PMCCPMCCPMCC

Search tips
Search criteria 

Advanced


 
Formats
Article sections
Authors
Related links
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Immunity. Author manuscript; available in PMC 2008 March 12.
Published in final edited form as:
PMCID: PMC2266981
NIHMSID: NIHMS41133
Copyright notice and Disclaimer
Aging of the Immune System: How Much Can the Adaptive Immune System Adapt?
Nan-ping Weng1*
1Laboratory of Immunology, National Institute on Aging, National Institutes of Health, 5600 Nathan Shock Drive, Baltimore, Maryland 21224
* Correspondence: wengn/at/mail.nih.gov
Small right arrow pointing to: The publisher's final edited version of this article is available at Immunity
Small right arrow pointing to: See other articles in PMC that cite the published article.
Summary
The competency of the adaptive immune function decreases with age, primarily because of the decline in production of naïve lymphocytes in the bone marrow and thymus as well as the expansion of incompetent memory lymphocytes. Here I discuss the recent progress on age-associated changes in lymphocytes and their effect on the adaptive immune system.
The adaptive immune system is equipped with two key weapons: enormously diverse antigen-recognizing lymphocyte populations (naïve lymphocytes) and very long-lived antigen-experienced lymphocytes (memory lymphocytes). The former ensures a specific response to any potential challenges from the universe of foreign antigens, and the latter guarantees a more rapid and robust response to subsequent encounters of a previously experienced antigen. Despite a continuous decline in the generation of naïve lymphocytes and an imperfect maintenance of memory lymphocytes after puberty, the adaptive immune system adjusts to these age-associated changes and protects the body successfully against most pathogens for nearly all of adult life. Only in the late phase of life does the progressive decline of immune function create vulnerability, with a resultant increase in morbidity and mortality due to infection in the elderly.
Immunosenescence describes a state of profound age-associated changes in the immune system; these changes are manifested by the overall decline of antigen-specific immunity. At the cellular level, the most prominent features of immunosenescence include a substantial decrease in the number of naïve lymphocytes, as a result of a reduction in thymic output of T cells (Linton and Dorshkind, 2004), as well as fewer bone marrow early progenitor B cells (Allman and Miller, 2005) and an accumulation of oligoclonally expanded and functionally incompetent memory lymphocytes (Figure 1). As the direct consequence of these changes, the diversity of the antigen-recognition repertoire is markedly decreased with age. Based on the analysis of TCR Vβ chain usage in human peripheral T cells, the antigen-recognition repertoire of T cells decreases from approximately 108 in young adults to 106 in the elderly (Goronzy and Weyand, 2005). The causes of age-associated decline in the generation of naïve cells are likely to be multifactorial and to involve changes in growth factors and/or hormones, hematopoietic progenitor cells, and their surrounding microenvironment. The accumulation of memory cells with age may reflect an adaptive response to the decline of production of naïve lymphocytes through homeostatic expansion, as well as the cumulative effect of past and persistent viral infections (Pawelec et al., 2005).
Figure 1
Figure 1
Age-Associated Changes of Lymphocytes at the Cellular Level
Functionally, the decline of adaptive immunity with age can be attributed to impairment at the systemic and microenvironmental levels in lymphoid and non-lymphoid organs involving multiple types of cells other than lymphocytes (extrinsic defects) and to the specific impairment in the function of lymphocytes (intrinsic defects). The age-associated extrinsic changes often affect multiple systems, e.g., the neuroendocrine system (Glaser and Kiecolt-Glaser, 2005) and even the innate immune system (Solana et al., 2006 [this issue]) influence the adaptive immune function; therefore, cause and effect are not always clear cut. Conversely, the age-associated defects of lymphocytes are better defined. Here, I shall summarize recent progress in understanding age-associated changes in lymphocytes at three levels, (1) cell, (2) transcription, and (3) telomere, and discuss the effect of these changes on adaptive immunity.
