Rodent diet aids and the fallacy of caloric restriction

Understanding the molecular mechanisms of normal aging is a prerequisite to significantly improving human health span. Caloric restriction (CR) can delay aging and has served as a yardstick to evaluate interventions extending life span. However, mice given unlimited access to food suffer severe obesity. Health gains from CR depend on control mice being sufficiently overweight and less obese mouse strains benefit far less from CR. Pharmacologic interventions that increase life span, including resveratrol, rapamycin, nicotinamide mononucleotide and metformin, also reduce body weight. In primates, CR does not delay aging unless the control group is eating enough to suffer from obesity-related disease. Human survival is optimal at a body mass index achievable without CR, and the above interventions are merely diet aids that shouldn't slow aging in healthy weight individuals.CR in humans of optimal weight can safely be declared useless, since there is overwhelming evidence that hunger, underweight and starvation reduce fitness, survival, and quality of life. Against an obese control, CR does, however, truly delay aging through a mechanism laid out in the following tumor suppression theory of aging.


Introduction
The last century has brought impressive gains in human life expectancy (Oeppen and Vaupel, 2002). Most of these gains have been achieved by reducing unnecessary deaths due to infectious disease, violence and injuries and massive improvements in standards of living, nutrition and sanitation. Most deaths in industrialized countries are now due to diseases of old age like cancer or cardiovascular disease, which cannot be prevented or delayed without understanding the underlying process of aging itself. Even though the phenomena of aging are familiar to all humans and can be well described, surprisingly little is known about the molecular mechanism(s) that drive the process of aging and, as a consequence, whether there is anything that might be done about it (Hayflick, 2021;Kirkwood, 2017).
Experimentally however, caloric restriction (CR) is hailed as a universal mechanism to delay aging (Madeo et al., 2019a). CR, the process of reducing caloric intake in a group of animals relative to the amount the animals would consume if given unlimited access to food but without malnutrition (i.e. while still maintaining and adequate supply of nutrients like vitamins and minerals) delays aging relative to a control group (which is given unlimited access to food) (Masoro, 2005). Discovered first in rodents more than a hundred years ago (McCay et al., 1935), it has since been reproduced in a variety of species ranging from invertebrates to higher mammals (Weindruch, 1996). This culminated in perhaps the longest and possibly most important experiment in aging research conducted so far: testing the validity of CR as an age-retarding intervention in primates (Colman et al., 2009(Colman et al., , 2014Masoro, 2003;Mattison et al., 2012a). The results of the experiment, despite being somewhat disappointing (Maxmen, 2012), are generally interpreted as confirming the effectiveness of caloric restriction as an intervention that delays the aging process (Mattison et al., 2017).
Over the last one or two decades, several small molecules have been found to increase life span in mice, the most famous and promising of which could be considered to be resveratrol, rapamycin, increasing cellular levels of nicotinamide mononucleotide and metformin. All of these are seen as well-established pathways relevant to the human aging process (Campisi et al., 2019). How these substances prolong life is an active area of research, but there is a broad consensus that all of them produce changes at least partially resembling those seen under caloric restriction, and this is interpreted as a validation of their anti-aging effects (de Cabo et al., 2014;Fontana and Partridge, 2015). Mice are evolutionary and metabolically close to humans; it is commonly assumed that translation of these findings to humans is feasible and merely a question of time and technology.
I will argue here that caloric restriction is merely preventing the excessive obesity-related mortality seen in most experimental animals given unlimited access to food. Furthermore, most of the above pharmacologic interventions that purportedly delay aging are mere weight loss compounds, and there is little evidence that they affect the root cause of aging other than through weight loss. Their positive effects on life span originate from their ability to reduce proliferative signaling and body weight gain and the excess mortality associated with it in mice. Weight loss is frequently downplayed in investigations of how these compounds work, but a uniting feature of resveratrol, rapamycin, metformin and NAD+ supplementation.

Obesity and health in rodents
In a natural environment, food scarcity is the primary constraint on survival and reproductive success. Food supply is highly variable in the wild, so the natural behavior of most animals is to consume and store as much energy as possible (Berry and Bronson, 1992). Laboratory animals with constant access to unlimited food eat much more than they would be able to in the wild. Mammals store available excess energy mostly in fat deposits, and laboratory mice and rats are grossly obese compared to their brethren living in the wild (Austad and Kristan, 2003;Martin et al., 2010). Mice taken from their natural environment weigh only about half as much as common laboratory strains (Mutze et al., 1991), and are much lighter than wild-derived mice (mice that have been caught in the wild but kept in a laboratory for a few generations) (Austad and Kristan, 2003). Laboratory mice have been bred in captivity for hundreds of generations. Captive breeding can lead to further adaption and selection for traits such as fast growth and early fertility even though mice caught from the wild or held in captivity for a short time are equally passionate eaters (Austad and Kristan, 2003). While mice caught from the wild had body fat levels between 3 and 5%, mice in captivity had between 9 and 22% body fat, with the popular C57BL/6J inbred strain having a whopping 22% body fat. Part of the obesity seen in laboratory mice might therefore be due to selection and adaption to the sedentary lifestyle in a laboratory cage (Harper et al., 2006), but their general behavior seems to have stayed the same. How far overeating in captive mice might shorten life span is therefore an important factor to consider in order not to over-extrapolate mechanism of aging in excessively obese rodents to rodents of a healthy weight or non-obese humans.
