Finding longevity genes is only one of many goals for gerontologists. An equally important mission is unraveling the genetic processes involved in age-related traits and diseases. IRGB and NIA are focusing their attention on the Mediterranean island of Sardinia. Why? We quote directly from the rationale and background for the initiation of the study in 2001, which also apply to follow-up studies in the course of the Project:
To identify genetic bases for prominent age-associated changes, including cardiovascular risk factors and determinants of personality traits, in a founder population. The results of the study will extend the studies of aging-associated conditions of outbred populations.
Recent reports form the Baltimore Longitudinal Study on Aging (BLSA) and other population studies have identified some epidemiological and genetic risk factors for aging-associated diseases; but the majority of studies have been observational, determining the range of values for a phenotype as a function of aging. These analyses have been increasingly supplemented in recent years by genetic studies based on powerful genome-wide association studies in numerous general populations. Nevertheless, considerable attention has focused on the promise of ‘founder populations’ for the simplification of analysis of complex traits and aging-related diseases. Such rare populations arise from a delimited group living in a defined region for many centuries with minimal admixture from outside populations. The Sardinian population is one of the few that is both numerous and accessible enough, and one of the most extreme in its relative lack of heterogeneity. It can thus be studied for a wide variety of both frequent and relatively rare traits. The targeted region for the present study, Ogliastra, has a particularly isolated population of 60,000, in an area enclosed by two mountain ranges and the sea; and for the study of diseases, the entire catchment of 1,500,000 Sardinians across the island can furnish enough cases and controls to investigate genetic factors for a wide range of conditions and maladies.
The ProgeNIA/SardiNIA study has now decisively established that analyses on the island can indeed help to analyze both complex traits and diseases, often providing information that was difficult or impossible to obtain in other population studies.
Concerning the choice and progress of the Project, its positive features include
- Relatively constant environment (in rural towns for the core study)
- High level of altruism and interest in sharing genetic patrimony
- Support of the local Bishop, Mayors, and Health Authorities
- Local clinic location that facilitates rapid and repeated visits: e.g., 3,500 were recalled in 1 year for fractionation of white cells into 95 subtypes within 90 min of blood draw, and for a subsequent second blood draw to extend studies to B cells and cytokines.
- Structures of more than 1,000 families embedded in a large cohort of individuals, permitting accurate estimates of genetic heritability.
- Lower cost sequencing and genotyping coverage: 2,000 individuals sequenced at 4-fold coverage yielded the imputed DNA sequence for 7,000 in the cohort.
- Reliable detection of rare DNA sequence variants Assessment of diagnostic/prognostic strength of genetic factors in a longitudinal study with medical outcomes.
- Analysis of mitochondrial variation and its transmission.
- Straightforward formulation and testing of imputation algorithms.
- Excellent coverage of genetic variation in Europe.
- Easier discovery of interactions/pleiotropy of traits.
In a first survey, the project team recruited over 6,100 subjects from a catchment area including four towns in east-central Sardinia and assessed a first list of >200 traits. The baseline survey has been followed by follow-up visits that collected longitudinal data on the same traits collected at baseline but added assessment of frailty-related traits, namely measures of bone density and geometry, muscle strength, and gait speed, and additional cardiovascular measures (see below). In the course of SardiNIA3, along with the expansion of the cohort and the addition of more traits (see below), increases in testing efficiency and additional cost-sharing funds from Sardinian sources permit the completion of Fourth Visits for the entire cohort. Also in current actions, DNA sequencing has recovered essentially all of the genetic variation in the cohort, and further arrangements for an Outcome Study have also been made to be implemented in “SardiNIA4”, a further 5-year continuation of the Project that also projects Fifth Visits for the cohort.
The infrastructure for the clinic and phenotypic testing has been stable, with stringent quality control, which is reflected in the high quality of the database. The initial sample cohort included over 62% of the eligible population living in the region (age 14-102 years), and at least 96% of the initial cohort have all grandparents born in the same province. The initial group included 4,933 phenotyped sib pairs, 4,266 phenotyped parent-child pairs, >4,069 phenotyped cousin pairs, and >6,459 phenotyped avuncular pairs. Additional recruitment has increased the cohort substantially, and results have consistently shown that for essentially every trait, most of the associated genes and variants would be involved in determining variance in both young and old and in men and women. Thus, genetic analyses can draw on data from all ages and both genders.
