- Email: [email protected]
Lifespan of human memory T-cells in the absence of T-cell receptor expression
Immunology Letters 62 (1998) 99 – 104
Lifespan of human memory T-cells in the absence of T-cell receptor expression Shigeko Umeki a, Yoichiro Kusunoki a, John B. Cologne b, Keisuke S. Iwamoto a, Yuko Hirai a, Toshio Seyama a, Kohzo Ohama c, Seishi Kyoizumi a,* a
Department of Radiobiology, Radiation Effects Research Foundation, 5 -2 Hijiyama Park, Minami-ku, Hiroshima 732, Japan b Department of Statistics, Radiation Effects Research Foundation, 5 -2 Hijiyama Park, Minami-ku, Hiroshima 732, Japan c Department of Obstetrics and Gynecology, Hiroshima Uni6ersity Medical School, 1 -2 -3 Kasumi, Minami-ku, Hiroshima 734, Japan Accepted 6 March 1998
Abstract To evaluate the intrinsic lifespan of human memory T-cells in the absence of T-cell receptor signaling, we used radiation-induced mutant CD4 + T-cells lacking surface expression of TCR/CD3 complex as an in vivo cell marker. We analyzed the long-term kinetics of TCR/CD3 − mutant T-cells among CD4 + CD45RA + naı¨ve and CD4 + CD45RA − memory T-cell fractions in peripheral blood of gynecological cancer patients receiving radiotherapy. Both the proportion and number of these mutant T-cells decayed exponentially with time following radiotherapy. The estimated half-life of mutant memory T-cells was 2 to 3 years and did not differ from that of mutant naı¨ve T-cells. These results indicate that the lifespan of mature CD4 + T-cells is limited regardless of their memory or naı¨ve phenotype in the absence of TCR/CD3 expression. This finding may suggest that continued T-cell receptor signaling is required for lifetime maintenance of human memory T-cells. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Lifespan; Human memory T-cell; TCR mutation; Radiation therapy
1. Introduction The mechanism of T-cell memory has remained controversial with two opposing views [1 – 3]. One postulates that T-cell memory is due to long-lived memory T-cells that do not require specific antigen stimulation for their survival [4 – 7]. The other hypothesizes that long-term memory results from continuous stimulation of T-cells by persisting antigens [8 – 10]. In animals, experimental approaches using appropriate transgenic and knockout mice have provided significant insights [6,7,11,12]. Although even using these mouse models it is difficult to completely exclude the possibility of exposure to endogenous or environmental cross-reactive Abbre6iations: Mf (TCR/CD3 − mutant frequency). * Corresponding author: Tel.: +81 82 2613169; fax: + 81 82 2637279; e-mail: [email protected] 0165-2478/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S0165-2478(98)00037-6
antigens, it is demonstrated that continued TCR-MHC interaction is required for long-term survival of memory T-cells [11,12]. In humans, because experimental approaches can not be applied, few data are available so one can only speculate on the mechanism of T-cell memory. Long-term kinetics of unstable-type chromosome aberrations in human T-cells induced by radiotherapy suggested that memory T-cells receive continuous growth stimuli [13–16]. Analysis of telomere length also suggested more frequent replication in human memory T-cells [17], but it is not clear whether such stimuli are derived from antigenic exposure through TCR/CD3 complex or antigen-independent intrinsic growth signals. To approach this issue, radiation-induced mutant CD4 + T-cells lacking surface expression of TCR/CD3 complex [18,19] is considered to be an useful marker. Using this marker we can
100
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
estimate the intrinsic lifespan of memory T-cells in the absence of TCR-mediated stimulation. Mutant T-cells with loss or alteration of surface TCR/CD3 expression are present among human peripheral blood CD4 + lymphocytes at a frequency of about 2.5 × 10 − 4 [18,19]. Isolated mutant T-cell clones from blood can be grown by stimulation with phytohemagglutinin and IL-2 in vitro and stably retain their CD3 − 4 + phenotype. Their abnormality can be attributed to various alterations in TCR a or b chains, including defects in expression or partial chain deletion. Other surface marker characteristics of these mutants are CD1 − , 2 + , 5 + , 8 − , 11b − , 19 − , and 57 − , indicating that these mutants have the surface phenotype of typically mature CD4 + T-cells, except for altered TCR/ CD3 expression. These mutations are sometimes associated with abnormal TCR gene rearrangements or stable chromosome translocations [18]. Because transgenic mice carrying mutant TCR genes can not produce mature T-cells [20,21], TCR defective mutant T-cells observed in blood are assumed to be generated in the periphery. Radiation exposure induces TCR mutations in a dose-dependent manner in vitro and in vivo [19,22,23]. Flow cytometric analysis of lymphocytes in gynecological cancer patients receiving localized radiotherapy showed that the TCR/CD3 − mutant frequency (Mf) in peripheral blood started to increase soon after therapy [22,23]. A half year after therapy, Mf reached a peak level and then decayed with time. However, it was unknown whether TCR mutations were equally induced in naı¨ve and memory T-cell fractions and whether rate of decay of Mf differed between them. In the present study, the analysis of long-term in vivo kinetics of these mutant cells after radiotherapy indicates that the lifespan of human CD4 + memory T-cells does not differ from that of CD4 + naı¨ve T-cells in the absence of TCR/CD3 expression.
2. Materials and methods
2.1. Blood sample Forty-one peripheral-blood samples were obtained from 29 women who had undergone a full course of radiation therapy by linear-accelerator X-rays following surgery for uterine or cervical cancer. The total doses, ranging from 50 to 60 Gy, were usually given in 2-Gy daily fractions, 5 days a week, for 5 to 6 weeks. Time of sampling following the last 2-Gy fraction varied from 6 to 138 months. Donor ages ranged from 33 to 81 years (mean9 S.D.= 629 11). For controls, peripheral blood samples were collected from 16 healthy women who were in the similar age range (mean9 S.D.= 569 6).
