biomedicines Article Impact of Behavioral Assessment and Re-Test as Functional Trainings That Modify Survival, Anxiety and Functional Profile (Physical Endurance and Motor Learning) of Old Male and Female 3xTg-AD Mice and NTg Mice with Normal Aging Lidia Castillo-Mariqueo 1,2 and Lydia Giménez-Llort 1,2, * 1 Institut de Neurociències, Universitat Autònoma de Barcelona, 08193 Barcelona, Spain; lidia.castillom@autonoma.cat 2 Department of Psychiatry and Forensic Medicine, School of Medicine, Universitat Autònoma de Barcelona, 08193 Barcelona, Spain * Correspondence: lidia.gimenez@uab.cat Abstract: Longitudinal approaches for disease-monitoring in old animals face survival and frailty limitations, but also assessment and re-test bias on genotype and sex effects. The present work investigated these effects on 56 variables for behavior, functional profile, and biological status of male and female 3xTg-AD mice and NTg counterparts using two designs: (1) a longitudinal de- sign: naïve 12-month-old mice re-tested four months later; and (2) a cross-sectional design: naïve 16-month-old mice compared to those re-tested. The results confirmed the impact as (1) improve- Citation: Castillo-Mariqueo, L.; ment of survival (NTg rested females), variability of gait (3xTg-AD 16-month-old re-tested and Giménez-Llort, L. Impact of naïve females), physical endurance (3xTg-AD re-tested females), motor learning (3xTg-AD and NTg Behavioral Assessment and Re-Test 16-month-old re-tested females), and geotaxis (3xTg-AD naïve 16-month-old males); but (2) worse as Functional Trainings That Modify anxiety (3xTg-AD 16-month-old re-tested males), HPA axis (3xTg-AD 16-month-old re-tested and Survival, Anxiety and Functional naïve females) and sarcopenia (3xTg-AD 16-month-old naïve females). Males showed more func- Profile (Physical Endurance and tional correlations than females. The functional profile, biological status, and their correlation are Motor Learning) of Old Male and discussed as relevant elements for AD-pathology. Therefore, repetition of behavioral batteries could Female 3xTg-AD Mice and NTg Mice be considered training by itself, with some variables sensitive to genotype, sex, and re-test. In the with Normal Aging. Biomedicines AD-genotype, females achieved the best performance in physical endurance and motor learning, 2022, 10, 973. https://doi.org/ 10.3390/biomedicines10050973 while males showed a deterioration in most studied variables. Academic Editors: Masaru Tanaka Keywords: Alzheimer’s disease; aging; survival; anxious profile; functional profile; motor perfor- and Nóra Török mance; frailty; training; gait; kyphosis Received: 14 February 2022 Accepted: 11 April 2022 Published: 22 April 2022 1. Introduction Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in Specific motor skills impaired in old age include a broad and varied spectrum that published maps and institutional affil- involves a reduction in gait speed, loss of strength and muscle mass, and decline of bal- iations. ance [1–3]. However, aging has become increasingly recognized as a potentially modifiable risk factor for chronic disease and frailty [4,5]. The deterioration of motor performance re- lated to cognitive dysfunction in Alzheimer’s disease (AD) has recently gained importance in clinical research [6–9]. Particularly, gait impairment and its association with cognitive Copyright: © 2022 by the authors. impairment [10] could shed light on potentialities to distinguish AD [1]. Inclusive, higher Licensee MDPI, Basel, Switzerland. levels of Aβ and tau are associated with more significant memory decline, but not with This article is an open access article changes in executive function [11]. The study by Sperling points out that these results distributed under the terms and could explain why some clinically active patients presented elevated tau and Aβ levels [11]. conditions of the Creative Commons Thus, Aβ and tau proteins can serve as markers of cognitive impairment; however, they are Attribution (CC BY) license (https:// insufficient and cannot detect all cases of dementia, especially in the early stages [11,12]. creativecommons.org/licenses/by/ For this part, gait speed, for example, is longitudinally associated with cognitive decline, 4.0/). Biomedicines 2022, 10, 973. https://doi.org/10.3390/biomedicines10050973 https://www.mdpi.com/journal/biomedicines Biomedicines 2022, 10, 973 2 of 28 dementia, and falls in older adults [13,14], with slower gait associated with increased fall risk and poor baseline cognition [6]. However, motor dysfunctions and deterioration remain poorly explored. Consequently, functional and cognitive decline comorbidity is a warning sign for increased disability [8], a