Naïve CD4+ T cells isolated from older humans and mice display decreased in vitro responsiveness to T cell receptor stimulation and altered profiles of cytokine secretion when they are compared to naïve CD4+ T cells isolated from young hosts. In parallel, the helper function of naïve CD4+ T cells for antibody production by B cells is also decreased (Swain et al., 2005). The decline in naïve CD4+ T cell function with age may in part be attributable to altered events in proximal signaling pathways; such changes may include delayed relocalization of signaling proteins to the immune synapse and decreased fluidity of lipid rafts with high levels of cholesterol in response to activation through the T cell receptor (TCR) (Sadighi Akha and Miller, 2005). These age-related defects in naïve CD4+ T cells are due to the chronologic age of naïve CD4+ T cells rather than the chronologic age of the individual (Swain et al., 2005). This finding suggests that long-term maintenance of naïve CD4+ T cells through homeostatic cytokines may be detrimental to their function. Indeed, naïve CD4+ T cells that have undergone homeostatic cell divisions proliferate less and produce less IL-2 in response to antigen stimulation than do naïve CD4+ T cells that have not undergone previous homeostatic division (Swain et al., 2005). However, the mechanism underlying homeostasis-associated dysfunction of naïve CD4+ T cells is not known. Whether homeostatic expansion reduces the replicative capacity of naïve CD4+ T cells remains to be determined. Newly generated naïve CD4+ T cells from old mice exhibit robust proliferation, IL-2 secretion, and helper functions in response to antigen ex vivo and in vivo. This raises the possibility that thymic regeneration and the production of newly generated naïve T cells may be a solution for reducing age-associated defects of immune function (Taub and Longo, 2005).
In contrast to naïve cells, memory CD4+ T cells are long-lived, maintained by homeostatic cytokines, and relatively competent with age. Isolated memory CD4+ T cells from healthy elderly humans and old mice are normal in antigen-induced proliferation in vitro (Kovaiou et al., 2005). Furthermore, memory CD4+ T cells that were generated at a young age respond well to antigens over time, whereas memory CD4+ T cells derived in old age respond poorly (Haynes et al., 2005). These findings suggest that age-associated defects in memory CD4+ T cells may stem from the defects of aged naïve CD4+ T cells that have reduced diversity and proliferative capacity. However, changes of the composition of memory CD4+ T cell subsets such as central and effector memory cells with age have also been implicated for the impaired immune response to viral infections such as the influenza virus and vaccines (Kang et al., 2004).
One of the hallmarks of age-associated changes in the human immune system is the accumulation of CD28− CD8+ T cells. CD28− CD8+ T cells are absent in the newborn and become the majority (80%–90%) of circulating CD8+ T cells in the elderly. The accumulation of CD28− CD8+ T cells was also found in viral infections such as cytomegalovirus (Almanzar et al., 2005), so CD28− CD8+ T cells may be derived from CD28+ CD8+ T cells after repeated antigenic stimulation. Functionally, CD28− CD8+ T cells have a reduced proliferative response to TCR crosslinking but exhibit normal or even enhanced cytotoxic capacity and are resistant to apoptosis (Azuma et al., 1993).
Clonal expansion of CD8+ T cells, especially those CD28− CD8+ T cells, presents another prominent age-associated change. As a result of the reduction of naïve CD8+ T cell-output, some degree of oligoclonal expansion of CD8+ T cells with age is commonly observed in the healthy elderly (Messaoudi et al., 2004). Such changes may reflect a compensatory mechanism to control latent viral infections or to fill available T cell space. When clonal expansion reaches a critical level, the diversity of the T cell repertoire is reduced, and the ability of immune protection to new infection is compromised. In fact, clonal expansion of CD28− CD8+ T cells appears to be directly responsible for increased infections and a failed response to vaccines in the elderly (Almanzar et al., 2005).
The numbers of B cells in the periphery decrease in old humans (Franceschi and Cossarizza, 1995) but appear to be normal in old mice (Johnson and Cambier, 2004). As a consequence of decreased generation of early progenitor B cells, the output of new naïve B cells decreases in old mice (Allman and Miller, 2005), and consequently antigen-experienced B cells are expanded. This causes a reduced antigen-recognition repertoire of B cells in both old humans and old mice (Johnson and Cambier, 2004). At the cellular level, alteration in immunoglobulin generation (through class switch) in B cells is observed in aged mice and humans (Frasca et al., 2005), which may also contribute to the decline of the quality of humoral response in the elderly. In addition, age is associated with the incidence of B cell malignancy in older adults with oligoclonally expanded B cells.
