Extracellular Spermine Activates DNA Methyltransferase 3A and 3B

We first demonstrated that long-term increased polyamine (spermine, spermidine, putrescine) intake elevated blood spermine levels in mice and humans, and lifelong consumption of polyamine-rich chow inhibited aging-associated increase in aberrant DNA methylation, inhibited aging-associated pathological changes, and extend lifespan of mouse. Because gene methylation status is closely associated with aging-associated conditions and polyamine metabolism is closely associated with regulation of gene methylation, we investigated the effects of extracellular spermine supplementation on substrate concentrations and enzyme activities involved in gene methylation. Jurkat cells and human mammary epithelial cells were cultured with spermine and/or D,L-alpha-difluoromethylornithine (DFMO), an inhibitor of ornithine decarboxylase. Spermine supplementation inhibited enzymatic activities of adenosylmethionine decarboxylase in both cells. The ratio of decarboxylated S-adenosylmethionine to S-adenosyl-L-methionine increased by DFMO and decreased by spermine. In Jurkat cells cultured with DFMO, the protein levels of DNA methyltransferases (DNMTs) 1, 3A and 3B were not changed, however the activity of the three enzymes markedly decreased. The protein levels of these enzymes were not changed by addition of spermine, DNMT 3A and especially 3B were activated. We show that changes in polyamine metabolism dramatically affect substrate concentrations and activities of enzymes involved in gene methylation.


