Stem Cells in Aging

Abstract

Aging is a genetically programmed decline in the functional effectiveness of the organism. It is manifested by a collective group of changes in cells or organs that occur over the course of a lifespan, limiting the duration of life. Longevity usually refers to long-lived members of a population within species. Organs develop and can involute according to specific timetables. Such timetables correlate with a preordained proliferative capacity of cells mediated by cell and organ clocks. In this review, we discuss different aspects related to genetic and environmental factors that are involved in determining life span. We discuss the influence of ontogenic, genetic and environmental factors in aging. The genetic factors can be studied in embryonic stem cells (ESC) and in niches (microenvironments) of stem cells (SC) using cellular or experimental animal models. We discuss molecular mechanisms involving genes and proteins associated with death pathways, niches, or hubs, on longevity. Moreover, we also discuss genes and proteins, associated with death pathways, on longevity. Unraveling these mechanisms may further our understanding of human aging leading to development of therapeutic interventions with the potential of prolonging life.

Introduction

Studies in animal models and in human populations have documented a decline in the functional effectiveness with age. Aging is likely to be mediated by several overlapping mechanisms that work simultaneously. The mechanisms commonly cited include progressive accumulation of damage to DNA leading to genomic instability [reviewed in ref. 1], telomere shortening due to telomerase suppression during embryogenesis [reviewed in ref. 2], damage to macromolecules by reactive oxygen species (ROS) [reviewed in ref. 3] and changes in gene expression patterns brought about by epigenetic alterations that include changes influenced by the environment [reviewed in ref. 4]. While the aging process may have these common mechanisms, pathways leading to death may follow different paths. Embryonic stem cells (ESC) and experimental animal models have been used to study the role of genetic factors in aging. Among the genetic changes occurring in the ESC, the p53, Rb, Akt and telomerase genes have been found to be important [5]. The genetic and environmental factors known to affect aging are related to the role of the immune system, and specifically the thymus, as well as dietary manipulations. In this regard, one of the most important "clocks" studied during aging has been the thymus, which is associated with immune dysfunction including infections, autoimmune phenomena [6-11] and cancer [reviewed in ref. 12]. Recombinant inbred (RI) strains of mice enabled identification of potential genetic regions associated with short or long life span as well as genes that control variability at death [13].

This review describes how genetics of animal models has aided in understanding the aging phenomena, with the potential prediction of those factors affecting lifespan in humans. These mouse strains, which have been generated several decades ago, as well as the more recent ones, continue to provide a source for addressing unresolved questions of aging paradigms.

We will discuss the difficulties of finding genes and proteins involved in aging because of the many interactions that can be produced in many microenvironments or niches, which can also be affected by environmental factors. However, these experimental models are useful to map genetic regions or markers to guide the discovery of encoded proteins involved in long life of cells, niches, organs and the organism using methods such as microarrays. The use of experimental animals such as the RI strains are a first step to map certain genetic regions and can be used to test their relevance of such functions in relation to lifespan. Such studies combined with searches of other proteins encoded by genes (for example using microarrays along with proteomics) would help to elucidate the complexities of molecules involved in aging.

Embryonic Stem Cells During Aging: Regulation of Senescence

Mammalian cells experience limited number of cell divisions when placed in culture before entering into a stage of proliferation arrest, also known as replicative senescence [14]. The number of cell divisions in culture shows variability based on the cell type or the species, and the phenomenon called the Hayflick limit [15], suggests the existence of a mitotic clock. Senescence is invariable in most cells, including transformed cells [14]. In contrast, ESC [16] and some somatic cells [17] do not have demonstrable senescence. Several pathways that are activated independently or together with others can allow the cells to bypass senescence: the telomerase pathway required to maintain telomere ends [2], the p53 and Rb pathways needed to direct senescence in response to DNA damage [5], telomere shortening and mitotic signals [2], and the insulin-like growth factor-Akt pathway that may regulate lifespan, cell proliferation and the mitochondrial/oxidative stress pathway [18, 19]. One experimental design that has been successful in defining these pathways is based on the fact that senescence is dominant over immortality [20, 21]. These pathways are summarized in Table 1. The senescence pathway in ESCs can be characterized by the presence of activity of p53 and Rb pathways, inactivity of Akt pathway, absence of telomerase and short telomere length. By contrast, ESCs express inactive p53 and Rb, active Akt, and maintain telomerase and telomere length. These pathways include a large number of proteins encoded by different genes. Several of them can be mapped using genetic mouse models or by the use of microarrays.

Table 1. Pathways of senescence of embryonic stem cells and genes involved (ESCs).

Senescence pathways	Senescent cells	ESC	ESC over-expressed genes*
Telomerase	absent	Present	HSPCA, TEBP, DKC1,
Telomere length	short	Maintenance	TERF1, RAD50, Rif1, MRE11A, TNKS,
P53 and Rb pathway	active	Inactive or unknown	MDM2, CDC25A, CDK2, CDK4, CCND1, CCND2
Akt pathway	inactive	Active or unknown	PTEN, FKHR

The protein p53 is a tumor suppressor that normally responds to DNA damage by inducing apoptosis and preventing cell transformation [22, 23]. It is also a regulator of senescence in response to several signals including telomere shortenings, DNA damage, oncogene activation and over-expression of tumor suppressor genes [24-29]. Significantly, suppression of telomerase activity in differentiated ESC requires histone deacetylation in early hTERT gene downregulation and DNA methylation for maintenance of silencing of hTERT gene [30]. In addition, over-expression of the catalytic subunit of telomerase (TERT) allows cells to maintain a normal karyotype and proliferative capacity for prolonged periods of time beyond the Hayflick limits [30]. By contrast, loss of proliferative potential or cell death is associated with loss of telomeres. [31]. Similar to p53, Rb is an important regulator of senescence that interacts with p53 [32, 33]. Another important regulator of senescence, demonstrated in flies and worms, is the IGF/Akt pathway [34, 35]. Recently, signaling pathways of insulin/IGF-1/phosphatidylinositol-3-kinase (P13K/Akt), also known as protein kinase B, has been shown to extend the lifespan of the nematode C. elegans [36-38]. In mammalian cells, where Akt is activated, proliferation and cell survival is induced as well. This pathway is important in regulating cell size, as suggested by the phenotypes of PTEN (a phosphatase that inhibits the Akt pathway [reviewed in 39]).

