja 1 2 3 4, Grace C Lee # 5 7 8 9, Lyle R McKinnon # 10 11, Justin A Meunier # 5 12, Maristella Steri # 13, Nathan Harper # 5 12, Edoardo Fiorillo # 13, Alisha M Smith # 5 14 12, Marcos I Restrepo # 5 7 6, Anne P Branum # 5 12, Matthew J Bottomley # 15 16, Valeria Orrù 13, Fabio Jimenez 5 12, Andrew Carrillo 5 12, Lavanya Pandranki 5 6, Caitlyn A Winter 5 6 12 17, Lauryn A Winter 5 6 12 17, Alvaro A Gaitan 5 12, Alvaro G Moreira 5 17, Elizabeth A Walter 5 7 6, Guido Silvestri 18, Christopher L King 19, Yong-Tang Zheng 20 21, Hong-Yi Zheng 20 21, Joshua Kimani 11, T Blake Ball 11, Francis A Plummer 11, Keith R Fowke 11, Paul N Harden 16, Kathryn J Wood 15, Martin T Ferris 22, Jennifer M Lund 23 24, Mark T Heise 22, Nigel Garrett 10, Kristen R Canady 5, Salim S Abdool Karim 10 25, Susan J Little 26 27, Sara Gianella 26 27, Davey M Smith 26 27 28, Scott Letendre 26, Douglas D Richman 27, Francesco Cucca 13 29, Hanh Trinh 7, Sandra Sanchez-Reilly 7 6, Joan M Hecht 7 12, Jose A Cadena Zuluaga 7 6, Antonio Anzueto 7 6, Jacqueline A Pugh 5 7 6; South Texas Veterans Health Care System COVID-19 team; Brian K Agan 30 31, Robert Root-Bernstein 32, Robert A Clark # 5 14 7 6 12, Jason F Okulicz # 30 33, Weijing He # 5 12

Collaborators, Affiliations

Immune resilience despite inflammatory stress promotes longevity and favorable health outcomes including resistance to infection

Sunil K Ahuja et al. Nat Commun. 2023.

Abstract

Some people remain healthier throughout life than others but the underlying reasons are poorly understood. Here we hypothesize this advantage is attributable in part to optimal immune resilience (IR), defined as the capacity to preserve and/or rapidly restore immune functions that promote disease resistance (immunocompetence) and control inflammation in infectious diseases as well as other causes of inflammatory stress. We gauge IR levels with two distinct peripheral blood metrics that quantify the balance between (i) CD8+ and CD4+ T-cell levels and (ii) gene expression signatures tracking longevity-associated immunocompetence and mortality-associated inflammation. Profiles of IR metrics in ~48,500 individuals collectively indicate that some persons resist degradation of IR both during aging and when challenged with varied inflammatory stressors. With this resistance, preservation of optimal IR tracked (i) a lower risk of HIV acquisition, AIDS development, symptomatic influenza infection, and recurrent skin cancer; (ii) survival during COVID-19 and sepsis; and (iii) longevity. IR degradation is potentially reversible by decreasing inflammatory stress. Overall, we show that optimal IR is a trait observed across the age spectrum, more common in females, and aligned with a specific immunocompetence-inflammation balance linked to favorable immunity-dependent health outcomes. IR metrics and mechanisms have utility both as biomarkers for measuring immune health and for improving health outcomes.

© 2023. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Fig. 1. Study concepts and cohorts.

a Immunologic resilience (IR) erosion-resistant and erosion-susceptible phenotypes and predicted outcomes. Phenotypes are defined by sexually dimorphic immune allostasis responses to antigenic (Ag) stimulation that links high or low immunocompetence (IC) and inflammation (IF) states to the indicated immunity-dependent health outcomes. Possible sources of Ag stimulation, outcomes, and cohorts/datasets are depicted. Arrows depict induction (red) and reversibility (blue) of IR states with Ag stimulation on and off respectively. *number of samples studied; **incident cancer in immunocompromised renal transplant recipients. RIX, recombinant inbred inter-cross. DILGOM dietary, lifestyle and genetic determinants of obesity and metabolic syndrome, SIV simian immunodeficiency virus. Abbreviations frequently used in this study are noted. n, number of individuals and/or samples studied. b Model. IC and IF changes during an immune injury-repair cycle in response to a single instance of Ag stimulation. c Ordinate, IC-IF states associated with the degree of deviation from optimal IR during increased Ag stimulation in individuals with the IR erosion-resistant versus -susceptible phenotypes. The alignment of optimal, suboptimal, and nonoptimal IR status with phenotypes is noted. Abscissa, time window overlapping with a period of increased Ag stimulation that could be acute, chronic, or repetitive irrespective of age. In this model, since age is a proxy, albeit imperfect, for antigenic experience, individuals with the IR erosion-susceptible phenotype may manifest suboptimal or nonoptimal IR with advancing age.

