Haploinsufficiency of the Myc regulator Mtbp extends survival and delays tumor development in aging mice

Mtbp^+/- mice have increased longevity

Since Myc^+/- mice have increased longevity [8] and we have shown that Mtbp is a positive regulator of Myc [22,23], we investigated the contribution of Mtbp to longevity using a cohort of littermate-matched Mtbp^+/+ and Mtbp^+/- mice. Mtbp heterozygous mice had increased longevity compared to wild-type controls, exhibiting a median survival of 785 days compared to 654 days (p=0.0013; Figure 1A, Supplemental Figure S1), a 20% increase. This significant difference in lifespan was represented in both male and female populations (Figure 1B and 1C). Mtbp heterozygous males had a median survival of 774 days, compared to 672 days for wild-type control males (p=0.0166, Supplemental Figure S1), a 15.2% increase. Mtbp^+/- females had a median survival of 790 days, compared to 650.5 days for Mtbp^+/+ females (p=0.0439), a 21.4% increase (Supplemental Figure S1).

In addition to median lifespan, Mtbp heterozygosity also increased maximum lifespan. Specifically, Mtbp^+/- mice were overrepresented in the longest living decile and quartile of mice with 9 of 11 (81.8%) and 19 of 26 (73.1%) of the mice, respectively (Figure 1D). The trend was not affected by gender, as the longest living decile and quartile of mice were 4 of 5 (80.0%) and 9 of 12 (75.0%) heterozygous males, respectively, and 4 of 5 (80.0%) and 10 of 13 (76.9%) heterozygous females, respectively. In contrast, Mtbp wild-type mice (all mice and both male and female) were disproportionally represented in the shortest lived decile and quartile of mice 90.9% and 80.8%, respectively (Figure 1D). This was also reflected in the observation that Mtbp^+/- mice have a significantly decreased instantaneous death rate compared to Mtbp^+/+ littermate-matched controls (log-rank p=0.0013, Chi-sq=10.27, df=1; Figure 1E). Both Mtbp heterozygous males and females have a significantly reduced instantaneous death rate (log-rank p=0.0166, Chi-sq=5.74, df=1 and log-rank p=0.0439, Chi-sq=4.06, df=1, respectively; Supplemental Figure S2). Therefore, an Mtbp haploinsufficiency confers increased median and maximum survival.

Delay in tumorigenesis and a change of tumor spectrum in Mtbp heterozygous mice

As is commonly seen in C56Bl/6 mice [26,27], gross and histopathological tissue analysis at time of death of representative mice demonstrated the majority had cancer (17 of 23 Mtbp^+/+ mice and 29 of 34 Mtbp^+/- mice). Notably, 32.4% (11 of 34) of Mtbp^+/- mice had lymphoma, which was twice the incidence of lymphoma in Mtbp^+/+ mice (17.4%, 4 of 23); (Figure 2A). The lymphomas were detected at an average age of 840 days in heterozygotes, compared to 682.3 days in wild-type controls, a significant delay (p=0.0320, Supplemental Figure S1). Similarly, Mtbp^+/- mice developed carcinomas later in life at 848 days (3 of 34, 8.8%) compared to 694 days for Mtbp^+/+ mice (3 of 23, 13.0%) and the tissue distribution of the carcinomas differed between the two genotypes. Specifically, two carcinomas that developed in the wild-type mice were hepatocellular carcinoma and one was a carcinoma of the small intestine, whereas two of the Mtbp heterozygous mice developed aural squamous cell carcinoma and one had pulmonary adenocarcinoma. The difference in age of carcinoma development and the type of carcinoma that emerged between the two genotypes was not statistically significant likely due to the small number of carcinomas that developed (p=0.1412, Supplemental Figure S1) (Figure 2B). In contrast, Mtbp^+/+ and Mtbp^+/- mice had a similar frequency and age of onset of sarcoma. Approximately half of the cancers that developed in both genotypes of mice were sarcoma [10 of 23 (43.5%) for Mtbp^+/+ and 15 of 34 (44.1%) for Mtbp^+/-) (Figure 2C)], occurring at a mean age of 806 days for Mtbp^+/+ and 743 days for Mtbp^+/- mice (p = 0.1547, Supplemental Figure S1). The vast majority of the sarcomas in both the wild-type and Mtbp heterozygous mice were histiocytic sarcomas. One Mtbp wild-type mouse developed both a sarcoma and a carcinoma, and all Mtbp^+/- mice that were diagnosed with a malignancy had only one tumor type. Although, twice the proportion of Mtbp wild-type control mice were cancer free at time of death (7 of 23, 30.4%) compared to Mtbp heterozygous mice (5 of 34, 14.7%); (Figure 2D), the Mtbp^+/- mice lived an average of 836.4 days compared to 640.3 days for wild-type controls (Supplemental Figure S1). This difference in Mtbp^+/- mice represents a significant delay in mortality among cancer free mice (p=0.03340). These data collectively indicate a decrease in Mtbp expression alters the tumor spectrum and age of onset as mice age, as well as extends overall survival independent of cancer development.

