Lifelong aerobic exercise protects against inflammaging and cancer

Biological aging is associated with progressive damage accumulation, loss of organ reserves, and systemic inflammation ('inflammaging'), which predispose for a wide spectrum of chronic diseases, including several types of cancer. In contrast, aerobic exercise training (AET) reduces inflammation, lowers all-cause mortality, and enhances both health and lifespan. In this study, we examined the benefits of early-onset, lifelong AET on predictors of health, inflammation, and cancer incidence in a naturally aging mouse model (C57BL/J6). Lifelong, voluntary wheel-running (O-AET; 26-month-old) prevented age-related declines in aerobic fitness and motor coordination vs. age-matched, sedentary controls (O-SED). AET also provided partial protection against sarcopenia, dynapenia, testicular atrophy, and overall organ pathology, hence augmenting the ‘physiologic reserve’ of lifelong runners. Systemic inflammation, as evidenced by a chronic elevation in 17 of 18 pro- and anti-inflammatory cytokines and chemokines (P < 0.05 O-SED vs. 2-month-old Y-CON), was potently mitigated by lifelong AET (P < 0.05 O-AET vs. O-SED), including master regulators of the cytokine cascade and cancer progression (IL-1β, TNF-α, and IL-6). In addition, circulating SPARC, previously known to be upregulated in metabolic disease, was elevated in old, sedentary mice, but was normalized to young control levels in lifelong runners. Remarkably, malignant tumours were also completely absent in the O-AET group, whereas they were present in the brain (pituitary), liver, spleen, and intestines of sedentary mice. Collectively, our results indicate that early-onset, lifelong running dampens inflammaging, protects against multiple cancer types, and extends healthspan of naturally-aged mice.

In the current study, we report that lifelong, voluntary wheel-running with onset at sexual maturity alleviated senescence and inflammaging, including the key drivers of the cytokine cascade and tumor progression (TNF-α, IL-6, and IL-1β), and protected against several different types of cancer (e.g., brain, liver, spleen, and intestinal) in naturally-aged mice. Although lifelong AET provided a pleiotropic stimulus, which attenuated the biological aging process of multiple organ systems (for example, heart, skeletal muscle, and testicles) and improved overall health, the median lifespan was only modestly enhanced in the runners. These results are integrated and discussed within the framework of major biological aging hypotheses, namely the entropic, garbage catastrophe, and inflammaging/remodelling models by Hayflick et al., Terman, and Franceschi, respectively [39, 53, 54].

Ethics
All methods and procedures described herein followed the guidelines published by the Canadian Council of Animal Care and were approved by McMaster University's Animal Research and Ethics Board under the Animal Utilization Protocol 12-03-09.

Animals and study design
Four-week-old C57BL/J6 mice were purchased from Jackson Laboratories (Bar Harbor, ME), matched according to sex and bodyweight, and assigned to either young baseline control (Y-CON; N = 40), old sedentary (O-SED; N = 32), or old lifelong aerobic exercise (O-AET; N = 36) conditions. Following a 3-week acclimation period, O-AET animals were independently housed in activity wheel chambers and engaged in lifelong voluntary wheel-running, while baseline and sedentary cohorts were kept in separate microisolator cages with standard environmental enrichment. All mice were maintained on a 12-h light/dark cycle in a temperature and humidity-controlled room with water and chow available ad libitum (Harlan Teklad 8640 22/5). Sexual maturity (56 d) and old age (806 d) were used as endpoints for young and old cohorts, respectively, with the latter representing the lower limit of the 95% confidence interval for the median lifespan of C57BL/J6 mice [55,56]. One week prior to necropsies, animals underwent a health examination and a battery of functional tests over two days to determine aerobic fitness, muscular strength and endurance, and motor coordination. A sub-set of mice from each group was terminally bled at rest or post exercise for assessment of age-and exercise-induced circulatory factors (Y-CON; n = 7, Y-CON-EX; n = 8, O-SED; n = 8, O-SED-EX; n = 6, O-AET; n = 9, and O-AET-EX; n = 8). To ensure a robust exercise stimulus in the young cohort, additional C57BL/J6 mice were obtained and exercised separately from the main-study animals until complete exhaustion and assessed independently for the secretory response (Y-CON 2 ; n = 5 and Y-CON-EX 2 ; n = 5).
were performed by a veterinary pathologist blinded to the group allocations. Carcasses were first evaluated externally prior to internal examination and collection of major organs of the respiratory (lungs), cardiovascular (heart), digestive (liver, gallbladder, and intestine), lymphatic (spleen), endocrine (pancreas), urogenital (kidney, urinary bladder, and gonads), integumentary (skin), musculoskeletal (skeletal muscle), and central nervous (brain and spinal cord) systems. Any suspected tumor or mass was formalin-fixed, embedded in paraffin, sectioned, and stained with haematoxylin and eosin (H&E) for histological examination. The H&E-stained slides were then sent to an anatomical pathologist for blinded microscope analyses. The tumors, based on histology, were classified as malignant, benign, or unclassified.

