The house mouse maintains constant telomere length throughout life

Abstract

Telomeres protect the chromosome ends from deleterious DNA damage response and repair activities. In humans, telomerase maintains telomere length in germ and stem cells, but not in most somatic cells. Consequently, telomeres shorten with cell division and age, limiting cell proliferation and protecting against cancer. When telomeres become critically short, they may also cause senescence, inflammation, and organ failure, which are major drivers of aging. Therefore, maintaining an optimal, age-appropriate telomere length is crucial for healthy aging. In the house mouse, Mus musculus, telomerase is active in most somatic tissues, yet its long telomeres were thought to shorten rapidly with age. We have followed telomere length over age in blood and tail of wild-type M. musculus and in two engineered mouse strains with shorter telomeres (Telomouse and HHS mouse). We also measured the precise length of single telomeres in blood leukocytes of these mouse strains by a long-read nanopore sequencing method, NanoTelSeq. We show that telomeres in blood and tail of these three mouse strains do not shorten with age. We conclude that M. musculus maintains long telomeres in blood and tail throughout life, excluding the possibility that global telomere shortening in these tissues contribute to aging-associated phenotypes.

Graphical Abstract

Graphical Abstract.

Introduction

Telomeres, the protective caps of eukaryotic chromosomes, consist of a conserved repetitive nucleotide sequence, TTAGGG in all vertebrates, bound by specialized proteins [1]. In mammals, telomeric restriction fragments (TRFs) range in length from a few kilobases (kb) to several tens of kb, and up to 150 kb in the house mouse Mus musculus [2–4]. Telomeres shorten with age due to oxidative damage and accidental breakage, and in mitotic cells also due to the processing of the telomeric 5′ end and the inability of the canonical replication machinery to fully replicate the telomere to its end [5, 6]. When telomeres reach a critical length, a cell with intact DNA damage response (DDR) would activate cell cycle arrest and enter senescence, setting a limit to cell proliferation and protecting against cancer [7]. To compensate for telomere shortening and extend the proliferative capacity, a special enzyme, telomerase, can be induced to elongate the G-rich 3′ overhang of the telomere [8]. In humans, telomerase is expressed in the germline and during early embryonic development, but upon differentiation it is silenced in most somatic cells, resulting in telomere shortening with age. Highly proliferative cells, such as stem cells, lymphocytes, and most cancer cells, maintain telomerase activity [5]. Telomerase expression, biogenesis, and action are tightly regulated to maintain an optimal age-appropriate telomere length [8]. Excessive telomere shortening may cause cellular senescence and aging-associated pathologies, and, in extreme cases, telomere biology disorders (TBDs) [9, 10]. On the other hand, overly long telomeres may cause clonal hematopoiesis and contribute to various types of cancer [11]. In contrast to humans, in M. musculus and some other species, telomerase activity is detectable in most somatic tissues [2, 12]. The roles of telomere length in replicative aging and cancer prevention have been studied mostly in humans, and less so in other organisms such as M. musculus [2].

Telomere length is highly variable not only among different mammals but also among individual humans, different cell types, and even different chromosome ends [2, 13, 14]. We have recently developed a method for measuring the length of single telomeres by long-read nanopore sequencing, called NanoTelSeq, which revealed a large heterogeneity in the precise length of the telomere repeat tracks in blood leukocytes of single human individuals, ranging from 2 to 28 kb [15]. Ultimately, the shortest telomeres are the ones that dictate cell fate and therefore are the most relevant to aging [16]. In humans, it was estimated that telomeres shorten by 30–150 bp per replication cycle in fibroblasts and lymphocytes [17]. A recent careful examination of the mechanism of telomere shortening in the absence of telomerase revealed a rate of 84–110 nt per cell division, reflecting the average length of the overhang [6]. Previous studies have suggested that the rate of accumulation of short telomeres, rather than the initial telomere length, is the critical variable that determines the lifespan of a species [18, 19]. In the same study, Whittemore et al. also estimated that telomeres in M. musculus shorten with age at a rate of 7 kb/year, about 100-fold faster than human telomeres, and proposed that this rapid erosion contributes to aging and limits lifespan in mice.