2.2. Measurement of Mf The frequency of TCR/CD3 − mutant cells (Mf) in CD4 + CD45RA + or CD4 + CD45RA − cells was measured by three color flow cytometry. The Mf is considered to be the total Mf of TCR a and b genes in each CD4 + T-cell subset. The principle behind and method for measuring the Mf has been described previously [18,19]. Briefly, lymphocytes from patients and healthy controls were isolated by Ficoll-Hypaque density centrifugation and incubated on ice for 30 min with fluorescein-isothiocyanate-labelled anti-CD45RA (FL1), phycoerythrin-labelled anti-CD4 (FL2) and biotinylated anti-CD3 (FL-3) antibodies obtained from Becton-Dickinson Immunocytometry Systems (BD), San Jose, CA. After being washed, the cells were stained with RED670-labelled streptavidin (GIBCOBRL, Gaithersburg, MD) (FL3). After an additional washing, the cell suspensions were analyzed by FACScan (BD). The lymphocyte fraction was gated by forward and right-angle light scatter, and more than 2× 105 events were aquired in 3-parameter (FL1,2 and 3) correlated mode. After gating CD45RA + or RA − cells in the FL1 histogram, the frequency of CD3 − cells among RA + or RA − CD4 + cells was obtained by FL2 and FL3 dual parameter analysis (Fig. 2). The mutant window was set to the region where the surface CD3 level (FL3) was B1/25th of that for normal CD4 + cells, as described previously [18,19]. The left and right limits of FL2 were set at values half of, and two times greater than, respectively, the mode intensity of FL2 for normal CD4 T-cells. The Mf was calculated as the number of events in the mutant cell window divided by the total number of CD4 + T cells in the flow cytogram. Complete pairs of measurements were available for 36 samples from 26 women.
2.3. Statistical analysis The data are paired observations (Mf in RA + and RA − cells) obtained from 29 donors at arbitrary lengths of time following radiotherapy. Repeated measurements at different times were obtained for 10 of the 29 patients. RA + T-cell Mf could not be measured in five samples from three donors due to insufficient number (less than 5000) of RA + CD4 + T − cells analyzed. Excluding pairs with missing RA + T-cell Mf and ignoring correlation among repeated observations, we computed a simple paired-sample t-test of the difference between log-transformed RA + and RA − CD4 + T-cell Mf (the differences were approximately normally distributed). Using all of the data, we fit an exponential-decay model and tested differences in specific mechanistic parameters using the generalized estimating equation (GEE) method [24] with a g distribution for Mf, because the data come from repeated measures of
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
101
Fig. 1. Detection and analysis of CD3 − mutant CD4 + cells among CD45RA + and RA − T-cells in peripheral blood lymphocytes from gynecological cancer patients after receiving radiotherapy. (A) CD3 − 4 + mutant cells detected in total lymphocyte fraction. (B) Expression of CD45RA in CD4 + T-cells. Regions for either CD45RA + or RA − cells are set for the analysis of CD3 − 4 + mutant T-cells among each T-cell subset. (C) CD3 − 4 + mutant cells among C45RA + cells. (D) CD3 − 4 + mutant cells among CD45RA − cells. Frequency of CD3 − 4 + mutant cells among total CD4 + (a), RA + (b) and RA − (c) cells were 28.3 × 10 − 4, 15.8×10 − 4 and 32.1× 10 − 4, respectively.
paired observations and are not normally distributed. The method allows nonlinear, non-normal model fitting without requiring exact specification of the correlations among observations. Plots of residuals revealed no inadequacies in the model assumptions. The basic exponential-decay model is: m(t)= m0 + De ln(2)t/t where m(t) is mean Mf at time t (in years) since radiotherapy, m0 is background Mf, (D) is radiation-induced increase in Mf, and t is decay half life. RA + versus RA − differences in specific parameters were tested by adding appropriate indicator variables to the basic model. The number of mutant cells per unit volume of blood was also analyzed based on this model after multiplying Mf by estimated total RA + and RA − lymphocyte counts. Total counts were estimated by an exponential model fitted to lmphocyte restoration data following radiotherapy. Total RA + or RA − counts were obtained using proportions of these cell fractions derived from the flow cytograms.
3. Results and discussion TCR/CD3 − mutants in peripheral blood CD4 + Tcells of gynecological cancer patients from 6 months to 12 years after radiotherapy were detected and measured by flow cytometry (Fig. 1). These mutant T-cells were clearly different from peripheral blood CD3 − 4 + T-cell precursors (pre-T-cells) in their functional and surface phenotypes [25]. The pre-T-cells can differentiate into mature T-cells expressing CD3 in vitro [25], but the mutant T-cells maintain their CD3 − phenotype in vitro [18]. Furthermore, the pre-T-cells express a low level of CD4 and a high level of HLA-DR [25], whereas mutant T-cells were found to express a normal level of CD4 but no HLA-DR (data not shown). Since the pre-T-cells exhibit higher forward and side scatter during flow cytometry than mature T-cells including the mutant T-cells, we were able to completely gate out the pre-Tcells from the mutant window. We analyzed the TCR/CD3 − mutant frequency (Mf) in naı¨ve and memory CD4 + T-cells (Fig. 1B–D). We
102
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
used the CD45 isoform phenotype as a surface marker for either naı¨ve or memory T-cells [26]. Despite some problems [16,27–29], this marker is generally believed to reflect the functional phenotype. Using three color flow cytometry with anti-CD3, anti-CD4 and anti-CD45RA antibodies, we measured Mf in CD4 + CD45RA + (naı¨ve) and CD4 + CD45RA − (memory) cell fractions (Fig. 1C,D). Although data are not shown, Mf in CD4 + CD45RA + and CD4 + CD45RA − cells was significantly correlated with Mf in CD4 + CD45RO − and CD4 + CD45RO + cells, respectively. Fig. 2A shows the Mf in RA + and RA − CD4 + T-cells post radiotherapy. Although individual values varied considerably among the patients, they demonstrated a discernible, gradual decline. By the pairedsample t-test, the Mf among RA − CD4 + T-cells was significantly higher than that of RA + CD4 + T-cells (p =0.014). The mean difference was 1.30×10 − 4, which is similar to that in nonexposed controls (Mf in RA cells was higher by 1.85×10 − 4 than that in RA +
Fig. 2. The number of TCR/CD3 − mutants per 104 cells (A) and the number of TCR/CD3 − mutants per unit volume (ml) of blood (B) in RA − () and RA + () CD4 + cells plotted against time post radiation treatment for gynecological cancers. Exponential-decay curves were fitted to the frequency and the number of mutants in RA − (- - -) and RA + (——) CD4 + T-cells.
