Genetic reduction of mTOR extends lifespan in a mouse model of Hutchinson‐Gilford Progeria syndrome

Abstract Hutchinson‐Gilford progeria syndrome (HGPS) is a rare accelerated aging disorder most notably characterized by cardiovascular disease and premature death from myocardial infarction or stroke. The majority of cases are caused by a de novo single nucleotide mutation in the LMNA gene that activates a cryptic splice donor site, resulting in production of a toxic form of lamin A with a 50 amino acid internal deletion, termed progerin. We previously reported the generation of a transgenic murine model of progeria carrying a human BAC harboring the common mutation, G608G, which in the single‐copy state develops features of HGPS that are limited to the vascular system. Here, we report the phenotype of mice bred to carry two copies of the BAC, which more completely recapitulate the phenotypic features of HGPS in skin, adipose, skeletal, and vascular tissues. We further show that genetic reduction of the mechanistic target of rapamycin (mTOR) significantly extends lifespan in these mice, providing a rationale for pharmacologic inhibition of the mTOR pathway in the treatment of HGPS.

of HGPS cases, death occurs as a result of complications of severe cardiac disease (myocardial infarction or heart failure) or cerebrovascular disease (stroke) at an average age of 14.6 years.
Classic HGPS is caused by a de novo mutation (c.1824C→T, p.G608G) in the LMNA gene, encoding the A-type lamins A, C, and AΔ10 (Eriksson et al., 2003;De Sandre-Giovannoli et al., 2003). Along with B-type lamins, A-type lamins comprise the nuclear lamina underlying the inner nuclear membrane. Prelamin A is subject to several post-translational processing events that include farnesylation of the cysteine in the C-terminal CAAX sequence, cleavage of the terminal AAX sequence with addition of a methoxy group to the terminal cysteine, and subsequent cleavage of the terminal 15 amino acid residues by the metalloprotease ZMPSTE24. Although the G608G mutation does not alter the LMNA coding sequence, activation of an exonic cryptic splice donor site produces an in-frame protein with an internal deletion, termed progerin, that excludes the ZMPSTE24 cleavage recognition site. Permanently farnesylated progerin functions in a dominant-negative manner to disrupt the roles of the nuclear lamina in regulating nuclear shape, DNA replication, transcription, cell division, and chromatin organization (Kubben & Misteli, 2017).
Development of a transgenic model that reproduces the human HGPS phenotype should be useful in testing therapeutic approaches to target either the mutation or mutant gene product in the context of the human sequence. We previously reported the creation of a transgenic mouse harboring the G608G mutated human LMNA on a 164-kb bacterial artificial chromosome (BAC), designated as C57BL/6-Tg(LMNA*G608G)HClns/J, and henceforth referred to as LMNA G/+ (Varga et al., 2006). Although the LMNA transgene expresses human lamin A, lamin C, and progerin in all tissues tested, mice carrying a single copy of the human BAC (LMNA G/+ ) exhibit normal growth and external phenotypic features compared to wild-type littermates or mice carrying a normal human LMNA transgene. The most dramatic finding was the progressive loss of VSMC, particularly in descending aorta and carotid artery, that was associated with thickening of the adventitia and medial layer, and proteoglycan accumulation by 5 months of age.
Arterial calcification was observed in older mice with severe VSMC loss and extracellular matrix deposition. Thus, while single-copy transgenic mice (LMNA G/+ ) lack the external phenotype seen in patients, this model demonstrates the progressive vascular abnormalities that closely resemble the most lethal aspect of HGPS (Varga et al., 2006).
Aging-related disorders such as cancers, diabetes, cardiovascular, and neurodegenerative diseases, as well as normal aging, have been linked to dysregulated signaling of the mechanistic target of rapamycin (mTOR) pathway (Dazert & Hall, 2011). While genetic reduction of mTOR, mLST8, or S6K1 in mice increases lifespan and inhibits the onset of aging-related diseases, inhibitors of mTOR signaling components have drawn interest as potential treatments for disorders associated with deposition of insoluble protein aggregates by means of enhanced clearance through autophagic-lysosomal pathway activation (Selman et al., 2009;Wu et al., 2013). In cell culture experiments, treatment with rapamycin, or its analogs everolimus and temsirolimus, has been demonstrated to reduce progerin aggregates and nuclear blebbing, enhance cell proliferation, and delay the onset of cellular senescence, thereby providing evidence that chemical inhibition of mTOR signaling might restore cellular homeostasis in HGPS (Cao et al., 2011;Cenni et al., 2011;DuBose et al., 2018). These findings have led to a clinical trial of everolimus in children with HGPS, with results expected within a year. Here, we present the phenotypic characterization of double-copy LMNA G608G transgenic mice, which faithfully recapitulate vascular, dermal, adipose, and skeletal features that develop in HGPS patients. Furthermore, we report the beneficial effects of genetic mTOR reduction in these mice as a rationale for pharmacologic intervention in HGPS therapy.