Introduction
Polyamines are linear aliphatic hydrocarbons with three or more primary amino groups. The representative polyamines are spermidine (three amino groups) and spermine (four amino groups). Putrescine, a precursor of polyamine, has two amines and is therefore referred to as a diamine, and its biological activities differ from those of polyamines [1,2]. For example, polyamines suppress the production of pro-inflammatory cytokines from immune cells upon stimulation with lipopolysaccharide and phorbol 12-myristate 13-acetate [2] and decrease the amount of lymphocyte function-associated antigen 1 (LFA-1) on the cell membrane of immune cells [1], while putrescine seem not to have such biological activities [1][2][3]. Polyamines are synthesized within cells and are essential for functions including differentiation and proliferation. The enzyme Figure 1. Polyamine synthesis and gene methylation. Ornithine produced from arginine is converted to putrescine by the action of ornithine decarboxylase (ODC), a rate-limiting enzyme in polyamine synthesis. Spermidine is synthesized by addition of an aminopropyl group supplied from decarboxylated S-adenosylmethionine (dcSAM) via the action of spermidine synthase. A second aminopropyl group can be added to spermidine by spermine synthase to produce spermine. When spermine is supplied from extracellular sources as a result of increased polyamine intake, spermidine is produced by the degradation of spermine via spermidine/spermine N 1 -acetyltransferase (SSAT) / acetyl coenzyme A (acetyl CoA) and N 1acetylpolyamine oxidase (APAO). dcSAM is synthesized from SAM by enzymatic activity of Sadenosylmethionine decarboxylase (AdoMetDC). SAM is synthesized from methionine and adenosine, and serves as a methyl-group donor in vivo. dcSAM concentrations have an inverse association with DNMT activity. AdoMetDC: S-adenosylmethionine decarboxylase, APAO: N 1acetylpolyamine oxidase, ATP: adenosine triphosphate, dcSAM: decarboxylated Sadenosylmethionine, DNMT: DNA methyltransferase, OAZ: ODC antizyme-1, ODC: ornithine decarboxylase, SAM: S-adenosylmethionine, SMO: spermine oxidase, SSAT: spermidine/spermine N 1 -acetyltransferase.
Aging is associated with enhanced demethylation of DNA in various organs and tissues in several animals and humans [20,21]. However, increased hypermethylation associated with age has also been reported in some genes [22,23]. The aging-associated changes in aberrant DNA methylation status, namely increased de-methylation in some areas and hypermethylation in other areas, are considered to be among the most important mechanisms underlying agingassociated pathologies. Our previous studies showed that D,L-alpha-difluoromethylornithine (DFMO)-induced ODC inhibition caused aberrant methylation in Jurkat cells, while spermine supplementation reversed this condition [16,24]. And, the site responsible for LFA-1 expression was demethylated and associated with increased LFA-1 protein levels after ODC inhibition, and spermine supplementation reversed the demethylation and increase of LFA-1 protein levels ( Figure 2). Aging is associated with decreases in ODC [5] and DNMT activities [25], increased aberrant methylation status of entire genome, and enhanced demethylation of the LFA-1 promoter area in association with increases in LFA-1 protein levels [1,16,26]. In a murine model involving chows with different polyamine concentrations, the methylation status of the entire genome in old mice fed regular chow showed an increase in aberrant methylation of entire genome in association with increases in LFA-1 protein levels. However, lifelong intake of high- Figure 1. Polyamine synthesis and gene methylation. Ornithine produced from arginine is converted to putrescine by the action of ornithine decarboxylase (ODC), a rate-limiting enzyme in polyamine synthesis. Spermidine is synthesized by addition of an aminopropyl group supplied from decarboxylated S-adenosylmethionine (dcSAM) via the action of spermidine synthase. A second aminopropyl group can be added to spermidine by spermine synthase to produce spermine. When spermine is supplied from extracellular sources as a result of increased polyamine intake, spermidine is produced by the degradation of spermine via spermidine/spermine N 1 -acetyltransferase (SSAT)/acetyl coenzyme A (acetyl CoA) and N 1 -acetylpolyamine oxidase (APAO). dcSAM is synthesized from SAM by enzymatic activity of S-adenosylmethionine decarboxylase (AdoMetDC). SAM is synthesized from methionine and adenosine, and serves as a methyl-group donor in vivo. dcSAM concentrations have an inverse association with DNMT activity. AdoMetDC: S-adenosylmethionine decarboxylase, APAO: N 1 -acetylpolyamine oxidase, ATP: adenosine triphosphate, dcSAM: decarboxylated S-adenosylmethionine, DNMT: DNA methyltransferase, OAZ: ODC antizyme-1, ODC: ornithine decarboxylase, SAM: S-adenosylmethionine, SMO: spermine oxidase, SSAT: spermidine/spermine N 1 -acetyltransferase.
Changes in polyamine metabolism influencing SAM and dcSAM concentrations and AdoMetDC activity may affect DNMT activities and gene methylation. However, how changes in polyamine metabolism affect substrate concentrations and enzymatic activities involved in gene methylation is not known in detail. In this study, we investigated the effects of decreased polyamine synthesis and of extracellular polyamine supply on substrate concentrations and enzyme activities involved in polyamine metabolism. In our previous studies, spermine concentrations in blood cells of humans and mice fed a high-polyamine diet increased by 1.1-to 1.5-fold, whereas there was no significant increase in spermidine concentration [15]. And, 500 µM spermine in culture supernatant of human peripheral blood mononuclear cells increased intracellular spermine concentrations similar to those in vivo studies [1]. Spermidine at this concentration showed similar bioactivities (e.g., suppression of pro-inflammatory cytokine synthesis and LFA-1 expression) [1,2], but this intracellular spermidine concentration markedly exceeds (by 3 to 4 times) the physiological range [1]. Therefore, in this study, we used spermine to replicate physiological effects in vivo.