To examine whether shorter telomere setpoint at birth would result in the accumulation of critically short telomeres and age-related phenotypes, we utilized two mouse models we have previously generated by introducing a single homozygous amino acid change into the helicase RTEL1 of M. musculus. In the first mouse model, termed Telomouse, we changed methionine 492 of RTEL1 to a lysine (Rtel1^M492K) [15]. This mutation in Telomouse resulted in progressive telomere shortening with each generation, eventually stabilizing at F15 and later generations. Telomice at the age of 6 months displayed median telomere length of 6.1 kb in tail tissue and 6.4 kb in the blood as measured by NanoTelSeq, which are comparable to human telomeres [15]. In the second mouse model, termed HHS mouse, we changed the same amino acid to an isoleucine (Rtel1^M492I), which is a mutation found in human patients of a severe TBD termed Hoyeraal–Hreidarsson syndrome (HHS) [20, 21]. HHS mouse displayed a milder telomere shortening than Telomouse and increased telomere damage not observed in Telomouse [21]. Telomere length stabilized at F8 and successive generations. F14 HHS mice at the age of 13 months displayed median telomere length of 17.8 kb in tail tissue and 19.5 kb in the blood, as measured by NanoTelSeq.

Given the shorter telomeres of Telomice and HHS mice, we expected that if the rate of telomere shortening in mice was indeed 7 kb/year, as previously suggested [18, 19], then telomeres in these strains should become critically short faster than those of wild-type (WT) M. musculus, potentially leading to age-related diseases. Therefore, we assessed telomere length in these mice strains at different ages, in blood and tail samples, using in-gel hybridization and NanoTelSeq. Surprisingly, we found no evidence of telomere shortening, neither in the short telomere strains nor in WT mice. On the contrary, based on the more accurate NanoTelSeq, telomeres in the blood of WT mice appeared to lengthen with age. We conclude that M. musculus maintains a constant telomere length setpoint throughout life and, at least in the tissues analyzed, aging phenotypes in the mouse cannot be attributed to overall telomere shortening.

Materials and methods

Ethics statement

The use of animals in this study followed the Guide for the Care and Use of Laboratory Animals, Laboratory Animal Ordinances, and the Animal Welfare Act. This study was approved by the University of Pennsylvania Institutional Animal Care and Use Committee, protocol number 805623, and by the Hebrew University Ethics Committee, protocol number MD-21-16590-3.

Genomic DNA extraction

DNA was extracted either by Monarch cells and blood HMW genomic DNA extraction kit (NEB Inc.) or by phenol extraction as follows: Leukocytes were obtained from mouse blood by lysing red blood cells in 155 mM NH₄Cl, 10 mM KHCO₃, and 0.1 mM ethylenediaminetetraacetic acid (EDTA; pH 8) followed by centrifugation. Mouse leukocytes were lysed in 10 mM Tris (pH 7.5), 10 mM EDTA, 0.15 M NaCl, 0.5% sodium dodecyl sulfate (SDS), and 100 μg/ml proteinase K overnight at 37°C. Mouse tail samples were lysed in 100 mM Tris–HCl (pH 8.5), 5 mM EDTA, 0.1% SDS, 200 mM NaCl, and 100 μg/ml proteinase K overnight at 50°C. Following cell lysis, high molecular weight genomic DNA was phenol extracted, ethanol precipitated, and dissolved in 10mM tris, 1mM EDTA pH 8.0 (TE) as previously described [15]. Genomic DNA samples were examined by agarose gel electrophoresis of undigested DNA, and samples suspected to be degraded were excluded from further analysis.