| RE SULTS
We have previously developed a transgenic mouse model of HGPS by using a 164Kb circular BAC containing 4 functional genes (UBQLN4, RAB25, MEX3A, and a recombineered LMNA) in addition to incomplete sequences of two additional genes, SSR2 and SEMA4A (Varga et al., 2006). Initial efforts to characterize the structure of the transgene expressed in our mouse model utilized hybridization capture and short-read NGS sequencing. Although this analysis was useful for obtaining the genomic location of the insertion sites on murine chromosome 4, it lacked the resolution required to resolve the actual transgene configuration and any recombination events that occurred during integration (Dubose et al., 2013). We therefore employed whole-genome sequencing of transgenic mouse DNA and comparison analyses with BAC sequence to determine the precise structure F I G U R E 1 Murine model harboring two copies of the human LMNA G608G transgene develops a severe progeroid phenotype. (a) Growth curves of wild-type (LMNA +/+ ; n = 16 males, 13 females), mice carrying one copy (LMNA G/+ ; n = 8 males, 8 females), or two copies (LMNA G/G ; n = 9 males, 8 females) of the human transgene containing the LMNA G608G mutation. By 5 weeks of age, LMNA G/G weigh significantly less than LMNA +/+ and LMNA G/+ littermates. Males, 5-15 weeks p < 0.01, 16-28 weeks p < 0.001; Females, 5-15 weeks p < 0.05, 16-23 weeks p < 0.01, 24-27 weeks p < 0.001. (b) Histologic analysis of skin from LMNA G/G and wild-type (LMNA +/+ ) littermates at 6 months of age demonstrates decreased subcutaneous fat in transgenic mice. Skin sections were stained with hematoxylin and eosin (H&E) and Masson's trichrome, which reveals the keratin and muscle (red), collagen (blue), and cellular cytoplasm (pink) and nuclei (dark brown). The unstained adipose tissue is located between the collagenous dermal layers (blue) and dark red-stained panniculus carnosus (fast twitch type IIB glycolytic fibers). (c) Micro-computed tomography (μCT)-derived images of femoral bone of wild-type (+/+), single-copy (G/+) and doublecopy (G/G) transgenic mice at 6 months of age. Both cortical (top images) and trabecular (bottom images) structural parameters demonstrate the reduced bone volume in LMNA G/G mice. (d) Lateral images of 5-month-old mice reveal kyphosis and growth deficiency observed in double-copy (LMNA G/G ) transgenic mice compared to wild-type (LMNA +/+ ) and single-copy (LMNA G/+ ) mice. (e) Kaplan-Meier plots illustrating the shortened lifespan of single-copy (LMNA G/+ , n = 79) and double-copy (LMNA G/G , n = 181) transgenic mice relative to wild type (LMNA +/+ ; C57BL/6J, n = 61) and contents ( Figure S1a). Fine mapping of the transgene identified three separate copies of human LMNA, the first of which includes sequence covering exon 1 and portions of the first intron, a second copy that includes exons 2 −12, and a third copy consisting of the full gene sequence with the G608G mutation. Partial gene sequences were identified for SSR2 and SEMA4A, as expected. A third gene, MEX3A, localized to an inversion junction and underwent multiple recombination events resulting in several fragments averaging ~400-500 bp each. Since UBQLN4 sequence was entirely absent from WGS reads we concluded that only two functional genes are encoded within the transgene, including the mutant LMNA and normal RAB25 genes.
Similar to HGPS patients, mice carrying either one (LMNA G/+ ) or two copies (LMNA G/G ) of the human LMNA G608G transgene are indistinguishable from wild-type littermates at birth. We noted, however, that the genotype distribution of offspring generated by matings between mice harboring one copy of the transgene slightly deviated from the predicted Mendelian ratio, with no observed lethality in LMNA G/+ mice but 12% lethality in LMNA G/G ( Figure S1b).
Although no difference was observed between male and female mice (n = 150 per gender), these numbers suggest a moderate level of either embryonic or perinatal lethality in LMNA G/G offspring.
We previously reported that LMNA G/+ transgenic mice lack pathologic changes outside of the vascular system prior to 20 months of age (Varga et al., 2006). In contrast to single-copy mice, LMNA G/G mice developed phenotypic features consistent with HGPS following the first month after birth, relative to wild-type (LMNA +/+ ) littermates. By 5 weeks of age, both male and female LMNA G/G mice were growth deficient (Figure 1a). From 5 to 15 weeks of age, double-copy transgenic mice weighed 10-20% less than LMNA +/+ and LMNA G/+ littermates. After 3.5-4 months of age, no additional weight gain was observed in LMNA G/G mice, with progressive wasting resulting in a 20-45% decrease in weight compared to LMNA +/+ and LMNA G/+ mice (p < 0.01 versus both genotypes) ( Figure 1a, Figure S1c).