The effect of obesity on mouse survival is illustrated in Supplementary Fig. 1. Mouse body mass index (BMI, at 13 weeks age) significantly correlates with life span (Supplementary Fig. 1b; p = 0.011, Pearson correlation coefficient -0.53) and survival in strains with a lower BMI is especially improved in mice that survived beyond the median lifespan ( Supplementary Fig. 1a), emphasizing the negative effect of obesity on mouse maximum life span. Life span is also significantly correlated to BMI at 6 and 12 months of age (both p < 0.05) . Similar plots have been created previously by . Measurements of body size to enable the calculation of mouse BMI are relatively rare, but weight measurements are common. Increased mouse body weight is associated with shorter life spans, and the correlation is significant for a range of ages and genders . Body weight at younger age is also negatively associated with life span in 31 strains of ILSXISS recombinant inbred mice ( Supplementary Fig. 2i) (Bennett et al., 2005;Liao et al., 2010). Beyond middle age however, the capacity to maintain body weight becomes a predictor of life span, and the negative correlation between body weight and life span disappears (Liao et al., 2011). This demonstrates a robust negative association of weight and obesity with life span in a wide variety of inbred mouse strains. Obesity is seldom appreciated as a determinant of longevity in mice, even though early weight gain has been found to be an important determinant of life span (with bigger mice dying earlier) (Goodrick, 1977;Miller et al., 2002). In rats, dietary behavior and growth responses early in life accurately predict longevity (Ross et al., 1976), and growth in general negatively impacts life span in rodents (Rollo, 2002).

Caloric restriction in mice
The original discovery of the life-prolonging effects of caloric restriction were reported about 85 years ago (McCay et al., 1935) and, over the years, confirmed in other rodents (McDonald and Ramsey, 2010). Also called dietary restriction, food restriction or energy restriction, CR fits well with the rate of living theory of aging (Pearl, 1928), which is supported by the both convincing and intuitively satisfying relation of body size and metabolic rate to life span (Austad and Fischer, 1991). As CR obviously reduces energy consumption per animal, a fixed life-long allowance for energy consumption should lead to the observed life span increase. The National Institute of Aging Biomarkers of Aging Program evaluated more than 60.000 rodents, and confirmed CR as a "gold standard" for evaluating intervention strategies in aging (Turturro et al., 1999). The magnitude of the life-span extending effects of CR, however, differed significantly between both strains and gender. When weight gain during adulthood is measured as the area under the weight curve, the magnitude of life span extension afforded by CR is beautifully correlated to the weight gain in control mice fed ad lib (Sohal and Forster, 2014). DBA/2 mice exhibit less weight gain after maturity even with unlimited access to food and respond little to CR (Hempenstall et al., 2010). Differences in body weight between ad libitum and calorie-restricted animals are small, and decrease with increasing age (Forster et al., 2003). Body fat was almost the same in older ad lib and calorically restricted DBA/2 mice, and it was concluded that life span extension by CR depends on a positive imbalance between energy intake and energy expenditure, i.e. significant overeating (Sohal et al., 2009).
Calorically restricted laboratory mice weigh significantly more than wild mice and have far higher body fat levels (Austad and Kristan, 2003). Mice consuming a normal rodent diet (which provides 10% of calories from fat) and kept under standard conditions are already morbidly obese and poor models of healthy human aging (Martin et al., 2010). High fat diets designed to maximize obesity deliver much more calories from fat (60% for example in (Baur et al., 2006)) (Speakman, 2019). The purpose of feeding high fat diets to mice is often purportedly to mimic a human diet high in fat, but high fat diets seem to activate the reward system and are a sure method to increase obesity to a level probably unimaginable for a wild mouse (Hu et al., 2018;Salmon and Flatt, 1985). Reducing this morbid obesity without serious toxicity must be expected to improve health and life span significantly.