Table 1 T-cell receptor mutant frequency of CD4+ T-cells in nonexposed female controls T-cells
Mutant frequency (×10−4) (9 S.D.)
Total CD4+ RA+CD4+ RA−CD4+
2.82 91.08 2.32 9 1.12 4.17 9 2.16 (p = 0.0096)*
* P-value of the Mf for RA−CD4+ T-cells compared with RA+ CD4+ T-cells by paired t-test.
cells in female controls; Table 1). This suggests that the difference between decays of RA + and RA − cells with time following radiotherapy can be explained by a difference in background Mf. We therefore fitted an exponential decay model with different background levels. The estimated half-life was 2.05 (9 0.44) years for both cell fractions; there was no evidence of a difference in half-life (P=0.68). The model-estimated background Mf was 2.15 (90.61)×10 − 4 and 4.38 (9 0.90)× 10 − 4 for RA + and RA − T cells, respectively, which are very similar to those in the nonexposed controls (Table 1). This supports the validity of the model used here. Similar conclusions were obtained from the analyses of the number of mutant cells per unit volume of blood (Fig. 2B), i.e., the estimated half-life was 3.38 ( 9 1.41) years for both naı¨ve and memory cells and does not differ between them (P= 0.66). The estimated background Mf calculated from this fitted model was 1.20 (9 1.48)× 10 − 4 and 3.85 (9 1.39)× 10 − 4 for RA + and RA − T-cells, respectively, which are again similar to those in the nonexposed. It is open to debate whether T-cell memory is mediated by long-lived lymphocytes or by a low-level immune response against persistent antigen [1–3]. In mice, experimental approaches have been used to examine the antigen dependence of CD8 + and CD4 + T-cell memory [1–7,9,10,28–30]. More extensive studies have been done on CD8 + memory against murine viruses [4– 7,10,29] and male-specific antigen [6,7,12], but the evidence supports either view—that CD8 + memory cells can persist in the absence of specific antigen [4–7,12] or that memory depends on the presence of antigen [10]. Few studies have investigated whether antigen is required for maintenance of CD4 + T-cell memory [9,30], and the findings also support either view—that T-cell helper function decayed within a short period in the absence of antigen [9], or that antigen is not essential for the persistence of CD4 + T-cell memory [30]. In humans, on the basis of unstable chromosome marker studies of radiotherapy patients, naı¨ve-phenotype Tcells can remain in interphase for years, whereas memory-phenotype T-cells divide more frequently (once every 22 weeks) [15,16]. This can explain why the TCR Mf in memory cells is higher than that in naı¨ve cells in normal individuals (Table 1); somatic mutations are, in
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
general, believed to be spontaneously induced by DNA replication errors. However, it is unclear whether the more rapid mitotic turnover of memory T-cells represents a specific mechanism for maintaining antigen-specific memory or is an antigen-independent mechanism for maintaining the total number of peripheral T-cells. Thus, critical questions on T-cell memory and lifespan still remain unresolved. To answer this question, it is important to clarify whether growth stimuli to memory T-cells are derived from TCR/CD3-mediated antigen stimulation or from antigen-independent growth stimuli such as signaling through adhesion and costimulatory molecules. This can be done experimentally with the use of TCR transgenic [6,7] and RAG2 knockout [7] mice. Such studies have suggested that immunological memory is due to long-lived memory cells that persist in the absence of antigen, but the possibility of stimulation by cross-reactive antigens cannot be neglected. In fact, a recent study using MHC class II − mice suggested that long-term maintenance of CD4 + T-cells requires interactions between TCR and self-peptides presented by MHC class II molecules [11]. Also, using MHC class I − mice long-term survival of CD8 + memory T-cells was found to require TCRMHC class I interactions [12]. These studies using MHC knockout mice [11,12] suggested that continued TCR signaling from possible endogenous cross-reactive antigens presented by MHC molecules is needed for memory T-cell maintenance. In humans we believe that the present study is the first to provide sufficient information to clarify this issue. Although our study fails to distinguish between memory cells and effector cells [3], which are both CD45RA − phenotype, TCR/CD3 − CD4 + T cells in both naı¨ve and memory cell-fractions disappeared with similar half-lives, as shown in Fig. 2. Combining this with the finding on unstable chromosome aberrations [15,16] as mentioned above, it suggests that TCR-mediated signaling is required for frequent replication and possibly for long-term persistence of memory CD4 + T-cells. Furthermore, the rate of decay of mutant naı¨ve CD4 + T-cells presumably reflect the in vivo replenishing of normal naı¨ve CD4 + T-cells, because the mutant naı¨ve cells can not convert to memory phenotype due to the lack of TCR/CD3 expression. The decay rate of mutant naı¨ve CD4 + T-cells is very similar to that of T-cells carrying unstable chromosome aberrations [15,16], suggesting virtually no antigenic exposure in naı¨ve T-cells of adult humans. Further studies are needed to clarify this possibility. A weak point of the present study is the lack of information on the lifespan of human normal memory CD4 + T-cells in the radiotherapy patients to compare with the lifespan of the TCR/CD3 − memory CD4 + T-cells, because there are few suitable cell markers for tracking the fate of normal T-cells in vivo. Stable type
103
of radiation-induced chromosome aberrations may be a useful marker, although it is unknown if such aberrant T-cells have some survival disadvantage in vivo. We did not measure the frequency of such aberrant T-cells due to the limitation of blood in the present study. The previous study on the frequency of T-cells carrying stable chromosome aberrations in radiotherapy patients estimated the half-life of the memory T-cells to be about 10 years [16]. This finding suggests that the lifespan of mutant memory T-cells lacking TCR/CD3 expression is much shorter than that of normal memory T-cells. In conclusion, our study demonstrates that the lifespan of the majority of memory CD4 + T-cells in blood is limited in the absence of surface expression of TCR/CD3 complex. This may suggest that continued TCR-mediated signaling is required for lifetime maintenance of human memory CD4 + T-cells, although it is unknown whether such signaling is derived from exogenous specific antigens or endogenous cross-reactive antigens.