growing public health problem [15], and it is already present at preclinical stages of Alzheimer’s disease [5]. On the other hand, aging is a frequent risk factor for different diseases, including dementia [16]. Recently, in a review of the literature that examined the pathophysiological basis and biomarkers of AD and other neurodegenerative diseases, it was pointed out that the predisposing factors for neuroinflammation are aging, metabolic diseases, hypertension, cerebrovascular accidents, depression and depression, dementia, among others [17]. In addition, healthy aging would be associated with chronic inflammation, contributing to a greater vulnerability to anxiety and depression [17]. Thus, cause–effect relationships can become bidirectional in the pathogenesis of multifactorial diseases, leading to a disease- prone state [18]. Age-related deficits in the ability to process contextual information and regulate responses to threat, addressing that structural and physiological alterations in the prefrontal cortex and medial temporal lobe determine cognitive changes in advanced aging, which may eventually cause patterns of cognitive dysfunction seen in patients with AD and mild cognitive impairment (MCI) [19]. Furthermore, it is known that AD is characterized by high heterogeneity in the dis- ease’s manifestation, progression, and risk factors [20]. Such a high phenotypic variability is considered one of the most significant obstacles in early diagnosis and clinical trial design [20]. Therefore, there is great interest in identifying factors driving variability used for patient stratification [20,21]. Additionally, the impact of sex on the disease varies throughout its progression [22,23]. It is important to identify the role of sex differences in the cognitive dimension if potentially more precise diagnoses and treatments should emerge [24,25], but few studies have reported differences in the psychomotor functional dimension of the disease. In the last decade, at the translational level, the impact of interventions on age-related disability, frailty, and the onset of AD has been investigated in animal models to develop clinically relevant measures that provide indications for the approach and management of disability, frailty, and illness [26,27]. In addition, genotype and sex differences in cognitive, emotional, and locomotor performance have been studied at the preclinical level to assess the effects of promising interventions before their application in clinical settings [28–30]. Recently, our laboratory has developed a study method to identify psychomotor impairments and deficits at different stages of Alzheimer’s disease [31]. Previously, we reported a functional impairment phenotype in male mice’s gait and physical performance of the 3xTg-AD transgenic model in the initial, intermediate, and advanced stages of the disease. The results showed that 3xTg-AD mice show a significant functional impairment in the quantitative variables of gait and exploratory activity, movement limitations, and muscle weakness related to functional decline in the different stages of severity of the disease intensify with increasing age. In addition, signs of frailty accompany functional deterioration, and sarcopenia is evident in an advanced stage of AD, with differences in the morphological characteristics of muscle fibers and the number of fat cells [31]. Furthermore, we differentiate the disorders and postural patterns into two types of kyphosis (postural and structural) that differ in severity and limit the exploratory activity. In addition, the results indicated that the presence of bizarre gait patterns accounts for behavior similar to anxiety when 3xTg-AD mice face novelty situations and recognition of places, with circling and backward movements being the most frequent, in an already frailty setting [32,33]. The present study was designed to investigate the impact of two factors, sex, and repeated test, assessing the behavioral outputs of 3xTg-AD mice and mice with normal aging in longitudinal and transversal experimental designs. According to our previous work, a battery of psychomotor tests: gait, exploration, muscle strength, motor learning, physical endurance, and frailty status, was used [31]. In addition, phenotype of frailty Biomedicines 2022, 10, 973 3 of 28 and biological status (HPA axis and sarcopenia index) was also included [32,33]. Thus, in the first place, we studied the sex factor by characterizing the psychomotor phenotype of middle—(12 months) and old—(16 months) age females, that in the 3xTg-AD mice corresponds to two neuropathological stages of the disease [31], as compared to that of aged-matched males. On the other hand, long-term studies provide better insights for assessing interventions with preclinical validity, but the administration of behavioral batteries is not exempt from carryover effects. In addition, behavioral batteries and repeated tests can be considered behavioral stimulation [34]. Therefore, we aimed to investigate the effects of repeated tests on the behavioral performance of animals assessed in two scenarios: 1) in a longitudinal design, with within-subjects analysis of a set of 12-month-old animals re-tested four months later, at 16 months of age; and 2) in a transversal design, when comparing 16-month-old animals that had (re-tested) or not (naïve) that experienced the battery of tests. 