Alternative dietary interventions in mice
Like CR, protein and particularly methionine restriction, but not carbohydrate or lipid restriction, increases maximum longevity in rodents (Sanchez-Roman and Barja, 2013). Amino acids, and branched-chain amino acids (BCAA) like leucine in particular (Wolfson et al., 2016), are key signals for mTOR activation and insulin release (Chotechuang et al., 2009;Jewell et al., 2013;Neishabouri et al., 2015;Proud, 2004;Yang et al., 2010), and high branched-chain amino acid intake is central to the development of obesity-associated insulin resistance via chronic phosphorylation of mTOR (Newgard et al., 2009). Low protein content induces compensatory overeating (Le Couteur et al., 2016;Solon-Biet et al., 2014) and the opposite, an exome-matched diet (which adjusts food protein content to the exome amino acid composition), reduced mouse ad libitum food intake despite improved growth and higher lean and fat mass (Piper et al., 2017). In rats, methionine restriction stops growth completely despite higher food and calorie intake on a body weight basis (Orentreich et al., 1993). Body weight was halved, and testis and seminal vesicle weights were reduced to around one fourth to one sixth of controls (Orentreich et al., 1993), but both median and maximum life span was increased significantly in 4 strains of rats when methionine was reduced from 0.86% to 0.17% (Orentreich et al., 1993;Zimmerman et al., 2003). Methionine restriction below 0.12% was lethal within 1 month (Orentreich et al., 1993). In mice, about 25% of methionine-restricted (0.1− 0.15% vs. 0.43% in controls, from 6 weeks of age) mice died within the first year, but the survivors managed to gain weight at about one fourth the rate of controls (Miller et al., 2005). The methionine-deficient diet lowered serum IGF-1, insulin and glucose, and body weights averaged ~65% of controls between 6 and 18 months of age. The surviving methionine-restricted mice lived significantly longer, with both median and maximum life span increased (Miller et al., 2005). Starting methionine restriction later prevented early mortality and increased longevity less, but body weight was unfortunately not reported (Sun et al., 2009). Similarly, a low BCAA diet (leucine, isoleucine and valine reduced to 33% of control) reduced weight and, consistent with weight loss, promoted metabolic health and decreased frailty in aged mice (Richardson et al., 2021). Low BCAAs also reduced weight in young mice, but longevity increased only in males, not in females or when BCAA restriction was started in middle age (Richardson et al., 2021).

Interventions in mice -resveratrol
The first pharmacological intervention to achieve widespread attention was resveratrol with its promise of "guilt-free gluttony" (Kaeberlein and Rabinovitch, 2006). Consumption of resveratrol at a not unreasonable dose improved health and survival in C57BL/6 mice fed a 60% fat diet (Baur et al., 2006). Previously known for its cancer-preventive activity (Jang et al., 1997) and shown to increase life span in yeast (Howitz et al., 2003), C. elegans (Viswanathan et al., 2005) and short-lived fish (Valenzano et al., 2006),resveratrol also came with the story of the "French paradox", a low incidence of coronary heart disease despite high dietary intake of cholesterol and saturated fats in France (Renaud and de Lorgeril, 1992).
Resveratrol-mediated weight loss was clearly shown, even though it was downplayed by pointing out that the weight loss lost statistical significance at old age (Baur et al., 2006). Weight loss due to resveratrol supplementation was replicated by Lagouge et al., who also showed a remarkable decrease in body fat (Lagouge et al., 2006). Since then, many others have also reported weight loss due to resveratrol supplementation (Andrade et al., 2014;Cho et al., 2012;Milton-Laskibar et al., 2017;Soyoung et al., 2011;Sung et al., 2017;Um et al., 2010;Wang et al., 2020), with weight loss generally bigger and/or dependent on a high fat diet (Milton-Laskibar et al., 2017). Moderate Sirt1 overexpression also reduced weight gain under a high fat diet, despite increased food intake (Pfluger et al., 2008). Sometimes weight or weight gain in resveratrol treatment groups is not reported (Price et al., 2012). Lifespan extension by resveratrol is not observed when nutrients are restricted in C. elegans and Drosophila melanogaster, which shows that the pathways of CR and resveratrol supplementation might overlap (Wood et al., 2004). The NlA Interventions Testing Program, established