Genetics of Aging Models

The difficulties of finding genes and proteins involved in aging are due to the multiple interactions occurring in numerous microenvironments or niches, which can also be affected by environmental factors. Experimental mouse models are useful for mapping genetic regions or markers to guide the discovery of encoded proteins involved in long life of cells, niches, organs and the organism, using methods such as microarrays. The use of experimental animals such as the RI strains is a first step in mapping certain genetic regions and testing their relevance in relation to lifespan. Such studies combined with searches for other proteins encoded by candidate genes using microarrays would help to elucidate the complexities of molecules involved in aging. Thymic involution that occurs earlier in some individuals than others may result from complex interactions between genetic factors and the environment. Such interactions may produce defects of thymus-dependent immune regulation associated with aging.

Using different experimental protocols to study the influence of the MHC in life span resulted in inconsistent findings, particularly following viral infections [40]. In the case of Sendai virus infection, T cells and the MHC were involved in susceptibility to acute infection and early death [41], with highest incidence in H-2^d/H-2^d mice. However, the H-2^d conferred longer life in mice not infected or not exposed to the virus, suggesting that mice exposed to virus can change the profile of life span [42]. Genes important for lifespan need to be studied against many genetic backgrounds and under different environmental conditions because of its complexity. Several genetic models have been used, in backcross and intercross mice. In (C57BL/6xDBA/2) F1xDBA/2 backcross the F1 lived longer than H-2^d homozygous in animals heterozygous for the brown locus b or homozygous b. The Bb mice lived longer than bb females; however, the dilute locus d on chromosome 9 did not influence life span. The dilute locus and the brown locus have shorter lifespan in females. The longest-lived mice were females heterozygous at the H-2 and Brown (b) loci. The shortest were males homozygous at the H-2 and Brown loci. The (C57BL/6XDBA/2) F1xF1 intercross revealed that females lived longer than males. The longest- lived females were homozygous H-2^d of dominant black phenotype at the Brown locus of chromosome 4 and homozygous at the dilute locus of chromosome 9. The longest and shortest- lived male genotypes were dilute brown H-2^d/H-2^d and dilute brown H-2^b/H-2^d, respectively [42].

In B10 congenic mice, males lived longer than females, H-2^d mice were not disadvantaged when compared to H-2^b that had not been exposed to Sendai virus infection, but the reverse was true when they were exposed [43]. In experiments using RI strains, known genes or genetic regions from the BxD (C57BL/6 and DBA/2) lifespan data were originally published by Gelman et al. [44]. These data can be retrieved from Web QTL Published Phenotypes database.

In 22 different strains BxD recombinant mice, lifespan was studied and we identified 3 (strains 2, 14, 32) that have a short survival and 4 (19, 24, 9 and 15) with long lifespan. COX models for multiple gene analyses of 1692 gene markers available demonstrated potential genes: HDC in chromosome 2, Rho in chromosome 6, CYP2A in chromosome 7, and IL-5 of chromosome 11. Of interest, the gene variants associated with longer lifespan came from DBA/2 in chromosomes 2, 14 and 6 and from C57B1/6 in chromosome 7 and 11. Figure 1, Tables 2 and 3 (unpublished data).

Figure 1.

Survival of 7 recombinant inbred strains of the F1 hybrids of C57BL/6 and DBA/2 mice. This figure shows separate survival curves using the shortest possible survival times for strains 2, 14 and 32 and the long lived span trains 9, 15, 19 and 24. The linear regression of survival on strain used ordinary least squares and analysis of variance. The linear regression of average survival in each strain on genes was based on the R-squared and Cp selection criteria and used the algorithm explained with details in Gelman et al. 1988 [44].

Table 2. Recombinant strains of mice (short and long lived), proportional hazards model.

	Strains	Median	Mean	SD	Range	Significance level Proportional hazards model
Short lived	2	479	490	128	531	<0.0001
	14	493	529	138	548	0.0002
	32	440	419	238	842	0.04
	9	816	884	176	610	0.05
	15	798	783	179	678	0.02
Long lived	19	904	939	151	647	0.003
	24	835	854	157	620	0.03
	11*	750	847	266	950	0.04

Summary statistics ** [44]

Table 3. Recombinant strains of mice (short and long lived) proportional hazards model.

Chromosome	Mapped genes	BXD content
2	Hdc (histidine decarboxilase)	D
	Gabpb1 (GA repeat binding protein, beta 1)	D
	Illa (Interleukin 1 alpha)	D
4	Iapls1-10 (intracisternal A particle lymphocyte specific 10)	D
	Adft (adipose differentiation related protein)	D
	Ifna (interferon alpha gene family, leukocyte	D
6	Rho (rhodopsin)	D
	Tpi1 (triosephosphate isomerase 1)	D
7	Cyp2b10 (cytochrome p450, family 2, subfamily b, polypeptide 10)	B
	Tam1 (tosyl arginine methylesterase 1)	B
	Iapls1-11 (intracisternal A particle, lymphocyte specific 1-11)	B
	Iapls3-4 (intracisternal A particle, lymphocyte specific)	B
	Zfp30 (zinc finger protein 30)	B
	Pmv18 (polytropic murine leukemia virus 18)	B
	Xmv30 (xenotropic murine leukemia virus 30)	B
	Ngfg (nerve growth factor, gamma)	B
11	IL5 (Interleukin 5)	B
Chromosome	Mapped sequences
2	D2Bir1 (DNA segment)	D
	D2Mit493 (DNA segment)	D
	D2Mit304 (DNA segment)	D
4	D4Rik108 (DNA segment)	D
	D4Rik109 (DNA segment)	D
	D4Rik110 (DNA segment)	D
	D4Bir5 (DNA segment)	D
	Ms6hm (minisatellite 6 hypermutable)	D
	D4Mc2 (DNA segment)	D
	D4Mit327 (DNA segment)	D
6	D6Nds2 (DNA segment)	D
7	D7Mit178 (DNA segment)	B
	Mr66-2 (middle repeat MR66-2)	B
	Tel7p (telomerase sequence)	B
	D7Mc2 (DNA segment)	B
	D7Mit114 (DNA segment)	B
	D7Rik79 (DNA segment)	B
	D7Rik80 (DNA segment)	B
	D7Mit145 (DNA segment)	B
11	D11Mit140 (DNA segment)	B
	D11Mit86 (DNA segment)	B
	D11Mit23 (DNA segment)	B
13	D13Mit18 (DNA segment)	B

The use of recombinant strains of mice with short or long life spans and the variability of age at death need to be studied by comparing adaptor proteins controlling stress response during life span. Another possible mechanism may involve the telomere lengths between these strains because telomere length was highly variable among genotypically identical siblings [45]. In this regard it is interesting that RI that are genotipically identical differ in one genetic region of chromosome 11, Gls-psl as was reported before [46], but in re-analyses of the data another region of genes associated with variability of age at death was discovered, including Hba, Hba-x, Lamrl-rs5 and Telrsl (Figure 2, Table 4).

