The Cbs Locus Affects the Expression of Senescence Markers and mtDNA Copy Number, but not Telomere Dynamics in Mice

Cystathionine β-synthase (CBS) is a housekeeping enzyme that catalyzes the first step of the homocysteine to cysteine transsulfuration pathway. Homozygous deletion of the Cbs gene in mice causes severe hyperhomocysteinemia and reduces life span. Here, we examined a possible involvement of senescence, mitochondrial DNA, and telomeres in the reduced life span of Cbs−/− mice. We found that senescence-related p21, Pai-1, Mcp1, and Il-6 mRNAs were significantly upregulated (2–10-fold) in liver, while p21 was upregulated in the brain of Cbs−/− mice (n = 20) compared with control Cbs+/− siblings (n = 20) in a sex- and age-dependent manner. Telomere length in blood (n = 80), liver (n = 40), and brain (n = 40) was not affected by the Cbs−/− genotype, but varied with sex and/or age. Levels of mitochondrial DNA tended to be reduced in livers, but not brains and blood, of Cbs−/− females (n = 20–40). The Cbs−/− genotype significantly reduced Tert mRNA expression in brain, but not liver, in a sex- and age-dependent manner. Multiple regression analysis showed that the senescence-related liver (but not brain) mRNAs and liver (but not brain or blood) mitochondrial DNA were associated with the Cbs genotype. In contrast, telomere length in blood, brain, and liver was not associated with the Cbs genotype or hyperhomocysteinemia, but was associated with sex (in brain and liver) and age (in brain and blood). Taken together, these findings suggest that the changes in senescence marker expression and mtDNA levels, but not telomere shortening, could account for the reduced life span of Cbs−/− mice.


Introduction
Cellular senescence irreversibly arrests proliferation and is an important contributor to aging and age-related disease [1]. Senescence results in a loss of tissue repair ability due to stabilization of p53 and upregulation of the downstream p53 targets such as p21 and PAI-1 [2,3], which prevents CDK2-mediated inactivation of RB, thereby preventing entry into the S-phase of the cell cycle [1]. Senescent cells produce the senescence-associated secretory phenotype, i.e., pro-inflammatory and matrix degrading molecules (IL-6, MCP1, PAI-1, etc.). Although well studied in cultured cells in vitro, senescence in vivo in humans and animals is poorly understood [4]. One of the factors that induces the p53-p21 senescence pathway is telomere shortening [1].
Telomeres are tandem DNA repeats of TTAGGG structures at the end of chromosomes important for genomic stability in all vertebrates. The telomere sequence is more sensitive to damage than
We hypothesized that the reduced life span of the Cbs −/− mice could be caused by premature senescence. To examine this hypothesis, we quantified the expression of senescence-related mRNAs (Pai-1, p21, Mcp1, Il-6, p16, and Il-1β) and an aging-related Kl mRNA relative to β-actin and Gapdh mRNAs in mouse brains and livers. We found that Pai-1, p21, Mcp1, and Il-6 mRNAs were significantly upregulated in livers of Cbs −/− mice compared to Cbs +/− siblings ( Figure 1A-C,E), while p16 and Il-1β mRNAs were not affected by the Cbs −/− genotype ( Figure 1D,F). The expression of the corresponding mRNAs in brains was similar between Cbs −/− mice and their Cbs +/− siblings (not shown).
Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 3 of 19 mRNAs in mouse brains and livers. We found that Pai-1, p21, Mcp1, and Il-6 mRNAs were significantly upregulated in livers of Cbs −/− mice compared to Cbs +/− siblings ( Figure 1A,B,C,E), while p16 and Il-1β mRNAs were not affected by the Cbs −/− genotype ( Figure 1D,F). The expression of the corresponding mRNAs in brains was similar between Cbs −/− mice and their Cbs +/− siblings (not shown). Box and whiskers represent the mean ± SEM and the mean ± 95% CI, respectively. Upper and lower pvalues refer to data normalized to β-actin (blank boxes) and Gapdh (gray boxes) mRNAs, respectively. NS, not significant.
Although the Cbs genotype had no effect on Il-1β and p16 mRNAs in the whole mouse cohort (Figure 1 D, F), subgroup analyses showed that p16 mRNA was downregulated in young and old female Cbs −/− mice relative to Cbs +/− females (young: 0.04 vs. 0.15, p = 0.035; old: 6.5 vs. 14.1, p = 0.046), but not in males. Il-1β mRNA was not affected by the Cbs genotype in any of the age and sex subgroups. Il-1β mRNA was significantly upregulated in old mice, similarly in Cbs −/− and Cbs +/− females and males (2.5-to five-fold; p = 0.015-0.039).