Figure 2.
Aging, ODC activity, DNA methyltransferase (DNMT) activity, gene methylation status, and progression of aging associated pathologies and senescence (summary of the results of our previous studies).(Upper) Aging is associated with decreases in ODC and DNMT activities, increased aberrant methylation status (increased demethylation in certain areas and increased hypermethylation in other areas) of entire genome, and enhanced pro-inflammatory status. D,Lalpha-difluoromethylornithine (DFMO)-induced ODC inhibition caused decreased DNMT activities, increased aberrant methylation, increased demethylation of the LFA-1 promoter area (ITGAL), and enhanced pro-inflammatory status (increased LFA-1 protein). (Lower) Spermine supplementation reversed changes induced by the inhibition of ODC by DFMO. Increased polyamine intake elevated blood spermine levels in mice and humans, and lifelong intake of high-polyamine chow inhibited aging-associated increase in aberrant DNA methylation and LFA-1 expression, inhibited aging-associated pathologies, and extended lifespan of mice. DFMO: D,L-alpha-difluoromethylornithine, ODC: ornithine decarboxylase, DNMT: DNA methyltransferase, ITGAL: promoter area of LFA-1, LFA-1: lymphocyte function-associated antigen 1.
Changes in polyamine metabolism influencing SAM and dcSAM concentrations and AdoMetDC activity may affect DNMT activities and gene methylation. However, how changes in polyamine metabolism affect substrate concentrations and enzymatic activities involved in gene methylation is not known in detail. In this study, we investigated the effects of decreased polyamine synthesis and of extracellular polyamine supply on substrate concentrations and enzyme activities involved in polyamine metabolism. In our previous studies, spermine concentrations in blood cells of humans and mice fed a high-polyamine diet increased by 1.1-to 1.5-fold, whereas there was no significant increase in spermidine concentration [15]. And, 500 µM spermine in culture supernatant of human peripheral blood mononuclear cells increased intracellular spermine concentrations similar to those in vivo studies [1]. Spermidine at this concentration showed similar bioactivities (e.g., suppression of pro-inflammatory cytokine synthesis and LFA-1 expression) [1,2], but this intracellular spermidine concentration markedly exceeds (by 3 to 4 times) the physiological range [1]. Therefore, in this study, we used spermine to replicate physiological effects in vivo.

Determination of Culture Condition
Polyamines are contained in all cells in high micromolar to low millimolar quantities. Our previous study showed that spermine concentrations of up to 1 mM for up to 80 h were not toxic

Determination of Culture Condition
Polyamines are contained in all cells in high micromolar to low millimolar quantities. Our previous study showed that spermine concentrations of up to 1 mM for up to 80 h were not toxic [1]. And, the previous studies have established optimal concentrations for DFMO of 3.0 mM in Jurkat cells [24].
Flow cytometric examination revealed that DFMO treatment (3 mM) for 3 days increased mean fluorescent intensity (MFI) of CD11a. Spermine treatment alone and DFMO plus spermine treatment decreased the MFI of CD11a staining. The percentage of cells negative for ViaProbe was not changed, indicating that these treatments decreased LFA-1 expression without affecting cell viability ( Figure A1).

DNMT Levels and Activities
Tumor cells such as Jurkat cells grow rapidly, however multiplication of normal cells such as HMEpCs was very slow. Therefore, it was difficult to secure enough HMEpCs to measure protein levels and DNMT activities. Thus Jurkat cells were cultured for examination of protein levels and activities of DNMTs because sufficient enzyme can be obtained from these cells for measurement. DNMT 1 (Figure 5a

