- Email: [email protected]
Age-dependent DNA methylation changes in the ITGAL (CD11a) promoter
Mechanisms of Ageing and Development 123 (2002) 1257– 1268 www.elsevier.com/locate/mechagedev
Age-dependent DNA methylation changes in the ITGAL (CD11a) promoter Zhiyong Zhang, Chun Deng, Qianjin Lu, Bruce Richardson * Department of Medicine, Veterans Affairs Hospital, 5310 Cancer Center and Geriatrics Center Building, Uni6ersity of Michigan, Ann Arbor, MI 48109 -0940, USA Received 13 September 2001; received in revised form 4 February 2002; accepted 5 February 2002
Abstract DNA methylation patterns change with age in a complex fashion, typically with an overall decrease in genomic deoxymethylcytosine (dmC) content, but with local increases in some promoters that contain GC-rich sequences known as CpG islands. While the consequences of age-dependent CpG island methylation have recently been studied in organs such as the colon, less is known about the functional significance of the progressive hypomethylation of promoters lacking CpG islands, and the significance of age-dependent changes in T cell DNA methylation is completely unexplored. We asked if age-dependent DNA hypomethylation might contribute to overexpression of the T cell ITGAL gene, which encodes CD11a, a subunit of LFA-1. CD11a mRNA increased with age as well as with experimentally induced DNA hypomethylation. This increase correlated with hypomethylation of sequences flanking the ITGAL promoter in vitro and in aging. ‘Patch’ methylation of the region suppressed promoter function. DNA methyltransferases 1 and 3a also decreased with aging. These results indicate that hypomethylation of regions flanking the ITGAL promoter may increase CD11a expression, and suggest that age-dependent hypomethylation of promoters lacking CpG islands, perhaps due to decreased DNA methyltransferase expression, may be one mechanism contributing to increased T cell gene expression with aging. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: DNA methylation; ITGAL; LFA-1; T lymphocytes; Aging
1. Introduction A number of biochemical changes occur in DNA over time, and may contribute to the alterations in gene expression associated with aging. These changes include oxidative damage, telomere shortening, and the accumulation of mutations * Corresponding author. Tel.: +1-734-763-6940; fax: + 1743-936-9220. E-mail address: [email protected] (B. Richardson).
(Guarente, 1996). Another process associated with aging is the progressive alteration of DNA methylation patterns. Methylation patterns are established during ontogeny, then change with aging in a complex fashion. In general, total genomic deoxymethylcytosine (dmC) content decreases with age, and involves genes as well as repetitive sequences. However, the transcriptional relevance of age-dependent DNA demethylation has generally not been examined. Paradoxically, this progressive demethylation is associated with
0047-6374/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S0047-6374(02)00014-3
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1258
abnormal hypermethylation of CpG islands. Agedependent CpG island hypermethylation typically results in suppression of the associated gene, and may contribute to carcinogenesis by inhibiting expression of genes with various growth suppressing functions (Issa, 1999). Immune abnormalities also occur with aging, and the mechanisms causing the changes are unknown. The abnormalities include the progressive development of T cell anergy, as well as the development of autoimmunity and anti-nuclear antibodies (Yung et al., 2001a,b). We have reported that human T lymphocyte DNA demethylates with age (Golbus et al., 1990), that by middle age approximately 15% of T cell CpG islands are variably methylated on a clonal basis, and that the increased CpG island methylation is associated with decreased transcription of the associated gene (Zhu et al., 1999). Further, using murine models, we have demonstrated an association between progressive DNA hypomethylation and the development of autoimmunity and some signs of immune senescence (Yung et al., 2001a,b). This suggests that some of the functional changes associated with immune senescence may be due to altered DNA methylation patterns. To further define the role that altered DNA methylation plays in age-associated changes in T cell gene expression, we asked if age-dependent hypomethylation occurs in transcriptionally relevant sequences of a gene that lacks a CpG island and that increases expression with aging. We selected CD11a, one chain of the b2 integrin lymphocyte function associated antigen 1 (LFA-1) also referred to as CD11a/CD18 or aLb2. T cell LFA-1 increases progressively throughout life (Pallis et al., 1993), and overexpression of T cell LFA-1 causes autoimmunity and anti-DNA antibodies (Yung et al., 1996). Further, inhibiting T
cell DNA methylation with DNA methyltransferase inhibitors like 5-azacytidine increases LFA1 expression through effects on CD11a (Kaplan et al., 2000; Richardson et al., 1992). Our results support the contention that progressive DNA hypomethylation contributes to changes in T cell gene expression with age.
2. Materials and methods
2.1. Donor populations Cord blood was obtained from placentas of healthy newborns. Peripheral venous blood was obtained from healthy middle aged donors, and from elderly donors 65 years of age and older recruited through the Geriatrics Center Research Participant Program, sponsored by the Claude D. Pepper Older Americans Independence Center (Table 1). Exclusions included conditions and drugs known to affect DNA methylation, such as rheumatoid arthritis, lupus, cancer, and the drugs procainamide and hydralazine. T cells were isolated by e-rosetting as described (Golbus et al., 1988). Where indicated the cells were stimulated with phytohemagglutinin (PHA) and cell numbers expanded for up to 7 days in media supplemented with interleukin 2 (IL-2; Richardson et al., 1990). Where indicated, the cells were treated with Trichostatin A (Sigma, St. Louis, MO).
2.2. AciI methylation analysis PHA stimulated T cells were treated with 1 mM 5-aza-2%deoxycytidine (5-azaC; Fluka, Milwaukee, WI) for 48 h and genomic DNA isolated as described (Kaplan et al., 2000). The DNA was digested with AciI and BglII, fractionated by
Table 1 Characteristics of the donor population Group
Number
Age range (mean 9 S.D.)
Percent CD4+(n) (mean 9S.D.)
Percent CD8+ (n) (mean 9S.D.)
Young Middle age Old
25 16 21
0 23–50 (34 98) 66–86 (77 9 6)
74 913 (10) 759 10 (8) 719 7 (10)
21 914 (10) 23 99 (8) 28 98 (10)
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1259
Fig. 1. ITGAL promoter structure. The 27 potentially methylated CpG pairs in the ITGAL promoter fragment are identified by the ‘lollipops’, and the AciI and BglII sites identified by lines. The transcription and translation start sites are identified by arrows, and the probe used for Southern analysis is identified by the shaded box.
electrophoresis, transferred to nylon filters and hybridized with a 380 bp 32P-labeled Bsm1-Xho1 CD11a promoter fragment using published protocols (Kaplan et al., 2000).
2.3. Bisulfite sequencing Purified T cell DNA (1– 5 mg) was treated with sodium bisulfite (Clark et al., 1994), then the 2.3 kb CD11a promoter fragment (bp 1– 2277, Fig. 1; Cornwell et al., 1993) was amplified, cloned into PBS + (Stratagene, La Jolla, CA), and sequenced by the University of Michigan Sequencing Core.
2.4. Patch methylation A 1.9 kb fragment containing the human ITGAL promoter (bp 151– 2130, kindly provided by Dr Dennis Hickstein) was cloned into pGL3-Basic (Promega, Madison, WI). An Nde1 site was engineered at bp 1568 using the QuikChange site-directed mutagenesis kit (Stratagene), and intact function confirmed by transfection into Jurkat cells (Richardson et al., 1994). The region from the beginning of the fragment to the Nde1 site was excised, methylated with Sss1 and S-adenosylmethionine (both from New England Biolabs, Beverly, MA) using instructions provided by the
manufacturer, then ligated back into the reporter construct and purified by gel electrophoresis. Completeness of methylation was tested by digestion with Aci1 (New England Biolabs). Controls included a mock methylated construct, prepared by omitting the Sss1.
2.5. Transient transfection Plasmid DNA was introduced into Jurkat cells by electroporation as described (Richardson et al., 1994; Kaplan et al., 2000). Twenty-four hour later cells were washed, suspended in Reporter Lysis Buffer (Promega), lysed, and luciferase assays performed (Kaplan et al., 2000). b-Galactosidase determinations were performed using the Galacto-Light system as per the manufacturer’s protocol (Tropix, Bedford, MA).
