Abstract
While many diseases of aging have been linked to the immunological system, immune metrics capable of identifying the most at-risk individuals are lacking. From the blood immunome of 1,001 individuals aged 8–96 years, we developed a deep-learning method based on patterns of systemic age-related inflammation. The resulting inflammatory clock of aging (iAge) tracked with multimorbidity, immunosenescence, frailty and cardiovascular aging, and is also associated with exceptional longevity in centenarians. The strongest contributor to iAge was the chemokine CXCL9, which was involved in cardiac aging, adverse cardiac remodeling and poor vascular function. Furthermore, aging endothelial cells in human and mice show loss of function, cellular senescence and hallmark phenotypes of arterial stiffness, all of which are reversed by silencing CXCL9. In conclusion, we identify a key role of CXCL9 in age-related chronic inflammation and derive a metric for multimorbidity that can be utilized for the early detection of age-related clinical phenotypes.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 digital issues and online access to articles
$119.00 per year
only $9.92 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
The cell subpopulation, immune protein and cell signaling data for the Stanford Aging and Vaccination studies are publicly available on ImmPort Bioinformatics Repository (http://www.immport.org/immport-open/public/home/home) under the following study IDs SDY311 (cytokines, phosphoflow assays and CyTOF surface phenotyping), SDY312 (cytokines, phosphoflow assays and flow cytometry surface phenotyping), SDY314 (flow cytometry surface phenotyping), SDY315 (cytokines, phosphoflow assays and CyTOF surface phenotyping) and SDY478 (cytokines and CyTOF surface phenotyping). The gene expression data utilized in this study to compute gene expression-iAge has been uploaded to the Gene Expression Omnibus under accession number GSE168753. Our study complies in full with the STROBE statement, STARD guidelines and GATHER statement.
Code availability
The code used for the identification of immunotypes and construction of the inflammatory clock has been deposited on GitHub (https://github.com/) and is available under: https://github.com/clingsz/GAE. For the immunological characterization of immunotypes, we used R programming (https://www.r-project.org/). The LASSO and Elastic Net regularized generalized linear models package (glmnet) for R programming can be found at: https://cran.r-project.org/web/packages/glmnet/index.html. Maximum likelihood estimation is a function of the STATS4 R package found at: https://www.rdocumentation.org/packages/stats4/versions/3.4.1.
Change history
26 July 2021
A Correction to this paper has been published: https://doi.org/10.1038/s43587-021-00102-x
References
Kotas, M. E. & Medzhitov, R. Homeostasis, inflammation, and disease susceptibility. Cell 160, 816–827 (2015).
Liu, F. et al. Mechanosignaling through YAP and TAZ drives fibroblast activation and fibrosis. Am. J. Physiol. Lung Cell Mol. Physiol. 308, L344–L357 (2015).
Crusz, S. M. & Balkwill, F. R. Inflammation and cancer: advances and new agents. Nat. Rev. Clin. Oncol. 12, 584–596 (2015).
Franceschi, C. & Campisi, J. Chronic inflammation (inflammaging) and its potential contribution to age-associated diseases. J. Gerontol. A Biol. Sci. Med. Sci. 69, S4–S9 (2014).
Furman, D. et al. Expression of specific inflammasome gene modules stratifies older individuals into two extreme clinical and immunological states. Nat. Med. 23, 174–184 (2017).
López-Otín, C., Blasco, M. A., Partridge, L., Serrano, M. & Kroemer, G. The hallmarks of aging. Cell 153, 1194–1217 (2013).
Cavadas, C., Aveleira, C. A., Souza, G. F. P. & Velloso, L. A. The pathophysiology of defective proteostasis in the hypothalamus — from obesity to ageing. Nat. Rev. Endocrinol. 12, 723–733 (2016).
Efeyan, A., Comb, W. C. & Sabatini, D. M. Nutrient-sensing mechanisms and pathways. Nature 517, 302–310 (2015).
Grivennikov, S. I., Greten, F. R. & Karin, M. Immunity, inflammation, and cancer. Cell 140, 883–899 (2010).
Hunter, R. L. et al. Inflammation induces mitochondrial dysfunction and dopaminergic neurodegeneration in the nigrostriatal system. J. Neurochem. 100, 1375–1386 (2007).
Jurk, D. et al. Chronic inflammation induces telomere dysfunction and accelerates ageing in mice. Nat. Commun. 5, 4172 (2014).