Acknowledgements The authors would like to acknowledge Dr N. Nakamura and N. Kodama for their helpful comments, M. Yamaoka for excellent technical assistance and Y. Katayama, M. Yonezawa and S. Nakamura for manuscript preparation. This publication is based on research performed at the Radiation Effects Research Foundation (RERF), Hiroshima and Nagasaki, Japan, and is supported in part by Grants-in-Aid (S-7) for cancer research from the Ministry of Health and Welfare. RERF is a private nonprofit foundation funded equally by the Japanese Ministry of Health and Welfare and the United States Department of Energy through the National Academy of Sciences.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
J. Sprent, D.F. Tough, Science 265 (1994) 1395. R.M. Zinkernagel, Science 271 (1996) 173. R. Ahmed, D. Gray, Science 272 (1996) 54. L.L. Lau, B.D. Jamieson, T. Somasundaram, R. Ahmed, Nature 369 (1994) 648. S. Hou, L. Hyland, K.W. Ryan, A. Portner, P.C. Doherty, Nature 369 (1994) 652. L. Bruno, J. Kirberg, H. von Boehmer, Immunity 2 (1995) 37. L. Bruno, H. von Boehmer, J. Kirberg, Eur. J. Immunol. 26 (1996) 3179. D. Gray, H. Skarvall, Nature 336 (1988) 70. D. Gray, P. Metzinger, J. Exp. Med. 174 (1991) 969. T.M. Ku¨ndig, M.F. Bachmann, U.W. Hoffmann, J.J.L. Simard, C.P. Kalberer, H. Pircher, P.S. Ohashi, H. Hengartner, R.M. Zinkernagel, Proc. Natl. Acad. Aci. USA 93 (1996) 9716.
104
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
[11] S. Takeda, H.-R. Rodewald, H. Arakawa, H. Bluethmann, T. Shimizu, Immunity 5 (1996) 217. [12] C. Tanchot, F.A. Lemonnier, B. Perarnau, A.A. Freitas, B. Rocha, Science 276 (1997) 2057. [13] A. Norman, M.S. Sasaki, R.E. Ottoman, A.G. Fingerhut, Science 147 (1965) 745. [14] K.E. Buckton, W.M.C. Brown, P.G. Smith, Nature 214 (1967) 470. [15] C.A. Michie, A. McLean, C. Alcock, P.C.L. Beverey, Nature 360 (1992) 264. [16] A.R. McLean, C.A. Michie, Proc. Natl. Acad. Sci. USA 92 (1995) 3707. [17] N.P. Weng, B.L. Levine, C.H. June, R.L. Hodes, Proc. Natl. Acad. Sci. USA 92 (1995) 11091. [18] S. Kyoizumi, M. Akiyama, Y. Hirai, Y. Kusunoki, K. Tanabe, S. Umeki, J. Exp. Med. 171 (1990) 1981. [19] S. Kyoizumi, S. Umeki, M. Akiyama, Y. Hirai, Y. Kusunoki, N. Nakamura, J. Endoh, J. Konishi, M.S. Sasaki, T. Mori, S. Fujita, J.B. Cologne, Mutat. Res. 26 (1992) 173.
.
[20] K.L. Philcott, A.R. Clark, M.A. Rudnicki, J. Iacomini, S. Itohara, J.J. Lafalle, L. Wong, Y. Ichikawa, R. Jaenisch, M.L. Hopper, S. Tonegawa, Science 256 (1992) 1448. [21] P. Mombaerts, J.L. Viney, G. Kay, S. Rastan, E.M. Gardiner, S. Chae, A.C. Hayday, M.J. Owen, Nature 360 (1992) 225. [22] K.S. Iwamoto, Y. Hirai, S. Umeki, Y. Kusunoki, S. Kyoizumi, K. Kodama, N. Ohama, N. Nakamura, M. Akiyama, J. Radiat. Res. 35 (1994) 92. [23] N. Ishioka, S. Umeki, Y. Hirai, M. Akiyama, T. Kodama, K. Ohama, S. Kyoizumi, Muta. Res. 390 (1997) 269. [24] K.Y. Liang, S.L. Zeger, Biometrika 73 (1986) 13. [25] L. Bruno, P. Res, M. Dessing, M. Cella, H. Spits, J. Exp. Med. 185 (1997) 875. [26] P.C.L. Beverley, Semin. Immunol. 4 (1992) 35. [27] E.B. Bell, S.M. Sparshott, Nature 348 (1990) 163. [28] D. Tough, J. Sprent, J. Exp. Med. 179 (1994) 1127. [29] D.F. Tough, P. Borrow, J. Sprent, Science 272 (1996) 1947. [30] D.J. Topham, R.A. Tripp, A.M. Hamilton-Easton, S.R. Sarawar, P.C. Doherty, J. Immunol. 157 (1996) 2947.