2.6. Semiquantitati6e RT-PCR Realtime semiquantitative reverse transcriptasepolymerase chain reaction (RT-PCR) was performed using a LightCycler (Roche, Indianapolis, IN) and previously published protocols (Yung et al., 2001a,b). The primers used were, CD11a: forward: 5%-AAATGGAAGGACCCTGATGCTC-3%, backward: 5%-TGTAGCGGATGATGTCTTTGGC-3%;
1260
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Dnmt1 : forward: 5%-GATTTGTCCTTGGAGAACGGTG-3%, backward: 5%-TGAGATGTGATGGTGGTTTGCC-3%; Dnmt3a: forward: 5%-CGAGTCCAACCCTGTGATGATTG-3%, backward: 5%-GCTGGTCTTTGCCCTGCTTTATG-3%; Dnmt3b: forward: 5%-TTGGAATAGGGGACCTCGTGTG-3%, backward: 5%-AGAGACCTCGGAGAACTTGCCATC-3%; Histone H4 : forward: 5%-AGACAACATTCAG G G C A T C A C C A A G C C T G C C A T T - 3 %, backward: 5%-TTGAAGCGCGTACACCACATCCATG-CTGTGA-3%; i-actin: forward: 5%-GCACCACACCTTCTACAATGAGC-3%, backward: 5%-GGATAGCACAGCCTGGATAGCAAC-3%.
2.7. Ribonuclease protection assays (RPAs) RPAs were performed as described (Yung et al., 2001a,b). The probe for L32 was kindly provided by Dr Monte Hobbs (Hobbs et al., 1993). The probes for the methyltransferases were prepared by isolating total RNA from normal human T cells, followed by reverse transcription and amplification with the following primers, Dnmt1 : forward: 5%- GGAGGAGAAGAGACGCAAAAC,G-3%, backward: 5%-AGACGGGTCATCATCATAGATTGG-3%; Dnmt3a: forward: 5%-CCACCAGAAGAAGAGAAGAATCCC-3%, backward: 5%-GTAACATTGAGGCTCCCACAGG-3%; Dnmt3b: forward: 5%-GGCACTGTGGTTTTG, GTATCTTAGC-3%, backward: 5%-TGATTCCTGAAGGCATCCCC-3%. The amplified fragments were cloned into PCRII-TOPO (Invitrogen, Carlsbad, CA) or PGEM-11Z( + ) (Promega), and orientation verified by sequencing. The plasmids were linearized with BamHI or NotI, gel purified, and antisense RNA probes generated using the RiboQuant™ in vitro transcription kit (PharMingen, San Diego, CA). Results are expressed relative to L32.
2.8. Immunoblotting DNA methyltransferase proteins were quantitated by immunoblotting using antibodies to
Dnmt3a from Santa Cruz Biotechnology (Santa Cruz, CA) and to Dnmt1 from Imgenex (San Diego, CA). Nuclear proteins were purified from stimulated T cells, fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), and detected by immunoblotting using previously published protocols (Scheinbart et al., 1991; Grolleau et al., 2000). Bands were quantitated by densitometry as before (Grolleau et al., 2000).
3. Results
3.1. Methylation of the ITGAL promoter in normal indi6iduals Fig. 1 shows a map of the ITGAL promoter, and identifies the CG pairs, restriction enzyme recognition sequences, transcription and translation start sites, and the fragment used in the Southern analyses (vide infra). Fig. 2a shows the methylation pattern of the ITGAL promoter and 5%-flanking region in DNA isolated from the T cells of three healthy middle aged controls, determined by bisulfite sequencing. The transcribed region (3% to bp 1950) was completely demethylated in all fragments from the subjects, while the majority of the sequences 5% to the start site were partially methylated. Of note is a region containing Alu elements, which was more heavily methylated in all the controls, consistent with previous reports demonstrating that repetitive DNA sequences are usually heavily methylated (Schmid and Rubin, 1995). We next asked if 5-azaC affects ITGAL promoter methylation. Since 5-azaC only inhibits DNA methylation during S phase (Jones, 1984), T cells from the donors shown in Fig. 2a were stimulated with PHA for 3 days, then the methylation status of the region from the Alu elements to the transcription start site (bp 689–1882) was determined (Fig. 2b). T cell stimulation had no significant effect on methylation patterns. However, treating the stimulated T cells with 5-azaC demethylated this region (fraction methylated 0.449 0.05 vs. 0.269 0.04, mean9 S.E.M., PHA vs. 5-azaC, P= 0.048, Fig. 2c). Surprisingly, there
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1261
appears to be a modest increase in DNA methylation in the region between bp 1000 and 1500 in the 5-azaC treated cells. This may represent compensatory mechanisms, as previously described by our group (Yang et al., 1997). CD11a mRNA increased 2.6-fold relative to bactin in the 5-azaC treated cells as measured by realtime RT-PCR (0.54 vs. 1.39). The bisulfite sequencing results were confirmed using Southern analysis. DNA from unstimulated cells shows 1.2, 1.0 and 0.4 kb bands, corresponding to 603 (BglII)— 1848/1882 (AciI), 829 (AciI)-1848/1882 (AciI), and 1848/1882 (AciI)— 2264 (AciI), respectively (Fig. 3). PHA stimulation caused no significant change in the bands. However, 5-azaC treatment caused a 39% decrease in the 1.2 kb band, and 35% increase in the 1.0 kb band, corresponding to demethylation of the AciI site at bp 829 in the heavily methylated Alu region. In four serial experiments, the 1.0 kb band density increased 219 5%, and the 1.2 kb band decreased 279 7% (mean9 S.E.M., PB 0.02). These results thus resemble those shown in Fig. 2.
3.2. Effect of methylation on ITGAL promoter function
Fig. 2. Effect of stimulation and 5-azaC on ITGAL promoter methylation. (A) T cell DNA was isolated from healthy donors, treated with sodium bisulfite, then bp 689 to the end of the fragment shown in Fig. 1 was amplified in four overlapping fragments. Five cloned fragments from each section were then sequenced from each donor, and the fraction methylated for each CpG pair averaged across the five cloned and sequenced fragments. The results are presented as the mean of these results from the three donors, such that each value represents a total of 15 determinations. The region containing Alu elements is denoted by the horizontal line. (B) The average methylation from 689 to 1882 was determined in PHA stimulated T cells from the three donors shown for this region in panel A. (C) The PHA stimulated cells shown in panel B were treated with 5-azaC and methylation determined as in panel A. In each case the results again represent the average methylation of five fragments from three individuals.
The region between the start of the promoter and 1568 was excised, methylated in vitro, ligated back into the luciferase reporter construct, and transfected into Jurkat cells, using cotransfection with b-galactosidase as a control. Methylation decreased promoter expression by 40% (Fig. 4, mean9 S.E.M. of four experiments, mock methylated vs. methylated, P= 0.02 by paired t-test). The results are consistent with a functional significance for methylation changes in this region.
3.3. Effect of age on CD11a expression We next compared CD11a mRNA expression in T lymphocytes from newborns, middle aged, and elderly donors. The characteristics of these populations are shown in Table 1. To obtain sufficient cell numbers, the lymphocytes were first stimulated with PHA and numbers ex-
1262
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Fig. 3. Methylation sensitive restriction analysis of ITGAL promoter methylation. T cells from a healthy donor were stimulated with PHA and treated with 5-azaC as in Fig. 2. DNA was isolated from unstimulated cells (C), PHA stimulated cells (P), and 5-azaC treated PHA stimulated cells (P + ), then digested with BglII and AciI, fractionated by agarose gel electrophoresis, transferred to nylon filters and hybridized with the probe fragment shown in Fig. 1, labeled with 32P. Fragment sizes are listed on the right in kb.
panded in IL-2. No methylation inhibitors were added. Table 1 shows no significant differences in the subset composition of the T cell populations between a representative subset of each
group under these conditions (P= 0.6362 for CD4 + and P= 0.2678 for CD8+ , both by analysis of variance (ANOVA)). Since short-term culture of stimulated T cells does not affect methylation levels or patterns (Richardson et al., 1990; Zhu et al., 1999), and the present studies demonstrate that activation and culture of T cells does not affect methylation of the ITGAL promoter and flanking region, it is unlikely that this approach significantly affects methylation of this region. CD11a mRNA levels were then compared between groups (Fig. 5). A 7.5-fold increase is seen from birth (relative expression 30.29 11.7) to middle age (227.29 56.7), and a 2.4-fold increase from middle age (227.29 56.7) to old age (540.79 81.8) (P B0.0001 overall by ANOVA, P= 0.061 newborn vs. middle age, P= 0.0035 middle vs. old age, PB0.0001 newborn vs. old age, by post-hoc testing). LFA-1 expression increases from birth to middle age (Pallis et al., 1993), and the increase in CD11a from birth to middle age is consistent with this. Similarly, LFA-1 increases further from middle age to old age, even on the ‘memory’ subset (Pallis et al., 1993), and our data is also consistent with this.
3.4. Effect of age on ITGAL promoter methylation
Fig. 4. Effect of methylation on ITGAL promoter function. The region between the start of the ITGAL promoter and 1568 was excised, methylated with SssI in vitro, ligated back into a luciferase reporter construct, and transfected into Jurkat cells, using cotransfection with b-galactosidase as a control. Controls (Mock methylated) consisted of similar preparations but without the addition of SssI. The results are presented as the ratio of luciferase/b-galactosidase expression in arbitrary units, and represent the mean 9S.E.M. of four experiments (*, P= 0.02 by paired t-test).