The added value of studying a founder population has also been demonstrated by the extension of DNA analysis to the full range of variation by sequencing. The population has proven to contain the great bulk of variation found in other populations, but during its isolation over many thousands of years, many variants rare elsewhere have risen to relatively high levels on the island – by drift or selection – and others have newly arisen as Sardinian-specific. These have provided extensive new information about a whole range of traits and pathways (e.g., Nature Genetics November, 2015 articles and Editorial).
Sardinia also offers a special entrée to the genetics of specific diseases that are especially prevalent in the founder population. This includes the anomalously high incidence of autoimmune diseases including Multiple Sclerosis and Type 1 diabetes, which interrupt the high to low gradient of incidence from Northern to Southern Europe. Again, this has fostered novel findings in causation and pathophysiology.
The founder population itself also contains within its DNA a record of human demography through history, which has permitted the inference of the timing of human population movements based on mitochondrial and Y chromosome analyses.
Regarding the course and mechanism of aging, the longitudinal study, now in its 15th year, focuses on residents of the cluster of towns to collect longitudinal information on more than 400 age-related quantitative traits (“endophenotypes” or “quantitative risk-related genetic or environmental factors”) that can be scored on a continuous scale, as well as >200 dichotomous traits (including major diseases and risk factors such as smoking). The use of quantitative traits permits the study of the entire range of allelic variation in a population, with particular interest in a range of cardiovascular risk factors, anthropometric measurements, blood test values, facets of personality, and bone-density and frailty-related variables.
The longitudinal study of a broad range of phenotypes in a founder population is distinctive in this study, and stable environmental/epidemiological factors combined with the simplification of genetic analyses also aid in proposed joint investigations of relative risk. Furthermore, because we are collecting risk factor data, we can also analyze, in an Outcome Study, the prognostic power and/or pathophysiological relevance of earlier predictors for the onset of serious risk factors [e.g., increases in pulse wave velocity as a function of earlier (predictor) lipid and inflammatory markers.
The governance of the Project began in 2001 with its implementation under the auspices of the National Institute on Aging (NIA), leading to the name SardiNIA (i.e., Sardi + NIA). A contract to Giuseppe Pilia, M.D., Principal Investigator for the Italian National Research Council (CNR) Institute in Cagliari (now the Institute of Genetic and Biomedical Research (IRGB), with David Schlessinger, Ph.D., Chief of the Laboratory of Genetics at NIA, as Principal Investigator for NIA and the indispensable support of Antonio Cao, M.D., the Dean of human genetics studies in Italy. During the initial 5 years (mid-2001 to mid-2005), a clinic was set up in the town of Lanusei and a team of Investigators (“Progenia”) assembled there to recruit participants and collect data. After the untimely death of Dr. Pilia, Dr. Manuela Uda, took over for the next 5-year period, and was succeeded in 2010 by Francesco Cucca, M.D, Professor of Medical Genetics at the University of Sassari and director of IRGB-CNR. Soon after the project commenced, Drs. Pilia and Schlessinger were joined by a senior statistical geneticist, Goncalo Abecasis, Ph.D., at the University of Michigan, to head up data analysis efforts and train several Sardinian geneticists, while other Sardinians grew to capable members of the team of Ph.D. analysts n Cagliari, Sardinia – they include Serena Sanna, Carlo Sidore, Mauro Pala, Maristella Steri and Eleonora Porcu. Currently, the collaborative group comprising the leadership of the Project includes those individuals working with investigators at several U.S. Universities (especially, Stephen Montgomery, Ph.D., at Stanford University, and John Novembre, Ph.D., at the University of Chicago, as well as Dr. Abecasis at the University of Michigan) and Myriam Gorospe head of the the Genetics Laboratory at the NIA in Baltimore, and since 2018 co-PI of the Project). Additional key collaborators include Edward Lakatta, M.D. (Cardiovascular Science) and Luigi Ferrucci (Translational Gerontology Branch at NIA).