In-gel analysis of TRFs

Genomic DNA was treated with RNase A, digested with the HinfI restriction endonuclease, and quantified by Qubit fluorometer. Equal amounts (1–5 μg) of the digested DNA samples were separated by pulsed-field gel electrophoresis (PFGE) as previously described [15], using a Bio-Rad CHEF DR-II apparatus in 1% agarose and 0.5× tris-borate-EDTA (TBE) buffer, at 14°C, 200 V, and pulse frequency gradient of 1 s (initial) to 6 s (final) for 18–22 h. Gels were ethidium stained, dried, and hybridized to analyze the native signals. DNA was then denatured by incubating the dried gel in 0.5 M NaOH and 1.5 M NaCl for 30 min, neutralized in 1.5 M NaCl and 0.5 M Tris–HCl (pH 7.0) for 30 min, rinsed with H₂O, and hybridized as previously described [21]. Probe mixtures contained a telomeric oligonucleotide (AACCCT)₃, and mid-range PFG ladder (NEB Inc.) and 1 kb ladder (GeneDirex Inc.) that were digested with HinfI, dephosphorylated by quick calf intestinal phosphatase (NEB Inc.), and heat inactivated. All probes were 5′ end labeled with [γ-³²P]-ATP and T4 polynucleotide kinase (NEB Inc.). The gels were exposed to a phosphor imager. TRF length was quantified using TeloTool (in the “corrected” mode) [22]. Native and denatured in-gel hybridization intensities were quantified using ImageQuant TL (GE Healthcare Inc.).

NanoTelSeq

Two to twelve samples were pooled in a sequencing library using barcoded telorette 3 oligonucleotides as described previously [15]. The barcoded telorette oligonucleotides were first annealed to a complementary teltail-tether oligonucleotide, and then each double-stranded telorette with a specific barcode (0.1 μM) was ligated to an individual DNA sample (0.5–3 μg each). The ligation reactions were stopped by adding EDTA to a final concentration of 20 μM, purified by AMPure XP beads (Beckman Coulter Life Sciences), pooled together and ligated to 4 μl sequencing adapter, and sequenced following the standard protocol for genomic DNA ligation sequencing with native barcoding (SQK-LSK109 with EXP-NBD104 or SQK-NBD114.24; Oxford Nanopore Technologies). 0.5–7.5 μg of the sequencing libraries was loaded onto a Nanopore R9 or R10 flow cell (for SQK-LSK109 or SQK-NBD114.24, respectively).

Oligonucleotides used for sequencing

Telorette3-oligos:

P-5′-TGCTCCGTGCATCTCC-AAGGTTAA-barcode-CAGCACCT-CTAACC-3′

“P” indicates 5′-phosphate and barcode is one of the 24-nt-long barcodes of the native barcoding kit (SQK-NBD114.24, Oxford Nanopore Technologies).

Teltail-tether:

5′-AACCTTGGAGATGCACGGAGCAAGCAAT-3′

Computational processing of nanopore sequencing reads

Deducing the nucleotide sequence (base calling) from the raw data (POD5 files) was performed with the Nanopore Dorado basecaller at super accurate mode. The sequences (FastQ files) were filtered and analyzed by the dedicated application TelomereAnalyzer (https://doi.org/10.5281/zenodo.8379363), searching for the patterns (after reverse complement) of YYAGGG (Y indicates C or T) for R9 chemistry and TTAGGG for R10 chemistry, with one mismatch allowed in each telomeric repeat. The telomere length (with and without mismatches) and read length were extracted and presented as plots of telomere density along the entire read. Telomere density was calculated in a moving window of 100 nt and describes the fraction of telomeric repeats within each 100-nt sequence (from 0, no telomeric sequence, to 1, fully telomeric). A telomere is identified by a summed telomere density of at least 2 for at least three consecutive windows, each having a telomere density of at least 0.3. Then, the telomere beginning and end are defined by the first and last telomeric repeats within the identified windows or flanking windows.

Mice

For in-gel analysis, DNA samples from a total of 33 WT (100% pure M. musculus C57BL/6), 87 K/K, and 32 I/I mice were analyzed multiple times in separate gels, following them at different ages (as detailed in Supplementary Tables S1–S3). The average age for I/I mice analyzed in the gels was 329 days, while the average age for M/M mice was 353 days, and for K/K was 339 days. At each generation (F) of breeding (for I/I and K/K), the number of mice and age-ranges are indicated below:

• WT mice (M/M): 33 mice, from 63 to 665 days old.