DNA was isolated from newborns, middle age, and old donors, then methylation changes in the heavily methylated region containing the Alu elements was analyzed by Southern analysis as in Fig. 3. Southern analysis was chosen because of its utility in identifying methylation changes in a relatively small region in large numbers of samples. Fig. 6a shows a representative autoradiogram illustrating the relative intensities of the 1.2, 1.0 and 0.4 kb bands from young, middle age, and old donors. A progressive decrease in the 1.2 kb band, and a corresponding increase in the 1.0 kb band is seen. Fig. 6b shows the mean9S.E.M. of the relative intensities of the 1.2 and 1.0 kb bands from six donors in each group plotted against age. The 1.2 kb band decreases markedly from birth to middle age, and somewhat less from middle age to
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1263
Fig. 5. Effect of age on T cell CD11a mRNA expression. RNA was isolated from the same number of T cells from seven newborns, six middle aged donors, and six elderly donors, and CD11a mRNA quantitated using realtime RT-PCR. Results are expressed as arbitrary units, and represent the mean 9S.E.M. of the six to seven determinations. (P B0.0001 overall by ANOVA, P=0.061 newborn vs. middle age, P =0.0035 middle vs. old age, PB 0.0001 newborn vs. old age, by post-hoc testing).
old age, while the 1.0 kb band shows a reciprocal increase. These results indicate a progressive demethylation of the AciI locus at bp 829 (P = 0.0002 overall by ANOVA, P= 0.0012 young vs. middle, P=0.0003 young vs. old) by post hoc testing for the decrease in the 1.2 kb band. Methylation of promoter flanking sequences can suppress gene expression in part by promoting chromatin inactivation. Methylcytosine binding proteins like MeCP2 attract a chromatin inactivation complex containing histone deacetylases, which condense chromatin into a structure inaccessible to transcription factors (Bird et al., 1998). Since newborns had high levels of ITGAL promoter methylation, we tested whether inhibiting histone deacetylases with Trichostatin A would increase ITGAL expression. Newborn T cells were stimulated and cultured as before, then
treated for 3 days with 1 mM 5-azaC or for 24 h with 12.5–100 nM Trichostatin A. CD11a mRNA, quantitated by real-time RT-PCR increased 32% in the 5-azaC treated cells relative to b-actin, and 27% in cells treated with 100 nM Trichostatin A, with lesser amounts at the lower concentrations. The increases were 49% (5-azaC) and 33% (Trichostatin A) in a confirming experiment.
3.5. Effect of age on DNA methyltransferase mRNA and protein The mechanism causing DNA hypomethylation with aging is unknown. One potential mechanism is passive demethylation, in which decreased expression of a maintenance DNA methyltransferase during replication leads to hypomethylation
1264
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
(Richardson and Yung, 1999). RPAs were used to compare expression of DNA methyltransferases 1, 3a and 3b (Dnmt1, Dnmt3a, and Dnmt3b) in individuals from each of the three experimental groups. Two isoforms of Dnmt1 have been described (Hsu et al., 1999), so RPAs quantitating both transcripts were used. Fig. 7a shows a representative RPA assay for Dnmt3a and both isoforms of Dnmt1, Dnmt1(L) and Dnmt1(S), using mRNA from four old donors. The levels of Dnmt3b were too low to permit accurate quantitation (not shown). Levels of Dnmt1 and Dnmt3a decreased with age (Fig. 7b), with a large decrease from young to middle age and a smaller decrease from middle to old age. The decrease from young to old for Dnmt1 was 40% (P B 0.001), and 52%
for Dnmt3a (PB 0.01). The decrease from middle to old age did not reach statistical significance. No differences were observed in the two Dnmt1 isoforms. Similar results were observed in a confirming experiment, with a 55% decrease in Dnmt1 (PB 0.05) and 69% in Dnmt3a (PB 0.001). Dnmt1 is believed to be tightly linked to cell cycle (Lee et al., 1996), so we considered the possibility that the decrease in the methyltransferase transcripts was due to altered cell cycle kinetics. Transcripts for both Dnmt1 and Dnmt3a were compared with histone H4 using RT-PCR. Interestingly, Dnmt1 decreased 58% relative to histone H4 by middle age (P = 0.038 overall by ANOVA). Similarly, Dnmt3a levels decreased 75% relative to histone H4 by middle age (P =
Fig. 6. Effect of age on ITGAL promoter methylation. (A) T cell DNA was isolated from a young (Y), middle age (M) and old (O) donor, then digested with BglII and AciI and analyzed as in Fig. 3. (B) The intensities of the 1.2, 1.0 and 0.4 kb bands were determined, and the relative intensity of the 1.0 and 1.2 kb bands determined as a fraction of the sum of the 3. The relative intensity of the 1.2 kb band (open squares) and the 1.0 kb band (closed diamonds) was then averaged among six young, six middle age and six old subjects. Results are presented as the mean 9 S.E.M. of the six determinations per age group (P = 0.0002 overall by ANOVA, P = 0.0012 young vs. middle, P= 0.0003 young vs. old by post hoc testing for the decrease in the 1.2 kb band).
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1265
Fig. 7. Effect of age on Dnmt 1 and Dnmt3a expression. (A) RPAs for the large and small forms of Dnmt1 (Dnmt1 L and Dnmt1 S, respectively) and Dnmt3a were used to analyze Dnmt expression in RNA from T cells of four old donors. (B) The expression of Dnmt1 (open squares) and Dnmt3a (closed diamonds) was determined by RPA relative to L32 mRNA for four young, four middle age, and four old donors. Results are presented as the mean 9 S.E.M. of the four determinations per age group. (C) The expression of Dnmt1 (open squares) and Dnmt3a (closed diamonds) mRNA was compared with histone H4 mRNA, using RNA from T cells of three young, three middle aged and three old subjects. Results are presented as the mean 9S.E.M. of the three determinations per age group. (D) Nuclear proteins were isolated from the same number of T cells from three young, three middle age, and three old individuals, and Dnmt1 (open squares) and Dnmt3a (closed diamonds) proteins quantitated by immunoblotting. Results are expressed as the mean 9 S.E.M. of the three determinations per group, and have been standardized to the mean of the young group for each enzyme.
0.022; Fig. 7c). In contrast to the RPA assays, no decrease from middle to old age was observed for either transcript, possibly due to differences in the control transcripts with age. No significant change in Dnmt3b transcripts was observed, although levels were low (not shown). The effects of age on Dnmt1 and Dnmt3a
were confirmed using immunoblotting, using nuclear proteins isolated from the same numbers of stimulated T cells from three young, three middle age, and three old donors. Both Dnnt1 and Dnmt3a decrease with age. Overall Dnmt1 decreased 65% (P = 0.034 by ANOVA), and Dnmt3a decreased 55% (P= 0.002). Dnmt1
1266
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
decreased most from birth to middle age, while Dnmt3a appears to decrease continuously over time.
4. Discussion The regulation of LFA-1 expression is complex. LFA-1 levels increase with T cell stimulation, differentiation from naı¨ve to memory subsets, and with aging (Richardson et al., 1992; Pallis et al., 1993; Kern et al., 1994). Rapid changes in gene expression such as those occurring following T cell stimulation are usually not mediated by DNA methylation (Bird and Wolffe, 1999), and our results from PHA stimulated T cells are consistent with this. In contrast, changes in gene expression associated with differentiation occur more slowly, and can involve changes in methylation patterns (Bird and Wolffe, 1999). Since progression from naı¨ve to memory is associated with an increase in LFA-1, a portion of the CD11a mRNA increase may be attributable to differentiation. However, LFA-1 expression continues to increase with age, even on the memory subset (Pallis et al., 1993), suggesting additional mechanisms. The present results suggest that changes in DNA methylation may contribute. We have previously reported that 5-azaC increases CD11a mRNA (Richardson et al., 1992), and the studies reported here show that 5-azaC also decreases methylation of the ITGAL promoter flanking sequences. Further, patch methylating the flanking sequences suppress ITGAL promoter function, indicating that methylation of this region can affect gene expression. While the region changing methylation status is upstream of the active promoter, located in the first 100 bp 5% to the transcription start site (Cornwell et al., 1993), methylation changes can affect gene expression from a distance through methylcytosine binding proteins like MeCP2. MeCP2 contains a transcription repression domain which interacts with a chromatin inactivating complex containing histone deacetylases, condensing the chromatin into an inactive configuration (Bird et al., 1998). The observation that the histone deacetylase in-
hibitor Trichostatin A increased CD11a mRNA levels is consistent with this. We also asked if a decrease in maintenance DNA methyltransferases might contribute to hypomethylation with aging. When compared with histone H4, transcripts for Dnmt1 and Dnmt3a decreased most significantly from birth to middle age, in contrast to reports that Dnmt1 levels are tightly linked to the cell cycle (Lee et al., 1996). The decreases were confirmed at the protein level. This raises the possibility that decreases in Dnmt1 and possibly Dnmt3a may lead to imperfect replication of methylation patterns during mitosis. It should also be noted that while the literature indicates that histone H4 is tightly linked to the cell cycle (Lee et al., 1996), the effect of aging on this transcript has not been studied. The magnitude of the changes in ITGAL promoter methylation, CD11a mRNA expression, and methyltransferase transcript levels were all greater from newborn to middle age groups than from middle to old age. One explanation is that a decrease in DNA methyltransferases is responsible, since the changes are parallel. However, it is also possible that maturational changes contribute between birth and middle age, while other mechanisms contribute to the change from middle to old age. It should also be noted that since T cells are continually replenished through life, the middle and old age groups will contain a mixture of cells of various ages, with the oldest group having the greatest number of old cells. This may contribute to heterogeneity in the populations as well. Together, these results show a relationship between the progressive age-dependent increase in CD11a and a progressive demethylation of the ITGAL flanking sequences with aging. The results also show that methylation of the flanking region can suppress ITGAL promoter function. These results thus support the hypothesis that age-dependent decreases in T cell DNA methylation, similar to increases, can contribute to the changes in T cell function and gene expression that occur with aging. Since LFA-1 overexpression can induce anti-DNA antibodies, these changes may also contribute to the development of anti-nuclear antibodies with aging.