Concerning distribution of effort in ongoing and planned work, data collection continues at the SardiNIA/ProgeNIA clinic in Lanusei, where the genotyping work at different level of resolution is also performed, and island-wide sites for the collection of disease-specific and blood bank control samples. Sequencing work is done in a joint effort at the Sardinian and U. Michigan collaborative sites.
The SardiNIA investigators operating in Sardinia, led by Dr. Cucca, are involved in choosing and analyzing traits, with strong emphasis on immunological and haematological traits, and are the primary producers and analysts of all genetic data, including the design and coordination of the genotyping and sequencing work (in close collaboration with the University of Michigan team led by Dr. Abecasis) as well as of the phenotypic data, with the supervision of all clinical activities performed in the Center in Lanusei. They are also the primary investigators in traits and test protocols for haematological, inflammatory, immunological (both cellular and humoral), and anthropometric (including those related to internal organs) phenotypes, as well as those related to kidney function. They also participate actively in the design and analysis of cardiovascular traits, with special emphasis in Echocardiographic cardiovascular measures. They also take the lead in writing papers and obtaining grants for the traits of paramount relevance for this component of the project.
NIA investigators, led by Dr. Schlessinger, are involved in choosing and analyzing traits and test protocols for frailty, cardiovascular, and psychological traits; in the maintenance of the master database and preparation of summaries of heritability and age- and gender-specific distributions for all traits; they also actively collaborate on immunosenescence studies performed in Sardinia; in the development of algorithms for targeted types of analyses, and in the preparation of all manuscripts
The Core Project is supported by the ongoing contract from the National Institute on Aging with cost-sharing by the IRGB, aided by cost-sharing from the Italian National Research Council and awards and collaborative grants from other funding sources for specific topics. They include support for 1) extensive study of immune cells and molecules; 2) extensive DNA sequencing, with assessment of DNA methylation patterns; 3) RNA and methylation sequencing for white cells from the cohort; and 4) refined study of immune system cell types in cohort members. More recently, the whole project was greatly boosted from an Horizon2020 project, ImmunoAgeing, coordinated by Dr Francesco Cucca and dedicated to the study of Immunosenescence (http://www.immunoageing.eu).
Auxiliary studies that have continued to benefit from the cohort and population structure and repeated visits include analyses of human demographic history, involving the groups of Dr. John Novembre at the University of Chicago and Antonio Torroni at the Universita degli Studi di Pavia, and analyses of genetic and environmental effects on RNA levels with the team of Dr. Stephen Montgomery.
Case-control studies led by Dr Francesco Cucca have concentrated on diseases strikingly common in Sardinia such as autoimmune multiple sclerosis and type 1 diabetes as well as thalassemia. They also include an assessment of diseases of more global high prevalence such as chronic kidney disease in collaboration with Dr. Antonello Pani from the Brotzu Hospital in Cagliari.
For these case-control studies, the entire island remains the catchment to recruit enough cases from among well characterized clinical series collected by Sardinian physicians, compared with unaffected Sardinians represented by healthy blood donors collected in the main Transfusion Centers on the island.
The results of the case-control studies and those of related quantitative traits generated in the SardiNIA project are analyzed jointly to reveal coincident associations, indicative of molecules, cells and pathways primarily involved in the disease process. The results of these studies as well as those focusing on aging have the potential to translate into substantial improvement in clinical diagnosis and identification of new targets for treatment.
The overall goals are
- Measures of genetic variation, quantitative traits, and of longitudinal trends for traits and outcomes, focusing on the longitudinal dimension of the study, aiming at determining individual trajectories and features of the aging process, as well as the identification of longer-term health-related outcomes; and
- Gene variation associated with phenotypic traits, prevalent diseases, and age-related trends, based on increasingly complete recovery of the range of trait variability and genetic variation in the population.