Figure 2.

The genes and genetic markers summarized in Table 2 are placed at the position in mouse chromosome (left). The genes known at the same region are listed (right). These genes are classified accordingly: a) angiogenesis b) immune functions c) nucleic and transcription d) signal transduction e) metabolism f) cell metabolism g) protein mutation h) disease markers i) for non coded markers. Description of genes displayed in the figure: Mapping of genes involved in a) Angiogenesis: Adra2b: adrenergic receptor, α 2b.b) Immune function: Gma3: granulocyte macrophage antigen 3; Ly23: lymphocyte antigen 23; Il1: interleukin 1 complex; Il1b: interleukin 1 β; Ifnal-13: interferon α family, genes 1-13; Ifnb1: interferon 1β, fibroblast; Ifne1: interferon epsilon 1; Lilrb3: leukocyte immunoglobulin-like receptor, subfamily B, with TM and ITIM domains; Il11: interleukin 11; Ptgir: prostaglandin I receptor (IP); Tgfb1: transforming grow factor, beta 1; Cd33: CD33 antigen; Cd37: CD37 antigen; Il4: interleukin 4; Il13: interleukin 13; Irf1: interferon regulatory factor 1; Csf2: colony stimulating factor 2 (granulocyte-macrophage).c) Nucleic acid transcription: Myef2: myelin basic protein expression factor 2, repressor; Gata2: GATA binding protein 2; Mitf: microphtalmia-associated transcription factor; Rnu 7: U7 small nuclear RNA; Phb2: prohibitin 2; Tcf7: transcription factor 7 T-cell specific; Zfp62: zinc finger protein 62; Zfp354a: zinc finger protein 354A. d) Signal transduction: Rab 7: RAB7, member RAS oncogene family; Wnt7a: wingless-related MMTV integration site 7A; Gpr162: G protein-coupled receptor 162; e) Metabolism: Paep: progestagen-associated endometrial protein; Trh: thyrotroponin releasing hormone; Atn1: atrophin 1; Ceacam 9: CEA-related cell adhesion molecule 9; Psg18: pregnancy specific glycoprotein 18. f) Cell metabolism: Cdca3: cell division cycle associated 3; Bax: Bcl2-associated X protein; Gdf9: growth differentiation factor 9. g) Protein metabolism: Usp5: ubiquitin specific peptidase 5 (isopeptidase T); Eno2: enolase 2, gamma neuronal; Zfp61: zink finger protein 61; Axl: AXL receptor tyrosine kinase; Hspa4: heat shock protein 4; P4ha2: procollagen-proline, 2-oxoglutarate 4-dioxigenase (proline 4-hydroxylase), alpha II polypeptide; h) Disease Markers: Cea: carcinoembryonic antigen gene family. i) Non coding susceptibility loci: Lyr: lymphoma resistance; H21: histocompatibility 21; Sle2: systemic lupus erythematosus susceptibility 2; Hlq2: heat loss QTL 2; Idd13: insulin dependent diabetes susceptibility 13; Lgth2: body length 2; Lmr14: leishmaniasis resistance 14; Nktcn 2: natural killer T cell numbers 2; Foc1: follicular center cell lymphoma 1; Tlsr4: thymic lymphoma suppressor region 4; Amp1: Amplitude circadian rhythm 1; Kidq4: kidney weight QTL4; Lbw2: lupus NZB × NZW 2; Ots1: ovarian teratoma susceptibility 1; Chab3: cholesterol absorption 3; Lxw6: lupus BXSB × NZW 6; Aod5: autoimmune ovarian dysgenesis 5; Abhr3: allergen-induced bronchial hyperresponsiveness 3; Cfm1: cystic fibrosis modifier 1; Dsk3: dark skin 3; Scpro 1: stem cell proliferation 1; Akt2: thymoma viral proto-oncogene 2; Bcl3: B-cell leukemia/lymphoma 3; Nhdlq6: non-HDL QTL 6; Bomd4: bone mineral density 4; Pol8: viral polymerase 8; Blmpf2: bleomycin-induced pulmonary fibrosis 2; Itnr2: lentinan responsive 2; Tcpm1: T-cell phenotype modifier 1; Tria 1: T-cell receptor induced activation 1; Lbw8: lupus NZB × NZW 8.

Table 4. Variability of age at death in recombinant strains of mice, B allele for broad variability, and D allele for narrow variability.

Broad	Strains	Median	Mean	SD	Range
	1	519	394	261	803
	5	452	528	317	920
	11	847	750	266	950
	12	774	738	209	907
	16	617	642	191	862
Narrow	6	641	680	115	366
	18	750	742	140	483
	28	775	765	104	331
	29	704	703	114	349

Summary statistics * [44].

This genetic control of phenotype variability could result from environmental effects such as infections or from nutrition and this together with other unknown genes may produce short or long life span. We believe that caloric restriction and or addition of antioxidants to the diets will correct the defective adaptor proteins controlling oxidative stress, the p53 stress response and the telomere length in mice with premature immunosenescence. For this, it will be necessary to study by microarrays the possible differences in expression of these genes in recombinant strains of mice.

Role of the Microenvironment and Niches for Stem Cells During Aging

The distribution of lymphoid cells in autoimmune susceptible and resistant mice was studied by the capacity to trace ⁵¹Cr-labeled lymphoid cells (Yunis et al., unpublished observations). Splenocytes of old NZB mice were distributed in abnormally large numbers in the liver, and in abnormally small numbers in the spleen and lymph nodes, as compared to the distribution of labeled cells from young donors given to young recipients. The cells from old donors did not home to the bone marrow (BM) of old mice, but homed better to the BM of young mice. Cells from young animals given to old animals with significant autoimmunity also were deployed excessively to the liver and poorly to the spleen and marrow, as compared to tagged cells of young animals injected intravenously into old animals. These findings indicate that in NZB mice there is a significant age-related pathology of the normal ecotaxis, which has both cellular and organ-determined components.

One of the most interesting findings was that spleen cells from young NZB donors homed well to lymph nodes of both young and old recipients, whereas spleen cells from old donors homed very poorly to lymph nodes of either young or old recipients [7]. By contrast, cells from young and old CBA/H mice, which are long lived, were able to home to lymph nodes very well. Also, in unpublished experiments we showed that BM cells from young and aging CBA/H and NZB mice were comparable in their capacity to repopulate spleen of irradiated syngeneic mice. More research is needed to test the microenvironment of spleen, lymph nodes and BM during aging. It appears that the microenvironment of the BM during aging is not as affected as that of the spleen or lymph nodes.