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Acknowledgements This work was supported by PHS grants AG08808, AG014783, AR42525, and AI42753, and a Merit grant from the Department of Veterans Affairs. The authors thank Dr Dennis Hickstein and Dr Monte Hobbs for providing critical reagents, Dr John Attwood for his critical review of this manuscript, and Janet Stevens for her expert secretarial assistance.
References Bird, A.P., Wolffe, A.P., 1999. Methylation-induced repression— belts, braces, and chromatin. Cell 99 (5), 451 –454. Bird, J.J., Brown, D.R., Mullen, A.C., Moskowitz, N.H., Mahowald, M.A., Sider, J.R., Gajewski, T.F., Wang, C.R., Reiner, S.L., 1998. Helper T cell differentiation is controlled by the cell cycle. Immunity 9 (2), 229 –237. Clark, S.J., Harrison, J., Paul, C.L., Frommer, M., 1994. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22 (15), 2990 –2997. Cornwell, R.D., Gollahon, K.A., Hickstein, D.D., 1993. Description of the leukocyte function-associated antigen 1 (LFA-1 or CD11a) promoter. Proc. Natl. Acad. Sci. USA 90 (9), 4221 – 4225. Golbus, J., Salata, M., Greenwood, J., Hudson, J., Richardson, B.C., 1988. Increased immunoglobulin response to g-interferon by lymphocytes from patients with systemic lupus erythematosus. Clin. Immunol. Immunopathol. 46 (1), 129 –140. Golbus, J., Palella, T.D., Richardson, B.C., 1990. Quantitative changes in T cell DNA methylation occur during differentiation and ageing. Eur. J. Immunol. 20 (8), 1869 – 1872. Grolleau, A., Kaplan, M.J., Hanash, S.M., Beretta, L., Richardson, B., 2000. Impaired translational response and increased protein kinase PKR expression in T cells from lupus patients. J. Clin. Invest. 106 (12), 1561 –1568. Guarente, L., 1996. Do changes in chromosomes cause aging. Cell 86 (1), 9 –12. Hobbs, M.V., Weigle, W.O., Noonan, D.J., Torbett, B.E., Koch, R.J., Cardenas, G.J., Ernst, D.N., 1993. Patterns of cytokine gene expression by CD4 + T cells from young and old mice. J. Immunol. 150 (8 Pt. 1), 3602 – 3614. Hsu, D.W., Lin, M.J., Lee, T.L., Wen, S.C., Chen, X., Shen, C.K., 1999. Two major forms of DNA (cytosine-5) methyltransferase in human somatic tissues. Proc. Natl. Acad. Sci. USA 96 (17), 9751 –9756. Issa, J.P., 1999. Aging, DNA methylation and cancer. Crit. Rev. Oncol. Hematol. 1 (32), 31 – 43. Jones, P.A., 1984. Gene activation by 5-azacytidine. In:
1267
Razin, A., Cedar, H., Riggs, A.D. (Eds.), DNA Methylation, Biochemistry and Biological Significance. Springer, New York, pp. 163 – 187. Kaplan, M.J., Deng, C., Yang, J., Richardson, B.C., 2000. DNA methylation in the regulation of T cell LFA-1 expression. Immunol. Invest. 29 (4), 411 – 425. Kern, F., Docke, W.D., Reinke, P., Volk, H.D., 1994. Discordant expression of LFA-1, VLA-4a, VLA-b 1, CD45RO and CD28 on T-cell subsets: evidence for multiple subsets of ‘memory’ T cells. Int. Arch. Allergy Immunol. 104 (1), 17 – 26. Lee, P.J., Washer, L.L., Law, D.J., Boland, C.R., Horon, I.L., Feinberg, A.P., 1996. Limited up-regulation of DNA methyltransferase in human colon cancer reflecting increased cell proliferation. Proc. Natl. Acad. Sci. USA 93 (19), 10366 – 10370. Pallis, M., Robins, A., Powell, R., 1993. Quantitative analysis of lymphocyte CD11a using standardized flow cytometry. Scand. J. Immunol. 38 (6), 559 – 564. Richardson, B., Yung, R., 1999. Role of DNA methylation in the regulation of cell function. J. Lab. Clin. Med. 134 (4), 333 – 340. Richardson, B., Scheinbart, L., Strahler, J., Gross, L., Hanash, S., Johnson, M., 1990. Evidence for impaired T cell DNA methylation in systemic lupus erythematosus and rheumatoid arthritis. Arthritis Rheum. 33 (11), 1665 – 1673. Richardson, B.C., Strahler, J.R., Pivirotto, T.S., Quddus, J., Bayliss, G.E., Gross, L.A., O’Rourke, K.S., Powers, D., Hanash, S.M., Johnson, M.A., 1992. Phenotypic and functional similarities between 5-azacytidine-treated T cells and a T cell subset in patients with active systemic lupus erythematosus. Arthritis Rheum. 35 (6), 647 – 662. Richardson, B., Powers, D., Hooper, F., Yung, R.L., O’Rourke, K., 1994. Lymphocyte function-associated antigen 1 overexpression and T cell autoreactivity. Arthritis Rheum. 37 (9), 1363 – 1372. Scheinbart, L.S., Johnson, M.A., Gross, L.A., Edelstein, S.R., Richardson, B.C., 1991. Procainamide inhibits DNA methyltransferase in a human T cell line. J. Rheumatol. 18 (4), 530 – 534. Schmid, C.W., Rubin, C.M., 1995. Alu: What is the use. In: The Impact of Short Interspersed Elements (SINEs) on the Host Genome. Maraia and Landes, pp. 105 – 123. Yung, R., Powers, D., Johnson, K., Amento, E., Carr, D., Laing, T., Yang, J., Chang, S., Hemati, N., Richardson, B., 1996. Mechanisms of drug-induced lupus. II. T cells overexpressing lymphocyte function-associated antigen 1 become autoreactive and cause a lupuslike disease in syngeneic mice. J. Clin. Invest. 97 (12), 2866 – 2871. Yang, J., Deng, C., Hemati, N., Hanash, S.M., Richardson, B.C., 1997. Effect of mitogenic stimulation and DNA methylation on human T cell DNA methyltransferase expression and activity. J. Immunol. 159 (3), 1303 – 1309. Yung, R., Kaplan, M.J., Ray, D., Schneider, K., Mo, R.-R., Johnson, K., Richardson, B., 2001a. Autoreactive murine Th1 and Th2 cells kill syngeneic macrophages and induce autoantibodies. Lupus. 10, 539 – 546.
1268
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Yung, R., Ray, D., Eisenbraun, J.K., Deng, C., Attwood, J., Eisenbraun, M.D., Johnson, K., Miller, R.A., Hanash, S., Richardson, B., 2001b. Unexpected effects of a heterozygous Dnmt1 null mutation on age-dependent DNA hypomethylation and autoimmunity. J. Gerontol. A Biol. Sci. Med. Sci. 56 (6), B268 – B276.
Zhu, X., Deng, C., Kuick, R., Yung, R., Lamb, B., Neel, J.V., Richardson, B., Hanash, S., 1999. Analysis of human peripheral blood T cells and single-cell-derived T cell clones uncovers extensive clonal CpG island methylation heterogeneity throughout the genome. Proc. Natl. Acad. Sci. USA 96 (14), 8058 – 8063.