The decreased weight observed in LMNA G/G mice relative to wild-type (LMNA +/+ ) and single-copy transgenic (LMNA G/+ ) littermates largely reflected growth deficiency, combined with the loss of adipose tissue and bone mass. Consistent with the lipodystrophic phenotype of HGPS patients, histopathologic analysis of skin from 6-month-old LMNA G/G mice demonstrated decreased dermal and subcutaneous adipose tissue resulting in tightened skin ( Figure 1b).
Although mild hair loss occurred infrequently in these mice, no obvious abnormalities of hair follicles were evident.
We also noted that LMNA G/G , but not LMNA G/+ mice, began to develop hindlimb contractures and moved with a waddling gait beginning at 6-7 months of age. To assess whether transgenic mice exhibited skeletal features similar to those observed in HGPS patients, bone morphology was evaluated in murine femurs at 6 months of age by μCT analysis (Figure 1c). All trabecular and cortical structural parameters in long bones of LMNA G/+ mice were comparable to wildtype (Lmna +/+ ) littermates, reflecting the normal weights and the absence of pathologic features outside of the vascular system. In contrast, bone structural parameters were altered in LMNA G/G mice relative to wild type. Trabecular bone volume (BV/TV) was half of wild-type values (p < 0.01), due to decreased trabecular thickness rather than a decrease in trabecular number, while cortical thickness (CT Th) was modestly reduced (~11%) in LMNA G/G femurs (p < 0.001).
Additional musculoskeletal features observed in LMNA G/G mice included kyphosis, which developed by 4 months and progressed until mice reached endpoint (Figure 1d, Figure S1d). Similar to LMNA G/+ mice but even more extensive, progressive loss of VSMCs, elastic fiber breakage, proteoglycan, and collagen accumulation in the media, and dramatic thickening of the adventitia occurred in vascular tissues of 5-month-old LMNA G/G mice (Figure 2ac). Arterial and arteriolar pathology was prominent in the aorta, but also included the carotid arteries with focal pathology in arteries and arterioles of the heart, kidneys, skeletal muscles, and adipose tissues. These vascular abnormalities were characterized by decreased numbers of smooth muscle nuclei, increased amounts of ground substance in the tunica media, and increased amounts of adventitial connective tissue. Adventitial expansion was associated with deposition of collagenous matrix, as noted by picrosirius red staining ( Figure S1e).
Most notable, however, was the generalized severe phenotype exhibited in LMNA G/G mice, characterized by reduction of activity around 6 months of age and lasting until death by age 7-8 months (Figure 1e).
The average lifespan of LMNA +/+ and LMNA G/+ mice was 897 and 485 days, respectively, compared to 212 days for LMNA G/G mice (p ~ 10 −137 vs LMNA +/+ ; p ~ 10 −101 vs LMNA G/+ ). Furthermore, there was no significant difference in average lifespan of male (216 ± 36 days) and female (208 ± 35 days) LMNA G/G mice. Despite the severe vascular alterations observed in LMNA G/G mice; however, post-mortem analyses were unable to identify a specific acute cause of death.
At the tissue level, molecular and biochemical analyses of endogenous and transgene-derived A-type lamin expression were performed. As expected, quantitative PCR of endogenous and transgene-derived transcripts demonstrated only murine lamin A/C transcripts in wild-type (LMNA +/+ ) tissues, including skin, inguinal fat, heart, and aorta ( Figure S2a). In all tissues that were tested, the levels of endogenous Lmna transcripts were comparable across all genotypes. In both LMNA G/+ and LMNA G/G tissues, however, we found variable levels of transgene expression.
Transgene expression levels correlated with copy number in skin and heart tissue. Conversely, transgene expression was significantly decreased in inguinal fat and aorta tissue of LMNA G/G versus LMNA G/+ mice which may reflect loss of viable adipocytes and VSMCs, respectively, that express high levels of A-type lamins (Uhlen et al., 2015).