Global gene-expression profiles have been analyzed in CD4+ T cells from young and old mice (Mo et al., 2003) and in human memory CD8+ T cells from CD28+ and CD28− CD8+ subsets (Fann et al., 2005) (Figure 2). Based on their known functions, the genes whose expression is altered in “aged” lymphocytes can be divided into four groups: (1) cell-surface receptors, (2) cytokines and their receptors, (3) effector molecules, and (4) transcriptional regulators.
Figure 2
Figure 2
Age-Associated Changes of Lymphocytes at the Molecular Level
Alteration of cell-surface receptor expression, for example loss of costimulatory receptor CD28 expression in CD8+ T cells, is one of the most consistent age-related changes in lymphocytes. One striking finding is the gain of expression of a variety of stimulatory NK cell receptors in CD28− CD8+ memory T cells. These NK cell receptors include (1) KIR2DL2 (NKAT6), KIR2DS2, and NCR1 of the immunoglobulin-like NK cell receptor family, (2) KLRC3, KLRC4, KLRD1 (CD94), KLRF1, KLRG1, and KLRK1 (NKG2D) of the C-type lectin-like NK cell receptor family, and (3) CD16 and CD244. The gain of NK receptor expression in CD28− CD8+ T cells may facilitate their effector functions as compensation for impaired proliferation (Tarazona et al., 2000).
Elevated expression of chemokines and cytokine receptors was also found in both human CD28− CD8+ T cells and CD4+ T cells from aged mice. Chemokine receptors (CX3CR1, CCRL1) and chemokine-like receptor 1 (CMKLR1) are substantially more highly expressed in human CD28− CD8+ memory T cells compared with their CD28+ counterparts. CCR1, CCR2, CCR4, CCR5, CCR6, CCR8, and CXCR2 to CXCR5 are highly expressed in CD4+ T cells from old mice as opposed to young mice. Reduced expression of interleukin-7 receptor and IL-12 receptor β2 was found in CD28− CD8+ memory T cells. CD28− CD8+ memory T cells expressed higher amounts of IL-13, chemokine (CCL4), and effector proteins such as perforin, granzyme B, and granzyme H (GAMB and GZMH) than their CD28+ counterparts. Finally, elevated expression of transcription factors in CD28− CD8+ memory T cells was also identified; such transcription factors included, T-bet (encoded by TBX21), which functions in initiating T helper 1 lineage development and immunoglobulin class switching, and eomesodermin (EOMES), which induces production of interferon-γ, perforin, and granzyme B in CD8+ T cells. On the other hand, MYC, an important regulator of cell proliferation, differentiation, and apoptosis, was downregulated in CD28− CD8+ memory T cells.
Global gene expression analyses between human CD28− and CD28+ CD8+ memory T cells and mouse young and old CD4+ T cells provide a first glimpse of age-associated changes in gene transcription. Extending microarray analysis and combining this approach with proteomic analysis of T and B cell subsets from young and old donors will offer unprecedented power in elucidating age-associated changes in lymphocytes.
Adaptive immune responses depend on the ability of lymphocytes to undergo cell divisions in response to antigenic challenge. Lymphocytes are thought to have a finite replicative lifespan. Telomere length may act as the ultimate limit for the number of divisions that a human lymphocyte can undergo.
Human telomeres are 10- to 15-kb-long tandem hexanucleotide (TTAGGG)n repeats, and their binding proteins are located at the ends of chromosomes. As a result of incomplete replication of chromosomal termini, telomeres shorten with each cell division. Because telomeres are essential for maintaining chromosomal integrity, cells with critically shortened telomeres cease division (senescence) and are prone to apoptosis. Thus, the length of telomeres serves not only as a record of cell division history but also as a limit for the number of divisions that a cell can undergo. Telomeres are synthesized by a specialized enzyme, telomerase, which compensates for the loss of telomere length from cell divisions. In germ line cells and cancer cells, telomere length is maintained by telomerase. However, most normal somatic cells express little or no telomerase and undergo progressive telomere shortening with cell division. Thus, telomerase preserves replicative lifespan by maintaining telomere length.