to evaluate agents hypothesized to increase life span, found that resveratrol at doses twofold (0.3 g/kg food) and eightfold (1.2 g/kg food) higher than those showing a beneficial effect in C57BL/6 mice on a high-fat diet did not significantly extend life span in genetically heterogeneous UM-HET3 on a normal diet, when started at 12 months age (Miller et al., 2011). There seemed to be, however, a clear trend towards increased median and 90th percentile survival, at the higher resveratrol concentration, especially in males (Miller et al., 2011). Unfortunately, body weight was not reported for both resveratrol concentrations (Miller et al., 2011). When the lower resveratrol concentration (0.3 g/kg food) was started earlier (at 4 months of age), it led to a small, insignificant increase in median survival, with no discernible change in survival at the 90th percentile (Strong et al., 2013). Body weights were only slightly and insignificantly decreased in males, but there was a clearer trend towards lower body weights in females, which was nevertheless significant only at 12 months (Strong et al., 2013). Others also report improvements resembling CR, but insignificant effects on weight and life span with a normal diet (Barger et al., 2008;Pearson et al., 2008), and a small increase in longevity and weight with a low dose of resveratrol in mice on a high fat diet (Pearson et al., 2008). Resveratrol neither causes weight loss nor acts as a CR mimetic when administered intraperitoneally, i.e. when the gut is bypassed (Pallauf et al., 2019), supporting evidence that obesity reduction by resveratrol might be partially mediated by gut microbiota (Wang et al., 2020), which can consume dietary fatty acids (Agans et al., 2018). Absorption of fats from the small intestine is dependent on the enzyme pancreatic (triglyceride) lipase (Lowe, 1997), and pancreatic lipase is effectively inhibited by resveratrol (Sergent et al., 2012), even though fecal lipid content is unchanged (Baur et al., 2006;Lagouge et al., 2006;Wang et al., 2020). Polyphenolic pancreatic lipase inhibitors contained in tea also reduce body weight and body weight gain in mice fed a high fat diet by decreasing fat uptake (Basu and Lucas, 2007;Chan et al., 1999;Grove et al., 2012;Yang et al., 2016). This possible explanation for the weight loss is supported by the finding that weight loss is, like life span extension, not observed when animals are fed normal chow (Milton-Laskibar et al., 2017), as for green tea (Kitani et al., 2007;Strong et al., 2013).
Resveratrol activates the fuel-sensing AMP-activated kinase (AMPK), which is activated under conditions of low energy, or more specifically, by starvation or physical exercise. Mice lacking AMPK did not show resveratrol-induced weight loss (Um et al., 2010). This led to the conclusion that many of the effects of resveratrol are due to AMPK activation, which, like resveratrol, reduces fat accumulation, and increases glucose tolerance, insulin sensitivity, mitochondrial biogenesis and physical endurance (Price et al., 2012;Um et al., 2010), supporting the interpretation that the starvation-like response seen with resveratrol is induced by exactly that: "relative" starvation compared to high fat diet controls. Resveratrol's mechanism of action remains controversial (Pacholec et al., 2010;Park et al., 2012), but activation of sirtuins still seems to be seen as the main mechanism and a promising strategy for testing in humans (Longo et al., 2015). A 2016 review lists dozens of clinical trials for resveratrol and other sirtuin-activating compounds (Bonkowski and Sinclair, 2016), none of which seem to have been successful (as of February 2021).

Interventions in mice -rapamycin
Rapamycin, an immunosuppressant drug approved for use in humans, extends mouse lifespan when fed late in life (Harrison et al., 2009). This created a lot of excitement, because all previously discovered interventions had to be started early in life for a significant effect. Rapamycin inhibits mammalian target of rapamycin(mTOR), a regulator of cell metabolism, growth and proliferation (Laplante and Sabatini, 2009). Its activation is crucial in rapid proliferation and clonal expansion of T cells in the adaptive immune response. Because of this, rapamycin is an immunosuppressant and used to prevent the rejection of transplanted organs (Saunders et al., 2001). Because of its antiproliferant action, it also inhibits angiogenesis and is approved for the treatment of advanced renal cell carcinoma (Motzer et al., 2008).