Thus, it appears that there are intrinsic and extrinsic defects in the generation of hematopoietic cells in autoimmune susceptible strains and in aging animals. Recently, it has been shown that long-term hematopoietic stem cells from old mice expressed elevated levels of many genes involved in leukemic transformation. The data supports the concept that there is an age-dependent alteration in gene expression at the stem-cell level presage that contributes to age-dependent immune decline in the elderly [47].

New evidence indicates that cells of the connective tissue and blood vessels are part of the microenvironments or niches, for hematopoietic stem cells (HSCs) in adult BM as well as stromal fibroblasts are associated with cancer cells with respect to the self-renewal activity [48]. The most protected niche is the BM. The skin is composed of units in which every hair follicle has a tiny niche of stem cells, responsible for generating a new hair and to generate sebaceous glands and epidermis [49, 50, 51]. Also, the intestine is composed of many units each containing a villous and a crypt, their stem cells are located above the vase of the crypt [52]. In the adult central nervous system, stem cells are in the sub-ventricular zone where they generate glia and neurons [53]. Stem cells depend on their surrounding environment to maintain their functions and proliferation potentials [50]. It is important to mention that whenever SCs exit the niche it must be replenished, perhaps by self-renewal. Leaving the niche is due to changes in the microenvironment or loss of certain important components of the niche. These changes could be different in individual niches. This is important because the administration of stem cells to correct a disease or to prolong lifespan is not sufficient. An example is related to the role of signaling of Wnt and Notch in BM morphogenesis; Notch has pleotropic effects on stem cells and their lineages, which are different in distinct SC [54].

It is noteworthy mentioning that infections, such as HIV, nutrition and other non-genetic factors can influence the microenvironment of the niches. For example, measurable amounts of cytokine responses could vary depending on the genetic background. It is possible that these measurements will be variable in some animals and less variable but decreased in others, which may be genetically controlled. Such studies can be done in RI strains of mice where a gene in chromosome 11 distinguishes strains with short variability at death as compared with other that have wide variability of age at death. It is generally known that aging is characterized by upregulation of genes involved in oxidative stress responses [55-57], suggestive of an increased need to cope with the accumulation of macromolecular abnormalities. The studies of one cell type cannot address the question of the effect of tissues, which can also be altered during aging.

The quiescence of HSCs is controlled in the individual by signaling of receptors-ligands and cell-adhesion molecules [58]. Stem cells are cells with self-renewal capacity and also have the capacity to differentiate into single or multiple lineages [59]. A small subset of HSCs can be isolated using cell phenotype markers (Kit receptor, Sca-1 and Thy-1 but negative for lineage-specific antigens) [60, 61]. They differentiate into different lineages with specific cytokines [62]. During postnatal life, the BM supports both self-renewal and differentiation of HSCs in specialized microenvironmental niches (ecotaxis). Of interest, in fish blood cells are not produced by BM but rather by the kidney. But, in other animals including mammals, BM is found in bone cavities where cells of the niche are composed of fat, stromal cells and other components including blood sinusoids. There is a balance between self-renewal and commitment of stem cells controlled by both cell-intrinsic and external regulatory mechanisms. There is a significant information about the intrinsic molecular properties of stem cells as well as the specific microenvironment in which they reside. Not only the niches have been studied for hematopoiesis, but also for epidermis, intestinal epithelium, nervous system and gonads [49, 50].

The quiescence of stem cells is maintained in part by cyclin-dependent kinase inhibitor such as (Cdkn1a) [58]. Also, shortening of telomere length of chromosomes occurs during every cell division. HSCs have high proliferative potential and telomerase activity to protect the ends of the chromosomes [63]. However, they show telomeric shortening during replicative aging. The gene encoding p21 is repressed directly by c-Myc and HSCs lacking c-Myc overexpresses p21, indicating that the c-Myc-p21 pathway could be important in regulating the switch from resting to active HSCs [59].

Experiments with the ataxia telangiectasia mutated protein (ATM), show that it regulates reconstitutive capacity and a key cell cycle checkpoint in response to DNA damage of HSCs but not the proliferation or differentiation into progenitors [64]. The ATM protein maintains genomic stability by activation of a key cell cycle checkpoint in response to DNA damage, telomeric instability or oxidative stress. Elevated radical oxygen species (ROS) is maintained without telomere dysfunction. Elevated ROS induces upregulation of the cyclin-dependent kinase (C d k) inhibitor p16NK4A, also named Cdkn2a, (maps together with Ifna and Interferon cluster 1-13) (Fig 2) and the retinoblastoma (Rb) gene in ATM negative HSCs. Treatment of cells with antioxidative reagents restores them to normal state. Also, polycomb group ring finger 4 (bmi-1) is essential for self-renewal via the p16NK4a/Rb pathway; antioxidants produce inactivation of p16 and inactivation of Rb that result in restoration of stem cell function. In addition, studies of interactions of stem cells and niche cells with each other will help the understanding of growth of cancer cells; stromal fibroblasts associated with cancer cells are capable of self-renewal. Also, quantitative genetic variation in hematopoietic stem cell and progenitor cell compartment and lifespan were found closely linked at multiple loci in B×D recombinant strain of mice. Genes on chromosome 4 and 7 are involved in the number of LSK cells [65].

The Thymus: A Temporary Niche

Thymic involution and involution of the thymus-dependent system of cells responsible for cell-mediated immunity occur in man as they do in all animals that possess a thymus. Specific thymic alterations related to aging have been described and extensively discussed. It has been known for many years that beginning at the time of sexual maturation, an apparent programmed involution begins in the central lymphoid organs [7]. The interaction between different hormones and the involution has been reviewed in 1976 by Fabris et al. [66, 67]. More recently, this subject has been reviewed and it will not be discussed in this review. Instead, we will summarize recent studies of molecular events occurring in the thymus during aging. In order to understand the basis for age-associated thymic involution, microarray analyses on the thymi from young, middle-aged, and old mice have been recently performed to identify differences in gene expression patterns, that may be attributed to aging (reviewed in ref. 68). For this analysis, a total of 67 mice were used, divided into multiple categories. There were four age groups: 1, 6, 12, and 24 months. Of the 17000 genes on the murine cDNA array, 788 genes demonstrated significant changes in gene expression with progressive thymic aging. Dramatic changes in gene expression were observed between mice aged 1 month and 6 months (107 genes changed; 26 up and 81 down) and those aged 12 months (203 genes changed; 95 up and 108 down) or 24 months (788 genes changed; 418 up and 370 down). The early changes in expression may possibly reflect the known peak of thymic decline between the age of onset of puberty to the midlife period in mice. Genes involved in various biological processes (e.g. cell-cycle progression, transcriptional regulators, maintenance and remodeling of extracellular matrix, protein binding and transport, proteasomal proteins, apoptosis, stress response, inflammation and immune function, growth factors, energy metabolism, and mitochondrial function) and molecular functions (e.g. ATP, DNA and chromatin-binding proteins, RNA and protein binding, and transcriptional regulators) were observed to change with thymus age. On the basis of the fact that the thymus is undergoing involution, one would expect to see changes in genes that are known to be involved with thymus and involution, and indeed such changes were observed. Examples include genes such as LIF, several thymosin family members, trkA, and BDNF [69, 70]. Some of the gene expression changes associated with certain biological and molecular functions observed in our thymic analysis were quite similar to other microarray studies profiling age-associated alterations in liver, kidney, brain, muscle, and fibroblasts [4]. However, the majority of the specific genes found to change with thymic aging were quite distinct from these other organ systems suggesting that there is only a limited degree of overlap in the major categories of genes associated with aging and that age-induced modifications are for the most part species, organ, and tissue-specific. In addition, effects of caloric restriction (CR) on the age-associated changes in thymic gene expression have been examined. CR is currently considered one of the major life extension interventions in primates and rodents, and it has been utilized in numerous microarray studies to examine its effects on organ and cellular aging [71, 72]. Interestingly, aged CR thymi demonstrated a significant reversal in their gene expression profile compared with their AL-fed aged counterparts and revealed a significant profile match with AL-fed young animals. The microarray analysis to uncover specific biological processes at play during thymic aging has provided some excellent gene targets for further examination.