Fig. 2

Fig. 2. Metrics of immunologic resilience (IR) and association of the immune health grade (IHG) metric in the SardiNIA cohort.

a IR metrics. IHGs are described in panel (b). Two gene expression (transcriptomic) signatures termed survival-associated signature-1 (SAS-1) and mortality-associated signature-1 (MAS-1) are prognosticators of survival and mortality, respectively, after controlling for age and sex. b CD8-CD4 profiles by IHGs and cutoffs of the CD4:CD8 T-cell ratio and CD4+ T-cell counts (cells/mm3) used to derive IHGs. c Predicted associations of expression levels of transcriptomic proxies for immunocompetence (IC) and inflammation (IF) with longevity/survival. d Hazard ratios adjusted for age and sex (aHR) with 95% confidence intervals (CIs) for the indicated gene signatures associated with all-cause mortality in the acute COVID-19 cohort (90-day mortality) and Framingham Heart Study (FHS; survival over 9 years since first sampling). Representative genes and gene ontology biological process (GO-BP) terms are shown. +, positive. e Model and study phases 1 to 4. Far right, figures specific to the outcomes are noted. During antigenic (Ag) stimulation, preservation of and/or rapid restoration of a primordial status defined by IHG-I and a higher IC and lower IF (IChigh-IFlow) state is associated with superior immunity-dependent health outcomes, including a longevity/survival advantage. f Distribution of IHGs in the HIV– SardiNIA cohort. ***P < 0.001. F, female; M, male. g Odds of having the indicated IHG (with 95% confidence bands) by age and sex in the SardiNIA cohort. P, for differences in odds by sex and age are depicted. Rest, all other IHGs. h Features of CD8-CD4 equilibrium and disequilibrium grades. Assignment of IHG-I as an indicator of the IR erosion-resistant phenotype. A non-IHG-I grade signifies the IR erosion-susceptible phenotype. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.2., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 3

Fig. 3. Shift from immune health grade (IHG)-I to non-IHG-I grades in settings associated with increased antigenic stimulation.

a %IHG (prevalence) in Kenyan children according to Schistosoma haematobium egg counts in urine. b–f Acute COVID-19 cohort. b %IHGs at baseline vs. convalescence (paired): overall, by age and cytomegalovirus (CMV) serostatus. c IHG degradation and reconstitution during COVID-19 by CMV serostatus. Baseline %IHG-II and %IHG-IV, higher during COVID-19, is overrepresented in CMV− and CMV+ patients, respectively. d Baseline %IHGs by CMV serostatus: overall, and age. e Baseline %IHGs, hospitalization (hosp.) rates, and CMV seropositivity rates by age strata. _P_IHG-I and _P_IHG-IV, for the change in %IHG-I vs. other grades and %IHG-IV vs other grades across age strata, respectively. f Baseline %IHGs overall and stratified by sex and outcomes. F female, M male. Disease severity status defined by World Health Organization (WHO) ordinal scale: 1-4 [mild]; 5 [moderate]; 6–8 [severe]. g %IHGs in renal transplant recipients (RTRs) and patients with systemic lupus erythematosus (SLE). h Primary HIV infection cohort (PIC). Left, Baseline %IHGs by HIV viral load (HIV-VL). Right, %IHGs before (pre-antiretroviral therapy [ART]) and during 4 years of ART. i Schema for panel h with %IHGs in year 4 of ART. j %IHGs during 5 years of therapy-naïve HIV disease course in the subset with IHG-I at entry into the early infection cohort (EIC). k Baseline %IHG (top) and subsequent HIV seroconversion rates (bottom) in female sex workers who were HIV− at baseline stratified according to behavioral and biological (sexually transmitted infection [STI]) risk factors. Behavioral risk factors: duration of sex work, condom usage (1, never; 2, <50%; 3, ≥50%; and 4, always), clients/week, and ∆ (clients – condoms) (the difference between the number of clients/wk and condoms used/wk). Behavioral acitivty score (BAS) is the sum of scores of these risk factors. STI scores were derived based on direct and indirect indicators of STI. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.3., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 4