Discussion
In this study, we have shown that changes in polyamine metabolism affect DNMT activities. Inhibitory effects of DFMO on polyamine concentrations and on activation of AdoMetDC activity varied between Jurkat cells and normal HMEpCs, which may be due to the difference of capability to maintain polyamine homeostasis. Changes in spermine concentration upon spermine supplementation also varied between the two cell types. However, changes in intracellular spermidine concentration were found in both cell types cultured with spermine, which shows that intracellular polyamine metabolism was influenced by the extracellular spermine supply. In Jurkat cells treated with DFMO, spermine supplementation increased both spermine and spermidine concentrations. Since spermidine levels were under detectable levels in cells co-cultured with 3 mM of DFMO, an increase in spermidine concentration by spermine supplementation in DFMO-treated Jurkat cells reflects increased spermine catabolism. Spermine can be converted to spermidine by the enzymatic activities of SSAT/Acetyl CoA and APAO. Spermine supplementation to control Jurkat cells (cultured in unsupplemented culture medium) decreased spermidine concentrations, suggesting ODC inhibition by a negative feedback mechanism. Decreases in spermidine concentrations by spermine supplementation were observed in HMEpCs whether cells were co-cultured with DFMO or not, suggesting that 24 mM of DFMO could not completely inhibit ODC activity. Furthermore, AdoMetDC activity was markedly decreased by spermine in both cells, suggesting that intake of spermine into cells induces a negative feedback mechanism to inhibit polyamine synthesis.
As observed in both Jurkat cells and HMEpCs, dcSAM concentration in vivo is generally a few percent of the SAM concentration [27]. The DFMO-induced changes in AdoMetDC activity and SAM and dcSAM concentrations observed in Jurkat cells were similar to the results of previous studies [28][29][30]. DFMO significantly increased AdoMetDC activity and dcSAM concentration, whereas SAM concentration decreased in Jurkat cells. Although AdoMetDC activity in HMEpCs did not change significantly on treatment with DFMO, the effect of spermine supplementation was similar to that observed in Jurkat cells and in a previous report in which spermine inhibited DFMO-induced increases in AdoMetDC activities in Ehrlich ascites-carcinoma cells [30].
dcSAM is likely to have inhibited activation of DNMT through competition with SAM [31,32]. Thus, in Jurkat cells, dynamic changes in dcSAM concentration occurred with DFMO and spermine, with a negative relationship between the dcSAM concentration and DNMT activity, especially of DNMT 3B. Changes in dcSAM concentration were not significant in HMEpCs, however, spermine suppressed DFMO-induced increases in dcSAM/SAM ratio. An inverse relationship between the dcSAM/SAM ratio and DNMT activity was reported previously [33]. AdoMetDC activities and the ratio of dcSAM to SAM were relatively similar among the three conditions tested (DFMO(-)spermine(-), DFMO(-)spermine(+), DFMO(+)spermine(+)) on Jurkat cells, though not in cells co-cultured with DFMO alone. However, the changes in DNMT activities were not necessarily the same among these three conditions. These findings indicate that some factor(s) other than the dcSAM concentration and dcSAM/SAM ratio may affect DNMT activities and gene methylation [34].
One of the very interesting findings in this study is that changes in polyamine metabolism influenced DNMT activities without affecting their protein levels. The effect of decreased ODC activity was examined previously, and it was reported that protein levels of DNMT 3B in human oral cancer cells decreased when the ODC antizyme-1 gene, which degrades ODC and inhibits its activity, was transfected [35]. The different effects on DNMTs between studies may be due to the different cell lines employed or methods used to inhibit ODC activity. In the present study, the activities of all DNMTs decreased significantly on DFMO treatment. In contrast, spermine activated DNMT 1 when cells were not treated with DFMO. Although such an effect was not observed for DNMT 3A and 3B when spermine alone was added to the culture supernatants, spermine markedly activated DNMT 3B and DNMT 3A when cells were treated with DFMO. Methylation patterns in genomes are considered to be stably inherited by cells in the next generation [36]; however, reversibly modified regions of methylation have also been found [35,37]. DNMT 1 mainly acts to maintain methylation in DNA replication, whereas DNMT 3A and DNMT 3B have important roles in de novo methylation [35]. A study showed that the role of DNMT 1 is gradually compensated, at least partially, by DNMT3 [38], suggesting the importance of spermine supplementation for the maintenance of gene methylation status.
SAM serves as a methyl group donor, therefore increases in SAM concentrations by spermine supplementation indicate increased availability of methyl group. Increased availability of methyl group and increased DNMT activity seem to be the key to maintain DNA methylation status. Supplementation of methyl group by either methionine or SAM affects the DNA methylation status [39][40][41], and decrease DNMT is associated with alteration of methylation status of the entire genome [35,42]. Generally, decreases in ODC activity with aging [4,5,33] is associated with decreases in DNMT activities [25,43] and changes in DNA methylation status [43][44][45]. Aging associated change in DNA methylation status seems to be a non-directional change as it involves both hypermethylation and hypomethylation events [46][47][48]. Alteration of methylation status with aging changes chromatin accessibility, resulting in aberrant gene transcription as well as genomic instability. These factors may be key regulators of the aging process and contributors to the development of aging-associated diseases [49][50][51], including neoplastic growth [52][53][54] and aging itself [55][56][57].
In the previous studies, we have shown that aberrant methylation status induced by inhibiting ODC was almost reversed by spermine supplementation [24], and that life-long consumption of high-polyamine chow by mice inhibited aging-associated changes in methylation status of the entire genome, inhibited aging associated pathological changes, and extended lifespan [14,16]. In addition, increased polyamine intake followed by repeated weak carcinogenic stimuli decreased tumorigenesis in animals [16,58]. The increases in spermine concentrations by continuously increased polyamine intake may compensate for aging-associated decreases in DNMT activities by decreasing dcSAM, and increased SAM availability and increased DNMT activity may attenuate progression of aberrant gene methylation (Figure 6).