Rapamycin (14 mg/kg food) was encapsulated in an enteric coating material that led rapamycin to be released in the small intestine rather than in the stomach and reach blood concentrations approximately tenfold higher than non-encapsulated rapamycin (Nadon et al., 2008). Body weight was reportedly not changed, but the authors probably meant "not significantly changed" and did not show the data (Harrison et al., 2009). Mice given rapamycin at the same dose but from 9 months age were lighter than controls (Miller et al., 2011). Rapamycin-induced weight loss was dose-dependent, with higher doses leading to more weight loss and larger increases in median and 90th percentile life span (Miller et al., 2014). Another study found reductions in weight, lean and fat mass when rapamycin treatment was started both in midlife or late in life (Neff et al., 2013). A rapid onset, significant decline in body weight was also apparent during transient (3 month) rapamycin treatment (Bitto et al., 2016).The weight loss remained significant for several months after cessation of rapamycin treatment and did not seem to disappear (Bitto et al., 2016). During the early phase of transient rapamycin treatment (2-6 weeks, where weight loss is not yet apparent) only the detrimental effects of rapamycin treatment are observed, but once rapamycin-induced weight loss becomes large enough, beneficial effects start to be observed (Fang et al., 2013). When rapamycin was administrated intermittently (2 weeks per month) from an early age, it increased lifespan and reduced body weight and tumorigenesis in female mice (Anisimov et al., 2011). In rats, rapamycin treatment reduced body weight gain by 86%, which was partly attributed to reduced food intake, but mostly due to increased energy expenditure, decreased food efficiency and decreased plasma leptin levels (Houde et al., 2010). Side effects of rapamycin treatment are not limited to immunosuppression. Rapamycin treatment promotes insulin resistance, severe glucose intolerance, and increased gluconeogenesis by impairing lipid deposition in rat adipose tissue (Houde et al., 2010). Cataracts were increased and testicular degeneration was observed in 83% of rapamycin-treated male mice (Wilkinson et al., 2012). All stages of spermatogenesis were lost and the location where sperm should be was filled with multinucleate giant cells, dead cells from different stages, and debris. The full effects was apparent even in those mice receiving the lowest rapamycin dose (Wilkinson et al., 2012).
Deletion of a component of the mTOR signaling pathway, ribosomal S6 protein kinase 1 (S6K1), led to increased lifespan and resistance to age-related pathologies such as bone, immune and motor dysfunction and loss of insulin sensitivity (Selman et al., 2009). S6K1 knockouts also weighed only two-thirds of controls, fat tissue was reduced to one third of controls, and the authors emphasize the parallels to changes seen with CR (Selman et al., 2009).
As both CR and rapamycin inhibit mTOR, it has been suggested that the beneficial effects of rapamycin might be attributed to its ability to mimic caloric restriction (Kaeberlein and Kennedy, 2009). In particular, the striking dose-response relationship for both life span extension and weight loss (Table 1) suggest weight loss might be central to the effects of rapamycin (Miller et al., 2014).

Interventions in mice -NAD+
Sirtuins are NAD+-dependent deacetylases, and raising the concentration of the cosubstrate NAD+ promotes sirtuin activation (Mouchiroud et al., 2013). NAD+ seems to decline with age and limit sirtuin activity (Imai and Guarente, 2014), and several strategies to raise NAD+ levels have been reported to activate sirtuins similar to resveratrol and other sirtuin-activating compounds (Imai and Guarente, 2016). NAD+ also plays a central role in redox reactions, where the reduced forms NADH and NADPH supply electrons to reduce a variety of metabolic substrates (Jones and Sies, 2015).
The NAD+ precursor nicotinamide riboside reduced body weight in both old and young mice, accompanied by improved mitochondrial and stem cell function over a short treatment period (Zhang et al., 2016a). The authors maintained that body mass remained comparable among all groups, demonstrated enhanced life span when NAD+ repletion started at 2 years of age, but weight and gender was not reported for this part of the study (Zhang et al., 2016a). Feeding mice the NAD+ precursor NMN also suppressed body weight gain by accelerating metabolism (Mills et al., 2016). The NIA Interventions Testing Program could replicate the weight loss, but not the life extension with NR supplementation (Harrison et al., 2021). The age-related decrease in NAD+ is mediated by increased expression of the NADase CD38 (Camacho-Pereira et al., 2016). CD38-deficient animals are smaller, lighter, have a higher metabolic rate and are protected against high-fat diet-induced obesity (Barbosa et al., 2007), pointing to weight reduction as the central mechanism of how NAD+-mediates mouse health improvements.
Overexpressing two NAD+-producing enzymes in mice also mimicked caloric restriction and promoted health (Diaz-Ruiz et al., 2018). The mice had lower body weight between the age of 2 and 12 months, but were heavier at old age, when age-related terminal weight loss started to appear (Diaz-Ruiz et al., 2018). Similarly, mice moderately overexpressing human glucose-6-phosphate dehydrogenase (G6PD), an enzyme producing the reductive fuel NADPH, had higher levels of NADPH, lower ROS-derived damage, lower body weight and were protected from aging-associated functional decline, including extended median lifespan in females (Nobrega-Pereira et al., 2016). In humans, G6PD deficiency is the most common gene mutation. An estimated 400 million affected individuals are mostly asymptomatic with no known effect on life span (Beutler, 2008).