Mtbp^+/+ and Mtbp^+/- mice move, reproduce, and develop similarly

Long-lived mouse models will often retain elevated motor function compared to controls, particularly as they age. To determine if Mtbp heterozygosity improved locomotor activity, open field testing was performed for 1 hour on two days with a cohort of old (1.5 year) littermate matched mice. Although there was a trend for Mtbp heterozygotes to travel a greater distance (5737.7 cm) compared to wild-type controls (4551.0 cm), this difference did not reach statistical significance (p=0.1142; Figure 3A). When locomotor function was actively challenged using a rota-rod endurance test, the Mtbp^+/- mice (78.0 seconds) performed similarly to Mtbp^+/+ mice (73.6 seconds) after training (Figure 3B; p=0.3923). Analogous results were also obtained with younger mice (Supplemental Figure S3).

In nature, many animal species with increased longevity have reduced reproductive capacity to limit overpopulation. This trend has been reported in some long-lived mouse models [28]. Thus, we compared the reproductive efficiency of Mtbp^+/+ and Mtbp^+/- female mice. Only the number of pups per birth from crosses between Mtbp^+/+ and Mtbp^+/- mice were quantified, as deletion of Mtbp is embryonic lethal and would artificially lower the number of pups birthed [29]. This examination did not reveal a significant difference in the average number of pups per litter birthed by Mtbp^+/+ (6.6) and Mtbp^+/- (7.3) females (p=0.2247; Figure 4A).

Some long-lived mouse models reported to have reduced growth, resulting in smaller body size [30]. We detected no size differences in mature Mtbp^+/- mice. Specifically, Mtbp^+/+ and Mtbp^+/- mice had similar nose-to-rump lengths of 9.89 cm and 9.74 cm, respectively (p=0.9999; Figure 4B) as well as femur lengths of 15.0 mm and 14.8 mm, respectively (p=0.7160; Figure 4C) at the time of sacrifice. Given this observation, it was not surprising that analysis of serum isolated and frozen at time of sacrifice did not show a statistically significant difference in the level of circulating insulin-like growth factor-1 (IGF-1; p=0.4175; Figure 4D), a major growth-promoting factor [31]. Therefore, an Mtbp haploinsufficiency did not impact locomotion, birth rates, or bone size.

Long-lived Mtbp^+/- exhibit signs of increased cellular metabolism

Many long-lived mouse models have changes in metabolism detectable at a systemic level. To determine if Mtbp expression modulates levels of circulating metabolic markers, serum was isolated and frozen immediately after sacrifice of long-lived mice starved for 5 hours. The analysis revealed Mtbp^+/+ and Mtbp^+/- mice had similar levels of serum cholesterol (p=0.3572; Figure 4E) and triglycerides (p=0.4037; Figure 4F). Moreover, the circulating level of glucose (p=0.7116; Figure 4G) and insulin (p=0.6963; Figure 4H) were also similar, reflecting no major changes in physiologic glucose regulation.

Although systemic changes in cholesterol, triglyceride, and glucose metabolism were not observed, we examined markers of cellular metabolism in tissues of long-lived mice to determine if decreased Mtbp expression modulates metabolism at a molecular level. Using flash frozen tissue at the time of sacrifice, mRNA was isolated from the liver, skeletal muscle (gastrocnemius) and brown fat pad of long-lived mice. In the liver, Mtbp^+/- mice exhibited a global trend toward, and at times significantly, increased expression of metabolic markers (Figure 5), similar to previous reports for Myc^+/- mice [8]. Mtbp heterozygosity resulted in a trend toward increased expression of basic metabolic genes such as Gls, Hk2, Ncl, Cad, and Odc that control cellular energy flux, protein translation, and macromolecule synthesis (Figure 5A-E). Most notably, the livers of Mtbp^+/- mice showed a nearly 2-fold and statistically significant increase in the expression of Pcg1α (p=0.0106; Figure 5F) and Pcg1β (p = 0.0499; Figure 5G) compared to wild-type controls. This increase suggests elevated mitochondrial biogenesis and function in long-lived Mtbp heterozygous mice. These increased levels of metabolic markers in Mtbp^+/- livers also coincided with a nearly 60% increase in the expression of Sirt1 (p=0.1529; Figure 5H), a well-known anti-aging gene linked to caloric restriction [32].