Lifespan of human memory T-cells in the absence of T-cell receptor expression Shigeko Umeki a, Yoichiro Kusunoki a, John B. Cologne b, Keisuke S. Iwamoto a, Yuko Hirai a, Toshio Seyama a, Kohzo Ohama c, Seishi Kyoizumi a,* a
Department of Radiobiology, Radiation Effects Research Foundation, 5 -2 Hijiyama Park, Minami-ku, Hiroshima 732, Japan b Department of Statistics, Radiation Effects Research Foundation, 5 -2 Hijiyama Park, Minami-ku, Hiroshima 732, Japan c Department of Obstetrics and Gynecology, Hiroshima Uni6ersity Medical School, 1 -2 -3 Kasumi, Minami-ku, Hiroshima 734, Japan Accepted 6 March 1998
Abstract To evaluate the intrinsic lifespan of human memory T-cells in the absence of T-cell receptor signaling, we used radiation-induced mutant CD4 + T-cells lacking surface expression of TCR/CD3 complex as an in vivo cell marker. We analyzed the long-term kinetics of TCR/CD3 − mutant T-cells among CD4 + CD45RA + naı¨ve and CD4 + CD45RA − memory T-cell fractions in peripheral blood of gynecological cancer patients receiving radiotherapy. Both the proportion and number of these mutant T-cells decayed exponentially with time following radiotherapy. The estimated half-life of mutant memory T-cells was 2 to 3 years and did not differ from that of mutant naı¨ve T-cells. These results indicate that the lifespan of mature CD4 + T-cells is limited regardless of their memory or naı¨ve phenotype in the absence of TCR/CD3 expression. This finding may suggest that continued T-cell receptor signaling is required for lifetime maintenance of human memory T-cells. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Lifespan; Human memory T-cell; TCR mutation; Radiation therapy
1. Introduction The mechanism of T-cell memory has remained controversial with two opposing views [1 – 3]. One postulates that T-cell memory is due to long-lived memory T-cells that do not require specific antigen stimulation for their survival [4 – 7]. The other hypothesizes that long-term memory results from continuous stimulation of T-cells by persisting antigens [8 – 10]. In animals, experimental approaches using appropriate transgenic and knockout mice have provided significant insights [6,7,11,12]. Although even using these mouse models it is difficult to completely exclude the possibility of exposure to endogenous or environmental cross-reactive Abbre6iations: Mf (TCR/CD3 − mutant frequency). * Corresponding author: Tel.: +81 82 2613169; fax: + 81 82 2637279; e-mail: [email protected] 0165-2478/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S0165-2478(98)00037-6
antigens, it is demonstrated that continued TCR-MHC interaction is required for long-term survival of memory T-cells [11,12]. In humans, because experimental approaches can not be applied, few data are available so one can only speculate on the mechanism of T-cell memory. Long-term kinetics of unstable-type chromosome aberrations in human T-cells induced by radiotherapy suggested that memory T-cells receive continuous growth stimuli [13–16]. Analysis of telomere length also suggested more frequent replication in human memory T-cells [17], but it is not clear whether such stimuli are derived from antigenic exposure through TCR/CD3 complex or antigen-independent intrinsic growth signals. To approach this issue, radiation-induced mutant CD4 + T-cells lacking surface expression of TCR/CD3 complex [18,19] is considered to be an useful marker. Using this marker we can
100
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
estimate the intrinsic lifespan of memory T-cells in the absence of TCR-mediated stimulation. Mutant T-cells with loss or alteration of surface TCR/CD3 expression are present among human peripheral blood CD4 + lymphocytes at a frequency of about 2.5 × 10 − 4 [18,19]. Isolated mutant T-cell clones from blood can be grown by stimulation with phytohemagglutinin and IL-2 in vitro and stably retain their CD3 − 4 + phenotype. Their abnormality can be attributed to various alterations in TCR a or b chains, including defects in expression or partial chain deletion. Other surface marker characteristics of these mutants are CD1 − , 2 + , 5 + , 8 − , 11b − , 19 − , and 57 − , indicating that these mutants have the surface phenotype of typically mature CD4 + T-cells, except for altered TCR/ CD3 expression. These mutations are sometimes associated with abnormal TCR gene rearrangements or stable chromosome translocations [18]. Because transgenic mice carrying mutant TCR genes can not produce mature T-cells [20,21], TCR defective mutant T-cells observed in blood are assumed to be generated in the periphery. Radiation exposure induces TCR mutations in a dose-dependent manner in vitro and in vivo [19,22,23]. Flow cytometric analysis of lymphocytes in gynecological cancer patients receiving localized radiotherapy showed that the TCR/CD3 − mutant frequency (Mf) in peripheral blood started to increase soon after therapy [22,23]. A half year after therapy, Mf reached a peak level and then decayed with time. However, it was unknown whether TCR mutations were equally induced in naı¨ve and memory T-cell fractions and whether rate of decay of Mf differed between them. In the present study, the analysis of long-term in vivo kinetics of these mutant cells after radiotherapy indicates that the lifespan of human CD4 + memory T-cells does not differ from that of CD4 + naı¨ve T-cells in the absence of TCR/CD3 expression.
2. Materials and methods
2.1. Blood sample Forty-one peripheral-blood samples were obtained from 29 women who had undergone a full course of radiation therapy by linear-accelerator X-rays following surgery for uterine or cervical cancer. The total doses, ranging from 50 to 60 Gy, were usually given in 2-Gy daily fractions, 5 days a week, for 5 to 6 weeks. Time of sampling following the last 2-Gy fraction varied from 6 to 138 months. Donor ages ranged from 33 to 81 years (mean9 S.D.= 629 11). For controls, peripheral blood samples were collected from 16 healthy women who were in the similar age range (mean9 S.D.= 569 6).