Fig. 4. Reconstitution of immune health grade (IHG)-I and associations of IHGs with immunity-dependent health outcomes.

a, b Left, reconstitution of IHG-I in female sex workers (FSWs) who remained HIV− (a) during 10-year follow-up and (b) for at least 4 years according to the IHG at baseline (base.). Right, behavioral activity score (BAS). c–f %IHGs in (c) Left, 43 FSWs (paired) pre- and post-seroconversion with HIV. Right, SIV− and SIV+ sooty mangabeys (SM) (unpaired); (d) sooty mangabeys by SIV serostatus, sex, and age; (e) SIV– Chinese rhesus macaques by sex and age; and (f) the uninfected counterparts of Collaborative Cross-RIX mice grouped by outcomes after Ebola infection. g Groups based on IHG at baseline and predicted IHG before COVID-19 (top). Adjusted odds ratio (aOR) with 95% confidence interval (CI) for hospitalization (middle), and adjusted hazard ratio (aHR) with 95% CI for all-cause, 30-day mortality (bottom) from two separate models, adjusted by age strata. Model 1, by baseline IHG; and model 2, by CMV serostatus. %IHG-III was low (Fig. 3b) and not included in the models. h Time to second occurrence of cutaneous squamous cell carcinoma (CSCC) by IHG at time of first occurrence of CSCC in renal transplant recipients. i Time to AIDS (CDC 1993 criteria) by baseline IHG with median CD4+ and CD8+ counts, CD4:CD8 ratio and HIV viral load (HIV-VL) values at entry to the early HIV infection cohort (EIC). P, for differences in HIV-VL vs. IHG-I is shown. j HIV-VL by entry IHG in the primary HIV infection cohort (PIC). k HIV-VL at entry and subsequent 5 years of therapy-naïve follow-up in EIC participants. Differences in HIV-VL are between participants with IHG-I vs. rest (i.e., IHG-II, IHG-III, or IHG-IV) at the indicated timepoints. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. For box plots: center line, median; box, interquartile range (IQR); whiskers, rest of the data distribution and outliers greater than ±1.5 × IQR are represented as points. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.4., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 5

Fig. 5. Inferior immunity-dependent health outcomes associated with immune health grades (IHGs) that correspond to the immunologic resilience (IR) erosion-susceptible phenotype.

a Female sex workers (FSWs) stratified first by baseline behavioral activity score (BAS) and then by subsequent HIV seroconversion status. b Odds ratio (OR) with 95% confidence interval (CI) of having IHG-III or IHG-IV at baseline (purple) or future HIV seroconversion (blue) in FSWs according to baseline BAS and total sexually transmitted infection (STI) score. Far right, OR for HIV seroconversion by baseline IHG. c Associations of CD8-CD4 disequilibrium grades IHG-III and IHG-IV with age and sex; inducers of these grades; and outcomes. Findings are from the literature survey (also see Supplementary Table 2 for details and references) and our primary datasets. Flu, influenza; CMV, cytomegalovirus; MSW, men who have sex with men; &, interquartile range for age. †The original data from stratified the CD4:CD8 ratio as ≤1.0 and >1.0. d–f Models depicting risk of indicated outcomes is lower in persons with the IR erosion-resistant phenotype (IHG-I). d HIV-AIDS, (e) COVID-19, and (f) recurrent cutaneous squamous cell cancer (CSCC) in renal transplant recipients. Pie charts depict relative proportions of the IHGs in the study group. Risk scaled from 1 to 3. Ag, antigenic; VL, viral load. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.5., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 6

Fig. 6. Immune health grade (IHG) repertoire across HIV− and HIV+ cohorts with parallels between aging, COVID-19, and HIV disease.

a Schema for defining the full repertoire of IHGs. Subgrades of IHG-II and IHG-IV defined by the CD4+ T-cell count thresholds. b Distribution of IHGs with subgrades in the SardiNIA cohort by age strata. c IHGs with subgrades in the overall acute COVID-19 cohort at baseline (n = 541) and the subset of 220 individuals with available IHG data at baseline and convalescence. d IHGs with subgrades in persons with systemic lupus erythematosus (SLE), renal transplant recipients (RTRs), participants from the primary HIV infection cohort (PIC) before initiation of antiretroviral therapy (ART) and following ART, and female sex workers (FSWs) by baseline behavioral activity score (BAS, <0 vs. ≥0). Age, median age at IHG assessment, baseline or pre-ART are shown. e Model depicting the enrichment of non-IHG-I grades during aging and at presentation with COVID-19 or HIV infection. ***P < 0.001. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.6., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 7