2. Materials and Methods 2.1. Animals A total of 191 male and female mice were included for the survival analysis, and ninety-six of them, homozygous 3xTg-AD (n = 54) and non-transgenic (NTg, n = 42) male and female mice of 12 to 16 months of age in a C57BL/6J background (after em- bryo transfer and backcrossing of at least ten generations), established at the Universitat Autònoma de Barcelona, were included in the experimental study. As previously described, the 3xTg-AD mice harboring transgenes were genetically created at the University of California at Irvine [35]. Animals were kept in groups of 3–4 mice per cage (Macrolon, 35 cm×35 cm×25 cm) filled with 5 cm of clean wood cuttings (Ecopure, Chips, Date Sand, UK; uniform cross-sectional wood granules with 2.8–1.0 mm chip size) and nesting ma- terials (Kleenex, Art: 08834060, 21 cm×20 cm, White). In all cases, standard home cages covered with a metal grid to allow the perception of olfactory and auditory stimuli from the rest of the colony. All animals were kept under standard laboratory conditions of food and water ad libitum, 20 ± 2 ◦ C, 12 h light cycle:dark with lights turned on at 8:00 a.m. and 50–60% relative humidity. The study complied with the ARRIVE guidelines developed by the NC3Rs and aimed to reduce the number of animals used [36]. 2.2. Experimental Design A longitudinal and a transversal study were carried out to evaluate the anxious-like and functional profiles of male and female 3xTg-AD and NTg mice. Biological variables (corticosterone and sarcopenia) of animals at the end point (16 months of age) were also included. For this purpose, animals were randomly assigned into two experimental batches (see Figure 1, Experimental design) 2.3. Behavioral Assessment and Biological Status The assessment consisted of four consecutive evaluation steps conducted during 5 days, as follows: Day 1, bodyweight, phenotype scoring system, and frailty; Day 2, gait and exploration; Day 3, geotaxis, muscle strength, and rotarod; Days 4–5, rotarod. The procedures and protocol were based on the protocol used by Castillo-Mariqueo and Giménez-Llort [31]. Assessments were performed under dim white light (20 lx) in the light cycle (10:00 a.m. to 1:00 p.m.). Behavioral evaluations were carried out in a counterbalanced way by two independent observers, blind to the genotype. Animals were habituated to the test room 30 min before the start of the tests. 2.3.1. Survival, Bodyweight, Phenotype Scoring System, Frailty Score, and Kyphosis Survival curves were analyzed considering the cohort of siblings from the same litter of mice included in the study, from birth to 16 months of age. A total of 191 male and female mice, NTg and 3xTg-AD were included in this analysis (NTg males = 49; NTg females = 58; 3xTg-AD males = 50; 3xTg-AD females = 34). All animals were weighed and Biomedicines 2022, 10, 973 4 of 28 evaluated with the phenotype scoring system that includes four subtests and scores: ledge, grip, gait, and kyphosis [37,38]. Individual measures were scored from 0 (the absence of the relevant phenotype) to 3 (the most severe manifestation) [37]. The measures can be analyzed individually or combined into a composite phenotype score [39,40]. On the other hand, frailty was assessed using an adaptation of the MCFI by Whitehead [41], which includes 30 assessment items from the clinical setting. The 12 elements with the highest incidence previously reported by our laboratory were selected [42]. Their incidence was reported through an absence (0), presence (1) score. The clinical evaluation included Biomedicines 2022, 10, x FOR PEER REVIEW 4 of 31 physical aspects, injuries and wounds, alopecia, piloerection, body and tail position, tremor, and urogenital alterations. Figure Figure 1. 