Age-dependent DNA methylation changes in the ITGAL (CD11a) promoter Zhiyong Zhang, Chun Deng, Qianjin Lu, Bruce Richardson * Department of Medicine, Veterans Affairs Hospital, 5310 Cancer Center and Geriatrics Center Building, Uni6ersity of Michigan, Ann Arbor, MI 48109 -0940, USA Received 13 September 2001; received in revised form 4 February 2002; accepted 5 February 2002
Abstract DNA methylation patterns change with age in a complex fashion, typically with an overall decrease in genomic deoxymethylcytosine (dmC) content, but with local increases in some promoters that contain GC-rich sequences known as CpG islands. While the consequences of age-dependent CpG island methylation have recently been studied in organs such as the colon, less is known about the functional significance of the progressive hypomethylation of promoters lacking CpG islands, and the significance of age-dependent changes in T cell DNA methylation is completely unexplored. We asked if age-dependent DNA hypomethylation might contribute to overexpression of the T cell ITGAL gene, which encodes CD11a, a subunit of LFA-1. CD11a mRNA increased with age as well as with experimentally induced DNA hypomethylation. This increase correlated with hypomethylation of sequences flanking the ITGAL promoter in vitro and in aging. ‘Patch’ methylation of the region suppressed promoter function. DNA methyltransferases 1 and 3a also decreased with aging. These results indicate that hypomethylation of regions flanking the ITGAL promoter may increase CD11a expression, and suggest that age-dependent hypomethylation of promoters lacking CpG islands, perhaps due to decreased DNA methyltransferase expression, may be one mechanism contributing to increased T cell gene expression with aging. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: DNA methylation; ITGAL; LFA-1; T lymphocytes; Aging
1. Introduction A number of biochemical changes occur in DNA over time, and may contribute to the alterations in gene expression associated with aging. These changes include oxidative damage, telomere shortening, and the accumulation of mutations * Corresponding author. Tel.: +1-734-763-6940; fax: + 1743-936-9220. E-mail address: [email protected] (B. Richardson).
(Guarente, 1996). Another process associated with aging is the progressive alteration of DNA methylation patterns. Methylation patterns are established during ontogeny, then change with aging in a complex fashion. In general, total genomic deoxymethylcytosine (dmC) content decreases with age, and involves genes as well as repetitive sequences. However, the transcriptional relevance of age-dependent DNA demethylation has generally not been examined. Paradoxically, this progressive demethylation is associated with
0047-6374/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S0047-6374(02)00014-3
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1258
abnormal hypermethylation of CpG islands. Agedependent CpG island hypermethylation typically results in suppression of the associated gene, and may contribute to carcinogenesis by inhibiting expression of genes with various growth suppressing functions (Issa, 1999). Immune abnormalities also occur with aging, and the mechanisms causing the changes are unknown. The abnormalities include the progressive development of T cell anergy, as well as the development of autoimmunity and anti-nuclear antibodies (Yung et al., 2001a,b). We have reported that human T lymphocyte DNA demethylates with age (Golbus et al., 1990), that by middle age approximately 15% of T cell CpG islands are variably methylated on a clonal basis, and that the increased CpG island methylation is associated with decreased transcription of the associated gene (Zhu et al., 1999). Further, using murine models, we have demonstrated an association between progressive DNA hypomethylation and the development of autoimmunity and some signs of immune senescence (Yung et al., 2001a,b). This suggests that some of the functional changes associated with immune senescence may be due to altered DNA methylation patterns. To further define the role that altered DNA methylation plays in age-associated changes in T cell gene expression, we asked if age-dependent hypomethylation occurs in transcriptionally relevant sequences of a gene that lacks a CpG island and that increases expression with aging. We selected CD11a, one chain of the b2 integrin lymphocyte function associated antigen 1 (LFA-1) also referred to as CD11a/CD18 or aLb2. T cell LFA-1 increases progressively throughout life (Pallis et al., 1993), and overexpression of T cell LFA-1 causes autoimmunity and anti-DNA antibodies (Yung et al., 1996). Further, inhibiting T
cell DNA methylation with DNA methyltransferase inhibitors like 5-azacytidine increases LFA1 expression through effects on CD11a (Kaplan et al., 2000; Richardson et al., 1992). Our results support the contention that progressive DNA hypomethylation contributes to changes in T cell gene expression with age.
2. Materials and methods
2.1. Donor populations Cord blood was obtained from placentas of healthy newborns. Peripheral venous blood was obtained from healthy middle aged donors, and from elderly donors 65 years of age and older recruited through the Geriatrics Center Research Participant Program, sponsored by the Claude D. Pepper Older Americans Independence Center (Table 1). Exclusions included conditions and drugs known to affect DNA methylation, such as rheumatoid arthritis, lupus, cancer, and the drugs procainamide and hydralazine. T cells were isolated by e-rosetting as described (Golbus et al., 1988). Where indicated the cells were stimulated with phytohemagglutinin (PHA) and cell numbers expanded for up to 7 days in media supplemented with interleukin 2 (IL-2; Richardson et al., 1990). Where indicated, the cells were treated with Trichostatin A (Sigma, St. Louis, MO).
2.2. AciI methylation analysis PHA stimulated T cells were treated with 1 mM 5-aza-2%deoxycytidine (5-azaC; Fluka, Milwaukee, WI) for 48 h and genomic DNA isolated as described (Kaplan et al., 2000). The DNA was digested with AciI and BglII, fractionated by
Table 1 Characteristics of the donor population Group
Number
Age range (mean 9 S.D.)
Percent CD4+(n) (mean 9S.D.)
Percent CD8+ (n) (mean 9S.D.)
Young Middle age Old
25 16 21
0 23–50 (34 98) 66–86 (77 9 6)
74 913 (10) 759 10 (8) 719 7 (10)
21 914 (10) 23 99 (8) 28 98 (10)
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1259
Fig. 1. ITGAL promoter structure. The 27 potentially methylated CpG pairs in the ITGAL promoter fragment are identified by the ‘lollipops’, and the AciI and BglII sites identified by lines. The transcription and translation start sites are identified by arrows, and the probe used for Southern analysis is identified by the shaded box.
electrophoresis, transferred to nylon filters and hybridized with a 380 bp 32P-labeled Bsm1-Xho1 CD11a promoter fragment using published protocols (Kaplan et al., 2000).
2.3. Bisulfite sequencing Purified T cell DNA (1– 5 mg) was treated with sodium bisulfite (Clark et al., 1994), then the 2.3 kb CD11a promoter fragment (bp 1– 2277, Fig. 1; Cornwell et al., 1993) was amplified, cloned into PBS + (Stratagene, La Jolla, CA), and sequenced by the University of Michigan Sequencing Core.
2.4. Patch methylation A 1.9 kb fragment containing the human ITGAL promoter (bp 151– 2130, kindly provided by Dr Dennis Hickstein) was cloned into pGL3-Basic (Promega, Madison, WI). An Nde1 site was engineered at bp 1568 using the QuikChange site-directed mutagenesis kit (Stratagene), and intact function confirmed by transfection into Jurkat cells (Richardson et al., 1994). The region from the beginning of the fragment to the Nde1 site was excised, methylated with Sss1 and S-adenosylmethionine (both from New England Biolabs, Beverly, MA) using instructions provided by the
manufacturer, then ligated back into the reporter construct and purified by gel electrophoresis. Completeness of methylation was tested by digestion with Aci1 (New England Biolabs). Controls included a mock methylated construct, prepared by omitting the Sss1.
2.5. Transient transfection Plasmid DNA was introduced into Jurkat cells by electroporation as described (Richardson et al., 1994; Kaplan et al., 2000). Twenty-four hour later cells were washed, suspended in Reporter Lysis Buffer (Promega), lysed, and luciferase assays performed (Kaplan et al., 2000). b-Galactosidase determinations were performed using the Galacto-Light system as per the manufacturer’s protocol (Tropix, Bedford, MA).
2.6. Semiquantitati6e RT-PCR Realtime semiquantitative reverse transcriptasepolymerase chain reaction (RT-PCR) was performed using a LightCycler (Roche, Indianapolis, IN) and previously published protocols (Yung et al., 2001a,b). The primers used were, CD11a: forward: 5%-AAATGGAAGGACCCTGATGCTC-3%, backward: 5%-TGTAGCGGATGATGTCTTTGGC-3%;
1260
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Dnmt1 : forward: 5%-GATTTGTCCTTGGAGAACGGTG-3%, backward: 5%-TGAGATGTGATGGTGGTTTGCC-3%; Dnmt3a: forward: 5%-CGAGTCCAACCCTGTGATGATTG-3%, backward: 5%-GCTGGTCTTTGCCCTGCTTTATG-3%; Dnmt3b: forward: 5%-TTGGAATAGGGGACCTCGTGTG-3%, backward: 5%-AGAGACCTCGGAGAACTTGCCATC-3%; Histone H4 : forward: 5%-AGACAACATTCAG G G C A T C A C C A A G C C T G C C A T T - 3 %, backward: 5%-TTGAAGCGCGTACACCACATCCATG-CTGTGA-3%; i-actin: forward: 5%-GCACCACACCTTCTACAATGAGC-3%, backward: 5%-GGATAGCACAGCCTGGATAGCAAC-3%.