Voluntary running-wheel training
O-AET mice were singly housed in 35.3L x 23.5W x 20H cm chambers equipped with freespinning exercise wheels (40 cm/revolution) and counters connected to an activity monitoring software (Lafayette Instruments, Model 80820). Age-related changes in activity patterns were similar to those previously observed in C57BL/J6 mice [57,58], and daily running distances were significantly reduced from middle adulthood onwards (S1 Fig). In a similar fashion to recent published work [59,60], sedentary animals were independently housed without running-wheels to limit the complexity of the environment, spontaneous physical activity, and escape behaviors (e.g., climbing). Assessment of non-wheel cage activity during light and dark cycles using the Comprehensive Lab Monitoring System indicated no differences between O-SED and O-AET, while old animals were significantly less active versus young (S1 Table).

Functional testing and maximal run protocol
One week prior to sacrifice, a battery of in vivo functional tests was administered to predict health and morbidity status of the mice, including muscular strength ('dynapenia 1') and endurance ('dynapenia 2'), motor coordination, and aerobic fitness. Our group has validated these outcomes extensively in various mouse models, and performance variables generally reflect organ dysfunction and total disease burden [61,62]. In brief, forelimb maximal strength (N) was tested using a digital grip strength meter equipped with a pull bar and a force gauge (Columbus Instruments, Columbus, OH). Motor coordination and dynamic balance were assessed using a conventional RotaRod system (AccuScan Instruments, Inc., Columbus, OH), in which the speed (RPM) and time (s) at fall were recorded. Muscular endurance was evaluated by a four limb, inverted grid-hang test until failure (s) (Paw Grip Endurance, PaGE). Lastly, maximal aerobic capacity was predicted in all mice using an incremental running test, which consisted of a ramped exercise protocol on a motorized treadmill starting at 10 m/min and increasing 1 m/min until volitional exhaustion (e.g., sitting/lying on the electrical grid). A separate cohort of young mice was exercised until complete exhaustion, defined as the inability to run on the treadmill despite 10 sec of mechanical prodding, to provide a robust stimulus for exercise factor release. All other parameters remained the same between tests.

Statistical analyses
Parametric data. All parametric data were tested for normality and homogeneity of variance by Shapiro-Wilk and Levene's tests, respectively. Two balanced, three-way ANOVAs were used to determine the independent and interactive effects of age, AET, acute exercise, and gender on all parametric data, including functional, morphometric, and secretory outcomes. Specifically, the 2 x 2 x 2 ANOVA designs were: 1. AGE x ACUTE EXERCISE x GEN-DER (young and old mice) and, 2: AET x ACUTE EXERCISE x GENDER (old mice). If omnibus F-test(s) were significant, a Fisher's LSD post hoc test was used to specify group differences. Independent T-tests were used to assess the exercise-induced secretory response in a sub-set of young animals. Statistical significance was set at P � 0.05 for all parametric tests.
Non-parametric data. For organ and skin pathology, frequencies were coded as binary data (i.e. present vs. non-present), evaluated by either Chi-square or Fisher's Exact tests, and graphed as % incidences for each cohort ([total frequency of pathology/number of necropsies] � 100). Cumulative health index ranks were analyzed by Kruskal Wallis and Dunn's multiple comparison tests. Significance was set at P � 0.05 for all non-parametric tests.
Graphs, tables, and denotations. For conciseness, groups were collapsed across gender and acute exercise and presented as Y-CON, O-SED, and O-AET when appropriate, with the complete statistical designs available in table format in the supplementary data section. Statistical differences are denoted alphabetically (a, b, and/or c) or symbolically ( � , †, and ‡) and if group means are significantly different they do not share the same letter(s) or symbol(s) (P > 0.05).