Lasry, A. & Ben-Neriah, Y. Senescence-associated inflammatory responses: aging and cancer perspectives. Trends Immunol. 36, 217–228 (2015).
Nathan, C. & Cunningham-Bussel, A. Beyond oxidative stress: an immunologist’s guide to reactive oxygen species. Nat. Rev. Immunol. 13, 349–361 (2013).
Oh, J., Lee, Y. D. & Wagers, A. J. Stem cell aging: mechanisms, regulators and therapeutic opportunities. Nat. Med. 20, 870–880 (2014).
Thevaranjan, N. et al. Age-associated microbial dysbiosis promotes intestinal permeability, systemic inflammation, and macrophage dysfunction. Cell Host Microbe 21, 455–466 (2017).
Alpert, A. et al. A clinically meaningful metric of immune age derived from high-dimensional longitudinal monitoring. Nat. Med. 25, 487–495 (2019).
Goldberg, E. L. & Dixit, V. D. Drivers of age-related inflammation and strategies for healthspan extension. Immunol. Rev. 265, 63–74 (2015).
Franceschi, C., Garagnani, P., Vitale, G., Capri, M. & Salvioli, S. Inflammaging and garb-aging. Trends Endocrinol. Metab. 28, 199–212 (2017).
Morrisette-Thomas, V. et al. Inflamm-aging does not simply reflect increases in pro-inflammatory markers. Mechanisms Ageing Dev. 139, 49–57 (2014).
Furman, D. et al. Chronic inflammation in the etiology of disease across the life span. Nat. Med. 25, 1822–1832 (2019).
Blazkova, J. et al. Multicenter systems analysis of human blood reveals immature neutrophils in males and during pregnancy. J. Immunol. 198, 2479–2488 (2017).
Brodin, P. et al. Variation in the human immune system is largely driven by non-heritable influences. Cell 160, 37–47 (2015).
Furman, D. et al. Systems analysis of sex differences reveals an immunosuppressive role for testosterone in the response to influenza vaccination. Proc. Natl Acad. Sci. USA 111, 869–874 (2014).
Furman, D. et al. Apoptosis and other immune biomarkers predict influenza vaccine responsiveness. Mol. Syst. Biol. 9, 659 (2013).
Furman, D. et al. Cytomegalovirus infection enhances the immune response to influenza. Sci. Transl. Med. 7, 281ra43 (2015).
Price, J. V. et al. Characterization of influenza vaccine immunogenicity using influenza antigen microarrays. PLoS ONE 8, e64555 (2013).
Roskin, K. M. et al. IgH sequences in common variable immune deficiency reveal altered B cell development and selection. Sci. Transl. Med. 7, 302ra135 (2015).
Shen-Orr Shai, S. et al. Defective signaling in the JAK–STAT pathway tracks with chronic inflammation and cardiovascular risk in aging humans. Cell Syst. 3, 374–384 (2016).
Wang, C. et al. Effects of aging, cytomegalovirus infection, and EBV infection on human B cell repertoires. J. Immunol. 192, 603–611 (2014).
Montoya, J. G. et al. Cytokine signature associated with disease severity in chronic fatigue syndrome patients. Proc. Natl Acad. Sci. USA 114, E7150–E7158 (2017).
Maecker, H. T. et al. New tools for classification and monitoring of autoimmune diseases. Nat. Rev. Rheumatol. 8, 317–328 (2012).
Ridker, P. M. et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. N. Engl. J. Med. 377, 1119–1131 (2017).
Barnett, K. et al. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 380, 37–43 (2012).
Zou, H. & Hastie, T. Regularization and variable selection via the Elastic Net. J. R. Stat. Soc. B Stat. Methodol. 67, 301–320 (2005).
Rockwood, K., Rockwood, M. R. & Mitnitski, A. Physiological redundancy in older adults in relation to the change with age in the slope of a frailty index. J. Am. Geriatr. Soc. 58, 318–323 (2010).
Mahmood, S. S., Levy, D., Vasan, R. S. & Wang, T. J. The Framingham Heart Study and the epidemiology of cardiovascular disease: a historical perspective. Lancet 383, 999–1008 (2014).
Bertsch, T. et al. C-reactive protein and the acute phase reaction in geriatric patients. Z. Gerontologie Geriatr. 48, 595–600 (2015).