Loss of telomeres in lymphocytes in vivo was first demonstrated during CD4+ differentiation from naïve to memory cells (Weng et al., 1995). Naïve CD4+ T cells have longer telomeres than memory CD4+ T cells, and the difference in telomere length between them may reflect the number of cell divisions memory cells have undergone in vivo. Subsequently, telomere shortening was also found to occur in the transition from naïve to memory CD8+ T cells (Rufer et al., 1999). Furthermore, aged CD28− CD8+ T cells had shorter telomeres than their young CD28+ counterparts (Monteiro et al., 1996). Progressive telomere shortening was found in long-term cultured T cells, and markedly shortened telomeres were found in long-term cultured T cells that had reached replicative senescence (Weng et al., 1995; Effros and Pawelec, 1997). Naïve CD4+ T cells that possess longer initial telomeres undergo a greater number of cell divisions than do memory cells before reaching senescence in vitro (Weng et al., 1995). Consistent with the in vitro findings, cross-sectional analysis of telomere length shows a decrease of telomere length in CD4+ and CD8+ T cells with age (Rufer et al., 1999; Son et al., 2000), suggesting that telomere-length-controlled replicative lifespan may play a role in the age-associated decline of T cell functions.
Unlike other somatic cells, telomerase is activation regulated in lymphocytes. Resting T cells have little to no telomerase activity, but substantial activity is induced upon stimulation via cross-linking TCR, by pharmacologic reagents, or by homeostatic cytokines (Hodes et al., 2002). This mode of regulated telomerase expression suggests that telomerase may compensate for telomere loss in activation-induced T cell proliferation. Consistent with this hypothesis, telomerase activity correlates with telomere-length changes in long-term cultures of T cells. When telomerase activity is high, after initial T cell stimulation, telomere length is maintained in robustly dividing T cells. However, when telomerase activity is less effectively induced in later rounds of stimulation, telomere shortening becomes obvious in cultured T cells. Furthermore, introduction of the telomerase catalytic component (hTERT) in T cells results in enhanced expression of telomerase activity and increased proliferative capacity of T cells in vitro (Hooijberg et al., 2000).
Like T cells, B cells also show telomere shortening with age but at a relatively slower rate than in T cells (Son et al., 2000). Studies of B cells during germinal center (GC) response revealed interesting differences in the maintenance of telomere length in T and B cells. In the course of naïve B cell differentiation to GC B cells, exponential numbers of cell divisions occur in the process of successfully selecting cells for further differentiation to become plasma cells and memory B cells. This process can be iterative and involve multiple cycles of mutation, selection, and clonal expansion. In contrast to predictions, the telomeres in GC B cells are not shorter but rather are substantially longer than those in either naïve or memory B cells from the same individual (Weng et al., 1997). Furthermore, naïve or memory B cells express little or no telomerase activity, whereas GC B cells express high telomerase activity. Consistent with these findings, no shortening of telomere length occurs during the transition from naïve to memory B cells (Son et al., 2003). Efficient telomere lengthening during GC response minimizes telomere loss during differentiation from naïve to memory B cell differentiation and explains the slow rate of telomere shortening in peripheral blood B cells with age.
Thus, it is conceivable that progressive telomere erosion in lymphocytes with successive rounds of cell divisions can eventually lead to the senescence of lymphocytes. If this happens, the cells that are most likely to be affected are memory lymphocytes. Indeed, marked telomere erosion, which occurs in CD4+ memory T cells in the course of responding to antigenic challenge, may in turn reduce the cells' ability to mediate a subsequent response (Reed et al., 2004). This hypothesis is also supported by the characterization of a genetic disorder, dyskeratosis congenita, which displays a telomerase defect that results from mutation in the telomerase RNA template gene; this mutation leads to the failure of hematopoietic regeneration in vivo (Mitchell et al., 1999). However, at the observed rate of telomere erosion in lymphocytes (<100 bp/year) in human adults, based on cross-sectional analyses, the initial telomere length (10–15 kb) may be sufficiently long to support sustained telomere function in lymphocytes even in centenarians. Thus, the fundamental issue of whether telomere length is actually responsible for the age-associated decline of lymphocyte function in vivo remains to be proven.