1. Experimental Experimental design. design.Longitudinal Longitudinaldesign: design:the thefirst firstgroup groupwas wasassessed assessedin inthe the behavioral behavioral battery at the age of 12 months and again when the animals reached 16 months of age. Transversal battery at the age of 12 months and again when the animals reached 16 months of age. Transversal design: the second group was housed in standard conditions without manipulation until they were design: the second group was housed in standard conditions without manipulation until they were tested at 16 months of age, so they could be compared to re-tested 16 months old animals. tested at 16 months of age, so they could be compared to re-tested 16 months old animals. 2.3. Behavioral 2.3.2. Assessment Quantitative and Biological Parameters of Gait, Status Neophobia, and Exploration The The assessment quantitativeconsisted parametersof of four theconsecutive evaluation gait and exploration steps were conducted recorded duringthe by filming 5 days, as follows: spontaneous gaitDay 1, mice of the bodyweight, for 1 min. phenotype Later the scoring system, videos were and frailty; analyzed usingDay KINOVEA2, gait 0.8.15 and free software exploration; Dayaccording to the 3, geotaxis, Castillo-Mariqueo muscle strength, and and Gimenez-Llort rotarod; protocol The Days 4–5, rotarod. [31]. Stride length, procedures andstride length protocol variability, were based onspeed, and cadence the protocol used bywere included according Castillo-Mariqueo to the and Gimé- methodology nez-Llort used by Wang [31]. Assessments et al. were [43]. Theunder performed examination dim white included light (20observation of cycle lx) in the light body position, limb support, and movement. In addition, neophobia (immediate (10:00 a.m. to 1:00 p.m.). Behavioral evaluations were carried out in a counterbalanced fear of a new place) was assessed by means of the corner test [44] and the recording way by two independent observers, blind to the genotype. Animals were habituated to of freezing (latency of movement), the test room 30 theminnumber of explorations before the on the horizontal axis (visited corners), the start of the tests. latency and number of explorations on the vertical axis (rearings). 2.3.1. Survival, Bodyweight, Phenotype Scoring System, Frailty Score, and Kyphosis Survival curves were analyzed considering the cohort of siblings from the same litter of mice included in the study, from birth to 16 months of age. A total of 191 male and female mice, NTg and 3xTg-AD were included in this analysis (NTg males = 49; NTg fe- Biomedicines 2022, 10, 973 5 of 28 2.3.3. Muscular Strength—Hanger Test and Geotaxis The muscle strength was measured in the forelimbs using the hanger test. Three trials were performed to observe the tendency of a mouse to instinctively grasp a rack or bar when suspended by the tail. In the first and second trials, grip strength was assessed by holding the animal with its front legs for 5 s at the height of 40 cm. In the third trial, the animal is suspended for 60 s in a single attempt to assess muscular endurance. This test allows discriminating grip strength and muscular endurance according to the suspension times used by mice [45]. A box with sawdust is placed under the animal to prevent a possible fall in each trial. The bar used is graduated in 5-cm blocks to obtain the distance covered when the animal moves through the bar. The latency and movement distance are recorded. Geotaxis was measured using a 10 cm×12cm grid. A single trial registered the time it took for the animal to reach the vertical position from an inverted position at a 90◦ angle on the grid. 2.3.4. Motor Performance: Learning and Physical Endurance—Rotarod Six micro training cycles were carried out during three consecutive days with a previous learning session and psychomotor coordination. The animals were trained in the Rotarod apparatus (Ugo basile® , Mouse RotaRod NG) according to a training volume established in our previous research laboratory investigations [31]. An incremental intensity of 5 to 48 rpm was applied according to individual tolerance with a maximum duration of 360 s in each microcycle with a 1-min recovery between trial. 2.3.5. Biological Status: HPA Axis and Sarcopenia Index The animals were euthanized and the muscle tissues were necropsied. Plasma from a blood sample was obtained by centrifugation and stored and −80 ◦ C until corticosterone analysis. Corticosterone content (ng/mL) was analyzed using a commercial kit (Corticos- terone EIA Immunodiagnostic Systems Ltd., Boldon, UK). Absorbance was read at 450 nm with Varioskan LUX ESW 1.00.38 (Thermo Fisher Scientific, Massachusetts, MA, USA) [42]. The weights of the quadriceps and triceps surae muscles of the right lower extremity of each animal were recorded and kept for future analysis. The sarcopenia index [46] was applied to obtain an indirect measure of sarcopenia as a biological marker of frailty. 