Immunoblots of tissue-derived A-type lamins also revealed variable levels of total lamin A, progerin, and lamin C ( Figure S2b). We found that, only in skin and aorta, total levels of A-type lamins correlated with endogenous and transgene copy number. In heart and inguinal adipose tissue derived from LMNA G/G mice, total A-type lamins were reduced approximately 25% and 50% in fat and heart, respectively, compared to LMNA G/+ mice. However, total levels of endogenous and transgene-derived A-type lamins relative to the reference protein were consistently higher in tissues derived from transgenic mice compared to wild type ( Figure S2b Two treatments currently being tested include farnesyltransferase inhibitors to reduce production of the toxic progerin protein and F I G U R E 2 Severe vascular phenotype in HGPS mice can be partially rescued by genetic reduction of Mtor. (a) Movat's pentachromestained ascending aorta sections from 5-month-old mice. (b) Hematoxylin and eosin (H&E)-stained aortae from LMNA +/+ and LMNA G/G mice. (c) Quantitation of adventitial area and VSMC number in aortae. Genetic reduction of Mtor partially restores the loss of vascular smooth muscle cells but does not rescue the adventitial expansion of aortae in LMNA G/G mice. *, p < 0.05; **, p < 0.01; ***, p < 0.001 autophagy activation with rapamycin analogs to promote its clearance (Evangelisti et al., 2016;Harhouri et al., 2018;Young et al., 2013). While the former approach has proven clinically beneficial, the latter is currently the subject of a clinical trial . We reasoned that, using our mouse model, a genetic approach could be undertaken to investigate mTOR inhibition as a potential strategy to alleviate HGPS pathology. Mice carrying an Mtor hypomorphic allele (Mtor Δ/+ ) were bred into the G608G transgenic mouse line, followed by matings of doubly heterozygous mice (Mtor Δ/+ LMNA G/+ ). The offspring distribution from these matings was significantly skewed toward Mtor +/+ , regardless of LMNA genotype, and the presence of the hypomorphic allele increased embryonic lethality from 12% to 61% in LMNA G/G offspring ( Figure   S3). As demonstrated previously, Mtor +/+ LMNA G/G mice consistently weighed significantly less than their wild-type littermates by 5 weeks (Figure 3a). Although Mtor Δ/+ LMNA G/G mice failed to attain weights similar to wild-type littermates, after 20 weeks of age both male and female mice harboring one copy of the Mtor hypomorphic allele retained 10-15% more of their weight compared to Mtor +/+ LMNA G/G mice (p < 0.05).
Using an antibody specifically recognizing human LMNA (JoL2), no difference in the levels of transgene-derived lamin A, progerin, or lamin C relative to alpha-smooth muscle actin (aSMA) or beta-actin were found by Western analyses of immunoprecipitated A-type lamins from heart and liver, respectively, of Mtor +/+ LMNA G/G and Mtor Δ/+ LMNA G/G mice (Figure 5a,b). Subsequent detection with an alternative antibody (4c11) reactive to both human and mouse Atype lamins also failed to demonstrate any discernible quantitative differences in tissue homogenates derived from transgenic mice with genetically reduced Mtor. We therefore considered the possibility that, in vivo, autophagic elimination of progerin might be affected by transgene-derived RAB25 expression due to its putative role in regulating autophagy by activation of AKT (Cheng et al., 2012).
Quantitative RT-PCR demonstrated that human RAB25 expression was 2.5-to 60-fold greater than endogenous mouse Rab25, depending on the tissue analyzed, in mice harboring two transgene copies ( Figure S5a). To determine if high levels of transgene-derived RAB25 were inhibiting autophagy, we next generated murine embryonic fibroblasts from the offspring of Mtor Δ/+ mice crossed with either Lmna G609G knock-in mice, which lack human RAB25 expression, and LMNA G608G transgenic mice. Western analyses from cell lysates demonstrated a modest reduction in total A-type lamins in both LMNA G608G/G608G transgenic (G/G) and Lmna G609G/G609G knock-in (KI/KI) cell lines treated with everolimus (RAD001), as well as in cell lines harboring the Mtor hypomorphic allele ( Figure S5b).
The decrease in A-type lamins relative to ACTB levels, which was not specific to progerin, was associated with increased LC3-II/LC3-I ratios and decreased p62 levels, indicating autophagic activation in cell lines where mTOR was either chemically inhibited or genetically reduced ( Figure S5c).
Analysis of mTOR signaling in tissues collected from adult mice contrasted with the results from cell cultures. Although Western immunoblots of tissue homogenates collected from 5 month-old LMNA G/G mice revealed significantly increased levels of phosphorylated mTOR in hearts (Figure 6a), and to a lesser extent in livers (Figure 6b), compared to wild-type littermates, downstream targets of mTOR signaling were found to be hypophosphorylated. In heart tissue from adult mice, relative phosphorylation of S6 was significantly reduced in LMNA G/G versus wild type, with further reduction in phosphorylation occurring in transgenic mice with genetically reduced Mtor (Figure 6a). A similar, but nonsignificant, trend in reduced S6 phosphorylation was found in livers of adult mice, where genetic reduction of Mtor was associated with hypophosphorylation of 4EBP1 (Figure 6b). In both tissues, we found that AKT activation was suppressed in the presence of progerin. From these data, we conclude that mTOR signaling is hyperactivated in LMNA G/G newborn and embryo-derived cell cultures, but in tissues of mice where the HGPS phenotype has progressed mTOR signaling is inhibited. | 9 of 16 CABRAL et AL.