Scheller, J., Chalaris, A., Schmidt-Arras, D. & Rose-John, S. The pro- and anti-inflammatory properties of the cytokine interleukin-6. Biochim. Biophys. Acta 1813, 878–888 (2011).
Krutzik, P. O., Irish, J. M., Nolan, G. P. & Perez, O. D. Analysis of protein phosphorylation and cellular signaling events by flow cytometry: techniques and clinical applications. Clin. Immunol. 110, 206–221 (2004).
Kisseleva, T., Bhattacharya, S., Braunstein, J. & Schindler, C. W. Signaling through the JAK/STAT pathway, recent advances and future challenges. Gene 285, 1–24 (2002).
Altara, R. et al. Left ventricular dysfunction and CXCR3 ligands in hypertension: from animal experiments to a population-based pilot study. PLoS ONE 10, e0141394 (2015).
Altara, R. et al. CXCL10 is a circulating inflammatory marker in patients with advanced heart failure: a pilot study. J. Cardiovasc. Transl. Res. 9, 302–314 (2016).
Altara, R., Mallat, Z., Booz, G. W. & Zouein, F. A. The CXCL10/CXCR3 axis and cardiac inflammation: implications for immunotherapy to treat infectious and noninfectious diseases of the heart. J. Immunol. Res. 2016, 4396368 (2016).
Hardison, J. L., Wrightsman, R. A., Carpenter, P. M., Lane, T. E. & Manning, J. E. The chemokines CXCL9 and CXCL10 promote a protective immune response but do not contribute to cardiac inflammation following infection with Trypanosoma cruzi. Infect. Immun. 74, 125–134 (2006).
Vlachopoulos, C., Aznaouridis, K. & Stefanadis, C. Prediction of cardiovascular events and all-cause mortality with arterial stiffness: a systematic review and meta-analysis. J. Am. Coll. Cardiol. 55, 1318–1327 (2010).
Ben-Shlomo, Y. et al. Aortic pulse wave velocity improves cardiovascular event prediction: an individual participant meta-analysis of prospective observational data from 17,635 subjects. J. Am. Coll. Cardiol. 63, 636–646 (2014).
Chirinos, J. A. Large artery stiffness, microvascular function, and cardiovascular risk. Circ. Cardiovasc Imaging 9, e005903 (2016).
Figueiredo, V. N. et al. Vascular stiffness and endothelial dysfunction: correlations at different levels of blood pressure. Blood Press. 21, 31–38 (2012).
Vanhoutte, P. M., Feletou, M. & Taddei, S. Endothelium-dependent contractions in hypertension. Br. J. Pharmacol. 144, 449–458 (2005).
Castellon, X. & Bogdanova, V. Chronic inflammatory diseases and endothelial dysfunction. Aging Dis. 7, 81–89 (2016).
Harvey, A., Montezano, A. C. & Touyz, R. M. Vascular biology of ageing: implications in hypertension. J. Mol. Cell Cardiol. 83, 112–121 (2015).
Kamo, T., Akazawa, H. & Komuro, I. Cardiac nonmyocytes in the hub of cardiac hypertrophy. Circ. Res. 117, 89–98 (2015).
Arnaoutova, I. & Kleinman, H. K. In vitro angiogenesis: endothelial cell tube formation on gelled basement membrane extract. Nat. Protoc. 5, 628–635 (2010).
Mukai, N. et al. A comparison of the tube forming potentials of early and late endothelial progenitor cells Exp. Cell. Res. 314, 430–440 (2008).
Tousoulis, D., Kampoli, A. M., Tentolouris, C., Papageorgiou, N. & Stefanadis, C. The role of nitric oxide on endothelial function. Curr. Vasc. Pharmacol. 10, 4–18 (2012).
Voyta, J. C., Via, D. P., Butterfield, C. E. & Zetter, B. R. Identification and isolation of endothelial cells based on their increased uptake of acetylated low-density lipoprotein. J. Cell Biol. 99, 2034–2040 (1984).
Boisen, L., Drasbek, K. R., Pedersen, A. S. & Kristensen, P. Evaluation of endothelial cell culture as a model system of vascular ageing. Exp. Gerontol. 45, 779–787 (2010).
Brandes, R. P., Fleming, I. & Busse, R. Endothelial aging. Cardiovascular Res. 66, 286–294 (2005).
Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126, 663–676 (2006).
Sayed, N. et al. Clinical trial in a dish using iPSCs shows lovastatin improves endothelial dysfunction and cellular cross-talk in LMNA cardiomyopathy. Sci. Transl. Med. 12, eaax9276 (2020).
Rivard, A. et al. Age-Dependent Impairment of Angiogenesis. Circulation 99, 111–120 (1999).
Minamino, T. & Komuro, I. Vascular cell senescence: contribution to atherosclerosis. Circ. Res. 100, 15–26 (2007).
Yin, H. & Pickering, J. G. Cellular senescence and vascular disease: novel routes to better understanding and therapy. Can. J. Cardiol. 32, 612–623 (2016).
Uryga, A. K. & Bennett, M. R. Ageing induced vascular smooth muscle cell senescence in atherosclerosis. J. Physiol. 594, 2115–2124 (2016).
Jia, G., Aroor, A. R., Jia, C. & Sowers, J. R. Endothelial cell senescence in aging-related vascular dysfunction. Biochim. Biophys. Acta, Mol. Basis Dis. 1865, 1802–1809 (2019).
Stojanovic, S. D., Fiedler, J., Bauersachs, J., Thum, T. & Sedding, D. G. Senescence-induced inflammation: an important player and key therapeutic target in atherosclerosis. Eur. Heart J. 41, 2983–2996 (2020).
Zieman, S. J., Melenovsky, V. & Kass, D. A. Mechanisms, pathophysiology, and therapy of arterial stiffness. Arter. Thromb. Vasc. Biol. 25, 932–943 (2005).
Kohn, J. C., Lampi, M. C. & Reinhart-King, C. A. Age-related vascular stiffening: causes and consequences. Front. Genet. 6, 112 (2015).
Lu, P., Takai, K., Weaver, V. M. & Werb, Z. Extracellular matrix degradation and remodeling in development and disease. Cold Spring Harb. Perspect. Biol. 3, a005082 (2011).
Kuzuya, M. et al. Glycation cross-links inhibit matrix metalloproteinase-2 activation in vascular smooth muscle cells cultured on collagen lattice. Diabetologia 44, 433–436 (2001).
González-Amaro, R., Diaz-González, F. & Sánchez-Madrid, F. Adhesion molecules in inflammatory diseases. Drugs 56, 977–988 (1998).
Khanamiri, S. et al. Marked vascular dysfunction in a case of peripartum cardiomyopathy. J. Vasc. Res. 56, 11–15 (2019).
Sayed, N. et al. Transdifferentiation of human fibroblasts to endothelial cells: role of innate immunity. Circulation 131, 300–309 (2015).
Hannum, G. et al. Genome-wide methylation profiles reveal quantitative views of human aging rates. Mol. Cell. 49, 359–367 (2013).
Horvath, S. DNA methylation age of human tissues and cell types. Genome Biol. 14, R115 (2013).
Holly, A. C. et al. Towards a gene expression biomarker set for human biological age. Aging Cell 12, 324–326 (2013).
Ishikawa, M. et al. Plasma and serum lipidomics of healthy white adults shows characteristic profiles by subjects’ gender and age. PLoS ONE 9, e91806 (2014).
Yu, Z. et al. Human serum metabolic profiles are age dependent. Aging Cell. 11, 960–967 (2012).
Menni, C. et al. Metabolomic markers reveal novel pathways of ageing and early development in human populations. Int J. Epidemiol. 42, 1111–1119 (2013).
Hertel, J. et al. Measuring biological age via metabonomics: the metabolic age score. J. Proteome Res. 15, 400–410 (2016).
Collino, S. et al. Metabolic signatures of extreme longevity in northern Italian centenarians reveal a complex remodeling of lipids, amino acids, and gut microbiota metabolism. PLoS ONE 8, e56564 (2013).
Lawton, K. A. et al. Analysis of the adult human plasma metabolome. Pharmacogenomics 9, 383–397 (2008).
Rahman, S. A. & Adjeroh, D. A. Deep learning using convolutional LSTM estimates biological age from physical activity. Sci. Rep. 9, 11425 (2019).
Tang X., Wang Z., Luo W. and Gao S. Face aging with identity-preserved conditional generative adversarial networks. Proc. IEEE/CVF Conf. Computer Vision Pattern Recognition https://doi.org/10.1109/CVPR.2018.00828 (2018).