Fig. 7. Immunologic resilience (IR) continuum defined by transcriptomic metrics of IR associated with immune health grades (IHGs).

a Schema for IR continuum. IR tiers and erosion phenotypes defined by the IR metrics IHGs, survival-associated signature (SAS)-1, and mortality-associated signature (MAS)-1. Higher expression of SAS-1 and MAS-1 serve as transcriptomic proxies for immunocompetence (IC) and inflammation (IF), respectively. Groupings of SAS-1 and MAS-1 based on higher or lower levels of these signatures are depicted. b–i Distribution of the SAS-1/MAS-1 groupings/profiles in (b) the Framingham Heart Study (FHS) stratified by sex and age; (c) the San Antonio Family Heart Study categorized by sex, age, and both; (d) a meta-analysis of persons without (controls) vs. with Alzheimer disease (AD) and other dementia disorders; (e) persons without (control) vs. with systemic lupus erythematosus (SLE) stratified by IHGs with subgrades and age strata; (f) the acute COVID-19 cohort stratified by IHGs with subgrades; (g) participants of the early HIV infection cohort (EIC) stratified by IHGs with subgrades reconstituted during virally suppressive antiretroviral therapy (ART) or in therapy-naïve spontaneous virologic controller (SVC); (h) the acute COVID-19 cohort sampled at baseline stratified by age, hospitalization, and survivor status; (i) acute COVID-19 cohort and patients with SLE stratified by IHGs, and healthy controls and therapy-naïve (without ART) HIV+ patients by disease stage. Asymp, asymptomatic. j Schema, proportions of SAS-1/MAS-1 groupings/profiles. In panels (b–i) the SAS-1/MAS-1 groupings are based on cohort-level higher or lower expression (above or below median, respectively) of SAS-1 and MAS-1. Cohort characteristics and sources of gene expression profile data are in Supplementary Data 13a. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.6., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 8

Fig. 8. Survival-associated signature (SAS)-1 and mortality-associated signature (MAS)-1 associate with mortality and acute respiratory viral infection outcomes.

a Proportion survived in the Framingham Heart Study (FHS) by SAS-1/MAS-1 groupings/profiles calculated at time 0. b Distribution of SAS-1/MAS-1 profiles in the FHS by survival status. c Model: age, sex, and immunologic resilience (IR) levels influence lifespan. d–g Representation of SAS-1/MAS-1 profiles. d Sepsis #1 comprises healthy controls and meta-analysis of patients with community-acquired pneumonia (CAP) and fecal peritonitis (FP) stratified by sepsis response signature groups (G1 and G2 associated with higher and lower mortality, respectively). Sepsis #2 comprises healthy controls and patients with systemic inflammatory response syndrome (SIRS), sepsis, and septic shock survivors (S) and nonsurvivors (NS). e Participants in a natural influenza season cohort (age: 18–49 years) sampled at pre and during acute respiratory infection (ARI) and at spring follow-up: overall (left) or according to the indicated SAS-1/MAS-1 profile during the pre-ARI season (right). P values (asterisks, ns) for participants with SAS-1low-MAS-1high at pre-ARI (right) are for their cross-sectional comparison to the profiles at the corresponding timepoints for participants with SAS-1high-MAS-1low at pre-ARI (middle). f Schema of the timing of gene expression profiling in experimental intranasal challenges with respiratory viral infection in otherwise healthy young adults with data presented in panels g and h. T, time. g Participants inoculated intra-nasally with respiratory syncytial virus (RSV), rhinovirus, or influenza virus stratified by symptom status and sampling timepoint. Symp. symptomatic, Asymp. asymptomatic. h Participants inoculated intra-nasally with influenza virus stratified by symptom status and sampling timepoint. i Individuals with severe influenza infection requiring hospitalization collected at three timepoints, overall, and by age strata and severity. Patients were grouped by increasing severity levels: no supplemental oxygen required, oxygen by mask, and mechanical ventilation. Cohort characteristics and sources of biological samples and gene expression profile data are in Supplementary Data 13a. *P < 0.05; **P < 0.01; ***P < 0.001; ns nonsignificant. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.8., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 9

Fig. 9. Associations of Immunologic Resilience (IR) erosion phenotypes and immune correlates of IR.