2.2. Measurement of Mf The frequency of TCR/CD3 − mutant cells (Mf) in CD4 + CD45RA + or CD4 + CD45RA − cells was measured by three color flow cytometry. The Mf is considered to be the total Mf of TCR a and b genes in each CD4 + T-cell subset. The principle behind and method for measuring the Mf has been described previously [18,19]. Briefly, lymphocytes from patients and healthy controls were isolated by Ficoll-Hypaque density centrifugation and incubated on ice for 30 min with fluorescein-isothiocyanate-labelled anti-CD45RA (FL1), phycoerythrin-labelled anti-CD4 (FL2) and biotinylated anti-CD3 (FL-3) antibodies obtained from Becton-Dickinson Immunocytometry Systems (BD), San Jose, CA. After being washed, the cells were stained with RED670-labelled streptavidin (GIBCOBRL, Gaithersburg, MD) (FL3). After an additional washing, the cell suspensions were analyzed by FACScan (BD). The lymphocyte fraction was gated by forward and right-angle light scatter, and more than 2× 105 events were aquired in 3-parameter (FL1,2 and 3) correlated mode. After gating CD45RA + or RA − cells in the FL1 histogram, the frequency of CD3 − cells among RA + or RA − CD4 + cells was obtained by FL2 and FL3 dual parameter analysis (Fig. 2). The mutant window was set to the region where the surface CD3 level (FL3) was B1/25th of that for normal CD4 + cells, as described previously [18,19]. The left and right limits of FL2 were set at values half of, and two times greater than, respectively, the mode intensity of FL2 for normal CD4 T-cells. The Mf was calculated as the number of events in the mutant cell window divided by the total number of CD4 + T cells in the flow cytogram. Complete pairs of measurements were available for 36 samples from 26 women.
2.3. Statistical analysis The data are paired observations (Mf in RA + and RA − cells) obtained from 29 donors at arbitrary lengths of time following radiotherapy. Repeated measurements at different times were obtained for 10 of the 29 patients. RA + T-cell Mf could not be measured in five samples from three donors due to insufficient number (less than 5000) of RA + CD4 + T − cells analyzed. Excluding pairs with missing RA + T-cell Mf and ignoring correlation among repeated observations, we computed a simple paired-sample t-test of the difference between log-transformed RA + and RA − CD4 + T-cell Mf (the differences were approximately normally distributed). Using all of the data, we fit an exponential-decay model and tested differences in specific mechanistic parameters using the generalized estimating equation (GEE) method [24] with a g distribution for Mf, because the data come from repeated measures of
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
101
Fig. 1. Detection and analysis of CD3 − mutant CD4 + cells among CD45RA + and RA − T-cells in peripheral blood lymphocytes from gynecological cancer patients after receiving radiotherapy. (A) CD3 − 4 + mutant cells detected in total lymphocyte fraction. (B) Expression of CD45RA in CD4 + T-cells. Regions for either CD45RA + or RA − cells are set for the analysis of CD3 − 4 + mutant T-cells among each T-cell subset. (C) CD3 − 4 + mutant cells among C45RA + cells. (D) CD3 − 4 + mutant cells among CD45RA − cells. Frequency of CD3 − 4 + mutant cells among total CD4 + (a), RA + (b) and RA − (c) cells were 28.3 × 10 − 4, 15.8×10 − 4 and 32.1× 10 − 4, respectively.
paired observations and are not normally distributed. The method allows nonlinear, non-normal model fitting without requiring exact specification of the correlations among observations. Plots of residuals revealed no inadequacies in the model assumptions. The basic exponential-decay model is: m(t)= m0 + De ln(2)t/t where m(t) is mean Mf at time t (in years) since radiotherapy, m0 is background Mf, (D) is radiation-induced increase in Mf, and t is decay half life. RA + versus RA − differences in specific parameters were tested by adding appropriate indicator variables to the basic model. The number of mutant cells per unit volume of blood was also analyzed based on this model after multiplying Mf by estimated total RA + and RA − lymphocyte counts. Total counts were estimated by an exponential model fitted to lmphocyte restoration data following radiotherapy. Total RA + or RA − counts were obtained using proportions of these cell fractions derived from the flow cytograms.
3. Results and discussion TCR/CD3 − mutants in peripheral blood CD4 + Tcells of gynecological cancer patients from 6 months to 12 years after radiotherapy were detected and measured by flow cytometry (Fig. 1). These mutant T-cells were clearly different from peripheral blood CD3 − 4 + T-cell precursors (pre-T-cells) in their functional and surface phenotypes [25]. The pre-T-cells can differentiate into mature T-cells expressing CD3 in vitro [25], but the mutant T-cells maintain their CD3 − phenotype in vitro [18]. Furthermore, the pre-T-cells express a low level of CD4 and a high level of HLA-DR [25], whereas mutant T-cells were found to express a normal level of CD4 but no HLA-DR (data not shown). Since the pre-T-cells exhibit higher forward and side scatter during flow cytometry than mature T-cells including the mutant T-cells, we were able to completely gate out the pre-Tcells from the mutant window. We analyzed the TCR/CD3 − mutant frequency (Mf) in naı¨ve and memory CD4 + T-cells (Fig. 1B–D). We
102
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
used the CD45 isoform phenotype as a surface marker for either naı¨ve or memory T-cells [26]. Despite some problems [16,27–29], this marker is generally believed to reflect the functional phenotype. Using three color flow cytometry with anti-CD3, anti-CD4 and anti-CD45RA antibodies, we measured Mf in CD4 + CD45RA + (naı¨ve) and CD4 + CD45RA − (memory) cell fractions (Fig. 1C,D). Although data are not shown, Mf in CD4 + CD45RA + and CD4 + CD45RA − cells was significantly correlated with Mf in CD4 + CD45RO − and CD4 + CD45RO + cells, respectively. Fig. 2A shows the Mf in RA + and RA − CD4 + T-cells post radiotherapy. Although individual values varied considerably among the patients, they demonstrated a discernible, gradual decline. By the pairedsample t-test, the Mf among RA − CD4 + T-cells was significantly higher than that of RA + CD4 + T-cells (p =0.014). The mean difference was 1.30×10 − 4, which is similar to that in nonexposed controls (Mf in RA cells was higher by 1.85×10 − 4 than that in RA +
Fig. 2. The number of TCR/CD3 − mutants per 104 cells (A) and the number of TCR/CD3 − mutants per unit volume (ml) of blood (B) in RA − () and RA + () CD4 + cells plotted against time post radiation treatment for gynecological cancers. Exponential-decay curves were fitted to the frequency and the number of mutants in RA − (- - -) and RA + (——) CD4 + T-cells.