• Telomice (K/K):

  • F5 (K/K): 7 mice, 341–895 days old;

  • F6 (K/K):13 mice, 318–787 days old;

  • F7 (K/K): 2 mice, 329–872 days old;

  • F8 (K/K): 7 mice, 233–778 days old;

  • F9 (K/K): 34 mice, 81–718 days old;

  • F10 (K/K): 7 mice, 42–412 days old;

  • F12 (K/K): 9 mice, 118 and 171 days old;

  • F14 (K/K): 8 mice, 44–205 days old.

• HHS mice (I:I):

  • F3 (I/I): 6 mice, 115–1014 days old;

  • F4 (I/I): 7 mice, 174–738 days old;

  • F5 (I/I): 5 mice, 149–505 days old;

  • F6 (I/I): 4 mice, 48–455 days old;

  • F10 (I/I):10 mice, 53–154 days old.

For nanopore analysis, DNA samples from six WT (C57BL/6), four F15 Telomice (from two litters of the same Telomouse parents), and three littermates F14 HHS mice were analyzed at different ages as detailed in Supplementary Table S4.

Statistical analysis

Statistical analysis was performed by Microsoft Excel and GraphPad Prism 8.0, using simple linear regression, and two-tailed paired or unpaired t-test, as indicated for each figure. P < 0.05 was considered statistically significant.

Results and discussion

Constant telomere length in the mouse throughout life

Telomeres in peripheral blood leukocytes of WT mice were reported to shorten at a rapid rate of 7 kb/year [18, 19]. The median telomere length in blood leukocytes of a 6-month-old Telomouse was 6.4 kb, and 4.6% of the telomeres were below 3 kb, as measured by NanoTelSeq [15]. If the same rate of telomere shortening as in the WT also occurred in Telomice, we would expect further shortening to accumulate critically short telomeres and cause bone marrow failure, as occurs in TBD patients. However, no apparent aging phenotypes or overt disease was observed in the mutant mice with advancing age. To examine the rate of telomere shortening with age, we measured the mean length of the TRF in tail and blood samples of Telomice, HHS mice, and pure WT M. musculus WT control mice of different ages by PFGE and in-gel hybridization with a telomeric probe. Surprisingly, following the TRF length in WT mice from age 2 to 23 months, we found no indication for telomere shortening in blood samples and even an apparent elongation in the tail (Fig. 1 and Supplementary Table S1). Furthermore, in eight different generations of Telomouse, ages 1–29 months, and in five different generations of HHS mice, ages 2–33 months, we did not detect any shortening of the mean TRF length (MTL) (Figs 2 and 3, Supplementary Figs S1 and S2, and Supplementary Tables S2 and S3). In addition to the overall telomere length, we also examined the telomeric 3′ overhang length by quantifying the native hybridization signal before denaturation of the DNA in the gel, and normalizing it to a WT sample in each gel. No difference in the overhang signal was found between WT and Telomouse, as reported previously [15]. Furthermore, no change with age was observed in the overhang signal in WT (from 11 to 16 months) and Telomouse (3 to 17 months) tissues (Supplementary Fig. S3 and Supplementary Tables S1 and S2). Altogether, these results indicate that blood and tail telomeres do not shorten with age, consistent with the detectable telomerase activity in murine somatic tissues [12].

Figure 1.

WT mouse telomeres do not shorten with age. Genomic DNA prepared from tail (A) or blood leukocytes (B) of pure Blk6 mice at the indicated ages was digested by HinfI and analyzed by PFGE and in-gel hybridization to the denatured DNA. MTL was quantified by TeloTool [22] and indicated below the lanes. MTL in tail (C) and blood (D) samples from 25 mice was repeatedly measured in independent gels and plotted. The percentage of TRF shorter than 15 kb for each tail (E) and blood (F) sample was calculated from the mean and standard deviation calculated by TeloTool. Each individual mouse in panels (A)–(F) is marked with a different color, and data points having the same color represent samples collected from the same individual mouse at different ages. Linear regression lines and formulas indicate the rates of change in MTL or in the accumulation of TRF < 15 kb with age. P-values indicate the deviation of the linear regression slope from 0. n, number of measurements of all samples. All data are summarized in Supplementary Table S1. (G) Samples were taken from the same individual WT mouse over the period of 9 months (tail) and 11 months (blood), and MTL was measured and plotted.