Specific Aim 1. Measures of genetic variation, quantitative traits, and of longitudinal data for traits and outcomes.
a) Measures of variation: genotyping and sequencing in an enlarged cohort
- Variation.Recruitment. The number of individuals in the overall datasets (combining the SardiNIA cohort and the various case-control studies) has increased to ~17,000.Genotyping and sequencing of DNA and RNA. The core group of ~6,000 individuals of the SardiNIA cohort has been supplemented with an additional ~700 volunteers from the same area to integrate those that could not participate in the various stages of the follow up with a targeted emphasis on individuals older than 90. Overall, ~6,700 individuals have had full-genome scans thorough 4 genotyping chips: immunochip, cardiometabochip, exome chip, and the Illumina OmniExpress Chip with 750,000 .common. SNPs that give comparable coverage across the genome. The case: control individuals have been studied with the OmniExpress chip to integrate their genomes into the cohort by imputation.In an extension of analyses, 3,400 individuals from the cohort, chosen to optimize the representation of cases and families, have had their DNA sequenced to 4-fold coverage across the whole genome. This has recovered a total of 17,500,000 single nucleotide variants, including the full range of 12,500,000 variants that occur with minor allele frequencies >0.4%. In addition, rare, personal, nonsense or nonsynonymous coding variants have also been recovered for further analyses. Using the family relationships, imputation has permitted the inference of all the variation useable for GWAS in all individuals in the cohort. DNA sequencing is thus completed for the cohort, and has permitted us to generate a Sardinian HapMap that increases the accuracy of imputation to > 99.9%.At the same time we have begun analysis of the levels of RNA transcripts in PBMC from the same core 3,400 individuals in the cohort. To date, poly(A)+ RNA from 605 individuals has been sequenced, and preliminary analyses working in collaboration with Dr. Steven Montgomery at Stanford University show the order of one cis-eQTL per locus across the genome.
- Quantitative traits.We have maintained a primary concentration on risk factors for age-related diseases and traits/risk factors. For traits, we have specifically profited from the experience in long-standing epidemiological studies at the NIA (for example, incorporating frailty-related tests), and have made trait lists concordant with the lists of traits studied in the Baltimore Longitudinal Study of Aging and the AGES and InCHIANTI studies to facilitate cross-replication and meta-analyses among the studies sponsored by the NIA.The focus is on levels of immune system cells, haematological traits, cardiovascular traits, personality traits, sensory capacity (retinal scans and hearing tests); railty-related traits (walking speed, grip strength, and bone density); and kidney ultrasound features.
- DiseasesData collection has been completed for multiple sclerosis, Type1 Diabetes, and breast cancer cases and controls and efforts continue to assemble data to assess chronic kidney disease.
b. Epidemiological and longitudinal trends: outcome study
To identify health related events that can be correlated with genetic and environmental risk factors and their interactions, a system has been set up to ascertain major health-related outcomes. The sources of information start with the self-reports of individuals at visits, but depend more on three more reliable sources: 1) medical records from local primary care physicians (every Sardinian has a local physician); 2) an extensive database maintained by the Sardinia Region for all health and social services delivered by the health care system, including hospitalization, treatment, and all pharmaceuticals taken by each individual; and 3) death certificates on file locally.
Specific Aim 2. Analyses of gene variation associated with phenotypic traits and/or prevalent diseases, and of age-related trends and outcomes.
a) Loci and gene variation associated with traits
- Heritability. This primary tool determines for us the relative contribution of epidemiological and genetic factors, including the ‘narrow’ heritability due to genetic variation. Based on the results, we determine which traits can be usefully followed up with further GWAS or other analyses
- Quantitative traits.Analyses. For all of the standard traits listed above (Specific Aim 1.b), including cardiovascular, personality, sensory capacity, frailty-related, and kidney ultrasound measures, full GWAS of all traits are being completed with the 12,500,000 SNP set and over 10,000 individuals.The analysis has proceeded furthest thus far for the sample trait of levels of immune system cells. GWAS analysis on 1,200 individuals with the 12,500,000 SNP set revealed 23 with a large effect on at least one cell type. For example, the SH2B3 variant, R262W, already found associated with several autoimmune diseases and negative regulation of hematopoiesis, was shown to particularly affect the numbers of CD4+ T cells, likely resulting in a loss of function. Extensions of this approach have now extended analyses to a much larger cohort, with up to 100 significant loci identified, and to additional types of cells, including B cell subtypes, and cytokines.Especially notable, a number of SNPs that affect the levels of particular cell types have also been associated with autoimmune diseases in Consortium studies; our analyses extend associations of variants with a disease to specify the cell types in which the alleles have their effects. This provides markers and potential cellular and molecular targets for further analyses and possible intervention.