Changes Produced in Neural and Hematopoietic Niches by HIV Infection: Niche Pathology or Defective Ecotaxis

Ontogenic control of the HSC microenvironment during human development [66] seems to be an accepted feature of the aging process. Comparison between fetal liver and adult BM derived stromal cells led to the information that these two microenvironments that support human HSC contain age dependent differences [73]. However, intra-species and ontogenic conservation of stem cell niches is also not an uncommon occurrence [74, 75]. The microenvironments or niches control the self-renewal and regeneration through cell division of neural and hematopoietic progenitor stem cells among others through cell cycle regulation and signaling processes involving transcription factors [76, 81]. Maintenance and differentiation of stem cells following their migration from the niches of “origin” to those niches, which are the sites of “development” suggests the existence of function specific microenvironments [82, 83]. Cytotoxic T lymphocytes also execute their protective action through available or designated niches and have to deal with invading organisms that are thriving in separate niches [84-87]. Thus, niches can be present in both extracellular and intracellular compartments within a living being [78, 79, 82].

The proliferative or maintenance stage in the niches of progenitor stem cells is followed by the migratory stage that brings about cell cycle changes [88]. C-Myc has been reported to regulate the interaction between the self-renewal and differentiation niches of the HSC [59]. C-Myc deficiency causes severe cytopenias arising from impaired differentiation and enforced c-Myc expression leads to loss of self-renewal activity and concomitantly supports differentiation [59].

Reproduction of in vivo niches ex vivo, that is, niche-independent simulation using putative molecular entities of normal physiology or pathology has been reported [89, 90] but is unclear and is fraught with deficiencies that support the multitude of functions that occur in vivo. Osteopontin is a stem cell niche component that maintains the size of the stem cell pool [91, 92] and therefore may reduce the risk of uncontrolled proliferation. Angiopoietin-1 is a regulator of cell cycle in the stem cell niche [93, 94]. Notch and Wingless (Wnt) are some of the signaling pathways that occur in the hematopoietic microenvironment to regulate self-renewal of the stem cells [95, 96]. Ontogenic differences in the human development include the Wnt signaling regulation in the human fetal liver stem cell niche compared to the control of Notch signaling in the adult bone marrow microenvironment [73]. The niches, in addition to molecular entities, also include cellular components such as the endothelial cells present in a neural stem cell niche secrete soluble factors that promote self-renewal and inhibit their differentiation [97]. Similarly, the AFT024 stromal cell line confers or delivers the properties of an in vivo stem cell niche [98, 99].

HIV infection of stem cell microenvironments or niches causes hematopoietic inhibition and hence cytopenias [100-105]. Hematopoietic CD34+ progenitor stem cells are reported to be resistant to HIV-1 infection, in vitro, or in vivo [106, 107]. Those cells that experienced the indirect effects of HIV-1 infection exhibit inhibition of their multilineage hematopoiesis as determined by colony forming activity ex vivo [106, 108-110]. It is reported that the hematopoietic stem cell microenvironment is damaged due to the indirect effects of HIV-1 infection of the thymocytes on the CD34+ progenitor stem cells but in a reversible manner, in the human fetal Thymus/Liver conjoint hematopoietic organ of the transplanted chimeric severe combined immunodeficiency mouse (SCID-hu) model system [108, 110]. It is therefore highly plausible that this implanted human organ in the SCID-hu mouse, which serves as a niche, not only for thymocyte expansion but also supports hematopoiesis, suffers niche dysfunction due to HIV-1 infection. Continued presence of the CD34+ progenitor stem cells in the infected niche seem to suffer due to exacerbation resulting from persistent virus mediated niche disruption via infection of thymocytes and consequent interactions and signaling network of the hubs. Thus cellular and molecular niches and involvement of their hubs might be at play in the pathology of this infected microenvironment within the human hematopoietic organ of SCID-hu mouse. Thus, this model is useful in understanding the cellular and molecular events that occur in human stem cell niches.

Mapping of Genes of Lifespan, Using RI, Stem Cells and Studies of Niches

In Fig 2 we placed candidate genes and genetic markers (life span genes) that have been mapped to the regions summarized in Table 2. The cytokine genes cluster in a region of chromosome 11 includes: growth differentiation factor 9, colony stimulating factor 2, IL-3 IL-5, IL-13, IL-4 [111-113]. It is of interest that recombinant strains of mice showed the IL-5 gene to be associated with lifespan. But the cytokines IL-3, IL-4, IL-13 and Irf1 (Interferon regulatory factor 1) have also been mapped in the same region. This is of interest because cytokines made by TH2 cells are spaced widely, but are kept primed for action by one master control region on the same chromosome [114]. It is suggested that there is an interaction between chromosome 11 and the initiation of activation of interferon-gamma on a different chromosome. But that this relationship falls apart when the T cells have differentiated into helper cells.

The regulation of the molecules involved in niches is more complex, since some of them could be upregulated while others are downregulated. Of interest, two genes identified thus far in the RI or in the studies of stems cells that are also involved in niches, are the Akt2 on chromosome 7, which is near Bcl3, and the other one is p16NK-4 which is on chromosome 4, near the IFN-gamma and -beta [115].