Interventions in mice -metformin
Mice treated with metformin (0.1% in food) weighed less than controls despite consuming more food, and a variety of beneficial effects, including increased lifespan, resembled caloric restriction (Martin--Montalvo et al., 2013). In apparently mammary tumor-prone short-lived HER-2/neu transgenic mice (Blagosklonny and Campisi, 2008) metformin (100 mg/l in drinking water) extended the life span and increased food consumption without changing body weight, but the weight data was not shown (Anisimov et al., 2005). In a second, identical study with fewer animals, metformin decreased body weight and extended life span, but maximum life span was decreased (Anisimov et al., 2010). In female SHR mice, another very short-lived, unknown outbred strain, metformin led to remarkable increases in both median and maximum longevity, without affecting body weight (Anisimov et al., 2008). On the other hand, the NIA Interventions Testing Program found no life span extension with 0.1% metformin in food, but did not report weight (Strong et al., 2016). In rats, metformin reduced body weight but did not increase life span (Smith et al., 2010). Intermittent high (1% in food) metformin reduced body weight and reduced food intake when mice were on metformin (Alfaras et al., 2017). Orally administered metformin also reduced mouse body weight by decreasing energy intake (Kim et al., 2013). The anorectic cytokine GDF15 (Emmerson et al., 2017;Mullican et al., 2017;Yang et al., 2017) is increased significantly by metformin in both normal chow and high fat diet groups (Day et al., 2019;Ouyang et al., 2020) and metformin reduced food intake, body mass, fasting insulin and glucose intolerance in mice consuming a high fat diet (Day et al., 2019;Matsui et al., 2010), in which the effects of metformin to reduce body weight were reversed by a GDF15 receptor-antagonist antibody (Coll et al., 2020).

Other treatments in mice
Besides the above "mainstream" pharmacologic treatments widely assumed to slow aging in mice, several other substances have also been reported to extend life span in mice. Combined statin and angiotensinconverting enzyme inhibitor treatment increased median but not maximum life span of male mice, but also decreased body weight, as did statin treatment alone (Spindler et al., 2016). Consumed calories were unchanged, suggesting that altered energy utilization was responsible for the weight loss (Spindler et al., 2016). Nordihydroguaiaretic acid (NDGA) extended the life span of flies and mice, in the latter in a narrow therapeutic window. Treatment was accompanied by weight loss and an increased incidence of tumors (Spindler et al., 2015;Strong et al., 2008). NDGA reduced weight previously (Miller et al., 2007), but some authors later insisted it was not . NDGA prevented weight gain in mice on a unhealthy diet (Chan et al., 2018) and caused weight loss in dyslipidemic rats (Zhang et al., 2016b). Life span extension with NDGA was replicated again by the NIA Interventions Testing Program, but weight data was, unlike for the other interventions reported at the same time, not shown Strong et al., 2016). In a rare case where an "antioxidant" was beneficial, N-acetylcysteine improved muscle function in a mouse model of Duchenne muscular dystrophy. But 2% in drinking water also reduced body weight gain both in the model mdx mice as well as in control C57Bl/10ScSn mice Table 1 Pharmacologic treatments extending life span in mice and their effects on body weight. Caloric (Weindruch et al., 1986) and protein restriction is shown for comparison.
A.M. Wolf (Pinniger et al., 2017), confirming previous results of weight loss and life span extension with 1% and 0.5% N-acetylcysteine (Flurkey et al., 2010). Chronic N-acetylcysteine treatment also reduced circulating IGF-1 by 20%-50%, depending on mouse strain (Ackert-Bicknell et al., 2004). The estrogen 17-α-estradiol extended male median lifespan (by 19%) when started late in life (16 and 20 months), but also caused a 20% weight drop (Harrison et al., 2021). Similar but smaller weight loss and life span extension was found at a lower dose of 17-α-estradiol (Strong et al., 2016). The α-glucosidase inhibitor acarbose reduces postprandial serum glucose and insulin concentrations and is used for the treatment of diabetes mellitus in humans (Chiasson et al., 2002). Acarbose also reduced weight, IGF-1 and insulin and extended life span, more in male than in female mice (Harrison et al., 2019. Mice eating a diet low in advanced glycation end products gain less weight than controls and live longer (Cai et al., 2007). Body weight reduction as the straightforward explanation for increased life span (as with the ACE inhibitor enalapril in rats), is rare (Santos et al., 2009).
For genetic interventions, reduced MYC (Hofmann et al., 2015), overexpressing FGF21 (Zhang et al., 2012) or the autophagosomal protein Atg5 (Pyo et al., 2013) and deletion of adenylyl cyclase type 5 (Yan et al., 2007) reduced weight and increased longevity. Liver-specific inactivation of the IGF-I gene decreased body weight and increased mean life span in mice, and the authors deserve recognition for openly stating that reduced body weight could be one mechanism for the prolonged life span (Svensson et al., 2011). Increasing FGF21 up to 17-fold using an adeno-associated virus-based gene therapy caused rapid weight loss and improved comorbidities of obesity compared to the control group consuming a high-fat died (Davidsohn et al., 2019). Positive effects in combinations of up to 3 genes were dependent on weight loss relative to the high-fat died consuming control mice. Gene therapies not resulting in weight loss were ineffective (Davidsohn et al., 2019). Table 1 summarizes studies that report both life span extension using pharmacological methods and body weight in mice. To limit the table size, genetic interventions, studies that did not report body weight or full life span data and studies not using mice were omitted. Calorie and protein restriction studies are included for comparison. Weight loss is ubiquitous, and the weight loss is generally larger in treatments with bigger effects on life span.