Cells and Culture Conditions
Jurkat cells (Human Science Research Resource Cell Bank, Tokyo, Japan) and human mammary epithelial cells (HMEpCs) (Cell Applications, San Diego, CA, USA) were used in this study. Jurkat cells were adjusted to a cell density of 1.0 × 10 6 cells/mL and incubated in RPMI-1640 culture medium (Sigma-Aldrich, St. Louis, MO, USA) containing 10% human inactivated serum (Cosmo Bio Co., Tokyo, Japan) for 72 h before use for subsequent experiments. HMEpCs were adjusted to a cell density of 5.0 × 10 4 cells/mL, incubated in serum-free medium

Cells and Culture Conditions
Jurkat cells (Human Science Research Resource Cell Bank, Tokyo, Japan) and human mammary epithelial cells (HMEpCs) (Cell Applications, San Diego, CA, USA) were used in this study. Jurkat cells were adjusted to a cell density of 1.0 × 10 6 cells/mL and incubated in RPMI-1640 culture medium (Sigma-Aldrich, St. Louis, MO, USA) containing 10% human inactivated serum (Cosmo Bio Co., Tokyo, Japan) for 72 h before use for subsequent experiments. HMEpCs were adjusted to a cell density of 5.0 × 10 4 cells/mL, incubated in serum-free medium (MammaryLife™ Comp Kit, Kurabo Industries, Tokyo), and subcultured according to the manufacturer's protocol. Cells were collected for use in experiments after sufficient passages resulted in a large enough number of cells, and then incubated for 72 h in the following conditions: 1. control culture (Jurkat: RPMI-1640 + 10% human inactivated serum; HMEpCs: serum-free medium); 2. culture mixed with D,L-alpha-difluoromethylornithine (DFMO) (Amine Pharma Institute, Chiba, Japan), an irreversible inhibitor of ODC; 3. culture with spermine (Sigma-Aldrich Japan, Tokyo); and 4. culture with DFMO and spermine. In Jurkat cells, 3.0 mM DFMO and 500 µM spermine were used in accordance with the protocol of a previous study [46]. HMEpCs were cultured with DFMO at different concentrations and examined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (In Vitro Toxicology Assay Kit, Sigma-Aldrich Japan) and the highest non-cytotoxic concentrations were determined. This resulted in the use of 24 mM DFMO in subsequent experiments.
HMEpCs attached to a culture dish were washed with 1× PBS(-), collected using a cell scraper, homogenized (150 W, 10 s, ×3), and centrifuged (13,964 × g, 4 • C, 30 min) after injecting the extraction buffer (300 µL) to obtain supernatants containing intracellular proteins. Protein contents in supernatants were measured by the Bradford method using bovine serum albumin as a standard.
The measuring principle for AdoMetDC activity is to determine CO 2 emitted when SAM (carboxyl-14 C) is converted to dcSAM by AdoMetDC. A filter paper was attached inside a 2 mL Eppendorf tube cap; 20 µL of 250 mM phosphate buffer (pH 7.5) and 15 mM DTT, 15 µL 25 mM putrescine dihydrochloride, and a sample (intracellular protein solution) of 80 µL were mixed; and 10 µL of a SAM cocktail of 4.8 mM cold (i.e., non-radioactive) SAM (New England Biolabs, MA, USA) and 0.2 mM SAM (carboxyl-14 C) (0.1 µCi/reaction) (American Radiolabeled Chemicals, St. Louis, MO, USA) were added. To adsorb CO 2 with alkali, 10% KOH was permeated in the filter paper on the cap. The cap was immediately shut to allow production of CO 2 at 37 • C for 30 min and then the tube was cooled on ice for 15 min. To release dissolved CO 2 , 50 µL 6 M HCl was added to the resulting solution and the cap was immediately shut to allow reaction with CO 2 at 37 • C for 15 min. Released 14 CO 2 was adsorbed on the filter paper. After ice-cooling, the filter paper was put into a vial containing 3.5 mL of scintillation solution and radioactivity was counted three times for 4 min each.