We have reanalyzed the data of genes of variability of age at death and discovered a new group of genes in chromosome 11 - Hba, Hba-x, Lamrl-rs5 and Tel-rs1. The chromosomal location of these genes is depicted in Figure 3. Other candidate genes involved were mapped to different chromosomes or regions involved in niches and include: bcl-2 and Akt3 on chromosome 1 [116, 117], bmi-1 [118] on chromosome 2, ATM on chromosome 9 [119], ROS-1 and LSK (also named Matk, (megakaryocyte-associated tyrosine kinase) on chromosome 10 [120, 121], p53 on chromosome 11 [122, 123], Akt-1 on chromosome 12 [124], Rb1 on chromosome 14 [125], c-Myc on chromosome 15 [126] and p21 on chromosome 17 [127]. Other genes mentioned were p66 (Shc1), a transforming factor on chromosome 3 [128], IL-2 and PLC-gamma on chromosomes 3 and 2 respectively [129, 130].

Figure 3.

Genes on chromosome 11 associated and life span variability in mice. Variability genes and the cytokine cluster. Glns-ps1: Glutamine sythetase pseudogene 1; Hba: hemoglobin alpha chain complex; Hba-x: hemoglobin X, alpha-like embryonic chain in Hba complex; Lamr1-rs5: laminin receptor 1 (ribosomal protein SA), related sequence 5; Tel-rs1: telomere, related sequence 1; Rcvrn: Recoverin.

Recently it was reported that a significant number of proteins are interactive with one another (interactomes) [131]. The interactome is composed of hubs that are interconnected via specific proteins [132]. Hubs are composed of highly interactive proteins. This concept of interactive proteins is reminiscent of the existence of niches. Is it possible that the anatomical concept of niches would be related to hubs and that protein components of them are needed for the integrity of hubs. Since there are many genetic interactions, particularly in niches, it would be important to know the relationship among such protein interactions with the proteins involved in the microenvironment of niches. It needs to be determined what are the proteins involved in the maintenance of the health of niches and ultimately which interactomes are involved in the process of aging resulting in the determination of life span.

Influence of Nutrition on the Decline of Immunity with Aging in Mice

Although the regulation of the immune response is predominantly under genetic control, environmental factors such as nutrition are also influential. Earlier studies have demonstrated that in the autoimmune-prone NZB mouse strain diets low in fat and high in protein and fiber content produced a delayed development of autoimmunity and was associated with prolonged life span in both males and females [133]. However, restriction of protein intake alone, while conferring beneficial influences on T-cell functions, did not significantly suppress the occurrence of autoimmune disease or prolong the life span or NZB mice. In contrast, B/W mice fed a normal diet in restricted amount (12 cal/day) lived at least twice as long as mice fed a normal diet (24 cal/day) [134]. This dramatic influence of nutrition was accompanied by prolonged maintenance of T cell-mediated functions, inhibition of the development of spontaneously active suppressor cells, and maintenance of inducible suppressor cells. Dietary restriction also inhibited immuno-complex-dependent renal injury and anti-DNA antibody production in B/W mice as well as preventing the development of circulating immune complexes. Furthermore, the high caloric diet was accompanied with cardiovascular diseases. In C3H mice, which develop spontaneous breast tumors, two methods, alternate-day feeding and daily caloric restriction, have been successful in the prevention or delay of tumor development [134]. In an unpublished study using the same alternate-day feeding schedule versus the ad libitum feeding in CBA/H (a long lived strain of mice), we have seen 70% survival of the experimental mice and only 42% survival of the control mice at 960 days of age. The difference in survival at 1,080 days of age was 40% and 5%, respectively, for the experimental and control mice. The final number of days was 1,260 for the mice with the diet restriction and 1,099 for those fed ad libitum.

MHC Genes in Aging

Studies of genetics of aging using mouse models that differ only in the MHC region demonstrated MHC association, but in recombinant strains of mice, non-MHC genes were found to be involved. Furthermore, it was of interest that MHC congenic strains of mice showed interaction between MHC alleles and production of IL-4 and low production of IFN-γ by NK and NKT cells [135]. It has been known for a long time that mice susceptible to autoimmunity showed deficient T cell functions [136]. Furthermore, studies of human aging using MHC markers have described associations with MHC alleles or haplotypes that are known to be markers for autoimmune disease, such as the HLA-DRB1*0301 and the haplotypes A1, B8, DR3, which were described in some ethnic groups [137]. Also, the TNFα -308 polymorphism is known to be a marker of such haplotypes [138, 139]. Our observations are relevant for ethnically matched aged and young individuals; for example, Caucasian markers such as TNF2 are associated with genetics of aging in Mexicans. In this regard, only 3 out of 12 elderly individuals carrying this mutation were of the B8, DR3 extended haplotype, suggesting that in Mexicans the association is not due to linkage disequilibrium (unpublished data). The TNF-α promoter polymorphisms have been associated with the pathogenesis of autoimmune disorders [138, 139]. In Mexicans, TNF2 has been associated with severity of RA [140]. The risk of developing autoimmune diseases such as systemic lupus or RA varies between populations [141]. Other studies suggest that TNF2 polymorphism may be a susceptible factor to bronchoalveolar infections in old hospitalized patients from Caucasian origin [142]. TNF2 is a marker for autoimmunity but we suggest that it could be a marker that promotes efficient immune inflammatory response against pathogens in older people with specific genetic background. However, this hypothesis does not support previous findings in Caucasians [143]. Since case control studies should take into account the possibility of genetic stratification, such studies should include information about the contribution of genes in both the elderly and young groups, as we showed in our studies. Mexican mestizos, who constitute about 90% of the country’s population (National Population Census 1992), can be defined as the descendants of the mixture of the autochthonous inhabitants with other genetic groups, mainly Spaniards. Genetic admixture between populations with high prevalence of RA, such as Amerindians, and populations with lower prevalence (Caucasians) has occurred, demonstrating the role of genetic admixed background and environmental factors that contributed to disease risk. Current approaches to mapping disease susceptibility genes within MHC, such as the analysis of DNA conserved blocks [144] or SNPs, and the methods to eliminate confounding have been important in defining the population stratification and genetic distance between admixed and ancestral populations [145].

The MHC class I and class II gene frequencies did not reveal significant associations with aging. However, the frequency of distribution of the DRB1*DQB1 blocks were used as a marker of genetic admixture and also as a measure of population diversity. There was higher diversity in the young group than in the elderly group, a proportion of 48.3% of Caucasian and Amerindian component of 27.5% compared with 40.4% and 30.8%, respectively, in the young group (unpublished observations). Although, these results showed comparable Amerindian admixture, there is a non-significant higher Caucasian component in the elderly group. In addition, it was remarkable that only 4 of the 120 DRB1*DQB1* block of the elderly group showed individual blocks with a frequency of less than 1% each, as compared to 45 of the young groups (p=0.0001). This finding suggests that multiple class II genes may be associated, reflecting a decreased diversity in aged individuals, as reported before [146, 147].