Aging, organ deterioration, and disability
Skeletal muscle (SM) comprises 30-50% of total bodyweight in mature adults and is lost at 3.7-4.7% of peak mass per decade between 18 and 80 years of age [63,64]. Age-related muscle loss is classified as sarcopenia ('flesh-deficiency') when falling two standard deviations below the young adult mean [65], and is estimated to affect more than 50 million individuals worldwide [66]. Sarcopenia and obesity are independent risk factors of NCDs, and act synergistically to worsen prognostic outcomes and disability [67]. Analogous to the human condition, aging in C57BL/J6 mice is characterized by a relative increase in fat mass versus lean mass [68]. Consistent with these observations, the O-SED cohort exhibited significant bodyweight gain concurrent with SM wasting (Fig 1A and 1B), which primarily affected the fast-twitch muscles of the lower leg, including the quadriceps and anterior crural muscle groups (S2 Table). As shown previously [56], whole brain (specific regions not assessed), heart, liver, kidney, and spleen were enlarged with aging, while testicles underwent significant atrophy (Fig 1C). Skeletal muscle deterioration was coupled with reduced locomotion, aerobic deconditioning, lower muscle strength and endurance (dynapenia; 'power deficiency'), and impaired motor coordination (Fig 2 and S1 Table). The disproportionate loss of functional capacity (-32%, -42%, and -71% in muscle mass vs. strength vs. endurance, respectively) likely reflects age-related degeneration of neural components (α-motor neurons, motor units etc.) which are integral for force/power production, and have previously been shown to contribute significantly to weakness and frailty at old age [63].

Lifelong AET preserves organ reserves
A sedentary lifestyle is very common among elderly and a globally occurring phenomenon, regardless of geographic area and social development index (6). This age-related decrease in PA levels also appears to be conserved across many species on Earth (from worms to humans), and accelerates the rate of organ deterioration in a positive feedback manner. Predictably, regular PA maintains organ reserves and prolongs health-and lifespan, particularly in modern humans [9][10][11][12][13][14][15]. Widely recognized as the foundation of cardiometabolic disease prevention, aerobic exercise training (AET) protects against obesity, T2DM, and death from cardiovascular events by enhancing oxidative capacity and maximal aerobic power [69]. Additional benefits include pro-anabolic effects in SM, enhanced musculoskeletal integrity, and improvements in organs that are not classically associated with exercise (such as skin, kidney, and brain) [70]. Our results are in general agreement with those of Garcia-Valles and colleagues (2015), who demonstrated that lifelong wheel running prevented several indices of frailty and improved overall health in C57BL/J6 mice, while it did not enhance median or maximal lifespan significantly [58]. Specifically, we report that lifelong AET mitigated bodyweight gain, muscle loss, strength deficits, and testicular atrophy, and even elevated motor function and aerobic fitness above young control levels (Figs 1 and 2). However, despite many favorable adaptations predicted to extend lifespan, the survival rate at 26 months (i.e. median lifespan; equivalent to 70-75 human years) was only modestly improved in lifelong runners vs. sedentary controls (+6% vs. O-AET vs. O-SED; NS). Notwithstanding, a 3-10% extension in average life-expectancy is consistent with large population-based studies in human runners [19,20].

Aging and systemic inflammation
In parallel with age-related declines in organ reserves, glucose tolerance, and sex-hormones, systemic inflammatory factors steadily rise from the third decade onwards in humans [71,72]. Imperfect repair, recycling, and removal of danger-associated molecular patterns (DAMPs) are postulated to trigger activation of evolutionary conserved cell danger response programs (CDR and DDR, respectively), which fuel a vicious cycle of unresolved cell damage, build-up of toxic debris, senescence, and inflammation [52]. Inflammaging is thereby multi-factorial, but principally attributed to DAMP-activation of the innate immune system, immunosenescence, and the senescence-associated secretory phenotype (SASP) [73].
Pro-inflammatory gerokines, such as TNF, IL-1, and IL-6, are positively correlated with morbidity and mortality in elderly [71,74], and predispose to a wide spectrum of NCDs, Lifelong aerobic exercise protects against inflammaging and cancer including cardiovascular diseases, neurodegenerative disorders, and cancers [33-37, 42, 75,76]. However, the biological effects of any cytokine in vivo depends on the presence of other cytokines with additive, synergistic, or antagonistic modes of action [77]. For example, IL-1 and TNF-α work in synergy to initiate and amplify the cytokine cascade, and blockade of either master regulator improves survival following infection [78,79]. The remodeling theory of aging, as outlined by Franceschi et al., stipulates a progressive accumulation of both proand anti-inflammatory gerokines occurring simultaneously with aging [39,41].