2.7. Ribonuclease protection assays (RPAs) RPAs were performed as described (Yung et al., 2001a,b). The probe for L32 was kindly provided by Dr Monte Hobbs (Hobbs et al., 1993). The probes for the methyltransferases were prepared by isolating total RNA from normal human T cells, followed by reverse transcription and amplification with the following primers, Dnmt1 : forward: 5%- GGAGGAGAAGAGACGCAAAAC,G-3%, backward: 5%-AGACGGGTCATCATCATAGATTGG-3%; Dnmt3a: forward: 5%-CCACCAGAAGAAGAGAAGAATCCC-3%, backward: 5%-GTAACATTGAGGCTCCCACAGG-3%; Dnmt3b: forward: 5%-GGCACTGTGGTTTTG, GTATCTTAGC-3%, backward: 5%-TGATTCCTGAAGGCATCCCC-3%. The amplified fragments were cloned into PCRII-TOPO (Invitrogen, Carlsbad, CA) or PGEM-11Z( + ) (Promega), and orientation verified by sequencing. The plasmids were linearized with BamHI or NotI, gel purified, and antisense RNA probes generated using the RiboQuant™ in vitro transcription kit (PharMingen, San Diego, CA). Results are expressed relative to L32.
2.8. Immunoblotting DNA methyltransferase proteins were quantitated by immunoblotting using antibodies to
Dnmt3a from Santa Cruz Biotechnology (Santa Cruz, CA) and to Dnmt1 from Imgenex (San Diego, CA). Nuclear proteins were purified from stimulated T cells, fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), and detected by immunoblotting using previously published protocols (Scheinbart et al., 1991; Grolleau et al., 2000). Bands were quantitated by densitometry as before (Grolleau et al., 2000).
3. Results
3.1. Methylation of the ITGAL promoter in normal indi6iduals Fig. 1 shows a map of the ITGAL promoter, and identifies the CG pairs, restriction enzyme recognition sequences, transcription and translation start sites, and the fragment used in the Southern analyses (vide infra). Fig. 2a shows the methylation pattern of the ITGAL promoter and 5%-flanking region in DNA isolated from the T cells of three healthy middle aged controls, determined by bisulfite sequencing. The transcribed region (3% to bp 1950) was completely demethylated in all fragments from the subjects, while the majority of the sequences 5% to the start site were partially methylated. Of note is a region containing Alu elements, which was more heavily methylated in all the controls, consistent with previous reports demonstrating that repetitive DNA sequences are usually heavily methylated (Schmid and Rubin, 1995). We next asked if 5-azaC affects ITGAL promoter methylation. Since 5-azaC only inhibits DNA methylation during S phase (Jones, 1984), T cells from the donors shown in Fig. 2a were stimulated with PHA for 3 days, then the methylation status of the region from the Alu elements to the transcription start site (bp 689–1882) was determined (Fig. 2b). T cell stimulation had no significant effect on methylation patterns. However, treating the stimulated T cells with 5-azaC demethylated this region (fraction methylated 0.449 0.05 vs. 0.269 0.04, mean9 S.E.M., PHA vs. 5-azaC, P= 0.048, Fig. 2c). Surprisingly, there
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1261
appears to be a modest increase in DNA methylation in the region between bp 1000 and 1500 in the 5-azaC treated cells. This may represent compensatory mechanisms, as previously described by our group (Yang et al., 1997). CD11a mRNA increased 2.6-fold relative to bactin in the 5-azaC treated cells as measured by realtime RT-PCR (0.54 vs. 1.39). The bisulfite sequencing results were confirmed using Southern analysis. DNA from unstimulated cells shows 1.2, 1.0 and 0.4 kb bands, corresponding to 603 (BglII)— 1848/1882 (AciI), 829 (AciI)-1848/1882 (AciI), and 1848/1882 (AciI)— 2264 (AciI), respectively (Fig. 3). PHA stimulation caused no significant change in the bands. However, 5-azaC treatment caused a 39% decrease in the 1.2 kb band, and 35% increase in the 1.0 kb band, corresponding to demethylation of the AciI site at bp 829 in the heavily methylated Alu region. In four serial experiments, the 1.0 kb band density increased 219 5%, and the 1.2 kb band decreased 279 7% (mean9 S.E.M., PB 0.02). These results thus resemble those shown in Fig. 2.
3.2. Effect of methylation on ITGAL promoter function
Fig. 2. Effect of stimulation and 5-azaC on ITGAL promoter methylation. (A) T cell DNA was isolated from healthy donors, treated with sodium bisulfite, then bp 689 to the end of the fragment shown in Fig. 1 was amplified in four overlapping fragments. Five cloned fragments from each section were then sequenced from each donor, and the fraction methylated for each CpG pair averaged across the five cloned and sequenced fragments. The results are presented as the mean of these results from the three donors, such that each value represents a total of 15 determinations. The region containing Alu elements is denoted by the horizontal line. (B) The average methylation from 689 to 1882 was determined in PHA stimulated T cells from the three donors shown for this region in panel A. (C) The PHA stimulated cells shown in panel B were treated with 5-azaC and methylation determined as in panel A. In each case the results again represent the average methylation of five fragments from three individuals.
The region between the start of the promoter and 1568 was excised, methylated in vitro, ligated back into the luciferase reporter construct, and transfected into Jurkat cells, using cotransfection with b-galactosidase as a control. Methylation decreased promoter expression by 40% (Fig. 4, mean9 S.E.M. of four experiments, mock methylated vs. methylated, P= 0.02 by paired t-test). The results are consistent with a functional significance for methylation changes in this region.
3.3. Effect of age on CD11a expression We next compared CD11a mRNA expression in T lymphocytes from newborns, middle aged, and elderly donors. The characteristics of these populations are shown in Table 1. To obtain sufficient cell numbers, the lymphocytes were first stimulated with PHA and numbers ex-
1262
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Fig. 3. Methylation sensitive restriction analysis of ITGAL promoter methylation. T cells from a healthy donor were stimulated with PHA and treated with 5-azaC as in Fig. 2. DNA was isolated from unstimulated cells (C), PHA stimulated cells (P), and 5-azaC treated PHA stimulated cells (P + ), then digested with BglII and AciI, fractionated by agarose gel electrophoresis, transferred to nylon filters and hybridized with the probe fragment shown in Fig. 1, labeled with 32P. Fragment sizes are listed on the right in kb.
panded in IL-2. No methylation inhibitors were added. Table 1 shows no significant differences in the subset composition of the T cell populations between a representative subset of each
group under these conditions (P= 0.6362 for CD4 + and P= 0.2678 for CD8+ , both by analysis of variance (ANOVA)). Since short-term culture of stimulated T cells does not affect methylation levels or patterns (Richardson et al., 1990; Zhu et al., 1999), and the present studies demonstrate that activation and culture of T cells does not affect methylation of the ITGAL promoter and flanking region, it is unlikely that this approach significantly affects methylation of this region. CD11a mRNA levels were then compared between groups (Fig. 5). A 7.5-fold increase is seen from birth (relative expression 30.29 11.7) to middle age (227.29 56.7), and a 2.4-fold increase from middle age (227.29 56.7) to old age (540.79 81.8) (P B0.0001 overall by ANOVA, P= 0.061 newborn vs. middle age, P= 0.0035 middle vs. old age, PB0.0001 newborn vs. old age, by post-hoc testing). LFA-1 expression increases from birth to middle age (Pallis et al., 1993), and the increase in CD11a from birth to middle age is consistent with this. Similarly, LFA-1 increases further from middle age to old age, even on the ‘memory’ subset (Pallis et al., 1993), and our data is also consistent with this.
3.4. Effect of age on ITGAL promoter methylation
Fig. 4. Effect of methylation on ITGAL promoter function. The region between the start of the ITGAL promoter and 1568 was excised, methylated with SssI in vitro, ligated back into a luciferase reporter construct, and transfected into Jurkat cells, using cotransfection with b-galactosidase as a control. Controls (Mock methylated) consisted of similar preparations but without the addition of SssI. The results are presented as the ratio of luciferase/b-galactosidase expression in arbitrary units, and represent the mean 9S.E.M. of four experiments (*, P= 0.02 by paired t-test).