2.4. Statistics Statistical analyses were performed using SPSS 15.0 software. Results were expressed as the mean ± standard error of the mean (SEM) for each task and trial. The variables recorded were analyzed with Student t-test, Chi-squared or Fisher’s exact test, one-way ANOVA, and multiple regression analysis (MRA). The split-plot ANOVA design with factors genotype (G), sex (S), previous experience either as a re-test (R) in the longitudinal approach or as naïve (N) in the transversal approach, were included. Their G×S, G×R, and S×N factor interactions were also studied. Post hoc comparisons were run with Bonferroni corrections. Pearson’s correlations were made to analyze the functional correlations with (1) corticosterone, (2) sarcopenia index, and (3) phenotype score system. The survival curve was analyzed with the Kaplan–Meier test (Log rank). In all cases, p < 0.05 was considered statistically significant. 3. Results 3.1. Survival, Bodyweight, Phenotype Scoring System, Frailty Score, and Kyphosis Figure 1 shows the data obtained for survival, frailty score and postural and structural kyphosis. Thus, the animal cohort was analyzed from birth to 16 months (16 months NTg and 16 months 3xTg-AD). Only siblings from the same litter belonging to mice meeting the end-points were considered. Log rank analysis showed statistically significant differences dependent on genotype and sex (G, χ2 (1) = 20.044, p < 0.001; S, χ2 (1) = 33.531, p < 0.001), see Figure 2A. In this way, it was possible to observe that females of both genotypes have higher mortality than males, and that of them the NTg is even higher (days of average Biomedicines 2022, 10, x FOR PEER REVIEW 7 of 31 Biomedicines 2022, 10, 973 6 of 28 survival: Males, NTg = 445.05 ± 15.75, CI: 414.16–475.94; 3xTg-AD = 505.57 ± 13.04, CI: 480.01–531.13. Females, NTg = 343.60 ± 15.39, CI: 313.42–373.78; 3xTg-AD = 442.76 ± 18.40, CI: 406.69–478.83). In addition, the female NTg cohort reached 79% (46/58) of mortality and 3xTg-AD the 50% (17/34) with ages 11 to 13 months having the greatest death. For their part, NTg males reached 37% (18/49) and 3xTg-AD 24% (12/50), with 15 to 16 months being the age of greatest death. During the follow-up of the animals that started the battery at 12 months, 20 deaths were detected, the NTg males had 29% (4/14), the 3xTg-AD males Biomedicines 2022, 10, x FOR PEER REVIEW 7 of 31 a 20% (4/20), the NTg females a 35% (7/20), and 3xTg-AD females 31% (4/13), see Figure 2A. Figure 2. Survival, frailty score, and postural and structural kyphosis. (A) Survival. Statistics: Figure 2.Figure 2. Survival, Survival, frailtyfrailty score, and and score,Kaplan–Meier postural postural test—Log and and rank, structural structural G, kyphosis. genotypekyphosis. effect: (A) (A) χ2, G*** Survival. p < Survival. 0.001***; Statistics: Statistics: S, sex Kaplan– effect: χ2, S*** p < Meier test—Log 0.001***. (B) Frailty score. 2 Statistics: χχ ,, G*** ANOVA, G, genotype pp << 0.001***; effect: G** p < 2 0.01**. (C) Postural and pp <<0.001***. Kaplan–Meier test—Log rank,rank, G, G, genotype genotype effect: effect: 2 G*** 0.001***; S, S, sex sex effect: effect: χχ ,, S*** 2 S*** structural kyphosis. Statistics: Fisher’s exact test, G, genotype effect in males, G** p < 0.01**; a, aging 0.001***.(B) (B)Frailty Frailtyscore. score.Statistics: Statistics: ANOVA, effect ANOVA, in G,G, females: a** pgenotype genotype < 0.01**. The effect: ⦸G** effect: symbol G**p

0.05. R, Re-test effect, R*** p < 0.001***, R** p < 0.01**, p < 0.05*, n.s. p > 0.05. G×S, genotype and sex interaction effect, G×S** p < 0.01**, p < 0.05*, n.s. p > 0.05. Bonferroni post hoc test: g, genotype, g* p < 0.05*; s, sex; $ expressed genotype differences between sex, s$$ p < 0.01$$ ; & expressed differences between re-test groups, r&& p < 0.01&& , p < 0.05& ; $ expressed genotype differences between re-test group, r& p < 0.05& , and r$$ , between sex differences. NR indicates the absence of the group. Biomedicines 2022, 10, 973 8 of 28 In bodyweight at 12 months was high in 3xTg-AD, and at 16 months it decreased in females. In addition, males naïve 16 months weighed more than re-test males at the same, see Tables S1–S3. 