Obesity and life expectancy in non-standard species
Besides laboratory animals, not many species have unlimited access to food, but some kept by humans as pets seem to do. Even though access to food is normally controlled by their owners, some cats and dogs are able to eat much more than what is healthy for them. Excess body weight in middle-aged cats is associated with increased health risks and early mortality (Scarlett and Donoghue, 1998). Obese cats also have a shorter life span and decreased survival when grouped according to a body condition score, which corresponds well to the percentage of body fat (Teng et al., 2018). Even moderately overweight dogs have a higher risk of early morbidity (Laflamme, 2012). Genetic selection for food motivation in Labrador retrievers, which results in good trainability, seems to be partially responsible for the high frequency of obesity in these dogs (Raffan et al., 2016). Middle-aged Labrador retrievers given unlimited access to food weighed about 32 kg, of which 10 kg was body fat, and lived a median 11.2 years. Litter-paired Labrador retrievers restricted to 75% of what their ad libitum sibling ate the previous day weighed about 24 kg, of which only 4 kg was body fat (Lawler et al., 2008). Nine of 24, or 37.5%, of CR dogs remained alive by the time all ad lib dogs had died, and median survival was increased to 13 years. Despite eating 25% less, CR retrievers had only about 10% less lean body mass as adults, suggesting that this level of restriction is not stunting growth (Lawler et al., 2008). And while Labrador retrievers are seen as especially motivated by food, other breeds do not seem to be significantly less motivated by it (Raffan et al., 2015). The tendency to overeat has become a major problem in pet health care (German, 2006), and loss of health and life span due to excess calorie intake is not limited to rodents and humans.

Caloric restriction in primates
Starting in the late 1980s and early 1990s, two long term studies in rhesus monkeys conducted by the Wisconsin National Primate Research Center (WNPRC) and the National Institute on Aging (NIA) sought to settle once and for all whether caloric restriction is beneficial in longerlived species (Masoro, 2003). In both studies, CR monkeys were kept on 30 percent fewer calories for more than 20 years. Both provided evidence that CR reduced the incidence of age-related conditions, such as cancer, heart disease, and diabetes. But in what was meant to be two independent replications of the same hugely important experiment, the WNPRC found an extension of lifespan (Colman et al., 2009), while the NIA did not (Mattison et al., 2012b). There were, in hindsight, several significant differences between the two studies, including monkey breed, type of food, and most importantly, the control groups (Mattison et al., 2017). Lifespan extension was found only when the control group was given unlimited access to food. If food was allotted to approximate ad libitum intake based on age and bodyweight without overfeeding, which means the control group was "restricted" to 100% of a normal food intake, lifespan extension disappeared. This means that rhesus monkeys (who live about 40 years maximum) do not live significantly longer when their calorie intake is reduced from 100% down to 70% of the amount an average monkey would consume. This finding was reinterpreted as to that the control group in the NIA study was "effectively undergoing caloric restriction" (Colman et al., 2014). I would absolutely agree. One group was served a balanced meal every day, but never ever allowed to eat more for their whole life. The other had an all-you-can-eat buffet every day. It is not at all surprising that this makes a big difference, given that (at least in worms) a psychological component is responsible for almost half the life extending effect of CR (Libert et al., 2007). But is there a point in salvaging caloric restriction as an anti-aging intervention by insisting that the control group has to be sufficiently obese? Without doubt, control group monkeys at both NIA and WNPRC were much heavier than what would be best for their survival.