| DISCUSS ION
Here, we describe a double-copy LMNA G/G transgenic mouse model of HGPS that replicates most of the features of the human disorder. Several other mouse models have been created to characterize the pathogenic mechanisms underlying HGPS, and the comparison of phenotypes is useful. Yang and colleagues generated a knockin model, Lmna HG , created by deletion of intron 10, intron 11 and the last 150 nucleotides of exon 11, so that only the mutant progerin transcript is produced without any normal lamin A or C (Yang et al., 2005). Lmna HG/+ mice, which have an average lifespan of approximately 27 weeks, develop growth deficiency, bone dysplasia, and kyphosis by 4 months of age. However, somewhat surprisingly these mice do not exhibit vascular smooth muscle cell loss and the intima, media, and adventitial layers of aortas appear normal out to 7 months of age (Yang et al., 2005). In contrast, mice homozygous for the progerin-only allele (Lmna HG/HG ) rapidly develop a more severe phenotype, characterized by spontaneous fractures and complete absence of adipose tissue with death at 3-4 weeks of age (Yang et al., 2006). Lmna G609G/G609G mice show a more pronounced and earlier progeroid phenotype than heterozygous Lmna G609G/+ mice and, prior to premature death at 14-15 weeks of age, display the main clinical manifestations of human HGPS such as osteoporosis, loss of principal fat deposits, and VSMC depletion, as well as aberrant hormonal profiles that manifest as hypoglycemia and IGF deficiency (Osorio et al., 2011).
In cell cultures from patients with HGPS, a minor fraction of transcripts derived from the mutant LMNA allele encode progerin, while the majority of transcripts yields mostly wild-type lamin A and lamin C (Eriksson et al., 2003). In contrast, we found that a significant proportion of murine transcripts derived from the mutant human LMNA transgene encode progerin. Nevertheless, the dominant-negative effects of progerin accumulation progress with age in both patients and murine models of HGPS. Despite the possibility that mice require a higher level of progerin expression to elicit the HGPS disease phenotype (Yang et al., 2005), it is apparent that the severity of the progeroid phenotype in LMNA G/G transgenic mice, as well as other mouse models of HGPS, correlates with gene dosage. Comparison to other mouse models of HGPS shows that LMNA G/G transgenic mice present with a less severe phenotype than Lmna G609G/G609G and Lmna HG/HG mice, which may be attributed to altered molecular interactions between human LMNA/progerin and endogenous murine nuclear proteins, or alternatively, retention of normal functioning of endogenous Lmna in transgenic mice.
We also noted that at the transcript level, inguinal fat and aorta from LMNA G/G mice expressed lower levels of progerin compared to LMNA G/+ , which may reflect the loss of specific cell populations with higher transgene expression in these tissues (Uhlen et al., 2015).
While a dramatic absence of subcutaneous adipose tissue was observed, LMNA G/G transgenic mice do not develop alopecia or glandular and follicular abnormalities as occurs in Lmna G609G mice (Osorio et al., 2011).
Although LMNA G/+ mice appear to lack any detectable bone phenotype at 6 months of age, micro-CT analysis of LMNA G/G mouse femora detected a modest decrease (~5%) in BMD with significant reductions in structural parameters of trabecular and cortical bone, consistent with previous reports (Cubria et al., 2020). This aspect of the HGPS phenotype appears to be due to defective modeling of bone tissue, since bone-and cartilage-specific Osx-driven expression of progerin leads to skeletal abnormalities associated with impaired osteoblast differentiation and loss of the osteocyte population in bone tissue (Schmidt et al., 2012). While there is no current evidence of increased resorption, the potential for reduced cortical and trabecular bone volumes in LMNA G/G femurs resulting from abnormal bone remodeling cannot be ruled out. Despite key differences existing between rodent and human skeletal systems, the LMNA G/G transgenic mouse model will be a useful model for further investigating the effects that progerin expression has on bone cell populations in HGPS bone dysplasia and, potentially, its role in aging-related osteoporosis.