a Schema of features associated with IR erosion phenotypes defined by immune health grade (IHG) status, and survival-associated signature (SAS)-1/mortality-associated signature (MAS)-1 profiles. Ag antigenic, F female, H high, IC immunocompetence, IF inflammation, L low, M male b IR erosion-resistant and IR erosion-susceptible phenotypes based on experimental models. c Correlation (r; Pearson) between expression levels of genes within SAS-1 and MAS-1 signatures with levels of an indicator for T-cell responsiveness, T-cell dysfunction, and systemic inflammation. Measures of T-cell responsiveness, T-cell dysfunction, and plasma IL-6 are from 55, 56, and 50 HIV+ individuals, respectively, on virally suppressive antiretroviral therapy from the early HIV infection cohort. d, e Levels of the indicated immune traits by IHGs in (d) sooty mangabeys seropositive for simian immunodeficiency virus (SIV) and (e) SIV-seronegative Chinese rhesus macaques. Comparisons were made between IHG-I vs. IHG-III and IHG-II vs. IHG-IV to mitigate the confounding effects of higher and lower CD4+ counts, respectively. *P < 0.05; **P < 0.01; ***P < 0.001. For box plots: center line, median; box, interquartile range (IQR); whiskers, rest of the data distribution and outliers greater than ±1.5 × IQR are represented as points. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.9., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Fig. 10

Fig. 10. Immune traits associated with immune health grade (IHG)-III or IHG-IV vs age vs both.

a In the HIV− SardiNIA cohort, 75 immune traits categorized into four groups. Within each group, traits were clustered into signatures according to whether their levels were higher or lower with IHG-III or IHG-IV, after controlling for age and sex; by age in older or younger persons with IHG-I or IHG-II, after controlling for sex; both; or neither. cDC, conventional dendritic cells. Arrows indicate significant difference at P < 1.67E-4; ND indicates no significant difference at P < 1.67E-4. Two arrows indicate both comparisons for IHG-I vs. IHG-III and IHG-II vs. IHG-IV or age within IHG-I and IHG-II are significant, one arrow indicates only one of the comparisons for IHG status or age is significant. b Representative traits by age in persons with IHG-I or IHG-II and by IHG status. Comparisons for the indicated traits were made between IHG-I vs. IHG-III and IHG-II vs. IHG-IV to mitigate the confounding effects of higher and lower CD4+ counts, respectively. Trait levels (_y_-axis) were normalized using inverse normal transformations with values ranging from −3 to 3; boxplots show covariate-adjusted residuals. Median number of individuals evaluated by IHG status and age within IHG-I or IHG-II. ns nonsignificant. c Linear regression was used to analyze the association between log2 transformed cell counts (outcome) with age and IHG status (predictors). The linear model was used to generate the fitted lines and 95% confidence bands and significance was determined by likelihood ratio test. FDR, false discovery rate P values adjusted for multiple comparisons. d Model differentiating features of processes associated with lower immune status that occur due to aging or via erosion of IR. SAS-1, survival-associated signature-1; MAS-1, mortality-associated signature-1. For box plots: center line, median; box, interquartile range (IQR); whiskers, rest of the data distribution and outliers greater than ±1.5 × IQR are represented as points. Two-sided tests were used. Statistics are outlined in Supplementary Information Section 11.3.10., P values are in Supplementary Data 14, and Source data are provided as a Source Data file.

Similar articles

Cited by

References

    1. Klunk J, et al. Evolution of immune genes is associated with the Black Death. Nature. 2022;611:312–319. doi: 10.1038/s41586-022-05349-x. - DOI - PMC - PubMed
    1. Klein SL, Flanagan KL. Sex differences in immune responses. Nat. Rev. Immunol. 2016;16:626–638. doi: 10.1038/nri.2016.90. - DOI - PubMed
    1. Austad SN, Fischer KE. Sex differences in lifespan. Cell Metab. 2016;23:1022–1033. doi: 10.1016/j.cmet.2016.05.019. - DOI - PMC - PubMed
    1. Takahashi T, et al. Sex differences in immune responses that underlie COVID-19 disease outcomes. Nature. 2020;588:315–320. doi: 10.1038/s41586-020-2700-3. - DOI - PMC - PubMed
    1. Gebhard C, Regitz-Zagrosek V, Neuhauser HK, Morgan R, Klein SL. Impact of sex and gender on COVID-19 outcomes in Europe. Biol. Sex. Differ. 2020;11:29. doi: 10.1186/s13293-020-00304-9. - DOI - PMC - PubMed

Publication types

MeSH terms

Grants and funding

LinkOut - more resources