Aging and myokines
In principle, all cells that undergo exchange with the extracellular environment may contribute to the myriad of blood-borne factors that exert autocrine, paracrine, and endocrine effects. For example, immune, muscle, and fat cells each provide a unique spectrum of proteins to the circulatory system (cytokines, myokines, and adipokines, respectively), while also exhibiting significant overlap between their secretomes (e.g., myocytokines, adipocytokines, adipomyokines etc.).
Skeletal muscle may be considered the largest endocrine organ in the body that actively secretes myokines at rest and in response to various stressors (pathological, contractile, metabolic etc.). To date, over 300 factors secreted by primary SM cells have been identified using targeted and exploratory proteomic approaches [64,[80][81][82]  Aging has well-documented adverse effects on SM cells, including build-up of DAMPs, chronic immune activation, and myofibrillar atrophy, all of which may alter the muscle-specific secretome. Resident and invading immune cells, adipocytes, and fibroblasts, also contribute to the SM secretory spectrum in vivo, especially during remodelling and pathological processes. Because muscle fiber degeneration may fuel systemic inflammation independently (in muscular dystrophies, for example), it is likely that age-related SM deterioration partly underlies inflammaging. Regardless of the tissue/cell origins, the key drivers of the cytokine cascade, TNF-α, IL-1β, and IL-6 (+128%, +159%, and +75%, respectively, O-SED vs. Y-CON), were chronically elevated at old age, and may have contributed to SM wasting in this study. Further indicative of anabolic dysregulation and an overall pro-catabolic systemic environment, myostatin, follistatin-like protein 1, FGF-21, and osteocrin/musclin, and to a lesser extent IL-15 and fractaline/CX3CL1, were significantly lower in O-SED vs. Y-CON cohorts (S5 and S6 Tables).
Conversely, the matricellular protein SPARC/osteonectin followed the same general pattern as pro-and anti-inflammatory gerokines and was elevated with aging ( Fig 5). SPARC is expressed during remodelling and repair in intestinal, glandular, skin, muscle (cardiac, smooth and skeletal), adipose, and neoplastic tissues [83]. In aggressive tumor phenotypes, SPARC may act as a positive regulator of metastatic potential by augmenting vascular permeability [84]. Persistent induction of SPARC promotes tissue fibrosis, nephropathy, retinopathy, and non-alcoholic fatty liver disease, while its deletion/inhibition protects against the same [85][86][87]. Considering that circulating SPARC is up-regulated in metabolic disease, accelerated aging conditions (Werner Syndrome etc.), and in sarcopenia [85,[88][89][90][91][92][93], it is conceivable that it may be used as a general biomarker of advancing age and/or pathological aging processes. Lifelong aerobic exercise protects against inflammaging and cancer