Zhang K. et al. Fine-grained age estimation in the wild with attention LSTM networks. IEEE Trans. Circuits Systems Video Technol. https://doi.org/10.1109/TCSVT.2019.2936410 (2019).
Putin, E. et al. Deep biomarkers of human aging: application of deep neural networks to biomarker development. Aging 8, 1021–1033 (2016).
Mamoshina, P. et al. Population specific biomarkers of human aging: a big data study using South Korean, Canadian, and Eastern European patient populations. J. Gerontol. A Biol. Sci. Med Sci. 73, 1482–1490 (2018).
Subramanian, N., Torabi-Parizi, P., Gottschalk, R. A., Germain, R. N. & Dutta, B. Network representations of immune system complexity. WIREs Syst. Biol. Med. 7, 13–38 (2015).
Bengio, Y. Learning deep architectures for AI. Found. Trends Mach. Learn. 2, 1–127 (2009).
Tanaka, T. et al. Plasma proteomic signature of age in healthy humans. Aging Cell. 17, e12799 (2018).
Fourati, S. et al. Pre-vaccination inflammation and B-cell signalling predict age-related hyporesponse to hepatitis B vaccination. Nat. Commun. 7, 10369 (2016).
Lehallier, B. et al. Undulating changes in human plasma proteome profiles across the lifespan. Nat. Med. 25, 1843–1850 (2019).
Bonfante, H. L. et al. CCL2, CXCL8, CXCL9 and CXCL10 serum levels increase with age but are not altered by treatment with hydroxychloroquine in patients with osteoarthritis of the knees. Int J. Rheum. Dis. 20, 1958–1964 (2017).
Roubenoff, R. et al. Monocyte cytokine production in an elderly population: effect of age and inflammation. J. Gerontol. A Biol. Sci. Med Sci. 53, M20–M26 (1998).
Torres, K. C. L. et al. Immune senescence and biomarkers profile of Bambui aged population-based cohort. Exp. Gerontol. 103, 47–56 (2018).
Shurin, G. V. et al. Dynamic alteration of soluble serum biomarkers in healthy aging. Cytokine 39, 123–129 (2007).
Zheng, Y. et al. Age-related pro-inflammatory and pro-angiogenic changes in human aqueous humor. Int J. Ophthalmol. 11, 196–200 (2018).
Sandovici, I. et al. Ageing is associated with molecular signatures of inflammation and type 2 diabetes in rat pancreatic islets. Diabetologia 59, 502–511 (2016).
de Araujo, F. F. et al. CXCL9 and CXCL10 display an age-dependent profile in Chagas patients: a cohort study of aging in Bambui, Brazil. Infect. Dis. Poverty 9, 51 (2020).
Zhou, L. et al. Age-specific changes in the molecular phenotype of patients with moderate-to-severe atopic dermatitis. J. Allergy Clin. Immunol. 144, 144–156 (2019).
Amorim, J. S. C. et al. Inflammatory markers and occurrence of falls: Bambui cohort study of aging. Rev. Saude Publica 53, 35 (2019).
Chiu, B. C. et al. Differential effects of ageing on cytokine and chemokine responses during type-1 (mycobacterial) and type-2 (schistosomal) pulmonary granulomatous inflammation in mice. Mech. Ageing Dev. 123, 313–326 (2002).
Janowski, A. M. et al. NLRC4 suppresses melanoma tumor progression independently of inflammasome activation. J. Clin. Investig. 126, 3917–3928 (2016).
Versari, D., Daghini, E., Virdis, A., Ghiadoni, L. & Taddei, S. Endothelial dysfunction as a target for prevention of cardiovascular disease. Diabetes Care 32, S314–S321 (2009).
Mezzaroma, E. et al. The inflammasome promotes adverse cardiac remodeling following acute myocardial infarction in the mouse. Proc. Natl Acad. Sci. USA 108, 19725–19730 (2011).
Nevers, T. et al. Left ventricular T-cell recruitment contributes to the pathogenesis of heart failure. Circulation Heart Fail. 8, 776–787 (2015).
Rawlings, J. S., Rosler, K. M. & Harrison, D. A. The JAK/STAT signaling pathway. J. Cell Sci. 117, 1281–1283 (2004).