Brain
The expression of senescence-related brain p21, Mcp-1, Pai-1, and Il-6 mRNAs appeared to be unaffected by the Cbs −/− genotype in the whole mouse cohort, as was Kl mRNA (not shown). However, examination of sex and age subgroups revealed that, in Cbs −/− female mice, brain p21 and Mcp-1 Box and whiskers represent the mean ± SEM and the mean ± 95% CI, respectively. Upper and lower p-values refer to data normalized to β-actin (blank boxes) and Gapdh (gray boxes) mRNAs, respectively. NS, not significant.
Although the Cbs genotype had no effect on Il-1β and p16 mRNAs in the whole mouse cohort ( Figure 1D,F), subgroup analyses showed that p16 mRNA was downregulated in young and old female Cbs −/− mice relative to Cbs +/− females (young: 0.04 vs. 0.15, p = 0.035; old: 6.5 vs. 14.1, p = 0.046), but not in males. Il-1β mRNA was not affected by the Cbs genotype in any of the age and sex subgroups. Il-1β mRNA was significantly upregulated in old mice, similarly in Cbs −/− and Cbs +/− females and males (2.5-to five-fold; p = 0.015-0.039).

Brain
The expression of senescence-related brain p21, Mcp-1, Pai-1, and Il-6 mRNAs appeared to be unaffected by the Cbs −/− genotype in the whole mouse cohort, as was Kl mRNA (not shown). However, examination of sex and age subgroups revealed that, in Cbs −/− female mice, brain p21 and Mcp-1 mRNAs were significantly upregulated relative to Cbs +/− females (1.9-fold in young, p = 0.035, and 1.4-fold in old females, p = 0.020, respectively; Supplementary Table S2). However, the expression of the senescence-related brain mRNAs was unaffected by the Cbs −/− genotype in males.
Brain Il-1β and p16 mRNAs were not affected by the Cbs genotype in any of the age and sex subgroups. Age and sex did not affect brain Il-1β mRNA, while age significantly affected brain p16 mRNA (six-to 10-fold higher in old vs. young mice, p = 0.009 to 0.000).

Blood TL
We quantified TL in the blood of Cbs −/− mice (n = 40) and their Cbs +/− siblings (n = 40) and found that blood TL was not affected by the Cbs genotype ( Figure 2A). In males, blood TL decreased with age over the span from 36 to 473 days, similarly in both Cbs −/− mice and their Cbs +/− siblings (Supplementary Figure S1A). In females, blood TL was also similar between Cbs −/− and Cbs +/− mice, and there was little change in blood TL with age (Supplementary Figure S1B). Blood TL tended to decrease with increasing plasma tHcy in male, but not in female mice (Supplementary Figure S1C). mRNAs were significantly upregulated relative to Cbs +/− females (1.9-fold in young, p = 0.035, and 1.4fold in old females, p = 0.020, respectively; Supplementary Table S2). However, the expression of the senescence-related brain mRNAs was unaffected by the Cbs −/− genotype in males. Sex significantly affected the expression of brain p21, Mcp-1, Il-6, and Kl mRNAs in Cbs −/− mice, (1.4-to 2.5-fold higher in females than in males, p = 0.040 to 0.004). In Cbs +/− mice, sex affected the expression of brain p21 in old mice (2.1-fold higher in females, p = 0.036) and Mcp-1 mRNAs (lower in females, p = 0.052; Supplementary Table S2).
Brain Il-1β and p16 mRNAs were not affected by the Cbs genotype in any of the age and sex subgroups. Age and sex did not affect brain Il-1β mRNA, while age significantly affected brain p16 mRNA (six-to 10-fold higher in old vs. young mice, p = 0.009 to 0.000).