The ability of the adaptive immune system to cope with age-associated changes is limited. The continuous decrease in diversity of the antigen repertoire and the accumulation of functionally impaired memory lymphocytes with age can lead to holes in the body's defenses, and pathogens can exploit these holes. Although substantial progress has been made in identification of age-associated changes in selected lymphocyte subsets, further studies are needed to identify age-associated changes in both T and B cells at the levels of gene expression, protein, and cellular organelles. It is necessary to definitively assess the role of telomeres in lymphocyte aging through the development of sensitive and reliable methods to measure telomere length in interphase chromosomes and to detect dying cells in vivo. It is also worth exploring the role of telomerase and telomere binding proteins for better maintenance of telomere length in lymphocytes with aging. Finally, it will be crucial to understand factors that regulate thymic involution and lymphocyte homeostasis, especially of naïve cells, with aging. In this regard, it is also important to understand age-associated changes in hematopoietic progenitor cells. Our knowledge of aging of the adaptive immune system is nevertheless limited. Therefore, a better understanding of the mechanisms underlying the age-associated changes of adaptive immunity will lead to new avenues for stimulating the output of naïve cells, maintaining a healthy memory-lymphocyte pool, and developing vaccines with improved efficacy in the elderly.
Acknowledgments
I am grateful to Drs. Ranjan Sen, Dan Longo, and Richard Hodes for critical reading of the manuscript. I apologize for not citing all relevant references because of space limitations. This work was supported by the Intramural Research Program of the National Institute on Aging, National Institutes of Health (NIH).
References
  • Allman D, Miller JP. B cell development and receptor diversity during aging. Curr Opin Immunol. 2005;17:463–467. [PubMed]
  • Almanzar G, Schwaiger S, Jenewein B, Keller M, Herndler-Brandstetter D, Wurzner R, Schonitzer D, Grubeck-Loebenstein B. Long-term cytomegalovirus infection leads to significant changes in the composition of the CD8+ T-cell repertoire, which may be the basis for an imbalance in the cytokine production profile in elderly persons. J Virol. 2005;79:3675–3683. [PMC free article] [PubMed]
  • Azuma M, Phillips JH, Lanier LL. CD28− T lymphocytes: antigenic and functional properties. J Immunol. 1993;150:1147–1159. [PubMed]
  • Effros RB, Pawelec G. Replicative senescence of T cells: Does the Hayflick Limit lead to immune exhaustion? Immunol Today. 1997;18:450–454. [PubMed]
  • Fann M, Chiu WK, Wood WH, III, Levine BL, Becker KG, Weng NP. Gene expression characteristics of CD28null memory phenotype CD8+ T cells and its implication in T-cell aging. Immunol Rev. 2005;205:190–206. [PubMed]
  • Franceschi C, Cossarizza A. Introduction: The reshaping of the immune system with age. Int Rev Immunol. 1995;12:1–4. [PubMed]
  • Frasca D, Riley RL, Blomberg BB. Humoral immune response and B-cell functions including immunoglobulin class switch are downregulated in aged mice and humans. Semin Immunol. 2005;17:378–384. [PubMed]
  • Glaser R, Kiecolt-Glaser JK. Stress-induced immune dysfunction: Implications for health. Nat Rev Immunol. 2005;5:243–251. [PubMed]
  • Goronzy JJ, Weyand CM. T cell development and receptor diversity during aging. Curr Opin Immunol. 2005;17:468–475. [PubMed]
  • Haynes L, Eaton SM, Burns EM, Randall TD, Swain SL. Newly generated CD4 T cells in aged animals do not exhibit age-related defects in response to antigen. J Exp Med. 2005;201:845–851. [PMC free article] [PubMed]
  • Hodes RJ, Hathcock KS, Weng NP. Telomeres in T and B cells. Nat Rev Immunol. 2002;2:699–706. [PubMed]
  • Hooijberg E, Ruizendaal JJ, Snijders PJ, Kueter EW, Walboomers JM, Spits H. Immortalization of human CD8(+) T cell clones by ectopic expression of telomerase reverse transcriptase. J Immunol. 2000;165:4239–4245. [PubMed]