Shuai, K. & Liu, B. Regulation of gene-activation pathways by PIAS proteins in the immune system. Nat. Rev. Immunol. 5, 593–605 (2005).
van Boxel-Dezaire, A. H. H., Rani, M. R. S. & Stark, G. R. Complex modulation of cell type-specific signaling in response to type I interferons. Immunity 25, 361–372 (2006).
Matsa, E. et al. Transcriptome profiling of patient-specific human iPSC-cardiomyocytes predicts individual drug safety and efficacy responses in vitro. Cell Stem Cell 19, 311–325 (2016).
Gu, M. et al. Patient-specific iPSC-derived endothelial cells uncover pathways that protect against pulmonary hypertension in BMPR2 mutation carriers. Cell Stem Cell 20, 490–504 (2017).
Lian, X. et al. Robust cardiomyocyte differentiation from human pluripotent stem cells via temporal modulation of canonical Wnt signaling. Proc. Natl Acad. Sci. USA 109, E1848–E1857 (2012).
Nagueh, S. F. et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. Eur. J. Echocardiogr. 10, 165–193 (2009).
Redfield, M. M., Jacobsen, S. J., Borlaug, B. A., Rodeheffer, R. J. & Kass, D. A. Age- and gender-related ventricular-vascular stiffening: a community-based study. Circulation 112, 2254–2262 (2005).
Lang, R. M. et al. Chamber Quantification Writing G, American Society of Echocardiography’s G, Standards C and European Association of E. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J. Am. Soc. Echocardiogr. 18, 1440–1463 (2005).
Godec, J. et al. Compendium of immune signatures identifies conserved and species-specific biology in response to inflammation. Immunity 44, 194–206 (2016).
Friedman, J., Hastie, T. & Tibshirani, R. Regularization paths for generalized linear models via coordinate descent. J. Stat. Softw. 33, 1–22 (2010).
Gold, L. et al. Aptamer-based multiplexed proteomic technology for biomarker discovery. PLoS ONE 5, e15004 (2010).
Behzadi-Khormouji, H. et al. Deep learning, reusable and problem-based architectures for detection of consolidation on chest X-ray images. Comput. Methods Prog. Biomed. 185, 105162 (2020).
Kingma, D. & Ba, J. A. A method for stochastic optimization. In 3rd International Conference for Learning Representations (2014).
Dai, H., Leeder, J. S. & Cui, Y. A modified generalized Fisher method for combining probabilities from dependent tests. Front. Genet. 5, 32 (2014).
Foroutan, M. et al. Single sample scoring of molecular phenotypes. BMC Bioinf. 19, 404 (2018).
Acknowledgements
We thank the study participants for their time and dedication and the staff of the Stanford-LPCH Vaccine Program for recruiting participants and conducting the studies. Support for the conduct of these studies was from The Buck Institute for Research on Aging (to D.F.), the Ellison Foundation, National Institutes of Health (NIH) U19 AI057229, U19 AI090019 (to M.M.D.) and NIH/NCRR CTSA award number UL1 RR025744. This work was also supported by grants to C.F. from the EU Horizon 2020 Project PROPAG-AGEING (grant 634821), the EU JPND ADAGE project, the Ministry of Education and Science of the Russian Federation Agreement (agreement no. 075-15-2020-808). We gratefully acknowledge additional funding support from the NIH K01 HL135455, Stanford TRAM scholar award (N.S.), the Paul F. Glenn Foundation and the NIH Stanford Alzheimer’s Disease Research Center P50AG047366.
Author information
Authors and Affiliations
Contributions
D.F. conceived, conceptualized and designed the study; coordinated the biological analysis of samples and contributed to the analysis of experimental data and interpretation of the results. D.F., M.M.D., C.L.D. and J.G.M. conceived the study, provided guidance and funding. J.C.W. and F.H. provided guidance for the experimental work. N.S. and L.C. conducted in vitro and in vivo mice and EC experiments. D.F., B.L., Y.H., K.N., A.A. and T.G. conducted deep-learning and statistical analyses. S.S.O., V.J., R.T. and T.H. provided guidance for the in silico analysis of experimental data. N.S., Y.R.-H., F.H. and H.T.M. carried out or supervised the human data measurements; T.K., A.G. and Z.K.-R. helped to edit the manuscript. C.F., T.W.-C., B.L., R.O., D.M. collaborated with the study in centenarians. N.S., Y.H. and D.F. wrote the manuscript. All authors approved the final version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
D.F. and M.M.D. are co-founders of Edifice Health, a company that utilizes iAge. The remaining authors declare no competing interests.