As was seen in LMNA G/+ mice, LMNA G/G mice were found to have a progressive and dramatic defect of large arteries, particularly the aorta. These pathologic alterations in the medial layer of large vessels of transgenic mice recapitulate findings in HGPS. Since SM2a-Cre-mediated VSMC-specific expression of progerin is sufficient to accelerate pathologic vascular alterations and death in atheroprone Apolipoprotein E-null mice (Hamczyk et al., 2018), progerin-induced F I G U R E 4 Reduction of mTOR levels in LMNA G608G newborn fibroblasts normalizes S6K activity. (a) Western analyses of A-type lamins and mTOR signaling pathway components in fibroblast cell lines derived from Mtor +/+ LMNA +/+ , Mtor Δ/+ LMNA +/+ , Mtor +/+ LMNA G/G , and Mtor Δ/+ LMNA G/G mice. (b) Quantitative analysis of mTOR signaling components from immunoblots demonstrates reduction of mTOR protein in cells heterozygous for the Mtor hypomorphic allele. Induction of autophagy is indicated by increased LC3-II/I ratios in Mtor +/+ LMNA G/G and Mtor Δ/+ LMNA G/G cells versus wild-type, and reduced levels of p62 in Mtor Δ/+ LMNA G/G cells versus Mtor +/+ LMNA G/G cells. Normalization of increased S6K phosphorylation in LMNA G/G cells occurs with reduced mTOR levels. *, p < 0.05 versus Mtor +/+ LMNA +/+ (c) Cell cycle analysis of proliferative fibroblasts in culture. M1, G 0 /G 1 phase; M2, S phase; M3, G 2 /M phase. Genetic reduction of Mtor in transgenic cells (Mtor Δ/+ LMNA G/G ) shifts the fraction of cells in G 0 /G 1 , S, and G 2 /M phases to levels seen in wild-type (Mtor +/+ LMNA +/+ ) cells. (d) LMNA G/G fibroblasts are less proliferative than LMNA +/+ cells in culture. Heterozygosity for the Mtor hypomorphic allele further inhibits proliferation of LMNA G/G murine fibroblasts. *, p < 0.05; **, p < 0.01; Mtor +/+ LMNA G/G significance versus Mtor +/+ LMNA +/+ ; Mtor Δ/+ LMNA G/G significance versus Mtor +/+ LMNA G/G | 11 of 16 CABRAL et AL. vascular pathology may be directly related to the cause of death in patients. Although the precise mechanism(s) underlying the vascular changes remains to be clarified, HGPS VSMCs and endothelial cells are more sensitive to mechanical stress, have been shown to be particularly susceptible to progerin accumulation and are exposed to the higher hemodynamic forces (pressure and shear stress) exerted on arterial walls (Varga et al., 2006;Verstraeten et al., 2008), which may be influenced by dysregulation of the glycocalyx components involved in flow shear stress sensing (Pitrez et al., 2020). Progerininduced VSMC vulnerability has also been attributed to dysregulation of the DNA damage response pathway, hyperactivation of ER stress (ER), and the unfolded protein response (UPR) (Hamczyk et al., 2019;Kinoshita et al., 2017;Zhang et al., 2014). Furthermore, it can be speculated that the adventitial expansion seen in LMNA G/G aortas is a fibrotic event mediated by VSMC-secreted inflammatory cytokines as part of an apoptotic program or, alternatively, driven by a senescence-associated secretory phenotype. Interestingly, disruption of the linker of the nucleoskeleton and cytoskeleton (LINC) complex in smooth muscle cells has been shown to reduce VSMC loss and adventitial fibrosis, suggesting that disruption of transmission of biomechanical forces to the nucleus may underlie these processes (Kim et al., 2018).
The alterations in tissues of mesenchymal origin, particularly skin, fat, bone and vascular tissues that we have characterized, make the LMNA G/G transgenic mouse a useful model for investigating therapeutic approaches to treat HGPS. Our primary goal in generating this mouse model was to enable testing of RNA-and DNAbased treatments targeting the classic HGPS mutation at the gene and transcript levels in the context of the human sequence. To date, however, clinical trials have been limited to small molecule applications. For example, we have previously demonstrated that the FTI tipifarnib not only prevents onset of the cardiovascular phenotype when treatment begins at weaning, it also may be capable of inducing disease regression in LMNA G/+ mice that already manifest severe arteriosclerotic changes (Capell et al., 2008). In human patients, results from a trial of the farnesyltransferase inhibitor lonafarnib demonstrated benefit to the vascular system with reduction in peripheral vascular resistance and a modest improvement in survival . A combination trial of lonafarnib, zoledronic acid, and pravastatin failed to show benefit beyond that obtained by lonafarnib alone (Gordon et al., 2016).  (Cao et al., 2011;DuBose et al., 2018). Considering this background, we have investigated inhibition of mTOR signaling as a potential treatment for HGPS.
In a genetic strategy to test the effect of mTOR inhibition in our mouse model, mice carrying an Mtor hypomorphic allele (Mtor Δ/+ ) were bred into the G608G transgenic mouse line. We found no significant improvement of nuclear membrane structure in newborn fibroblasts derived from offspring of these matings, nor could we verify any reduction of progerin in tissues of Mtor +/+ LMNA G/G mice in the presence of activated autophagic activity in adult tissues. Importantly, however, genetic reduction of Mtor in HGPS transgenic mice led to a 30% increase in lifespan versus Mtor +/+ LMNA G/G littermates.