  • Johnson SA, Cambier JC. Ageing, autoimmunity and arthritis: Senescence of the B cell compartment—implications for humoral immunity. Arthritis Res Ther. 2004;6:131–139. [PMC free article] [PubMed]
  • Kang I, Hong MS, Nolasco H, Park SH, Dan JM, Choi JY, Craft J. Age-associated change in the frequency of memory CD4+ T cells impairs long term CD4+ T cell responses to influenza vaccine. J Immunol. 2004;173:673–681. [PubMed]
  • Kovaiou RD, Weiskirchner I, Keller M, Pfister G, Cioca DP, Grubeck-Loebenstein B. Age-related differences in phenotype and function of CD4+ T cells are due to a phenotypic shift from naive to memory effector CD4+ T cells. Int Immunol. 2005;17:1359–1366. [PubMed]
  • Linton PJ, Dorshkind K. Age-related changes in lymphocyte development and function. Nat Immunol. 2004;5:133–139. [PubMed]
  • Messaoudi I, LeMaoult J, Guevara-Patino JA, Metzner BM, Nikolich-Zugich J. Age-related CD8 T cell clonal expansions constrict CD8 T cell repertoire and have the potential to impair immune defense. J Exp Med. 2004;200:1347–1358. [PMC free article] [PubMed]
  • Mitchell JR, Wood E, Collins K. A telomerase component is defective in the human disease dyskeratosis congenita. Nature. 1999;402:551–555. [PubMed]
  • Mo R, Chen J, Han Y, Bueno-Cannizares C, Misek DE, Lescure PA, Hanash S, Yung RL. T cell chemokine receptor expression in aging. J Immunol. 2003;170:895–904. [PubMed]
  • Monteiro J, Batliwalla F, Ostrer H, Gregersen PK. Shortened telomeres in clonally expanded CD28−CD8+ T cells imply a replicative history that is distinct from their CD28+CD8+ counterparts. J Immunol. 1996;156:3587–3590. [PubMed]
  • Pawelec G, Akbar A, Caruso C, Solana R, Grubeck-Loebenstein B, Wikby A. Human immunosenescence: is it infectious? Immunol Rev. 2005;205:257–268. [PubMed]
  • Reed JR, Vukmanovic-Stejic M, Fletcher JM, Soares MV, Cook JE, Orteu CH, Jackson SE, Birch KE, Foster GR, Salmon M, et al. Telomere erosion in memory T cells induced by telomerase inhibition at the site of antigenic challenge in vivo. J Exp Med. 2004;199:1433–1443. [PMC free article] [PubMed]
  • Rufer N, Brummendorf TH, Kolvraa S, Bischoff C, Christensen K, Wadsworth L, Schulzer M, Lansdorp PM. Telomere fluorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood. J Exp Med. 1999;190:157–167. [PMC free article] [PubMed]
  • Sadighi Akha AA, Miller RA. Signal transduction in the aging immune system. Curr Opin Immunol. 2005;17:486–491. [PubMed]
  • Solana R, Pawelec G, Tarazona R. Aging and innate immunity. Immunity. 2006;24:491–494. this issue. [PubMed]
  • Son NH, Joyce B, Hieatt A, Chrest FJ, Yanovski J, Weng NP. Stable telomere length and telomerase expression from naive to memory B-lymphocyte differentiation. Mech Ageing Dev. 2003;124:427–432. [PubMed]
  • Son NH, Murray S, Yanovski J, Hodes RJ, Weng N. Lineage-specific telomere shortening and unaltered capacity for telomerase expression in human T and B lymphocytes with age. J Immunol. 2000;165:1191–1196. [PubMed]
  • Swain S, Clise-Dwyer K, Haynes L. Homeostasis and the age-associated defect of CD4 T cells. Semin Immunol. 2005;17:370–377. [PubMed]
  • Tarazona R, DelaRosa O, Alonso C, Ostos B, Espejo J, Pena J, Solana R. Increased expression of NK cell markers on T lymphocytes in aging and chronic activation of the immune system reflects the accumulation of effector/senescent T cells. Mech Ageing Dev. 2000;121:77–88. [PubMed]
  • Taub DD, Longo DL. Insights into thymic aging and regeneration. Immunol Rev. 2005;205:72–93. [PubMed]
  • Weng NP, Granger L, Hodes RJ. Telomere lengthening and telomerase activation during human B cell differentiation. Proc Natl Acad Sci USA. 1997;94:10827–10832. [PubMed]
  • Weng NP, Levine BL, June CH, Hodes RJ. Human naive and memory T lymphocytes differ in telomeric length and replicative potential. Proc Natl Acad Sci USA. 1995;92:11091–11094. [PubMed]