- Disease associationsCase:control analyses are being completed for 2,000 cases of Type 1 diabetes, and 1,000 cases of breast cancer in the cohort. Multiple sclerosis, by contrast, has revealed several additional novel risk loci. GWAS with 3,000 patients and 3,000 controls studied with several million SNPs identified CBLB as associated with MS. We will also continue to analyze several age-related conditions that affect substantial numbers of the Core cohort (in conjunction with Consortium analyses; e.g., of blood pressure/hypertension and CKD).
b. Epidemiological and longitudinal trends: outcome study
With the completion of the fourth visit data analyses can start estimating age-related trajectories for most of the major traits and ask definitively 1) whether differential rates of change are predicted by polymorphisms in specific genes and 2) whether any such genes are the same that were found associated with same traits in the cross-sectional analysis. In a first example that used aggregate groups of individuals, a genetic score was generated from thirteen reported eGFR- and related loci. Both uni- and multivariable analyses were then assessed to query the relationship between clinical, ultrasound, and genetic variables with three outcomes: CKD, change in eGFR, and fast eGFR decline. Genetic risk score, in addition to classical factors, was determined to impose independent risk.
The Project has included 3 major interacting groups from its inception:
- The ProgeNIA Team, in Sardinia. The work of the team is funded by a contract to the Italian National Research Council, with Dr. Francesco Cucca as Principal Investigator. The team comprises:
- Physicians: Alessandro Delitala, Liana Ferreli, Francesco Loi, Nicola Curreli, Elen Balalci, Maria Grazia Pilia (visiting the volunteers and collecting medical records and data in the Center in Lanusei)
- Cardiologists: Marco Orru, Antonello Ganau and his team.
- Nurses: Monica Lai, Paola Loi
- Receptionists: Susanna Murino, Roberta Corda
- Biologists: Edoardo Fiorillo, Valeria Orru, Mariano Dei, Sandra Lai, Maria Grazia Piras, Monia Lobina, Francesca Virdis, Valentina Serra (study of the immune system, high-resolution analysis of cells and molecules of interest, processing of biological samples, extraction of DNA, preparation of cell lines, storage and maintaining of all samples), Mara Marongiu, Francesca Crobu (preparation of RNA samples and functional studies), Magdalena Zoledziewska, Antonella Mulas, Fabio Busonero, (genotyping and validation of genetic variants, analysis of traits of interest), Manuela Uda, Silvia Naitza, Laura Crisponi (functional analyses of traits of interest).
- Sequencing team in Pula at the Sardinia Ricerche Bioark: Andrea Angius, Andrea Maschio (high throughput sequencing of DNA and RNA).
- Statisticians and bioinformaticians: Carlo Sidore, Maristella Steri, Mauro Pala, Gabriella Sole, Marco Masala, Michele Marongiu, Stefano Onano
- Intramural Program investigators at the NIA, with Dr. David Schlessinger and Myriam Gorospe as co-Principal Investigators (see home page https://irp.nih.gov/pi/myriam-gorospe). Major participants include
- Drs. Edward Lakatta, Majd Al Ghatrif, James Strait, Kirill Tarasov, and Angelo Scuteri, Laboratory of Cardiovascular Sciences.
- Drs. Luigi Ferrucci and Toshiko Tanaka, Translational Gerontology Branch.
- Drs. Jun Ding, Osorio Meirelles, Yong Qian, Ramaiah Nagaraja, and Christopher Coletta, Laboratory of Genetics.
- The Statistical Genetics team led by Dr. Goncalo Abecasis at the University of Michigan (see home page https://sph.umich.edu/faculty-profiles/abecasis-goncalo.htm.
In the last 5 years two additional groups where included as key partners of the study:
- The Population genetics team led by Dr. John Novembre, University of Chicago (see home page http://jnpopgen.org)
- And the Gene regulation and expression team led by Dr. Stephen Montgomery, Stanford University (see home page http://montgomerylab.stanford.edu)