The global increase in metabolic markers observed in the livers of Mtbp heterozygous mice was largely not recapitulated in skeletal muscle (Figure 6) and brown fat (Figure 7). In skeletal muscle for example, there was only a statistically significant increase in the level of Cad in Mtbp^+/- mice (p=0.0202), but analogous levels of Gls, Hk2, Odc, and Ncl compared to Mtbp^+/+ mice (Figure 6A-E). There was a trend toward significantly decreased expression of Pcg1α (p=0.0736; Figure 6F) and Pcg1β (p=0.0710; Figure 6G), but similar levels of Sirt1 (Figure 6H) in Mtbp heterozygous skeletal muscle. In brown fat, levels of Gls, Hk2, Odc, Cad, and Ncl were similar between wild-type and heterozygous Mtbp mice (Figure 7A-E). However, the brown fat of Mtbp^+/- mice exhibited a trend toward increased expression of Pcg1α (p=0.0531; Figure 7F), but equivalent levels of Pcg1β (p=0.8492; Figure 7G). The levels of Sirt1 were analogous between the two genotypes (Figure 7H). Therefore, the data show Mtbp heterozygosity alters markers of cellular metabolism in disparate tissues, although the effects are more pronounced in the liver.

Mice

Mtbp^+/- [29] and littermate-matched Mtbp^+/+ C57Bl/6 mice of both genders were generated through interbreeding and were housed together. For survival analysis, mice were sacrificed after meeting humane end-of-life criteria. Necropsy with tissue collection was performed and tissues were evaluated by a board-certified veterinary pathologist (K.B.) in a blinded manner. For analysis of healthy aged mice (29 months-old), the mice were starved for 5 hours and crown-to-rump length was measured. Blood was collected and analyzed for blood glucose levels with Accu-chek test strips (Roche Diagnostics, Indianapolis, IN, USA [41], centrifuged and plasma frozen for later analysis. Mice were sacrificed by cervical dislocation. Liver and gastrocnemius muscle were frozen with a Wallenburg clamp pre-cooled in liquid nitrogen as previously described [42]. Brown fat pads were collected and frozen. Tissues were kept at -80^oC until analysis. Femurs were collected and measured with electronic calipers. Experiments were approved by the Vanderbilt Institutional Animal Care and Use Committee and followed all federal and state rules and regulations.

Quantitative real-time PCR (RT-PCR)

Total RNA was isolated from frozen tissues, cDNA was generated, and qRT-PCR was performed as previously described [23]. Primer sequences are listed in supplemental material.

Serum analysis

Analysis of serum was performed by the Mouse Metabolic Phenotyping Center in the Hormone Assay and Analytical Services and the Lipids and Lipoproteins Subcores at Vanderbilt University. Insulin levels were measured with a radioimmunoassay (SRI-13K, EMD Millipore, Billerica, MA, USA). IGF-1 levels were measured using a magnetic Luminex screening assay (LXSAMSM-01, R&D Systems, Minneapolis, MN, USA). Triglyceride and cholesterol levels were measured using Raichem reagents (R80035 and R84098, Cliniqa, San Marco, CA, USA).

Locomotion analysis

Explorative locomotion or open field testing was performed using a 48 channel IR controller (ENV-520, Med Associates Inc., St. Albans, Vermont) on young (6 months-old) and aged (21 months-old) mice. On two different days, mice were placed in the open field, the total distance traveled was recorded for 1 hour and the average of the two measurements was reported. Additionally, on three consecutive days, mice were placed on a five-lane accelerating rota-rod (47600, Ugo Basile, Varese, Italy) for three sequential trials separated by 10 minutes of rest. The rota-rod accelerated from 4 to 40 rpm over 3 minutes with a cut-off time of 3 minutes. Time running on the rota-rod was determined by using pressure sensors to detect falls or by observing >3 consecutive rotations of the mouse around the rod. The average times from day 3 were recorded [43].

Statistical evaluation

A log-rank test was used to calculate significance in Figure 1 and Supplemental Figure S1. Analysis for Figure 1E and Supplemental Figure S1 was performed with JMP statistical software (v12.2.0). Student’s t-test was used to calculate significance for data in Figures 2-7 and Supplemental Figures S2 and S3.