DNA was isolated from newborns, middle age, and old donors, then methylation changes in the heavily methylated region containing the Alu elements was analyzed by Southern analysis as in Fig. 3. Southern analysis was chosen because of its utility in identifying methylation changes in a relatively small region in large numbers of samples. Fig. 6a shows a representative autoradiogram illustrating the relative intensities of the 1.2, 1.0 and 0.4 kb bands from young, middle age, and old donors. A progressive decrease in the 1.2 kb band, and a corresponding increase in the 1.0 kb band is seen. Fig. 6b shows the mean9S.E.M. of the relative intensities of the 1.2 and 1.0 kb bands from six donors in each group plotted against age. The 1.2 kb band decreases markedly from birth to middle age, and somewhat less from middle age to
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1263
Fig. 5. Effect of age on T cell CD11a mRNA expression. RNA was isolated from the same number of T cells from seven newborns, six middle aged donors, and six elderly donors, and CD11a mRNA quantitated using realtime RT-PCR. Results are expressed as arbitrary units, and represent the mean 9S.E.M. of the six to seven determinations. (P B0.0001 overall by ANOVA, P=0.061 newborn vs. middle age, P =0.0035 middle vs. old age, PB 0.0001 newborn vs. old age, by post-hoc testing).
old age, while the 1.0 kb band shows a reciprocal increase. These results indicate a progressive demethylation of the AciI locus at bp 829 (P = 0.0002 overall by ANOVA, P= 0.0012 young vs. middle, P=0.0003 young vs. old) by post hoc testing for the decrease in the 1.2 kb band. Methylation of promoter flanking sequences can suppress gene expression in part by promoting chromatin inactivation. Methylcytosine binding proteins like MeCP2 attract a chromatin inactivation complex containing histone deacetylases, which condense chromatin into a structure inaccessible to transcription factors (Bird et al., 1998). Since newborns had high levels of ITGAL promoter methylation, we tested whether inhibiting histone deacetylases with Trichostatin A would increase ITGAL expression. Newborn T cells were stimulated and cultured as before, then
treated for 3 days with 1 mM 5-azaC or for 24 h with 12.5–100 nM Trichostatin A. CD11a mRNA, quantitated by real-time RT-PCR increased 32% in the 5-azaC treated cells relative to b-actin, and 27% in cells treated with 100 nM Trichostatin A, with lesser amounts at the lower concentrations. The increases were 49% (5-azaC) and 33% (Trichostatin A) in a confirming experiment.
3.5. Effect of age on DNA methyltransferase mRNA and protein The mechanism causing DNA hypomethylation with aging is unknown. One potential mechanism is passive demethylation, in which decreased expression of a maintenance DNA methyltransferase during replication leads to hypomethylation
1264
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
(Richardson and Yung, 1999). RPAs were used to compare expression of DNA methyltransferases 1, 3a and 3b (Dnmt1, Dnmt3a, and Dnmt3b) in individuals from each of the three experimental groups. Two isoforms of Dnmt1 have been described (Hsu et al., 1999), so RPAs quantitating both transcripts were used. Fig. 7a shows a representative RPA assay for Dnmt3a and both isoforms of Dnmt1, Dnmt1(L) and Dnmt1(S), using mRNA from four old donors. The levels of Dnmt3b were too low to permit accurate quantitation (not shown). Levels of Dnmt1 and Dnmt3a decreased with age (Fig. 7b), with a large decrease from young to middle age and a smaller decrease from middle to old age. The decrease from young to old for Dnmt1 was 40% (P B 0.001), and 52%
for Dnmt3a (PB 0.01). The decrease from middle to old age did not reach statistical significance. No differences were observed in the two Dnmt1 isoforms. Similar results were observed in a confirming experiment, with a 55% decrease in Dnmt1 (PB 0.05) and 69% in Dnmt3a (PB 0.001). Dnmt1 is believed to be tightly linked to cell cycle (Lee et al., 1996), so we considered the possibility that the decrease in the methyltransferase transcripts was due to altered cell cycle kinetics. Transcripts for both Dnmt1 and Dnmt3a were compared with histone H4 using RT-PCR. Interestingly, Dnmt1 decreased 58% relative to histone H4 by middle age (P = 0.038 overall by ANOVA). Similarly, Dnmt3a levels decreased 75% relative to histone H4 by middle age (P =
Fig. 6. Effect of age on ITGAL promoter methylation. (A) T cell DNA was isolated from a young (Y), middle age (M) and old (O) donor, then digested with BglII and AciI and analyzed as in Fig. 3. (B) The intensities of the 1.2, 1.0 and 0.4 kb bands were determined, and the relative intensity of the 1.0 and 1.2 kb bands determined as a fraction of the sum of the 3. The relative intensity of the 1.2 kb band (open squares) and the 1.0 kb band (closed diamonds) was then averaged among six young, six middle age and six old subjects. Results are presented as the mean 9 S.E.M. of the six determinations per age group (P = 0.0002 overall by ANOVA, P = 0.0012 young vs. middle, P= 0.0003 young vs. old by post hoc testing for the decrease in the 1.2 kb band).
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
1265
Fig. 7. Effect of age on Dnmt 1 and Dnmt3a expression. (A) RPAs for the large and small forms of Dnmt1 (Dnmt1 L and Dnmt1 S, respectively) and Dnmt3a were used to analyze Dnmt expression in RNA from T cells of four old donors. (B) The expression of Dnmt1 (open squares) and Dnmt3a (closed diamonds) was determined by RPA relative to L32 mRNA for four young, four middle age, and four old donors. Results are presented as the mean 9 S.E.M. of the four determinations per age group. (C) The expression of Dnmt1 (open squares) and Dnmt3a (closed diamonds) mRNA was compared with histone H4 mRNA, using RNA from T cells of three young, three middle aged and three old subjects. Results are presented as the mean 9S.E.M. of the three determinations per age group. (D) Nuclear proteins were isolated from the same number of T cells from three young, three middle age, and three old individuals, and Dnmt1 (open squares) and Dnmt3a (closed diamonds) proteins quantitated by immunoblotting. Results are expressed as the mean 9 S.E.M. of the three determinations per group, and have been standardized to the mean of the young group for each enzyme.
0.022; Fig. 7c). In contrast to the RPA assays, no decrease from middle to old age was observed for either transcript, possibly due to differences in the control transcripts with age. No significant change in Dnmt3b transcripts was observed, although levels were low (not shown). The effects of age on Dnmt1 and Dnmt3a
were confirmed using immunoblotting, using nuclear proteins isolated from the same numbers of stimulated T cells from three young, three middle age, and three old donors. Both Dnnt1 and Dnmt3a decrease with age. Overall Dnmt1 decreased 65% (P = 0.034 by ANOVA), and Dnmt3a decreased 55% (P= 0.002). Dnmt1
1266
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
decreased most from birth to middle age, while Dnmt3a appears to decrease continuously over time.
4. Discussion The regulation of LFA-1 expression is complex. LFA-1 levels increase with T cell stimulation, differentiation from naı¨ve to memory subsets, and with aging (Richardson et al., 1992; Pallis et al., 1993; Kern et al., 1994). Rapid changes in gene expression such as those occurring following T cell stimulation are usually not mediated by DNA methylation (Bird and Wolffe, 1999), and our results from PHA stimulated T cells are consistent with this. In contrast, changes in gene expression associated with differentiation occur more slowly, and can involve changes in methylation patterns (Bird and Wolffe, 1999). Since progression from naı¨ve to memory is associated with an increase in LFA-1, a portion of the CD11a mRNA increase may be attributable to differentiation. However, LFA-1 expression continues to increase with age, even on the memory subset (Pallis et al., 1993), suggesting additional mechanisms. The present results suggest that changes in DNA methylation may contribute. We have previously reported that 5-azaC increases CD11a mRNA (Richardson et al., 1992), and the studies reported here show that 5-azaC also decreases methylation of the ITGAL promoter flanking sequences. Further, patch methylating the flanking sequences suppress ITGAL promoter function, indicating that methylation of this region can affect gene expression. While the region changing methylation status is upstream of the active promoter, located in the first 100 bp 5% to the transcription start site (Cornwell et al., 1993), methylation changes can affect gene expression from a distance through methylcytosine binding proteins like MeCP2. MeCP2 contains a transcription repression domain which interacts with a chromatin inactivating complex containing histone deacetylases, condensing the chromatin into an inactive configuration (Bird et al., 1998). The observation that the histone deacetylase in-
hibitor Trichostatin A increased CD11a mRNA levels is consistent with this. We also asked if a decrease in maintenance DNA methyltransferases might contribute to hypomethylation with aging. When compared with histone H4, transcripts for Dnmt1 and Dnmt3a decreased most significantly from birth to middle age, in contrast to reports that Dnmt1 levels are tightly linked to the cell cycle (Lee et al., 1996). The decreases were confirmed at the protein level. This raises the possibility that decreases in Dnmt1 and possibly Dnmt3a may lead to imperfect replication of methylation patterns during mitosis. It should also be noted that while the literature indicates that histone H4 is tightly linked to the cell cycle (Lee et al., 1996), the effect of aging on this transcript has not been studied. The magnitude of the changes in ITGAL promoter methylation, CD11a mRNA expression, and methyltransferase transcript levels were all greater from newborn to middle age groups than from middle to old age. One explanation is that a decrease in DNA methyltransferases is responsible, since the changes are parallel. However, it is also possible that maturational changes contribute between birth and middle age, while other mechanisms contribute to the change from middle to old age. It should also be noted that since T cells are continually replenished through life, the middle and old age groups will contain a mixture of cells of various ages, with the oldest group having the greatest number of old cells. This may contribute to heterogeneity in the populations as well. Together, these results show a relationship between the progressive age-dependent increase in CD11a and a progressive demethylation of the ITGAL flanking sequences with aging. The results also show that methylation of the flanking region can suppress ITGAL promoter function. These results thus support the hypothesis that age-dependent decreases in T cell DNA methylation, similar to increases, can contribute to the changes in T cell function and gene expression that occur with aging. Since LFA-1 overexpression can induce anti-DNA antibodies, these changes may also contribute to the development of anti-nuclear antibodies with aging.