Additional information
Peer review information Nature Aging thanks M. Luisa Iruela-Arispe and the other, anonymous, reviewers for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data
Extended Data Fig. 1 1000 Immunomes Study design: systematic analysis of immune systems via ‘OMICS’ approaches.
The Stanford 1000 Immunomes Project consist of 1001 ambulatory subjects age 8 to 96 (34% males, 66% females) recruited during the years 2007 to 2016 for a longitudinal study of aging and vaccination, and for an independent study of chronic fatigue syndrome from which only healthy controls were included. For all samples of the Stanford 1KIP, deep immune phenotyping was conducted at the Stanford Human Immune Monitoring Center, where peripheral blood specimens were isolated and analyzed using standard procedures. Peripheral blood samples were obtained by venipuncture and peripheral blood mononuclear cells or whole blood samples were used for determination of cellular phenotypes and frequencies (N = 935) and for investigation of in vitro cellular responses to a variety of cytokine stimulations (N = 818); serum samples were obtained and used for protein content determination (including a total of 50 cytokines, chemokines and growth factors) (N = 1001). Clinical characterization was assessed via clinical questionnaire in a total of 902 subjects who completed the full set of 53 clinical items. From a total of 97 healthy young and older adults, comprehensive cardiovascular phenotyping was also conducted
Extended Data Fig. 2
Age distribution of the Stanford 1KIP cohort.
Extended Data Fig. 3 Estimation of the GAE code length and accuracy of age prediction.
We used 5-fold cross-validation to identify the best code length, among lengths from 1 to 10. We selected the length of code k, whose performance was not statistically significantly worse than that of longer codes (paired t-test p-value > 0.05). Within each fold we performed nested 3-fold cross-validation to select hyper-parameters (depth, weight decay and guidance-ratio). In our experiment, the best code length is 5 (a) as adding one more code (6) does not significantly improve the total loss (p = 0.18). After obtaining the best code length as 5, we used the 5-fold-cross-validation to select the best hyper-parameter setting (depth = 2, guidance-ratio = 0.2, L2 = 0.001) on all GAE with code length 5. Finally, we trained the GAE on the whole dataset with the selected best hyper-parameter setting and obtained the predictive function as the inflammatory clock predictor. GAE was compared to other machine learning methods such as autoencoder, neural networks, PCA, and RAW in (b). For the neural network, 2 fully connected layers with 5 nodes in each layer and tanh activation function were used. For PCA and RAW, we used elastic net to predict age. The GAE method outperforms linear methods for protein data reconstruction and prediction of chronological age (b). In (c), we found that the predictive performance of gradient boosting decision tree (GBDT) has similar performance as PCA. We conclude that GAE is superior to traditional machine learning methods.
Extended Data Fig. 4 Elimination of batch effect for serum immune protein data.
. Immune protein data from serum samples were subjected to normalization and batch correction procedures (See Methods) to ensure data from different sources can be combined and used as a whole. a, Spearman correlation between immune protein features and batch ID shows a strong dependency of data source on top 4 components (raw data, green line), which reaches a steady state after component 5. Data normalization and batch correction removes batch effect as indicated by lower mean absolute Spearman correlation between all features and batch id (blue line), which indicates impossibility to distinguish sample source from corrected data. b, Upper panel: immune protein expression heatmap of uncorrected data, Lower panel: immune protein expression heatmap of corrected data. The two batches come from two study cohorts, the Chronic Fatigue Syndrome Study (CFS) and Aging and vaccination study cohort (Flu).
Extended Data Fig. 5 iAge predictive of multi-mordity.
To select for predictors of comorbidity without bias, based on available data for all 902 subjects while controlling for the age effect, age-adjusted cross-validation was performed (a). By applying differential penalty values for each regressor, age variable is ‘forced in’, while imposing a stringent penalty (the lasso penalty) to all other features, so that selected variables do not correlate with age. A Mean Absolute Error (MAE) for the prediction of comorbidity of 0.41 is observed (b). Eighteen features are selected including inflammatory clock, high cholesterol and BMI (c) and immune parameters such as total CD8 (+) T cells, plasmablasts and transitional B cells (negative predictors) and IgD+CD27- and IgD-CD27- B cells, effector CD8 (+) T cells, total lymphocytes and monocytes, and central memory T cells (positive predictors) (d)
Extended Data Fig. 6 Univariate Regression between Age and CXCL9.