SAM and dcSAM Assay
Intracellular SAM and dcSAM concentrations were determined using reversed-phase HPLC. SAM and dc SAM were extracted from cells by the same method as that used for determination of the intracellular polyamine concentration. Using a Capcell pak C18 SG120 column (4.6 mm I.D. × 150 mm; Shiseido Co., Tokyo, Japan), 20 µL of sample was injected per assay for reversed-phase HPLC. The reversed-phase HPLC conditions were: column oven at 40 • C; flow rate, 0.5 mL/min; two solvent linear gradient, where solvent A consisted of 90% (v/v) 0.1 M sodium acetate (adjusted to pH 4.50 with acetate) with 10 mM sodium 1-octanesulfonate (Tokyo Chemical Industry Co., Tokyo) and 10% (v/v) methanol, and solvent B consisted of 90% (v/v) 0.2 M sodium acetate (pH 4.50)-acetonitrile (10:3) with 10 mM sodium 1-octanesulfonate and 10% (v/v) methanol; gradient, solvent B 0% to 100% over 50 min; UV detection at 250 nm; data were processed using LC Workstation chromatography software (Shimadzu Corp., Kyoto, Japan).

DNMT Assay and Activity Determination by Subtype
Since multiplication of HMEpCs was very slow, it was difficult to secure enough cells to measure protein levels and DNMT activities. Thus, these experiments were performed using Jurkat cells. An EpiQuik Nuclear Extraction Kit I (Epigentek Group, Farmingdale, New York (NY), USA) was used to extract nuclear proteins, using 1 µL of extraction buffer per 1.0 × 10 6 cells. Protein concentrations of DNMT 1, 3A and 3B were determined using EpiQuik Assay Kits for each protein (all from Epigentek) (Epigentek Group Inc., NY, USA). After reacting nuclear protein solutions extracted from cells on plates covered with substances with high affinity for the respective DNMTs (37 • C for 2 h), the concentration of each DNMT was detected by a colorimetric method using a multiplate reader (DTX 880 Multimode Detector, Beckman Coulter Inc., Brea, California, USA) at 450 nm.
DNMT 1, 3A and 3B activities were determined using an EpiQuik DNA Methyltransferase 1 Activity/Inhibitor Screening Assay Core Kit (Epigentek Group Inc., NY, USA), a DNMT3A Direct Activity Assay Kit (BPS Bioscience, San Diego, CA, USA), and an EpiQuik DNA Methyltransferase 3B Activity/Inhibitor Screening Assay Core Kit (Epigentek Group Inc., New York (NY), USA), respectively. Nuclear protein solutions were adjusted with assay buffer to make the protein concentrations of all DNMTs equivalent. Nuclear protein solutions were incubated with assay buffer on cytosine-rich DNA-covered plates (37 • C for 2 h). DNMT 1 and 3B were detected by a colorimetric method (450 nm) and DNMT 3A was detected by horseradish peroxidase chemiluminescence (1000 ms) using a multiplate reader. Analysis was conducted using Multimode Analysis Software ver. 3.2.0.5 (Beckman Coulter Inc. California, USA). DNMT activities are reported using indexes that were calculated by dividing measured activity by that of the control culture for each enzyme subtype.

Statistical Analysis
Statistical analyses were conducted using EZR software (Jichi Medical University Saitama Medical Center, Saitama-city, Japan) [60]. For analysis of differences between two groups, an unpaired t-test was used for homoscedasticity and the Mann Whitney test was used for heteroscedasticity. p < 0.05 was considered to indicate a significant difference in all analyses.

Conclusions
Decreases in ODC activity with aging is associated with decreases in DNMT activities and changes in DNA methylation status. Increased polyamine intake elevates blood spermine levels in mice and humans, and lifelong consumption of polyamine-rich chow inhibited aging-associated increase in aberrant DNA methylation, inhibited aging-associated pathological changes, and extend lifespan of mouse. The current study addresses the fundamental background of spermine-induced regulation of gene methylation leading to lifespan extension. In cells with decreased ODC activity, dcSAM/SAM ratio increased significantly. When ODC is suppressed, inhibition of AdoMetDC activity by spermine supplementation decreased dcSAM/SAM ratio. The decrease in dcSAM/SAM ratio is associated with an activation of DNMT 3a and 3b.