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Acknowledgements This work was supported by PHS grants AG08808, AG014783, AR42525, and AI42753, and a Merit grant from the Department of Veterans Affairs. The authors thank Dr Dennis Hickstein and Dr Monte Hobbs for providing critical reagents, Dr John Attwood for his critical review of this manuscript, and Janet Stevens for her expert secretarial assistance.
References Bird, A.P., Wolffe, A.P., 1999. Methylation-induced repression— belts, braces, and chromatin. Cell 99 (5), 451 –454. Bird, J.J., Brown, D.R., Mullen, A.C., Moskowitz, N.H., Mahowald, M.A., Sider, J.R., Gajewski, T.F., Wang, C.R., Reiner, S.L., 1998. Helper T cell differentiation is controlled by the cell cycle. Immunity 9 (2), 229 –237. Clark, S.J., Harrison, J., Paul, C.L., Frommer, M., 1994. High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22 (15), 2990 –2997. Cornwell, R.D., Gollahon, K.A., Hickstein, D.D., 1993. Description of the leukocyte function-associated antigen 1 (LFA-1 or CD11a) promoter. Proc. Natl. Acad. Sci. USA 90 (9), 4221 – 4225. Golbus, J., Salata, M., Greenwood, J., Hudson, J., Richardson, B.C., 1988. Increased immunoglobulin response to g-interferon by lymphocytes from patients with systemic lupus erythematosus. Clin. Immunol. Immunopathol. 46 (1), 129 –140. Golbus, J., Palella, T.D., Richardson, B.C., 1990. Quantitative changes in T cell DNA methylation occur during differentiation and ageing. Eur. J. Immunol. 20 (8), 1869 – 1872. Grolleau, A., Kaplan, M.J., Hanash, S.M., Beretta, L., Richardson, B., 2000. Impaired translational response and increased protein kinase PKR expression in T cells from lupus patients. J. Clin. Invest. 106 (12), 1561 –1568. Guarente, L., 1996. Do changes in chromosomes cause aging. Cell 86 (1), 9 –12. Hobbs, M.V., Weigle, W.O., Noonan, D.J., Torbett, B.E., Koch, R.J., Cardenas, G.J., Ernst, D.N., 1993. Patterns of cytokine gene expression by CD4 + T cells from young and old mice. J. Immunol. 150 (8 Pt. 1), 3602 – 3614. Hsu, D.W., Lin, M.J., Lee, T.L., Wen, S.C., Chen, X., Shen, C.K., 1999. Two major forms of DNA (cytosine-5) methyltransferase in human somatic tissues. Proc. Natl. Acad. Sci. USA 96 (17), 9751 –9756. Issa, J.P., 1999. Aging, DNA methylation and cancer. Crit. Rev. Oncol. Hematol. 1 (32), 31 – 43. Jones, P.A., 1984. Gene activation by 5-azacytidine. In:
1267
Razin, A., Cedar, H., Riggs, A.D. (Eds.), DNA Methylation, Biochemistry and Biological Significance. Springer, New York, pp. 163 – 187. Kaplan, M.J., Deng, C., Yang, J., Richardson, B.C., 2000. DNA methylation in the regulation of T cell LFA-1 expression. Immunol. Invest. 29 (4), 411 – 425. Kern, F., Docke, W.D., Reinke, P., Volk, H.D., 1994. Discordant expression of LFA-1, VLA-4a, VLA-b 1, CD45RO and CD28 on T-cell subsets: evidence for multiple subsets of ‘memory’ T cells. Int. Arch. Allergy Immunol. 104 (1), 17 – 26. Lee, P.J., Washer, L.L., Law, D.J., Boland, C.R., Horon, I.L., Feinberg, A.P., 1996. Limited up-regulation of DNA methyltransferase in human colon cancer reflecting increased cell proliferation. Proc. Natl. Acad. Sci. USA 93 (19), 10366 – 10370. Pallis, M., Robins, A., Powell, R., 1993. Quantitative analysis of lymphocyte CD11a using standardized flow cytometry. Scand. J. Immunol. 38 (6), 559 – 564. Richardson, B., Yung, R., 1999. Role of DNA methylation in the regulation of cell function. J. Lab. Clin. Med. 134 (4), 333 – 340. Richardson, B., Scheinbart, L., Strahler, J., Gross, L., Hanash, S., Johnson, M., 1990. Evidence for impaired T cell DNA methylation in systemic lupus erythematosus and rheumatoid arthritis. Arthritis Rheum. 33 (11), 1665 – 1673. Richardson, B.C., Strahler, J.R., Pivirotto, T.S., Quddus, J., Bayliss, G.E., Gross, L.A., O’Rourke, K.S., Powers, D., Hanash, S.M., Johnson, M.A., 1992. Phenotypic and functional similarities between 5-azacytidine-treated T cells and a T cell subset in patients with active systemic lupus erythematosus. Arthritis Rheum. 35 (6), 647 – 662. Richardson, B., Powers, D., Hooper, F., Yung, R.L., O’Rourke, K., 1994. Lymphocyte function-associated antigen 1 overexpression and T cell autoreactivity. Arthritis Rheum. 37 (9), 1363 – 1372. Scheinbart, L.S., Johnson, M.A., Gross, L.A., Edelstein, S.R., Richardson, B.C., 1991. Procainamide inhibits DNA methyltransferase in a human T cell line. J. Rheumatol. 18 (4), 530 – 534. Schmid, C.W., Rubin, C.M., 1995. Alu: What is the use. In: The Impact of Short Interspersed Elements (SINEs) on the Host Genome. Maraia and Landes, pp. 105 – 123. Yung, R., Powers, D., Johnson, K., Amento, E., Carr, D., Laing, T., Yang, J., Chang, S., Hemati, N., Richardson, B., 1996. Mechanisms of drug-induced lupus. II. T cells overexpressing lymphocyte function-associated antigen 1 become autoreactive and cause a lupuslike disease in syngeneic mice. J. Clin. Invest. 97 (12), 2866 – 2871. Yang, J., Deng, C., Hemati, N., Hanash, S.M., Richardson, B.C., 1997. Effect of mitogenic stimulation and DNA methylation on human T cell DNA methyltransferase expression and activity. J. Immunol. 159 (3), 1303 – 1309. Yung, R., Kaplan, M.J., Ray, D., Schneider, K., Mo, R.-R., Johnson, K., Richardson, B., 2001a. Autoreactive murine Th1 and Th2 cells kill syngeneic macrophages and induce autoantibodies. Lupus. 10, 539 – 546.
1268
Z. Zhang et al. / Mechanisms of Ageing and De6elopment 123 (2002) 1257–1268
Yung, R., Ray, D., Eisenbraun, J.K., Deng, C., Attwood, J., Eisenbraun, M.D., Johnson, K., Miller, R.A., Hanash, S., Richardson, B., 2001b. Unexpected effects of a heterozygous Dnmt1 null mutation on age-dependent DNA hypomethylation and autoimmunity. J. Gerontol. A Biol. Sci. Med. Sci. 56 (6), B268 – B276.
Zhu, X., Deng, C., Kuick, R., Yung, R., Lamb, B., Neel, J.V., Richardson, B., Hanash, S., 1999. Analysis of human peripheral blood T cells and single-cell-derived T cell clones uncovers extensive clonal CpG island methylation heterogeneity throughout the genome. Proc. Natl. Acad. Sci. USA 96 (14), 8058 – 8063.