Significant correlation between age and CXCL9 using univariate regression analysis. We used linear regression where CXCL9 were regressed onto age. Correlation coefficient (R2) and p-value of F-test of overall significance are reported.
Extended Data Fig. 7 Luminex data for cardiovascular validation cohort.
In a validation study, 97 healthy adults (aged 25–90) well matched for cardiovascular risk factors, were selected from a total of 151 recruited subjects. Immune protein analysis was conducted in samples from these subjects. CXCL9, HGF, CXCL1, and LIF were found to change in the same direction in both the Stanford 1KIP and the validation cohort.
Extended Data Fig. 8 Human blood endothelial progenitor cells and mice endothelial cells.
a, Representative images of human blood progenitor endothelial cells from young (left) and old (right) individuals. b, Representative images of capillary-like networks show impaired tube formation by human BECs of old individuals compared to young. To further confirm the potential contribution of CXCL9 in cardiovascular aging, we assessed its expression in young (3–4 month) and old mice (2 yr.) endothelial cells (c). ECs isolated from old mice showed higher levels of CXCL9 (P value = 0.023) (d), while at the same time showed impaired EC function as evident by decreased tube formation (P value = 0.042) (a, f). Figure S8: All data represented as mean ± SEM, n = 3, *P < 0.05. Statistical analyses were performed using Student’s t-test (paired). Scale bar: 50 μm.
Extended Data Fig. 9 Expression of CXCR3 RNA in different tissue types.
CXCR3 was not expressed in iPSC induced cardiomyocytes (iPSC-CM), Fibroblast, or iPSC. However, it is highly expressed in iPSC induced endothelial cells and Human Umbilical Vein Endothelial Cells (HUVEC). All data represented as mean ± SEM.
Extended Data Fig. 10 Validation of the effects of CXCL9 on endothelial function.
Representative images of capillary-like networks from scramble- and CXCL9-KD hiPSC-ECs show that CXCL9-KD hiPSC-ECs retain their capacity to form tubes even at later passages when compared to scramble that showed impaired tube formation towards later passages of hiPSC-ECs. Scale bar: 50 μm. Experiment was repeated 3 times.
Supplementary information
Supplementary Information
Supplementary Figs. 1 and 2, and Tables 1–3.
Source data
Source Data Fig. 1
Age, number of comorbidities, frailty source data.
Source Data Fig. 2
CyTOF, cytokine source data.
Source Data Fig. 3
Cytokine data,CXCL9, arterial stiffness, PWV raw data.
Source Data Fig. 4
CXCL9 expression in blood EC (relative fold change). Tube formation in blood EC (number of tubes). NO production (μmol l−1). LDL uptake (mg protein h−1). Tube formation in iPSC-EC (number of tubes). LDL uptake (mg protein h−1). NO production (μmol l−1).
Source Data Fig. 5
iPSC-KO and Scramble gene expression data. Count data of specific pathways. Enrichment scores of pathways.
Source Data Fig. 6
iPSC-KO and Scramble gene expression data. Percentage relaxation of aortas to acetylcholine.
Source Data Fig. 7
Cell count for proliferation assay. SA-b-Gal activity. count for CD31+ capillaries.
Source Data Extended Data Fig. 2
Demographics source data.
Source Data Extended Data Fig. 3
Raw data used in predictive models.
Source Data Extended Data Fig. 4
Batch corrected before and after cytokine expression.
Source Data Extended Data Fig. 5
CyTOF source data.
Source Data Extended Data Fig. 6
Age and CXCL9 source data.
Source Data Extended Data Fig. 7
Raw data of cytokine expression in validation dataset.
Source Data CXCL9 expression in mouse EC (relative fold change).
Tube formation in mouse ECs (number of tubes).
Source Data Extended Data Fig. 9
Expression of CXCR3 RNA source data.
Rights and permissions
About this article
Cite this article
Sayed, N., Huang, Y., Nguyen, K. et al. An inflammatory aging clock (iAge) based on deep learning tracks multimorbidity, immunosenescence, frailty and cardiovascular aging. Nat Aging 1, 598–615 (2021). https://doi.org/10.1038/s43587-021-00082-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s43587-021-00082-y