Blood TL
We quantified TL in the blood of Cbs −/− mice (n = 40) and their Cbs +/− siblings (n = 40) and found that blood TL was not affected by the Cbs genotype ( Figure 2A). In males, blood TL decreased with age over the span from 36 to 473 days, similarly in both Cbs −/− mice and their Cbs +/− siblings (Supplementary Figure S1A). In females, blood TL was also similar between Cbs −/− and Cbs +/− mice, and there was little change in blood TL with age (Supplementary Figure S1B). Blood TL tended to decrease with increasing plasma tHcy in male, but not in female mice (Supplementary Figure S1C).

Figure 2.
Blood telomere length (TL) (A) and plasma total Hcy (tHcy) levels (B) in Cbs −/− and Cbs +/− mice. Blood TL and plasma tHcy were quantified in Cbs −/− mice (n = 40) and sex-and age-matched control Cbs +/− siblings (n = 40) as described in the Materials and Methods. Box and whiskers represent the mean±SEM and the mean ± 95% CI, respectively. Upper and lower p-values refer to data normalized to β-actin (blank boxes) and Gapdh (gray boxes) mRNAs, respectively.
Effects of Age on Blood TL in Cbs +/− and Cbs −/− Mice P <0.000 P = 0.438 A. B.

Figure 2.
Blood telomere length (TL) (A) and plasma total Hcy (tHcy) levels (B) in Cbs −/− and Cbs +/− mice. Blood TL and plasma tHcy were quantified in Cbs −/− mice (n = 40) and sex-and age-matched control Cbs +/− siblings (n = 40) as described in the Materials and Methods. Box and whiskers represent the mean±SEM and the mean ± 95% CI, respectively. Upper and lower p-values refer to data normalized to β-actin (blank boxes) and Gapdh (gray boxes) mRNAs, respectively.

Brain TL and Tert mRNA
To examine whether Cbs genotype affects TL in a tissue-specific manner, we isolated DNA from brains and livers of Cbs −/− mice (n = 19) and their Cbs +/− siblings (n = 24) and quantified brain and liver TL. To examine whether TL is affected by telomerase expression, we isolated RNA and examined Tert mRNA levels in these tissues.
We found that brain TL was not affected by the Cbs genotype (Supplementary Figure S2A). However, brain TL was affected by sex in a Cbs genotype-and an age-dependent manner. Specifically, in old, but not young, Cbs −/− mice, brain telomeres were significantly longer in males compared to females (1.53 ± 0.34 vs. 0.95 ± 0.22, p = 0.010). In contrast, in young and old Cbs +/− mice, brain TL was not affected by sex and was similar between males and females (Supplementary Table S4).
We also found that brain Tert mRNA expression was not affected by the Cbs genotype in the whole mouse cohort (Supplementary Figure S3A). However, Tert mRNA appeared to be affected by the Cbs genotype in some age-and sex-stratified subgroups. Specifically, young, but not old, Cbs −/− male, but not female, mice had significantly reduced expression of brain Tert mRNA compared to Cbs +/− animals (0.55 ± 0.17 vs. 1.26 ± 0.39, p = 0.035) (Supplementary Table S4).
Brain Tert mRNA expression was also affected by age in a sex-dependent manner. Specifically, brain Tert mRNA levels were significantly reduced in old compared to young male (0.66 ± 0.21 vs. 1.26 ± 0.39, p = 0.022), but not female, Cbs +/− mice; however, this effect of age was absent in Cbs −/− mice (Supplementary Table S4).

Liver TL and Tert mRNA
Liver TL, similar to blood TL ( Figure 2, Supplementary Table S3) and brain TL (Supplementary Figure S2A), was not affected by the Cbs genotype (Supplementary Figure S2B). However, liver TL was affected by sex in an age-dependent manner. For example, in old mice, liver TL were significantly longer in females compared to males, both in Cbs −/− (1.11 ± 0.30 vs. 0.75 ± 0.19, p = 0.014) and Cbs +/− mice (1.50 ± 0.58 vs. 0.78 ± 0.23, p = 0.018) (Supplementary Table S5). In contrast, in young Cbs −/− and Cbs +/− mice, there was no significant difference in liver TL between females and males (Supplementary Table S5).
Similar to brain TL (Supplementary Table S4), but in contrast to blood TL ( Figure 2, Supplementary  Table S3), liver TL was not affected by age in Cbs −/− and Cbs +/− mice (Supplementary Table S5).