Figure 2.

Telomice telomeres do not shorten with age. Genomic DNA prepared from tail (A) or blood leukocytes (B) of Telomice from generation nine (F9) at the indicated ages was digested by HinfI and analyzed by PFGE and in-gel hybridization to the denatured DNA. MTL was quantified for each sample by TeloTool [22] and indicated below the lanes. MTL in tail (C) and blood (D) samples from 34 mice was repeatedly measured in independent gels and plotted. The percentage of TRF shorter than 15 kb for each tail (E) and blood (F) sample was calculated from the mean and standard deviation calculated by TeloTool. Each individual mouse in panels (C)–(F) is marked with a different color, and data points having the same color represent samples collected from the same individual mouse at different ages. Linear regression lines and formulas indicate the rates of change in MTL or in the accumulation of TRF < 15 kb with age. P-values indicate the deviation of the linear regression slope from 0. n, number of measurements of all samples. All data are summarized in Supplementary Table S2. (G) Tail and blood samples were taken from one individual Telomouse over the period of 3–20 months, and MTL was measured and plotted.

Figure 3.

HHS mice telomeres do not shorten with age. Genomic DNA prepared from tail (A) or blood leukocytes (B) of HHS mice at the indicated generations and ages was digested by HinfI and analyzed by PFGE and in-gel hybridization to the denatured DNA. MTL was quantified for each sample by TeloTool [22] and indicated below the lanes. MTL in tail samples from 6 F3 HHS mice (C) and in blood samples from 10 F10 HHS mice (D) was repeatedly measured and plotted. The percentage of TRF shorter than 15 kb for each tail (E) and blood (F) sample was calculated from the mean and standard deviation calculated by TeloTool. Each individual mouse in panels (C)–(F) is marked with a different color, and data points having the same color represent samples collected from the same individual mouse at different ages. Linear regression lines and formulas indicate the rates of change in MTL or in the accumulation of TRF < 15 kb with age. P-values indicate the deviation of the linear regression slope from 0. n, number of measurements of all samples. All data are summarized in Supplementary Table S3. (G) Tail DNA was extracted from four individual HHS mice at the indicated ages, and MTL was measured and plotted. Each mouse is indicated with a different color.

Measuring the exact length of single telomeres in blood leukocytes by NanoTelSeq reveals no shortening with age

The MTL of the three mouse strains did not shorten with age (Figs 1–3). However, the length of the shortest telomeres is more physiological relevant than MTL since they are the ones to induce telomere uncapping and DDR once reaching a critical short length [16]. We calculated the percentage of TRF shorter than 15 kb for each sample, from the mean and standard deviation of TRF lengths obtained by TeloTool analysis of the PFGE and in-gel hybridization [15]. These calculations showed no significant increase in TRF shorter than 15 kb with the age of the mice (Figs 1E and F, 2E and F, and 3D). However, this method cannot determine with confidence the abundance and precise length of the shortest telomeres due to the low hybridization signal of short telomeres, the variable length of the sub-telomeric regions included in the TRF, and the limited resolution of the PFGE. Therefore, we sequenced and measured the precise length of individual telomeres by the recently developed long-read sequencing method NanoTelSeq [15]. Consistent with the TRF analysis, the median length of the telomeric repeat tracts in blood leukocytes of six WT mice ages 3–31 months did not shorten with age (Fig. 4A and Supplementary Table S4). In addition, we performed NanoTelSeq on four Telomice generation 15 (F15), ages 3–22 months, and three HHS mice F14, ages 10–18 months, and no shortening of median telomere length was observed (Fig. 4B and C and Supplementary Table S4).

Figure 4.