Analysis of mTOR signaling in fibroblasts derived from newborn mice demonstrated that increased activation of the downstream effector S6 protein kinase (S6K) in Mtor +/+ LMNA G/G was normalized in Mtor Δ/+ LMNA G/G , consistent with lifespan extension observed in S6K-null mice (Selman et al., 2009). Genetic reduction of Mtor has previously been reported to normalize increased mTOR/S6K1 activity, partially restore autophagy and postpone the premature aging phenotype in Zmpste24-null mice (Pan et al., 2020). Hyperactivation of S6K has also been observed in adipocyte precursor and osteosarcoma cell lines expressing LMNA mutations associated with Type 2 Familial Partial Lipodystrophy (FPLD2) and Mandibuloacral Dysplasia (MADA), respectively (Evangelisti et al., 2015;Pellegrini et al., 2019). Moreover, there is evidence to suggest that S6K mediates the pathogenesis and premature death of Lmna-null mice, and reducing its activity improves muscle function as well as lifespan (Liao et al., 2017).
In contrast to cells from embryos and newborn mice, however, signaling downstream of mTOR was found to be inhibited in LMNA G/G adult tissues, particularly in the heart. These findings suggest that hyperactivation of mTOR signaling may be an early event in the onset of HGPS pathology in Mtor +/+ LMNA G/G mice, but its activity is age-, phenotype-, and tissue-dependent, similar to what has been observed for Zmpste24 −/− mice (Pan et al., 2020), and which may be informative to the rational design of a therapeutic approach involving chemical inhibition of mTOR signaling.
The mechanism by which reduction of S6K activity increases lifespan in our mouse model of HGPS remains to be determined. F I G U R E 5 Analysis of A-type lamins in tissues. (a, b) Western analysis of A-type lamins extracted from heart tissue and livers of Mtor +/+ LMNA +/+ , Mtor Δ/+ LMNA +/+ , Mtor +/+ LMNA G/G , and Mtor Δ/+ LMNA G/G mice (n = 2 per genotype, 1 male and 1 female). Samples in upper immunoblots were immunoprecipitated with antibodies to human lamin A/C [JoL2] and smooth muscle actin (SMA) or beta-actin (ACTB), then probed with an alternative antibody recognizing both mouse and human Lamin A/C and reference proteins SMA or ACTB. Lower immunoblots contain tissue homogenates probed with antibody recognizing both mouse and human Lamin A/C [4c11] and GAPDH. No significant difference in lamin A, progerin, or lamin C levels was observed in mice with genetic reduction of mTOR regardless of sample preparation procedure, antibody used for detection, or reference protein used for normalization | 13 of 16 CABRAL et AL.
Interaction of S6K1 with various substrates promotes several broad cellular processes such as protein production, cell growth and size, cell survival, and gene transcription. For example, S6K1 regulates protein production by phosphorylating substrates that function in translational initiation, elongation, and protein folding, including eIF4B, PDCD4, SKAR, CBP80, eEF2K, CCTβ, and RPS6, a component of the 40S ribosome. S6K1 also regulates transcription by activating CREM, estrogen receptor, and SREBP (Duvel et al., 2010;De Groot et al., 1994;Yamnik et al., 2009). Furthermore, S6K1 participates in the DNA damage response by phosphorylating MDM2, thereby blocking its nuclear import and ability to ubiquitinate the tumor suppressor p53 (Lai et al., 2010). Finally, S6K1 promotes negative feedback on PI3K signaling through IRS1, thus suppressing insulin and IGF sensitivity (Tremblay et al., 2007). It will be important, therefore, to determine which specific signaling events downstream of S6K1 are altered in response to genetic reduction of Mtor in our mouse model of HGPS.
These data provide further evidence, based on an animal model carrying the human progeria mutation, that inhibitors of mTOR might provide benefit for children with HGPS. It should be kept in mind, however, that the mTOR protein takes part in two different complexes, mTORC1 and mTORC2. Only the first of these is classically stated to be sensitive to rapamycin and its homologs. So the genetic approach we describe here, which reduces function of both mTORC1 and mTORC2, is not a perfect analog of the pharmacologic blockage of mTORC1 that is being pursued with an everolimus trial.
In summary, we have generated a murine model that is well suited for testing therapeutic interventions to treat Hutchinson-Gilford progeria syndrome in the context of the human gene and protein.
Furthermore, this study supports a strategy for pharmacological inhibition of a specific component of the mTOR signaling pathway as a potential treatment of HGPS.