Cbs Genotype and mtDNA Copy Number in Blood, Liver, and Brain
In the whole mouse cohort, blood, liver, and brain mtDNA levels were not affected by the Cbs genotype (Supplementary Figure S4). However, mtDNA was affected by age and sex (see below).

Blood mtDNA
Relationships between blood mtDNA and age for Cbs −/− and Cbs +/− mice of each sex are shown in Figure 3. Overall, blood mtDNA decreased with age over the span from 36 to 390 days, similarly for Cbs −/− and Cbs +/− mice and for both females and males.  Table  S5).

Cbs Genotype and mtDNA Copy Number in Blood, Liver, and Brain
In the whole mouse cohort, blood, liver, and brain mtDNA levels were not affected by the Cbs genotype (Supplementary Figure S4). However, mtDNA was affected by age and sex (see below).

Blood mtDNA
Relationships between blood mtDNA and age for Cbs −/− and Cbs +/− mice of each sex are shown in Figure 3. Overall, blood mtDNA decreased with age over the span from 36 to 390 days, similarly for Cbs −/− and Cbs +/− mice and for both females and males.  Table S6).
We found that brain and liver mtDNAs were affected by sex (higher in males) and age (Supplementary Table S6). However, in contrast to blood mtDNA, which decreased with age, mtDNA significantly increased in brains of old compared to young Cbs −/− males (1.31±0.23 vs. 0.95±0.14, p=0.002) and Cbs +/− males (1.40 ± 0.60 vs. 0.89 ± 0.13, p = 0.048). Increased mtDNA levels were also found in livers of old compared to young Cbs −/− and Cbs +/− mice of both sexes. Further analyses showed that blood mtDNA was not affected by the Cbs genotype, but was affected by sex and age: higher in young compared with old female  Table S6).
We found that brain and liver mtDNAs were affected by sex (higher in males) and age (Supplementary Table S6). However, in contrast to blood mtDNA, which decreased with age, mtDNA significantly increased in brains of old compared to young Cbs −/− males (1.31±0.23 vs. 0.95 ± 0.14, p = 0.002) and Cbs +/− males (1.40 ± 0.60 vs. 0.89 ± 0.13, p = 0.048). Increased mtDNA levels were also found in livers of old compared to young Cbs −/− and Cbs +/− mice of both sexes.

Determinants of Blood TL in Mice
Associations between blood TL and independent variables are shown in Tables 1 and 2. Pearson correlation analyses of a whole cohort including Cbs −/− and Cbs +/− mice showed that blood TL was significantly associated with age, but not with the Cbs genotype or tHcy (Table 1). Multiple regression analysis for the whole cohort including Cbs −/− and Cbs +/− mice of both sexes showed that blood TL was not associated with the Cbs genotype, plasma tHcy, age, or sex (Table 1). However, in male mice, blood TL was negatively associated with age, which explained 18-34% of the TL variation (Table 1).

Determinants of TL in Mouse Brain and Liver
Comparative multiple regression analysis of determinants of blood TL, brain TL, and liver TL is shown in Table 2. Similar to blood TL, brain and liver TL were also found not to be associated with the Cbs genotype. However, in contrast to blood TL, brain TL was associated with sex and age, while liver TL was associated with sex, explaining 39% and 26% of the TL variation in brain and liver, respectively ( Table 2).

Determinants of Tert mRNA Expression in Mouse Brain and Liver
Multiple regression analysis in a model involving sex, age, the Cbs genotype, and TL showed that the brain and liver Tert mRNAs were not associated with the Cbs genotype (Supplementary  Table S7). However, brain Tert mRNA expression was associated with sex (Supplementary Table S7), young Cbs −/− female mice having higher expression than males (1.29 ± 0.33 vs. 0.55 ± 0.17, p = 0.017, Supplementary Table S4), explaining 28% of the variability in Tert mRNA (Supplementary Table S7). Liver Tert mRNA was positively associated with TL and age, which explained 69% variability in Tert mRNA expression (Supplementary Table S7).

