Vol. 52 No. 4/2005, 931–935 Communication

Immunosenescence is viewed as a remodeling process with the exhaustion of naïve T cells and filling up of the immunological space with memory cells. In this study some phenotypic changes of CD8(+) human cells during in vivo ageing were compared with those observed in long term cultures of lymphocytes derived from cord blood or from peripheral blood from donors of different age. Both in vivo and in vitro a significant decrease of the fraction of CD8(+)CD28(+) cells was observed. Comparing the proportions of other T cell subpopulations (the CD4/CD8(+) ratio, CD56, CD57, CD27) made it possible to conclude that replicative senescence in vitro partially reflects in vivo ageing.

on-line at: www.actabp.plImmunosenescence is a complex remodeling process during which some parameters decrease, some increase, and others remain unchanged (the remodeling theory of ageing) (Franceschi et al., 2000b).Ageing of the immune system is described as exhaustion of naïve cells pool due to thymic involution and filling up of the immunological space by memory and effector T cell clones as a consequence of chronic antigenic stress (Franceschi et al., 2000a).A large body of evidence indicates that after stimulation by antigens CD8 + cells are induced to proliferate, undergo clonal expansion and change their phenotypic and functional characteristics.This process is described as maturation or differentiation (Appay & Rowland-Jones, 2004;Papagno et al., 2004).The replicative exhaustion of lymphocytes and a decrease of immune function occur, at least partially, as a result of differentiation (Papagno et al., 2004).Lymphocytes isolated from whole blood and cultured in vitro undergo replicative senescence, which is believed reflect to some extent in vivo ageing a�er antigenic stimulation.However, so far there is no data about the phenotypic changes of lymphocytes in long term cultures in vitro.Thus, this study was undertaken to compare the phenotypic changes of human T lymphocytes which occur during in vivo ageing with those appearing during replicative senescence in vitro.

MATERIALS AND METHODS
Cell isolation.Peripheral blood was collected from healthy young (25-35 years, group II) and middle-aged (45-55 years, group III) volunteers, as well as from elderly ones (65 years, group IV) and centenarians (taking part in the Polish Centenarian Project, group V).Lymphocytes were also isolated from cord blood (group I).The number of donors tested in each experiment is indicated in figure legend.Mononuclear cells were prepared by standard centrifugation over Ficoll-Paque.
Lymphocytes isolated from cord blood and from peripheral blood of donors from II and III age groups were cultured in vitro for 30-35 days in the presence of IL-2.Lymphocyte activation was induced with lectin PHA at 5 µg/ml (Sigma) added at the time of seeding (day 0).All cell cultures were conducted in RPMI 1640 medium supplemented with 10% FCS, 2 mM �-glutamine and antibiotics, at a starting density of 0.8 × 10 6 cells/ml.From day 3, cells were washed with PBS and 60 U/ml of recombinant IL-2 (rIL-2; Peprotech EC Ltd, UK) with fresh medium added.
Cell phenotyping.Cells from blood of all subjects were phenotyped on a flow cytometer using Simultest and other sets of antibodies from Beckton-Dickinson (anti-CD27-FITC, anti-CD56-FITC,
Statistical analysis.Statistical analysis of data were performed using Statistica 6.0.Non-parametric tests were applied.Parameters between age groups were compared with Kruskal-Wallis tests for paired observations.All results are given as mean ±S.E.M. P values < 0.05 were considered statistically significant, *P < 0.05, **P < 0.01 and ***P < 0.005.

RESULTS AND DISCUSSION
During ageing continuous phenotypic changes of T cell are observed.The best documented changes concern an increase of memory and a decrease of naïve cells as well as a dramatic drop of CD8 + T cells with the CD28 + co-receptor and concomitant accumulation of CD8 + CD28 -.In in vitro replicative senescence a very well known fact is an increase in the proportion of CD8 + CD28 -cells with diminished proliferation capacity.However, so far there is no data comparing other phenotypic parameters occurring during ageing in vivo and senescence in vitro.Here, I present data showing lymphocyte subpopulation changes during in vivo and in vitro senescence.