As described above (Table 1), the HLA associations with elderly individuals showed significant ethnic and nationality effects. In some cases a combination of HLA class II alleles was involved. Also, other genetic polymorphisms have been found to be important. For example, the potential relevance of IL-10 promoter region polymorphisms in longevity has been suggested [148]. In some of the studies, inflammatory cytokines were increased as well as genetic polymorphisms of cytokines, and it has been suggested that the infections in elderly individuals could explain such findings.

Our findings suggested that genetic admixture with Caucasian ethnic groups might promote the incorporation of genes associated with autoimmune disorders and that these genes could be associated with immunity against pathogens that frequently affect aged people, and are important for protection from infection. The mechanism by which the MHC genes participate in the longevity can be partially explained by the function of the product of these genes in the regulation of the specific immune response, although, the role of MHC genes in the susceptibility to infections and the potential connection with the innate immunity is not well understood. This study suggests that genes located within the MHC cluster, specifically those, which are close to TNF region, could influence longevity in Mexican individuals. Also, genetic diversity studies are important for the future attempts to understand the genetics of longevity in admixed populations. Our findings support the hypothesis that specific genes can have different roles in different stages of life and that they could be involved both in autoimmunity and longevity.

In summary, studies of genetics of longevity has yielded complex results, this is due in part to the need to assess the degree of genetic admixture and to determine the degree of genetic diversity of the elderly and control individuals. Good candidates for the studies of genetics of cytokines, related to serum secretion for determination of the role of pro-inflammatory cytokines, are TNF-α, IFN-γ and IL-6, as compared to cytokines involved in regulatory immune function such as IL-10. Our data, showing increased incidence of TNF-α mutation associated with increased levels of production of this cytokine in elderly Mexicans that have comparable Amerindian and Caucasian genes, support the above prediction. Our results are consistent with the concept that phenotypic interactions of genes within the MHC are operative in the elderly and that genes associated with autoimmunity may also be markers of longevity in populations with admixture, suggesting the possibility that one group of non-MHC genes can interact with the same MHC gene (i.e. TNF2) to produce autoimmunity, while other non-MHC genes could be associated with the same MHC gene (i.e. TNF2) with longevity. Associations with MHC and cytokine genes in different ethnic groups are shown in Table 4.

Discussion

In mammals, the mechanisms that regulate stress responses and increased life span involving reactive oxygen species (ROS) are poorly understood. Mice that are genetically deficient for the p66shc gene are less sensitive to the toxic effects of ROS and live 30% longer than their wild-type littermates. Ablation of p66shc enhances cellular resistance to apoptosis induced by H₂O₂. A serine-phosphorylation defective mutant of p66shc cannot restore the normal stress response in p66shc-/- cells. The p53 and p21 stress response is also impaired in p66-/- cells and p66-/- mice have a 30% increase of life span [149]. We have described the early immune abnormalities in the B10.AKM mice [40] similar to those reported by others in p53 knock-out mice with the accumulation of memory T cells and increased production of IL-4 by NKT cells [135]. Our results suggested that young B10.AKM mice have a primary immune abnormality that ultimately results in an early senescence phenotype.

Our studies in another murine experimental model (B×D recombinant mice) identified several potential candidate genes in short- and long-lived strains. Among them were in chromosome 2, histone deacetylase (HD), in chromosome 6 the Rho, and in chromosome 7, p450 gene (CYP2A) and apoptosis related BAX genes. In this regard, the human HD showed its involvement in cell cycle progression and activation in one report [150]. In addition, the inhibition of HDC was associated with early abrogation of IFN-γ production by Th1 lymphocytes and with p53 mRNA downregulation. Thus, HDC participates in IFN-γ gene regulation and cell survival. Interestingly, p53 deficient (p53-/-) mice have accelerated aging of the immune system (increased accumulation of memory cells, stronger Th2 cytokine profile). This has been suggested to be due to dysregulation of cell cycle, DNA repair and apoptosis in the lymphocyte population. In normally dividing cells, p53 is highly unstable with a short half-life. DNA damage leads to accumulation of p53 and transcription of target genes such as Bax and p21. It can also be hypothesized that a relationship between increased oxidative stress and transcriptional activation of Bax could lead to increased loss of naïve cells and accumulation of apoptosis-resistant memory cells in these animals.

CYP2A subfamily includes enzymes that catalyze oxidation of several compounds of chemical or toxicological interest. Little is known about associations between variant CYP2A alleles and diseases. Thus, we have the opportunity to study the role of ROS in cell death and life span using an experimental model with one genetic factor in chromosome 7. Also, we hypothesize that HDC may be defective in the DBA/2 background, since the contribution of recombinant short-lived BxD strains carry the D marker. Therefore, it is of importance to investigate the role of oxygen in mice and to investigate the possible role of genes involved in the downregulation of p53.

We believe that these defects could be corrected with low caloric diets or diets rich in antioxidants. We will test this hypothesis in congenic strains and recombinant strains of mice with short and long survival. In addition, we have identified candidate genes on chromosome 2, chromosome 6 and chromosome 7 which could provide additional information on molecules involved in oxidation such as a gene in chromosome 7 (CYP2A), and a gene on chromosome 2 involved in apoptosis. In this regard, two candidate genes mapped to different regions of chromosome 11 were Hsp-a4, important in longer life span in the cM 28, and tel-rs1m in cM 16 associated with variability at the age of death. A possible mechanism may involve the telomere lengths between these strains since telomere length was highly variable among genotypically identical siblings [151]. We believe that modifying caloric restriction and or addition of antioxidants to the diets will correct the defective adaptor proteins controlling the oxidative stress and the p53 stress response, as well as the telomere length in mice with premature immunosenescence. For this purpose, it will be necessary to study the possible differences between these proteins in recombinant strains of mice by microarrays. However, in the case of viral infection such as HIV, the defects of niches cannot be corrected by replacing stem cells without treating the viremia and defects produced in the niche of hematopoietic cells because the continued presence of the CD34+ stem cells in the infected niche seems to maintain the disruption by the infection of thymocytes [100].