Pearson Correlations
Pearson correlation analysis of the whole cohort including Cbs −/− and Cbs +/− mice of both sexes showed that blood mtDNA was associated with the Cbs genotype (p = 0.044), sex (p = 0.021), and age (p = 0.025) ( Table 3). Brain mtDNA was associated with sex (p = 0.007) and age (p = 0.005). Liver mtDNA was associated with age.
Pearson correlation analysis stratified by sex showed that blood mtDNA was significantly associated with the Cbs genotype (p = 0.014) and age (p = 0.005) in males. Brain and liver TL were associated with age in males and females.

Multiple Regression
In multiple regression analysis for the whole mouse cohort in models including sex, age, and the Cbs genotype or plasma tHcy, liver mtDNA was significantly associated with the Cbs genotype (β = 0.24, p = 0.041) and age (β = 0.70, p = 0.000) in the whole mouse cohort including both sexes. The association with the Cbs genotype was somewhat stronger in the female subgroup (β = 0.35, p = 0.030). In contrast, in the male subgroup, age (p = 0.003), but not the Cbs genotype, was the only significant determinant of liver mtDNA. The Cbs genotype and age explained 44% of the liver mtDNA variability in the whole cohort and 63 % in the female subgroup (Table 3).
In contrast, brain and blood mtDNAs were not associated with the Cbs genotype, but were associated with sex and age, which explained 35% (brain) and 24-29% (blood) of the mtDNA variability (Table 3). Brain mtDNA was also significantly associated with age in the female (β = 0.49, p = 0.023) and male (β = 0.49, p = 0.032) subgroups.
Multiple regression analysis of subgroups in models involving age and the Cbs genotype or plasma tHcy showed that most of the blood mtDNA variability occurred in males (R 2 = 0.35-0.48), not females (R 2 = 0.01-0.1). The blood mtDNA was strongly associated with age in males (β = −0.58, p = 0.000) and tended to be associated with the Cbs genotype (p = 0.058), but not plasma tHcy (p = 0.419). Age and the Cbs genotype explained 48% of the variability in mtDNA in males (Table 3).

Determinants of Senescence-Related mRNAs in Mice
Associations between senescence-related mRNAs and independent variables in mouse brain and liver are presented in Table 4. Table 3. Determinants of mtDNA in mouse blood, brain, and liver.

Liver
Multiple regression analysis for the whole cohort of mice in models including sex, age, the Cbs genotype, TL, Tert mRNA, and mtDNA showed that the Cbs genotype was the strongest determinant of the liver Pai-1, p21, and Mcp1 mRNAs (p = 0.001), while liver mtDNA was the strongest determinant of liver Il-6 (p = 0.000, Table 4) and Il-1β mRNAs (p = 0.000, Table 5).

Brain
Multiple regression analysis for the whole cohort of mice showed that the Cbs genotype was not a determinant of any senescence-related mRNAs in brain (Table 4).
In multiple regression analysis for sex subgroups, age was a positive determinant of brain Mcp1 and Il-6 mRNAs in male mice and a negative determinant of brain Kl mRNA in female mice ( Table 4). The Cbs genotype tended to be associated with brain p16 mRNA (p = 0.065, Table 5).