The CD4 and CD8 subpopulations in in vivo and in vitro ageing
The absolute number and the percentage of CD4 and CD8 lymphocytes in peripheral blood, as well as the CD4/CD8 ratio, seem to be an important parameter of health status.The lower CD4/CD8 ratio, together with some other immune parameters predicts a higher risk of death (Pawelec et al., 2004).However, there are contradictory reports showing either a decrease in absolute number of CD4 + and increase of CD8 + cells (Wikby et al., 2002;Huppert et al., 2003), decrease of CD8 + (Naumova et al., 2004), or no changes in either parameter during ageing (Peres et al., 2003).In this study lower mean values of CD8 + than of CD4 + cells were observed in donors from groups I-V.The CD4/CD8 ratio in peripheral blood of all donors ranged between 1.8-2, being significantly greater in cord blood (3.0).The percentage of centenarian donors with CD4/CD8 ratio below 1 was higher (25%) than in group IV (5%) (Fig. 1A).There was only a single donor with the CD4/CD8 ratio < 1 in group III.Thus, the older the age-group, the more donors exhibited a relative domination of the CD8 subpopulation, which might reflect the previously described age-related immunodeterioration (Pawelec et al., 2004).In in vitro culture of PBL and CBL a predominance of CD4 + cells was observed at the beginning of culture; a�er 7th day of PBL culture and a�er 14th day of CBL culture, CD8 + cells prevailed.This picture persisted until the end of CBL culture (about 35th day of culture).In PBL culture, the CD4/CD8 ratio changed again at the end of culture (Fig. 1B, C).These results suggest that in PBL culture CD8 + cells divided slower or underwent less division rounds, comparing to CD4 + cells.A similar observation of CD4 + lymphocyte expansion and a decrease of CD8 + cells in vitro was made by Denne� et al. (2002).Contrary to this, an expansion of CD8 + population was observed in culture with continuous restimulation (Spaulding et al., 1999).Thus, CD8 + cells cultured in vitro, in contrast to CD4 + cells, need to be repeatedly stimulated with exogenous antigen to maintain high proliferative activity.
Similar changes in CD8 + CD28 + and CD8 + CD28 - proportions were observed during in vitro ageing.The percentage of the CD8 + CD28 + subpopulation gradually decreased with time.The decrease of the CD8 + CD28 + subpopulation was faster in CBL than in PBL cultures; a dramatic drop from almost 100% of CD8 + CD28 + cells at the beginning of culture to 40% a�er two weeks was observed.However, at the end of culture irrespective of the age of donor some cells still displayed the CD28 + phenotype and the average percentage of CD8 + CD28 + cells at the 35th day of culture was 12, 35, 42 and 20% in age groups I, II, III and V, respectively (Fig. 2B).
In all donors tested I observed decrease of the number of CD8 + CD28 + and a concomitant increase of the number of CD8 + CD28 -cells both in vivo and in vitro, thus proving that in vitro replicative senescence reflects -at least considering this parameter -the ageing process occurring in vivo.
The question is whether the CD8 + CD28 -subpopulation is the progeny of CD8 + CD28 + cells.Actually, there are some data showing that CD28 -cells appear as progeny of CD28 + cells, a�er down-regulation of CD28 receptor expression (Fiorentini et al., 1999;Borthwick et al., 2000).In cultures of isolated fractions or clones of CD8 + CD28 + cells, CD28 -cells have been reported to appear (Fiorentini et al., 1999;Posne� et al., 1999;Borthwick et al., 2000).Moreover, in CD28 + and CD28 -subpopulations the same TCRVβ clones were expanded (Horiuchi et al., 2001).On the other hand, there are data about an extrathymic lineage of these cells (Tsukahara et al., 1997).Azuma et al. (1993) showed CD8 + clones with the CD28 + phenotype lasting for 4 weeks in culture, arguing that the loss of CD28 expression contributed insignificantly to the appearance of CD28 -cells.The results presented in this study show that CBL, being virtually all the CD8 + CD28 + , are gradually replaced with CD8 + CD28 -cells.In CBL the decrease of CD8 + CD28 + percentage was more intense (even 80% during four weeks of in vitro culture, Fig. 2B) and was accompanied by a higher proliferative activity of these cells comparing to PBL (not shown).This data is consistent with the conclusion that CD28 -cells are the progeny of CD28 + cells and appear in the process of antigen-driven differentiation.Presumably, proliferation of CD28 + cells cause the loss of CD28 receptor (Posne� et al., 1999;Borthwick et al., 2000).