Lifelong AET mitigates inflammaging
In contrast to chronic stress, the intermittent application of a physiological challenge, such as AET, minimizes perturbations to the intracellular milieu and allows for sufficient recovery for the cell to evolve into a more stress-resistant, homeostatic level, thereby attenuating biological aging and protecting against inflammation [51, 52]. The anti-inflammatory effects of exercise training are well-documented (with higher IL-10, and concomitantly lower TNF-α, IL-6 and CRP levels), and benefits appear to be independent of age and chronic disease [94]. As expected, lifelong AET effectively dampened inflammaging in the C57BL/J6 model, with a restoration of both pro-and anti-inflammatory gerokines to young control levels (Fig 3A and 3C; NS vs. Y-CON), including the pivotal regulators of the cytokine cascade and tumor development (Fig 3E; NS vs. Y-CON). In the full ANOVA design, 10 out of 17 age-induced markers were AET-responsive, specifically IL-17a (Fig 3B), IL-13 (Fig 3D), TNF-α (Fig 3F), IL-6, GM-CSF, IFN-γ, IL-2, IL-10, IL-12 (p70), and CCL-2/MCP-1 (S3 and S4 Tables). Although a majority of gerokines were significantly attenuated and even normalized by lifelong running (notably IL-13, GM-CSF, IFN-γ, IL-17, TNF-α, and IL-1β), basal levels of IL-10 and CXCL1/ KC were moderately enhanced by AET (+25% vs. O-SED).
IL-10 inhibits the activity of TH1 cells, NK cells, and macrophages and the production of wide range of pro-inflammatory cytokines and chemokines, including TNFα, IL-1α/β, IL-6, and MCP-1 [95]. By counteracting the proteolytic effects of IL-1β, the anti-inflammatory actions of IL-10 are integral for muscle repair and growth following strenuous exercise [96]. CXCL1/KC (GRO-α) is a chemoattractant that stimulates neutrophil recruitment and phagocytosis of injured tissue, but the functional significance of elevated serum KC levels following endurance training is unclear. Overexpression of CXCL1/KC in SM promotes its systemic release and induces an oxidative phenotype in high-fat fed mice, which protects against fat gain and glucose intolerance [97]. CXCL1/KC may therefore play an integral role in mediating the whole-body metabolic benefits of aerobic exercise training and be of therapeutic interest as an exercise mimetic. Lastly, circulating SPARC/osteonectin was restored to young control levels by lifelong running (Fig 5), further suggestive of SPARC being a general biomarker of advancing age and/or pathological aging. We did not observe any significant effects of lifelong AET on basal FSTL-1, musclin, or fractalkine; while MSTN ("), FGF-21 (#) and IL-15 (#) were moderately affected in our study (S5 and S6 Tables).

Lifelong AET alters the inflammatory response to acute exercise
Aerobic exercise is a single, hormetic stress stimulus that encompasses transitory spikes in intracellular danger signals (Ca 2+ , ROS, pH, and hypoxia), metabolic intermediates (NAD+/ NADH and AMP/ATP), and a myriad of neurotransmitters, hormones, and organokines, which act synchronously to stimulate mitochondrial biogenesis, antioxidant defense, cellular repair and recycling, and immunity [45-50]. We recently coined the term 'exerkine' to describe any biomolecule released into circulation in response to exercise, including non-coding RNAs, mRNAs, metabolites, peptides, and proteins [49]. Approximately 50% of all SMderived secretory factors are contractile activity-mediated [64], and other cell types likely contribute significantly to the total exerkine pool in vivo, principally liver, fat, and immune cells [77].
Several basic parameters may affect the systemic response to acute contractile activity, such as exercise variables (e.g., mode, intensity, and duration), blood sampling (time and fraction), training status, gender, and age. As reviewed by Peake at el., the exercise-mediated increase in circulating cytokines is highly variable and typically reflects the relative exercise intensity and degree of intracellular stress/damage [77]. In our study, an acute bout of aerobic exercise elicited a significantly higher inflammatory response in old sedentary C57BL/J6 mice (O-SED-EX) compared to aerobically trained (O-AET-EX) and young counterparts (Y-CON-EX), suggestive of a protective adaptation induced by lifelong AET to intracellular stress (S3 and S4 Tables). Statistical significance was almost exclusively driven by the robust exercise effect in O-SED-EX (AGE x ACUTE EXERCISE x GENDER; Table 1a) and included many classical pro-and anti-inflammatory cytokines previously identified as exercise factors. To illustrate this further, we analyzed the secretory response from old cohorts independently (AET x ACUTE EXERCISE x GENDER; Table 1b), and observed that the vast majority of inflammatory factors (14/18) and myokines (5/7) were contractile activity-regulated at old age ( Fig 6A). A second cohort of young animals was vigorously exercised until complete exhaustion (Y-CON-EX 2 ), and this intense physiological stimulus was insufficient to induce the same magnitude and breadth of cytokines as in O-SED (Fig 6B; Table 1c; S7 Table). Other notable age discrepancies in the secretory response were SPARC (" young vs. $ old), FSTL-1 (" young vs. $ old), and MSTN (# young vs. " old), while musclin and FGF-21 were elevated by acute exercise regardless of age (Table 1b and 1c).