Concluding Remarks: Perspectives for Future Research in Aging

Pathways leading to death are affected by multiple genetic and environmental factors. We have been able to partially define two such pathways. The first involves increased sensitivity to infections due to early onset of immune senescence in a congenic mouse strain, and the second defines decreased life span in recombinant strains of mice that develop lymphomas due to possible deregulation of apoptotic pathways and non-limitation of proliferative capacity of cells (that is a component of cellular senescence). In the future, the use of microarrays will be beneficial in analysis of differentially expressed genes between three recombinant mouse inbred (RI) strains (2, 14 and 32) that are short-lived and four (9, 15, 19 and 24) that are long-lived. Several thousand genes can be surveyed in the spleen, and we have observed that a small proportion displayed greater than three-fold increase or decrease in expression levels between them. Such analyses combined with the results of studies in the mapping of the genes of long lifespan could give important results of proteins that are unique or part of niches associated with survival.

We have recently completed DNA oligonucleotide microarray studies of differentially expressed genes between two recombinant inbred strains, strain 2 (short-lived) and 15 (long-lived). Of the 11,406 genes surveyed in the spleen, only 109 (0.9%) displayed greater than three-fold increase or decrease in expression levels between the two strains. This frequency is in agreement with a differential display analysis of gene expression and a study involving caloric restriction gene profiling in mice. While a variety of genes were differentially expressed between the two strains, the most significant difference observed was in the expression of a gene that has high homology to the Bc12-like gene Boo/Diva/Bcl-2l11. This sequence might represent a new member of the BH-domain family of proteins involved with mediating apoptosis and involved in proliferative disorders. Interestingly, in a separate system where we have been studying human immune deficiency disorders, we have observed upregulation of survival genes associated with abnormal B cell proliferation. It is proposed that mice that are short-lived may develop lymphomas following upregulation of a similar BH-family survival gene that inhibits apoptosis by inhibiting caspase activity. The lack of regulated expression of this gene, may be involved in lymphomagenesis in mice.

Disruption of the normal molecular processes within the niches such as those due to infection can result in abnormal functioning of the systems that are affected and lead to pathological consequences within the niche, or ecotaxic defects or failure, which is a disruption from the normal cell migration to physiological distribution. Abnormal transition or failure of ecotaxis including a premature, disruptive, delay, or failure can result in disease pathology. HIV infection of stem cell microenvironments or niches causes hematopoietic abnormalities and associated cytopenias [100-105]. Therefore, alterations of niches produced by environmental factors including infections should be taken into account when considering how important it is to correct defects produced in the niches together with those of the stem cells. Such alterations explain why the autoimmune susceptible strains of mice have abnormalities of homing of transplanted cells as reported before [152].

Our observations presented in this report suggested that the age at death is variable in the majority of the BxD recombinant strains of mice; this variability being controlled by a gene in chromosome 11. This gene confirmed by experiments in C57/B/L/6 was significant and produced age at death with higher variability than that of DBA/2 [46]. Genes affecting variability in phenotype expression of genetically identical populations have been suggested previously [46]. It is possible that these genes invoke sensitivity to environmental factors. In spite of the difficulties of cloning of this gene, it is reasonable to use experimental mice to correct the variability of age at death by calorie restriction or antioxidant administration to strains with extreme variability at age at death (strains 1, 5,11,12 and 16) compared to strains with short variability at death (strains 6, 18, 28 and 29). In such studies comparisons of not only the role of caloric restriction or treatment with antioxidants in the duration of life, but also correlation with functions of stem cells and other cell types are essential.

The use of RI strains is limited by the number of genomes used in the generation of the strains as well as the number of individual strains generated for given genomes. The probability of mapping accurately such genetic regions or individual genes is proportional to those variables. Also, such discoveries cannot be generalized for the animal or human population at large. Nevertheless, the identification of genes, proteins or niches of mortality or immortality in these experimental models should be necessary in designing strategies to determine important genes or proteins associated with long or short life in outbreed populations. But, ultimately it would require the use of multiple genetic and environmental manipulations to correct the defects that occur during aging. In this regard, the interacting proteins within hubs and their interconnections should be compared with those described in relation to the genes and proteins associated with embryonic stem cells, and stem cells and their niches associated with life span as summarized in this report. Therefore, since the genetics of life span and aging are complex, we suggest that it will be necessary in the future to categorize the genes and proteins of hubs and niches into groups that need to be investigated in relation to aging.

Although we did not discuss it in this review, aging is even more complex than the mechanisms controlled by genetics or the environment, because it is possible that embryogenesis, development which includes organs, niche development and germ cells result from genetic programs of patterns of DNA methylation in multiple cell types. This is the science of epigenetics, the study of epigenotypes, which provide the basis for switching of gene activities and the maintenance of stable phenotypes. A challenge is to determine the extent of epigenetic defects during aging, a most difficult task because such defects will be heterogeneous or random in any one tissue [153].

Table 5. Reported associations with MHC and cytokine genes in different ethnic groups.

Gene	Finding	Population	Ref
A1, B8, DR3	Increased in elderly men, decreased T cell function and decreased survival in women.	Caucasian, NAm	[154]
A1, B8, Cw7, DR3	Increased in Nonagenarian men	Caucasian, Ireland	[155]
DRB1*1401, DQB1*0503, DQA1*0101	Increased in Centenarians	Oriental, Japan	[146]
B16	Increased in elderly	Greek	[147]
DR7	Increased in elderly	Greek	[147]
B15	Decreased in elderly	Greek	[147]
DR4	Decreased in elderly	Greek	[147]
DR11	Increased in elderly women	Caucasian, French	[156]
DR7	Increased in elderly men	Caucasian, French	[156]
DR13	Increased in elderly men and women	Caucasian, French	[156]
A31, B7, Cw7, DQ1	Increased in elderly	Caucasian, Italy	[157]
A1, B8, DR3, TNF	No association	Caucasian, Ireland	[143]
DRB1*15	Increased in Centenarians	Sardinian	[148]
HSP70-1 -110A>C	Association of heterozygosity in aged twins	Caucasian, Danish	[158]
TNF -308 A	Increased in elderly with bronchoalveolar infections	Caucasian, Italy	[142]
IL-6-174 C	Increased in Centenarian men	Sardinian	[159]
IFN-γ+874A	Increased in centenarian women	Sardinian	[159]
IL-6-174 G/G	Homozygous genotype G/G increased in elderly	Caucasian, Denmark	[160]
IL-10-1082G/A,-819C/C,-592C/C	Increased in elderly	Caucasian, Bulgaria	[161]
IL-10-1082G	Increased in centenarian men	Caucasian, Italy	[162]
TGF-β1 +915G/C	Decreased in centenarians	Caucasian, Italy	[163]

Abbreviations

BM

bone marrow

ESC

embryonic stem cells

HSC

hematopoietic stem cells

RA

rheumatoid arthritis

ROS

reactive oxygen species

RI

recombinant inbreed

Niches

microenvironment of stem cells