The CD27, CD56 and CD57 subpopulations in in vivo and in vitro ageing
Differentiation of lymphocytes is induced by antigenic stimulation and is marked by decreasing expression of CD28 and CD27 receptors.The following phases of CD8 + T cell differentiation, based on the CD28 and CD27 phenotype, were proposed: early CD28 + CD27 + , intermediate CD28 -CD27 + and late CD28 -CD27 - (Appay & Rowland-Jones, 2004).In fact, during in vivo ageing a decrease of the percentage of CD8 + cells expressing the CD27 receptor was observed (Fig. 3A, P < 0.001), and the most important phenotypic changes concerned the percentage of the CD28 + CD27 + and CD28 -CD27 -subpopulations.The CD8 + CD28 -CD27 -cells expanded significantly (up to 60.9% in centenarians, not shown).The intermediate subpopulation (CD28 -CD27 + ) constituted almost the same percentage of the CD8 + pool in all age groups.Only in donors with acute infections or a�er transplantation was a great increase of the percentage of this subpopulation reported (Roos et al., 2000;Wills et al., 2002).I obtained a similar result in cultures senescing in vitro, in which CD8 + cells were losing CD28 expression without concomitant down-regulation of CD27 expression.Until the 30th day of culture more than 90% of CD8 + cells were CD27 positive (Fig. 3B).Thus, unlike during ageing in vivo, no relationship was found between the appearance of the CD8 + CD28 -and CD8 + CD27 -subpopulations in vitro.Similary to in vitro stimulation, decrease of CD28 expression on CD8 + cells without CD27 downregulation is a common feature of acute viral infection, during which multiple rounds of division in relatively short period occur (Roos et al., 2000;Papagno et al., 2004).This suggests, that -comparing to CD27 expression -the down-regulation of CD28 expression is more dependent on completed divisions.
The proportion of T lymphocytes bearing NK markers (CD56, CD57) increases with ageing (Tarazona et al., 2000;Wikby et al., 2002).In this study, a higher percentage of cells with the CD56 and CD57 receptors was observed in older donors than in   younger ones (Fig. 4A).In CBL all cells lacked the cytotoxic markers, which confirms the 'naïve phenotype' of newborns' blood.The percentage of the CD8 + CD57 + subpopulation increased during in vivo ageing to a higher level than of the CD8 + CD56 + subpopulation.The most significant differences were observed between I and III age groups, in further life there was a less significant increase of the percentage of CD56 + and CD57 + cells.
Lymphocytes with NK markers have been reported to have shorter telomeres, which reflects progress in replicative senescence (Tarazona et al., 2000).Acute viral infection induces high proliferation and clonal expansion in the CD8 + subpopulation and the expanded cells tend to express cytotoxic markers (Looney et al., 1999;Khan et al., 2002).The common CMV infection is noted more frequently in older people (Khan et al., 2002) and correlates with an increased number of CD8 + CD57 + cells (Ouyang et al., 2003).Published data suggest, that the relationship between CMV infection, CD8 + CD57 + expansion and reversed CD4/CD8 ratio exist (Wikby et al., 2002).In centenarians, I observed a large amount of CD8 + CD57 + cells in donors with the CD4/CD8 ratio below 1, as well as with a ratio above 1.Thus, the appearance of CD57 + cells is probably more dependent on age than on the CD4/CD8 ratio.Similar conclusions were offered by Merino et al. (1998).
During in vitro culture, cells isolated from CBL and PBL showed a different tendency in CD56 and CD57 expression changes, comparing to ageing in vivo.In PBL culture, in contrast to in vivo ageing, the percentage of CD8 + CD57 + cells decreased gradually (P < 0.05 in day 14 and day 30 vs day 0) (Fig. 4B).
The CD8 + CD56 + subpopulation changes were not statistically significant.Thus, in respect to acquiring the NK phenotype, CD8 + lymphocyte ageing in vivo differs from in vitro replicative senescence of CD8 + cells isolated from peripheral blood.However, during culture of CBL there was a significant increase of CD8 + CD56 + and CD8 + CD57 + percentage, similarly to changes observed during in vivo ageing (P < 0.05 in day 14 and day 30 vs day 0) (Fig. 4C).In CBL, the NK expressing cells were formed de novo during culture in vitro.The reason of these differences is unknown.
Altogether, this study shows that some phenotypic changes are common for ageing in vivo and in vitro.The evolution of cells with the naïve phenotype into highly differentiated effector/memory cells seems to be driven, first of all, by stimulation and proliferation.An understanding of the mechanisms of lymphocyte maturation may contribute to the development of strategies protecting from age-related decline in immune functions.For this reason, in vitro culture of human lymphocytes is useful for studies concerning particular parameters of immunosenescence undergoing in vivo.

Figure 1 .
Figure 1.The CD4 + /CD8 + subpopulation ratio in in vivo ageing (A) and in vitro culture (B, C).A. Percentage of donors displaying different CD4/CD8 ratios (< 1, 1-1.5, 1.5-2, >2) in group IV (elderly and V, centenarians).B, C. Percentage of CD4 + and CD8 + subpopulations isolated from cord blood (B) and peripheral blood (C) during in vitro culture.Results were obtained from seven CBL and nine PBL cultures from donors from groups II and III.

Figure 2 .
Figure 2. The percentage of CD8 + CD28 + subpopulation in vivo (A) and in vitro (B).Analyses were performed: A, in whole blood of 13, 15, 22, 57 and 108 donors from age groups I-V, respectively; B, during in vitro culture of lymphocytes isolated from seven donors from I group, eight donors from II group, seven donors from group III and ten from group V.

Figure 3 .
Figure 3.The percentage of CD8 + CD27 + subpopulation in ageing in vivo (A) and in vitro of PBL (B) and CBL (C) cultures.Analyses were performed: A, in the whole blood of 5, 22, 13, 48 and 47 donors from age groups I-V, respectively; B, during in vitro culture of lymphocytes isolated from five donors from group I and 12 donors from groups II and III.

Figure 4 .
Figure 4.The percentage of CD8 + CD56 + and CD8 + CD57 + subpopulations in ageing in vivo (A) and in vitro of PBL (B) and CBL (C) cultures.The analyses were done as described in the legend to Fig. 3.