Discussion
Cbs deficiency has been known to reduce life span both in humans and mice. However, the underlying mechanism of the reduced life span in Cbs deficiency was unknown. In the present work, we examined a possible involvement of telomere shortening, mtDNA copy number, and accelerated senescence in the reduced life span of Cbs −/− mice.
We used a Cbs −/− mouse model that recapitulated the reduced life span and pathological phenotypes generally associated with aging observed in CBS-deficient patients. We found that the expression of senescence-associated mRNAs (two targets of p53: Pai-1 and p21; and three components of the senescence-associated secretory phenotype: Mcp-1, Pai-1, and Il-6) were significantly upregulated in livers of Cbs −/− mice compared to control Cbs +/− siblings ( Figure 1). However, the expression of liver p16 and Il-1β and the antiaging Kl protein mRNAs was not affected by the Cbs genotype. These expression patterns were specific to liver and the Cbs −/− genotype had no effect on the expression of any of these mRNAs in brain. It should be noted that p16 mRNA, which is considered to be a more specific marker for senescence, is not expressed by all senescent cells [30] and is expressed in certain non-senescent cells [31]. The senescence phenotype is known to be variable [4,30,31], and our findings in Cbs −/− mice provided another example of such variability.
The present findings suggested that accelerated senescence could provide a plausible molecular mechanism contributing to hepatic steatosis observed in Cbs −/− mice [26,27,32]. Importantly, our findings also suggested that accelerated senescence in liver could also explained the reduced life span associated with Cbs deficiency in mice.
We found that the Cbs genotype and HHcy did not affect TL in blood (Figure 2), brain, and liver (Supplementary Figure S2). We also found that Tert mRNA expression was not associated with the Cbs genotype, but was associated with TL in liver and not in brain. The lack of association of TL with the Cbs genotype suggested that the reduced life span in Cbs −/− mice was caused by accelerated senescence in liver and was unlikely to be mediated by HHcy or telomere shortening.
Although we found no association between the Cbs genotype and TL in any of the threes mouse tissues that were examined, we did find that age and sex, known to affect TL in other biological systems [4,5], were associated with TL in our mouse cohort (Tables 1 and 2). For example, blood TL was associated with age in male mice and brain TL with sex and age, while liver TL was associated with sex. These effects of age/sex on TL provided positive controls and indicated that our assays were sufficiently sensitive to allow identification of variables known to be associated with TL, thereby increasing confidence in the present findings.
We also found that the Cbs genotype was a significant determinant of liver mtDNA copy number (Table 3). Specifically, liver mtDNA copy number was reduced in Cbs −/− mice compared with Cbs +/− sibling controls, suggesting the involvement of hepatic respiratory chain deficiency in the reduced life span of Cbs −/− mice. In contrast, the Cbs genotype did not affect the brain mtDNA copy number, showing that the Cbs −/− genotype-induced reduction in mtDNA copy number was liver specific. Because mtDNA copy number is essential for maintaining oxidative capacity, ATP generation, and ultimately cell survival, our findings suggested that reduced liver mtDNA copy number could be responsible for the reduced life span in Cbs −/− mice.
Our findings also showed that mtDNA copy number in Cbs −/− and Cbs +/− mice changed with age in a tissue-dependent manner: decreases in blood and increases in brain and liver (Table 3;  Supplementary Table S6). Similar tissue-dependent changes have been previously reported in aging wild-type mice [15,33] and rats [34] by other investigators. It should be noted, however, that both decreased and increased copy number of mtDNA can cause mitochondrial dysfunction [19]. In fact, increased mtDNA copy number is found in human patients with bipolar disorder, which is associated with accelerated aging [35]. However, similar to the present findings in Cbs-deficient mice, TL was not affected in the bipolar disorder patients.
The limitations of this study should be noted. First, we did not study the expression of senescence-associated genes at the protein level. Measuring senescence-related protein levels in the mouse (e.g., p16) can be challenging mostly due to the availability of antibodies [36]. However, as regulation of these genes occurs at the transcriptional level [4,30,31], one can expect to observe similar changes at the protein level caused by the Cbs genotype. Second, we did not have tissues suitable for immunohistochemistry of markers such as senescence-associated β-galactosidase (SA-β-gal). However, although SA-β-gal is prominent in senescent sells, it is neither required, nor a determinant of the senescence phenotype [4,37]. Third, although the number of Cbs −/− and Cbs −/− mice was 20 to 40 per each Cbs genotype group, the number of animals in the sex-and age-stratified subgroups was accordingly much lower. However, despite the relatively low number of animals in these subgroups, we found that age and sex, which are known to affect TL [13,15], were associated with TL in the present study (Tables 1 and 2). We also found that age, known to affect mtDNA copy number in wild-type mice [15,33] and rats [34], was associated with mtDNA in the present study ( Figure 3 and Table 3).
In conclusion, we showed that Cbs −/− mice had significantly upregulated expression of some senescence-related mRNAs (Pai-1, p21, Mcp-1, and Il-6) in liver, but not in brain, and reduced mtDNA copy number in liver, but not in brain and blood. We also showed that TL in any of those tissues was not affected by the Cbs −/− genotype. These findings suggested that accelerated senescence, mostly affecting liver, and impaired mitochondrial function, but not telomere shortening, could contribute to the reduced life span of Cbs −/− mice.