Table 1 T-cell receptor mutant frequency of CD4+ T-cells in nonexposed female controls T-cells
Mutant frequency (×10−4) (9 S.D.)
Total CD4+ RA+CD4+ RA−CD4+
2.82 91.08 2.32 9 1.12 4.17 9 2.16 (p = 0.0096)*
* P-value of the Mf for RA−CD4+ T-cells compared with RA+ CD4+ T-cells by paired t-test.
cells in female controls; Table 1). This suggests that the difference between decays of RA + and RA − cells with time following radiotherapy can be explained by a difference in background Mf. We therefore fitted an exponential decay model with different background levels. The estimated half-life was 2.05 (9 0.44) years for both cell fractions; there was no evidence of a difference in half-life (P=0.68). The model-estimated background Mf was 2.15 (90.61)×10 − 4 and 4.38 (9 0.90)× 10 − 4 for RA + and RA − T cells, respectively, which are very similar to those in the nonexposed controls (Table 1). This supports the validity of the model used here. Similar conclusions were obtained from the analyses of the number of mutant cells per unit volume of blood (Fig. 2B), i.e., the estimated half-life was 3.38 ( 9 1.41) years for both naı¨ve and memory cells and does not differ between them (P= 0.66). The estimated background Mf calculated from this fitted model was 1.20 (9 1.48)× 10 − 4 and 3.85 (9 1.39)× 10 − 4 for RA + and RA − T-cells, respectively, which are again similar to those in the nonexposed. It is open to debate whether T-cell memory is mediated by long-lived lymphocytes or by a low-level immune response against persistent antigen [1–3]. In mice, experimental approaches have been used to examine the antigen dependence of CD8 + and CD4 + T-cell memory [1–7,9,10,28–30]. More extensive studies have been done on CD8 + memory against murine viruses [4– 7,10,29] and male-specific antigen [6,7,12], but the evidence supports either view—that CD8 + memory cells can persist in the absence of specific antigen [4–7,12] or that memory depends on the presence of antigen [10]. Few studies have investigated whether antigen is required for maintenance of CD4 + T-cell memory [9,30], and the findings also support either view—that T-cell helper function decayed within a short period in the absence of antigen [9], or that antigen is not essential for the persistence of CD4 + T-cell memory [30]. In humans, on the basis of unstable chromosome marker studies of radiotherapy patients, naı¨ve-phenotype Tcells can remain in interphase for years, whereas memory-phenotype T-cells divide more frequently (once every 22 weeks) [15,16]. This can explain why the TCR Mf in memory cells is higher than that in naı¨ve cells in normal individuals (Table 1); somatic mutations are, in
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
general, believed to be spontaneously induced by DNA replication errors. However, it is unclear whether the more rapid mitotic turnover of memory T-cells represents a specific mechanism for maintaining antigen-specific memory or is an antigen-independent mechanism for maintaining the total number of peripheral T-cells. Thus, critical questions on T-cell memory and lifespan still remain unresolved. To answer this question, it is important to clarify whether growth stimuli to memory T-cells are derived from TCR/CD3-mediated antigen stimulation or from antigen-independent growth stimuli such as signaling through adhesion and costimulatory molecules. This can be done experimentally with the use of TCR transgenic [6,7] and RAG2 knockout [7] mice. Such studies have suggested that immunological memory is due to long-lived memory cells that persist in the absence of antigen, but the possibility of stimulation by cross-reactive antigens cannot be neglected. In fact, a recent study using MHC class II − mice suggested that long-term maintenance of CD4 + T-cells requires interactions between TCR and self-peptides presented by MHC class II molecules [11]. Also, using MHC class I − mice long-term survival of CD8 + memory T-cells was found to require TCRMHC class I interactions [12]. These studies using MHC knockout mice [11,12] suggested that continued TCR signaling from possible endogenous cross-reactive antigens presented by MHC molecules is needed for memory T-cell maintenance. In humans we believe that the present study is the first to provide sufficient information to clarify this issue. Although our study fails to distinguish between memory cells and effector cells [3], which are both CD45RA − phenotype, TCR/CD3 − CD4 + T cells in both naı¨ve and memory cell-fractions disappeared with similar half-lives, as shown in Fig. 2. Combining this with the finding on unstable chromosome aberrations [15,16] as mentioned above, it suggests that TCR-mediated signaling is required for frequent replication and possibly for long-term persistence of memory CD4 + T-cells. Furthermore, the rate of decay of mutant naı¨ve CD4 + T-cells presumably reflect the in vivo replenishing of normal naı¨ve CD4 + T-cells, because the mutant naı¨ve cells can not convert to memory phenotype due to the lack of TCR/CD3 expression. The decay rate of mutant naı¨ve CD4 + T-cells is very similar to that of T-cells carrying unstable chromosome aberrations [15,16], suggesting virtually no antigenic exposure in naı¨ve T-cells of adult humans. Further studies are needed to clarify this possibility. A weak point of the present study is the lack of information on the lifespan of human normal memory CD4 + T-cells in the radiotherapy patients to compare with the lifespan of the TCR/CD3 − memory CD4 + T-cells, because there are few suitable cell markers for tracking the fate of normal T-cells in vivo. Stable type
103
of radiation-induced chromosome aberrations may be a useful marker, although it is unknown if such aberrant T-cells have some survival disadvantage in vivo. We did not measure the frequency of such aberrant T-cells due to the limitation of blood in the present study. The previous study on the frequency of T-cells carrying stable chromosome aberrations in radiotherapy patients estimated the half-life of the memory T-cells to be about 10 years [16]. This finding suggests that the lifespan of mutant memory T-cells lacking TCR/CD3 expression is much shorter than that of normal memory T-cells. In conclusion, our study demonstrates that the lifespan of the majority of memory CD4 + T-cells in blood is limited in the absence of surface expression of TCR/CD3 complex. This may suggest that continued TCR-mediated signaling is required for lifetime maintenance of human memory CD4 + T-cells, although it is unknown whether such signaling is derived from exogenous specific antigens or endogenous cross-reactive antigens.