Lifelong AET protects against cancer
Inflammaging predisposes for a wide range of NCDs, including multi-systemic cancers. Chronic inflammation directly underlies~20% of all cancers and a vast majority of solid  (Fig 6A). C) Exerkines according to an additional cohort of young C57BL/J6 mice exercised until exhaustion and analyzed by independent t-tests (Fig 6B). The arrows depict the directionality of the contractile activity-induced response of each exerkine.  Main effects of exercise were mainly attributed to old cohorts as shown above and in Table 1a (left column). By removing young animals from the statistical analyses (Table 1b middle column), several additional exerkines were uncovered in old mice. B) Maximal exercise response of young C57BL/J6 mice expressed relative to unexercised controls (Y-CON-EX 2 vs. Y-CON 2 ; Table 1c right column). � Significantly different from unexercised, age-matched controls (P � 0.05). Groups that do not share the same letter(s) are significantly different (P � 0.05). The anti-tumorigenic effects of regular PA may be attributed to attenuating the loss of organ reserves, DAMP removal, and mitigation of the pro-inflammatory, systemic environment. Acutely, aerobic exercise is a stress stimulus that induces a CDR-driven inflammatory response, which is rapidly resolved and favors tumor clearance [98]. Specifically, voluntary wheel running has previously been shown to inhibit cancer development across a range of tumor models in mice, and involves an epinephrine-dependent mobilization of IL-6Rα-positive NK cells that require IL-6 for efficient tumor homing and infiltration [32].
In this study, we found malignant tumors in the liver (N = 3), brain (pituitary; N = 2), intestines (N = 1), and spleen (N = 1) of O-SED C57BL/J6 mice, while lifelong runners were cancer-free (Table 2). In keeping with the Jackson Laboratory Mouse Tumor Biology Database [99], we report that organs most likely to become malignant were the liver, pituitary, and intestines, and that these tumors were mainly hematopoietic malignancies (e.g., lymphomas) and adenocarcinomas (Fig 7A-7C). In contrast to the exclusive finding of malignant tumors in the O-SED cohort, there was a roughly equal distribution of benign tumors in the reproductive (ovary and testicle), digestive (gallbladder and intestine), and urinary (kidney, Fig 7D) systems of non-runners vs. runners. Overall, lifelong running protected against total organ pathology (22% vs. 64% in O-AET vs. O-SED, respectively), multi-systemic cancers (0% vs. 21%), and skin deterioration (3% vs. 15%) (Fig 8).

Lifelong AET extends healthspan
Cell aging is characterized by progressive malfunctioning of pivotal systems that regulate energy production, growth, and quality control (repair, folding, recycling/degradation, and synthesis), progressively depleting organ reserves and eventually causing death. Physical inactivity may accelerate this process by antagonizing a 'vicious cycle' of unresolved damage, persistent CDR/DDR activation, and inflammation. In this study, we found that early onset, lifelong AET dampened inflammaging, including master cytokines IL-1, TNF-α, and IL-6 (in circulation and tissue-specific), protected against multi-systemic cancers, and significantly enhanced healthspan of C57BL/J6 mice (cumulative health-index rank; Fig 9). These data support the contention that lifelong running decelerates the 'vicious cycle' and protects against the Table 2 Lifelong aerobic exercise protects against inflammaging and cancer most salient features of the aging process, namely critical organ deterioration/pathology, frailty, and premature death (Fig 10).

Conclusions
Cumulative evidence over the last 50 years unequivocally shows that sedentary living accelerates organ deterioration and increases all-cause mortality risk. In contrast, aerobic exercise training promotes healthy aging and extends average life expectancy by 3-10% [20]. Our data confirm that lifelong running provides significant health benefits and protects against cancer, while only modestly improving the median lifespan of naturally-aged mice. For more Lifelong aerobic exercise protects against inflammaging and cancer substantial pro-longevity benefits 'of mice and men', a combination of lifestyle modification, rejuvenation biotechnology, and intermittent pharmacotherapy may be necessary (e.g., antiaging polytherapy). Given that the global burden of chronic disease is steadily rising (population aging and physical inactivity pandemic), the development of innovative strategies to increase daily PA levels at the societal level, and/or alternative means to harness the multi-systemic benefits of exercise (e.g., exercise mimetics), will be of paramount importance in the 21 st century.  Table 1 were coded as binary data (present vs. non-present), evaluated by either Chi-square or Fisher's Exact tests, and graphed as % incidences for each group [(total frequency of pathology/number of necropsies) � 100)]. Symbols indicate significantly higher (P < 0.05) vs. Y-CON ( � ) or Y-CON and O-AET ( � †).