Mice
Transgenic Tg-I278T Cbs −/− mice on a C57BL/6J genetic background [27] were mated with their Tg-I278T Cbs +/counterparts to generate sufficient numbers of Tg-I278T Cbs −/− and Tg-I278T Cbs +/− animals required for the experiments. The mice were bred and housed at the New Jersey Medical School Animal Facility [29,38]. In these animals, the human CBS-I278T variant was under the control of the zinc-inducible metallothionein promoter, which prevents the neonatal lethality of the mouse Cbs −/− genotype by supplementation of the drinking water with 25 mM ZnCl 2 [27]. The zinc water was replaced by plain water after weaning at 30 days. The mice were fed a normal rodent chow (LabDiet5010, Purina Mills International, St. Louis, MO, USA). The Cbs genotype was established by PCR using the following primers: forward 5'-GGTCTGGAATTCACTATGTAGC-3', Cbsreverse 5'-GAGGTCGACGGTATCGATA-3' (affording a 176 bp product), Cbs+ reverse 5'-CGGATGACCTGCATTCATCT-3' (affording a 300 bp product). Mice (n = 40 to 80) of both sexes, 63 to 408 days old, were used in experiments. Animal procedures were approved by the Institutional Animal Care and Use Committee at the New Jersey Medical School.

Blood and Tissue Collection
Blood was collected from cheek veins into Eppendorf tubes containing 1% (v/v) 0.5 M EDTA. After centrifugation (2000× g, 10 min, 4 • C), separated plasma and cells were frozen at −80 • C. Mice were euthanized using CO 2 and the organs collected and frozen on dry ice. Brains and livers were powdered with dry ice using a mortar and pestle and stored at −80 • C.

Total Hcy Assays
Plasma tHcy was quantified by the conversion to Hcy-thiolactone, which was then separated by cation exchange HPLC using a Poly CAT A column, 35 × 2.1 mm, 5 µM, 300 Å (PolyLC), post-column derivatized with o-phthalaldehyde, detected and quantified by fluorescence as previously described [28,39]. An Agilent Infinity 1260 system, containing a degasser, binary pump, high performance auto-sampler, thermostated column compartment, and fluorescence detector, was used.

DNA Extraction
Total DNA, extracted from brain, liver, or whole mouse blood using the phenol method, was treated with RNase A (Thermo Scientific, Warsaw, Poland) and diluted with ultrapure water. The integrity and quality of the DNA were tested by agarose gel electrophoresis and spectroscopy (Thermo Scientific, Warsaw, Poland).
The master mix for each PCR amplification tube was prepared with iTaq™ Universal SYBR ® Green Supermix (Bio-Rad Polska Sp. z o.o., Warsaw, Poland), 900 nM of telomere-or albumin-specific primers, and 160 ng DNA as a template. The optimized thermal cycling profile for each reaction mixture was as follows: The relative telomere length (TL) was calculated from the Ct values as the ratio of the telomere signal to the single copy albumin gene signal (T/S ratio) [40,41]. To calculate ∆Ct, the Ct for the single-copy albumin gene was subtracted from the telomere Ct (∆Ct = Ct telomere − Ct albumin ). To calculate ∆∆Ct, the ∆Ct for the DNA from control Cbs +/− individuals was averaged and subtracted from the ∆Ct for the DNA from each Cbs −/− mouse and Cbs +/− control animal (∆∆Ct = ∆Ct Cbs−/− − ∆Ct Cbs+/− ). The T/S values were calculated according to the T/S = 2 −(∆∆Ct) equation [42]. Each sample was measured in duplicate. The reproducibility of the measurements was 6%.
All reactions were performed in 10 µL volumes in duplicate using 20 ng DNA, 3 µM primers, and 5 µL iTaq Universal SYBR Green Supermix (Bio-Rad Polska Sp.z o.o.) using a CFX96 Touch Real Time PCR Detection System (Bio-Rad). Cycling parameters: 95 • C (10 min), 40 cycles at 95 • C (15 s), 60 • C (10 s), and 72 • C (15 s). The homogeneity of qPCR products was confirmed by melting curve analysis (65 • C and a progressive increase up to 95 • C at 0.5 • C/min). Analysis of the data was performed with the CFX Manager™ Software and Microsoft Excel. Calculations were similar as those for telomere length above [42].