Telomice accumulate short telomeres with age. Genomic DNA samples, prepared from blood leukocytes of six WT mice (A), four F15 Telomice (B), and three F14 HHS mice (C) at the indicated ages was sequenced by NanoTelSeq as described in the “Materials and methods” section. Scatter plots show the length of individual telomeres in the indicated mouse samples. Median and interquartile ranges are indicated by horizontal lines. Median values in kb are indicated to the right of each scatter plot, and the percentages of telomeres below 2 and 3 kb are indicated to the left of each plot. n, number of telomeric reads. Each individual mouse is marked with a different color, and scatter plots with the same color represent samples collected at different ages from the same individual mouse. All Telomice analyzed were siblings (the ones plotted in red and pink are littermates, and the ones in green and yellow are also littermates), and all HHS mice analyzed were siblings and littermates. The data of all telomere reads are summarized in Supplementary Table S4.

To study whether there was an increase in the number of critically short telomere with age, we calculated the percent of telomeres shorter than 3 and 2 kb. No increase in the percentage of short telomeres was observed in WT and HHS mice with age (Fig. 4A and C). While two of the four Telomice analyzed accumulated increasing number of short telomeres—6.4% and 8.5% below 3 kb, and 3.9% and 6.5% below 2 kb, at 19 and 22 months of age, respectively—the other two did not show this accumulation (Fig. 4B). Further analysis is needed to clarify whether the accumulation of short telomeres in Telomice blood is associated with variable rates of aging or pathological condition, and whether it can model human aging physiology.

NanoTelSeq is the first method that measures directly, by sequencing, the exact length of single telomeres, shortest and longest, without any apparent bias [15]. Other commonly used methods can only estimate telomere length indirectly by measuring the length of TRFs or polymerase chain reaction (PCR) amplification products, or by comparing fluorescence in situ hybridization (FISH) intensities, and thus are far less accurate and reproducible. TRF analysis is limited by the range of electrophoresis separation, the presence of subtelomeric regions of variable and unknown lengths within the TRF, and hybridization signal biases toward longer telomeres; FISH and flow-FISH have limited sensitivity and linear range of signal quantification and can only estimate length by comparing the signal intensity to a reference sample; and PCR-based methods are prone to amplification biases [7, 23, 24]. As demonstrated here, NanoTelSeq can accurately measure the ultralong telomeres of the house mouse. Importantly, it can also measure the shortest telomeres, which are the most relevant to physiology since they may activate DDR upon shortening below a critical length, which is estimated around 1 kb [15]. NanoTelSeq can also identify telomere variant repeats [25] and enable studying their influence on the binding of telomere proteins and on telomere structure and function. We anticipate that long-read nanopore sequencing methods such as NanoTelSeq and others will soon become the methods of choice for analyzing telomere length, sequence, and functionality, for both clinical and basic research purposes [14, 15, 26, 27].

Conclusions

We have shown that M. musculus telomeres do not shorten and may even elongate with age (Figs 1C and 4A). These observations contradict previous reports of rapid telomere shortening in M. musculus [18, 19]. However, they are consistent with the abundance of telomerase activity in mouse somatic tissues [12]. Therefore, short telomeres per se cannot explain aging phenotypes or the limited lifespan of the house mouse. It is possible, though, that short and stable telomeres affect aging physiology differently than long telomeres do, by influencing the accumulation of other aging effectors. One possible way is by affecting gene expression through the telomere position effect over long distance (TPE-OLD) mechanism [28]. Short telomeres may also be more vulnerable to other types of damage, such as oxidized guanines (8-oxo-guanines) that may accumulate at telomeres with age, compromising their function and promoting senescence and aging [29]. Future research using single telomere long-read sequencing methods such as NanoTelSeq will elucidate specific features of telomeres associated with aging in various tissues, and whether/how Telomouse can be used as a valuable model for studying the roles of telomeres in human aging [30].

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was funded by the Israel Science Foundation (ISF) grants 2071/18 and 1342/23 to Y.T., by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) through grants R37-DK053839 and R01-CA249929 to K.H.K., and by the Israel–UK–Palestine GROWTH Fellowship to R.S. Funding to pay the Open Access publication charges for this article was provided by the ISF grant 1342/23.

Data availability

All measured and calculated values are provided in Supplementary Tables S1–S4. Additional gel replicates and data are available upon request to the corresponding authors.