3.2. Quantitative Parameters of Gait, and Neophobia and Exploration Quantitative parameters of gait are shown in Figure 3. For naïve 12 months, statistically significant differences were observed in all quantitative gait variables. Stride length showed differences in G and S, with the longest stride length in NTg males and 3xTg-AD in females. This interaction was also observed in gait speed, high in NTg mice. At the same time, the variability of gait presented differences associated with S, with females showing less than males’ variability and, therefore a gait with more homogeneous steps in its trajectory. Additionally, a genotype-dependent difference was observed in cadence, where NTg mice show better performance in this variable with marked differences between males. In the re-test at 16 months, this group registered a gait performance that shows the interaction between the G×S effect in stride length and speed, with the performance of 3xTg-AD females being the one with the best performance in both variables. The re-test of this group at 16 months showed differences in cadence, increasing its performance in the group of 3xTg-AD mice of both sexes and decreasing in the NTg group, see Figure 3A–D and Tables S1–S3. On the other hand, mice at 16 months did not show significant differences in quan- titative variables of gait. Differences could only be observed between the re-test and a naïve group of males at 16 months in stride length, cadence, and speed, where the re-test NTg mice presented a high performance in speed and cadence compared to the naïve NTg mice, 3xTg-AD re-test, and naïve, but lower performance in stride length than naïve mice. In addition, differences were detected between the 3xTg-AD males and females in the variable’s variability and gait speed, with the 3xTg-AD naïve and re-test females showing less variability than the 3xTg-AD males. This difference was also present in gait speed, with a better performance of re-test females followed by naïve 16 months females over 3xTg-AD males in both conditions, see Figure 3A–D and Tables S1–S3. For its part, the neophobia and exploratory activity presented sex differences in the ratio visited corners/rearings, being higher in females of both genotypes, see Figure 4. This difference was maintained at re-test 16 months, with the higher ratio in females. In addition, the ratio in MRA of the groups showed an interaction between G×S, indicating a lower performance in 3xTg-AD re-test males at 16 months and higher in re-test females, see Figure 4B and Tables S1–S3. As in gait, no significant differences were detected in exploratory activity between the naïve 16 months group. However, in contrast to the re-test males at 16 months, the naïve males presented high vertical activity than the re-test, see Figure 4C,D. Between the group of 3xTg-AD mice, S effect was identified in vertical activity, where naïve males presented higher activity. Movement latency was also lower in naïve males, but the same was not observed in 3xTg-AD females, see Figure 4A and Tables S1–S3. 3.3. Muscular Strength: Forelimb Grip Strength and Muscular Endurance—Hanger Test and Response to Gravity: Geotaxis Lower muscle strength can be observed in the resistance distance in the 3xTg-AD animals at the age of 12 months, which, despite not showing statistical differences, shows a trend with less strength in the 3xTg-AD males. No statistically significant differences were detected in the rest of the variables, although a worse performance of the animals was observed, see Table S1. At 16 months re-test and 16 months naïve, no significant differences were detected. On the other hand, significant differences in geotaxis were detected. The group of 3xTg-AD male took longer to complete the test in the re-test 16 months group in contrast to the 12 months and naïve 16 months group, see Figure 5A. Notably, at 16 months in the re-test group, an interaction was observed between the G×S of the animals, showing Biomedicines 2022, 10, 973 9 of 28 VIEW 7 of 31 more significant latency in 3xTg-AD males and NTg females. G×S interaction was also observed between the group of naïve 16 months 3xTg-AD mice. In addition, there was an interaction between the G×R was detected among male mice at 16 months in contrast to Biomedicines 2022, 10, x FOR PEER REVIEW naïve mice of the same age, with the test time being shorter in naïve mice, see Figure 5A10 of 31 and Tables S1–S3. Figure 3. Quantitative parameters of gait. (A) Stride length, (B) variability of stride, (C) speed, (D) Figure 3. Quantitativecadence. parameters of ANOVA, Statistics: gait. (A)G, Stride length, genotype (B) effect, variability G** ofpstride, p < 0.01**, G* < 0.05*. (C) speed, S, sex effect, (D) S* p < 0.05*. cadence. Statistics: G×S, genotype ANOVA, G, and sex interaction genotype effect, G**effects, p < G×S* pG* 0.01**, < 0.05*. p < R, re-test 0.05*. S, effect, sex R* p < S* effect, 0.05*. p < N, naïve effect, 0.05*. igure 2. Survival, frailty score, and postural and structural kyphosis. (A) N* p < 0.05*. a, Survival. aging, Statistics: a** p < 0.01**. Bonferroni post hoc test: g, genotype; s, sex; $ expressed genotype Kaplan–Meier test—Log rank, G, genotypeG× S, genotype effect: χ2, G***and p