RNA Extraction
Total RNA was extracted using a column-based Total RNA Purification Kit (Novazym, Poznań, Poland). Three-zone reagent was added to the frozen powdered brain or liver tissue. After the extraction step, the supernatant was mixed with 70% ethanol, loaded onto a total RNA miniprep column, washed with 70% ethanol, and eluted with ultrapure water. The RNA was DNase treated using the DNase I (Thermo Scientific, Warsaw, Poland). The integrity and quality of the RNA were confirmed by agarose gel electrophoresis and spectroscopy (Thermo Scientific, Warsaw, Poland).

Statistical Analyses
Data are expressed as the means ± SEM (Figures 1 and 2) and the means ± SD (Supplementary Tables S1-S6). Comparisons between groups were analyzed using an unpaired 2-sided t-test. Associations between telomere length and other variables were analyzed by Pearson's correlations and linear regression. The interaction effects of age, gender, tHcy, or the CBS genotype on TL and mtDNA were examined by analysis of covariance and/or multiple regression models where appropriate. Statistical analysis was performed using Statistica, Version 13 (TIBCO Software Inc., Palo Alto, CA, USA, http://statistica.io).
Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/7/2520/ s1. Figure S1. Relationships between blood TL, age, and tHcy in Cbs −/− and Cbs + /− mice. Blood TL vs. age in (A) female and (B) male mice. (C) Blood TL vs. tHcy; Figure S2. Telomere length (TL) in brains (A) and livers (B) of Cbs −/− and Cbs +/− mice. TL was quantified by qPCR in Cbs −/− mice (n = 21) and sex-and age-matched control Cbs +/− siblings (n = 22) as described in the Materials and Methods. Box and whiskers represent the mean ± SEM and the mean ± 95% CI, respectively; Figure S3. Tert mRNA levels in brains (A) and livers (B) of Cbs −/− and Cbs +/− mice. Tert mRNA was quantified by qPCR in Cbs −/− mice (n = 21) and sex-and age-matched control Cbs +/− siblings (n = 22) as described in the Materials and Methods. Box and whiskers represent the mean ± SEM and the mean ± 95% CI, respectively; Figure S4. mtDNA levels in blood (A), livers (B), and brains (C) of Cbs −/− and Cbs +/− mice. mtDNA was quantified by qPCR in Cbs −/− mice (blood, n = 40; liver and brain, n = 21) and sex-and age-matched control Cbs +/− siblings (blood, n = 40; liver and brain, n = 22) as described in the Materials and Methods. Box and whiskers represent the mean ± SEM and the mean ± 95% CI, respectively; Table S1. Aging/senescence-related mRNAs in the liver of Cbs −/− and Cbs +/− mice stratified by sex and age; Table S2. Aging/senescence-related mRNAs in the brain of Cbs −/− and Cbs +/mice stratified by sex and age; Table S3. Blood TL, plasma tHcy levels in Cbs −/− and Cbs +/− mice stratified by sex and age; Table S4. Brain TL and Tert mRNA expression in Cbs −/− and Cbs +/− mice stratified by sex and age; Table S5. Liver TL and Tert mRNA expression in Cbs −/− and Cbs +/− mice stratified by sex and age; Table S6. mtDNA levels in the blood, brain, and liver of Cbs −/− and Cbs +/− mice stratified by sex and age; Table S7. Determinants of Tert mRNA in the brain and liver of Cbs −/− and Cbs +/− mice: Multiple regression analysis; Table S8. List of primers used for mouse mRNA quantification by RT-qPCR.