| Mouse strains and animal care
A transgenic mouse model of HGPS was developed by retrofitting a human bacterial artificial chromosome (BAC) harboring the LMNA gene containing the classic G608G mutation (c.1824C > T). The 164 kb transgene was incorporated into the germline of C57BL/6J mice as previously described (Varga et al., 2006). Single-copy mice, designated LMNA G/+ , were bred within the C57BL/6J line for twenty generations prior to generation of double-copy mice, designated as LMNA G/G , for experimental use. Mice harboring the Mtor hypomorphic allele (Mtor Δ/+ ) were a generous gift from Beverly Mock, National Cancer Institute, and have been previously characterized (Wu et al., 2013). Mice harboring the Lmna G609G knock-in mutation were a generous gift from Carlos López-Otín, Universidad de Oviedo (Osorio et al., 2011). Murine growth curves were determined by weekly weights. Kaplan-Meyer plots were generated from inhouse data combined with published wild-type C57BL/6J data from Jackson Laboratories (https://www.jax.org/jax-mice-and-servi ces/ strai n-data-sheet -pages/ body-weigh t-chart -000664). Animal care and experiments were performed in accordance with a protocol approved by the NHGRI Animal Care and Use Committee.

| Cell cultures
Primary fibroblast cultures (FB) were derived from dermal tissue dissected from the abdomens of newborn mice. FB were allowed to grow out from dermal samples for 2 weeks before releasing by trypsin digestion. Murine embryonic fibroblasts (MEFs) were isolated from day 14-15 (E14-E15) embryos generated by crossing G608G transgenic or G609G knock-in mice with mice expressing the Mtor hypomorphic allele. After skin and viscera were removed, embryos were minced and digested in 0.25% trypsin for 15 min. Cells were plated and allowed to adhere for 24 h at 37℃. The following day, non-adherent cells were removed, and the adherent cell population was expanded for collection as MEFs. Cells were cultured in DMEM containing 10% fetal bovine serum, 2 mM glutamine, and 1% penstrep at 37℃ in 5% CO 2 . Cells cultured with everolimus (RAD001) were treated at 0.1 μM for 4 h prior to collection.

| Proliferation assays
Cells were plated at 10,000 cells per well in 24-well culture plates (day 0) in triplicate. Culture medium was replaced daily during the course of the assay. Each cell line was collected by trypsinization at 24-h increments. Cell concentrations were determined in Via-1 cassettes using a Nucleocounter with Nucleoview-3000 software (Chemometec). Cell number was calculated and plotted using Microsoft Excel software.

| Cell cycle analysis
Subconfluent fibroblast cultures were collected by trypsinization and pelleted at 500 × g for 10 min. Cells were then washed once with PBS and centrifuged again at 500 × g for 10 min. Cell pellets were resuspended in PBS and fixed in a final concentration of 70% EtOH overnight at 4℃. The following day cells were centrifuged at 500 × g for 5 min, resuspended in PBS for 1 min to rehydrate, then centrifuged again for an additional 5 min. Cells were gently resuspended in propidium iodide staining solution (20 μg/ml PI, 0.1% triton X-100, 200 μg/ml RNase A in PBS) and incubated at 37℃ for 30 min. Stained cells were analyzed using the Fixed Cell Cycle-PI Assay on a NucleoCounter NC-3000 (Chemometec).

| Histology
Tissues were fixed in 2% paraformaldehyde for 24 h prior to dehydration with graded alcohols and embedding in paraffin. Cross-sections (4μm thick) were cut and mounted on charged slides and visualized by Masson's trichrome, hematoxylin/eosin, Movat's pentachrome, or Picrosirius red staining (CV Path Inc). Images were captured on a Axioscan imaging system (Zeiss) at 20× magnification and were further processed by trimming using ZenBlue microscopy (Zeiss) and Photoshop CC (Adobe) software.
Sequential transaxial images were generated using an integration time of 250ms and a tube voltage and current of 70kVp and 114mA, respectively. Images were reconstructed at 10 μm isotropic voxel size, then assessed for volumetric bone mineral density (vBMD, mg/cm 3 ) using calibrated hydroxyapatite (HA) phantoms to convert X-ray attenuation to a known mineral density. Morphometric analyses were performed using native Scanco software. The trabecular region of interest (ROI) was located proximal to the distal femoral growth plate and extended 150 slices upward toward the proximal end. The diaphyseal cortical ROI started on the 55 percentile femur height from the proximal end, spanning 50 slices toward the distal end.

ACK N OWLED G M ENTS
We acknowledge Adam Phillippy, Winston Timp, Isac Lee, and Norah Hilger for contributions to PromethION WGS sequencing and analysis of transgene structure. This work was supported by NHGRI Intramural Research Funds to WAC, ULT, YDB, MAE, MRE, and FSC (NHGRI HG200305).

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the authors upon request.