Acknowledgements The authors would like to acknowledge Dr N. Nakamura and N. Kodama for their helpful comments, M. Yamaoka for excellent technical assistance and Y. Katayama, M. Yonezawa and S. Nakamura for manuscript preparation. This publication is based on research performed at the Radiation Effects Research Foundation (RERF), Hiroshima and Nagasaki, Japan, and is supported in part by Grants-in-Aid (S-7) for cancer research from the Ministry of Health and Welfare. RERF is a private nonprofit foundation funded equally by the Japanese Ministry of Health and Welfare and the United States Department of Energy through the National Academy of Sciences.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
J. Sprent, D.F. Tough, Science 265 (1994) 1395. R.M. Zinkernagel, Science 271 (1996) 173. R. Ahmed, D. Gray, Science 272 (1996) 54. L.L. Lau, B.D. Jamieson, T. Somasundaram, R. Ahmed, Nature 369 (1994) 648. S. Hou, L. Hyland, K.W. Ryan, A. Portner, P.C. Doherty, Nature 369 (1994) 652. L. Bruno, J. Kirberg, H. von Boehmer, Immunity 2 (1995) 37. L. Bruno, H. von Boehmer, J. Kirberg, Eur. J. Immunol. 26 (1996) 3179. D. Gray, H. Skarvall, Nature 336 (1988) 70. D. Gray, P. Metzinger, J. Exp. Med. 174 (1991) 969. T.M. Ku¨ndig, M.F. Bachmann, U.W. Hoffmann, J.J.L. Simard, C.P. Kalberer, H. Pircher, P.S. Ohashi, H. Hengartner, R.M. Zinkernagel, Proc. Natl. Acad. Aci. USA 93 (1996) 9716.
104
S. Umeki et al. / Immunology Letters 62 (1998) 99–104
[11] S. Takeda, H.-R. Rodewald, H. Arakawa, H. Bluethmann, T. Shimizu, Immunity 5 (1996) 217. [12] C. Tanchot, F.A. Lemonnier, B. Perarnau, A.A. Freitas, B. Rocha, Science 276 (1997) 2057. [13] A. Norman, M.S. Sasaki, R.E. Ottoman, A.G. Fingerhut, Science 147 (1965) 745. [14] K.E. Buckton, W.M.C. Brown, P.G. Smith, Nature 214 (1967) 470. [15] C.A. Michie, A. McLean, C. Alcock, P.C.L. Beverey, Nature 360 (1992) 264. [16] A.R. McLean, C.A. Michie, Proc. Natl. Acad. Sci. USA 92 (1995) 3707. [17] N.P. Weng, B.L. Levine, C.H. June, R.L. Hodes, Proc. Natl. Acad. Sci. USA 92 (1995) 11091. [18] S. Kyoizumi, M. Akiyama, Y. Hirai, Y. Kusunoki, K. Tanabe, S. Umeki, J. Exp. Med. 171 (1990) 1981. [19] S. Kyoizumi, S. Umeki, M. Akiyama, Y. Hirai, Y. Kusunoki, N. Nakamura, J. Endoh, J. Konishi, M.S. Sasaki, T. Mori, S. Fujita, J.B. Cologne, Mutat. Res. 26 (1992) 173.
.
[20] K.L. Philcott, A.R. Clark, M.A. Rudnicki, J. Iacomini, S. Itohara, J.J. Lafalle, L. Wong, Y. Ichikawa, R. Jaenisch, M.L. Hopper, S. Tonegawa, Science 256 (1992) 1448. [21] P. Mombaerts, J.L. Viney, G. Kay, S. Rastan, E.M. Gardiner, S. Chae, A.C. Hayday, M.J. Owen, Nature 360 (1992) 225. [22] K.S. Iwamoto, Y. Hirai, S. Umeki, Y. Kusunoki, S. Kyoizumi, K. Kodama, N. Ohama, N. Nakamura, M. Akiyama, J. Radiat. Res. 35 (1994) 92. [23] N. Ishioka, S. Umeki, Y. Hirai, M. Akiyama, T. Kodama, K. Ohama, S. Kyoizumi, Muta. Res. 390 (1997) 269. [24] K.Y. Liang, S.L. Zeger, Biometrika 73 (1986) 13. [25] L. Bruno, P. Res, M. Dessing, M. Cella, H. Spits, J. Exp. Med. 185 (1997) 875. [26] P.C.L. Beverley, Semin. Immunol. 4 (1992) 35. [27] E.B. Bell, S.M. Sparshott, Nature 348 (1990) 163. [28] D. Tough, J. Sprent, J. Exp. Med. 179 (1994) 1127. [29] D.F. Tough, P. Borrow, J. Sprent, Science 272 (1996) 1947. [30] D.J. Topham, R.A. Tripp, A.M. Hamilton-Easton, S.R. Sarawar, P.C. Doherty, J. Immunol. 157 (1996) 2947.