Caloric restriction in humans
Voluntary caloric restriction in humans increases life span by 8.4 years in men and 6.1 years in women when started at between 20 and 39 years of life (Grover et al., 2015). The caveat is that these improvements are only achieved against a control group of humans with a BMI above 35 (Grover et al., 2015). By the same reverse logic, voluntary CR (just a healthy food intake) increases health span by two to four times the increase in life span, for up to twenty years (Grover et al., 2015). Is this proof that caloric restriction slows aging in humans? Or is it not OK to use an obese control group? Compared against humans of a healthy weight, the life-shortening effects of obesity and overweight are striking. Overweight decreased life expectancy by 3 years and obesity by 6 and 7 years in men and women, respectively (Peeters et al., 2003). Years of life lost increase further with more severe obesity (Fontaine et al., 2003). Obesity affects so many hallmarks of aging that it may be apt to call it accelerated aging (Cutler, 1982;Salvestrini et al., 2019). Obesity accelerates the onset of cardiovascular disease (Silva et al., 2006), and epigenetic aging in liver (Horvath et al., 2014), thymus (Yang et al., 2009) and blood of middle aged, but notably not old humans (Nevalainen et al., 2017). CR-induced weight loss improves cardiovascular health also in healthy normal weight humans (Kraus et al., 2019). Weight loss alone improves conduit and resistance artery endothelial function in young and old overweight or obese adults (Pierce et al., 2008). Very large scale human studies have established that risk of death is lowest in humans with a body mass index around 25 (Adams et al., 2006;Berrington de Gonzalez et al., 2010;Di Angelantonio et al., 2016;Zheng et al., 2011). Whether slight overweight (a body mass index of between 25 and 30) is beneficial or detrimental relative to a lower body mass index seems to be controversial (see for example (Flegal et al., 2013)), but seems to protect against communicable disease, a significant cause of mortality in the elderly (Bhaskaran et al., 2018). The BMI associated with lowest mortality risk is higher in older individuals (Bhaskaran et al., 2018), when loss of body mass due to sarcopenia (age-related loss of muscle mass) and other age-related illnesses is called terminal weight loss (Masoro, 2011). Within the healthy range, a slight increase in BMI with age improves survival over staying skinny throughout life (Zheng et al., 2021). Nir Barzilai, even though advocating weight loss through metformin use (Kulkarni et al., 2020), actually finds centenarians "a chubby bunch" (Maxmen, 2012).
Effective CR is also almost impossible to achieve in humans (Kraus et al., 2019). Laboratory animals undergoing CR are "convinced" that food is scarce. Perception of food scarcity seems crucial for CR to work, as worms given the impression of abundant food by the smell of yeast lost almost half of the life extension provided by CR (Libert et al., 2007). Even if they would be able to adhere to CR (which they are almost certainly not (Kraus et al., 2019)), humans would still know that food is abundant and therefore loose the beneficial effects regulated by psychology. Dieting is therefore a particularly unattractive and ineffective intervention to increase lifespan in humans. To partially circumvent the unbearable constant hunger that dooms every effort to achieve laboratory animal levels of CR in humans, alternate day fasting has been promoted to mimic caloric restriction. Alternate day fasting reduced fat mass (particularly trunk fat) and depleted methionine on fasting days (Stekovic et al., 2019). Another fasting-mimicking diet improved almost all parameters in participants at risk for disease (i.e. obese patients), but far less in healthy individuals (Wei et al., 2017). Extremely obese patients often struggle to reduce their weight through changes in their diet or behavior. Bariatric surgery is a dangerous and drastic intervention to reduce weight in morbidly obese patients with no other viable option to improve their seriously compromised health (Buchwald et al., 2004). Compared to dietary interventions, loss of body weight is rapid and sustained, and resembles forced reduction in calorie intake in laboratory animals. As such, bariatric surgery can be viewed as a human equivalent of CR in rodents, and the improvement of many parameters of metabolic youth in patients after bariatric surgery underlines the similarities between CR in rodents and bariatric surgery in humans: mortality due to disease was decreased almost by half (Adams et al., 2007). Bariatric surgery reduced markers of metabolic syndrome (Rega-Kaun et al., 2020), senescence-associated secretory proteins, telomere oxidation and increased circulating leucocyte telomere length (Hohensinner et al., 2018). Drastic improvements not only in metabolic but also in many social factors like appearance, quality of life and social and economic opportunities (Buchwald et al., 2004) illustrate that surgically enforced weight loss constitutes a profound rejuvenation in several not only metabolic biomarkers of aging. But despite massive improvements, most patients still remain in the obese weight category and the doubling of suicide risk after gastric bypass surgery suggests that this intervention is often not a happy one (Adams et al., 2007).

Caloric restriction mimetics in humansrapamycin
Rapamycin derivatives are standard therapy for organ transplant recipients, some cancers, tuberous sclerosis complex (a genetic disease causing aberrant mTOR activation) as well as several other diseases. As a consequence, the side effects of long term rapamycin treatment are well known. Various mild or moderate infections as a side effect of immunosuppression are common, but life-threatening sepsis can also be observed (Trelinska et al., 2015). Wound-healing complications are common, as wound healing requires expansion of cells for tissue regrowth. Similar to mice, hyperglycemia, diabetes, hypercholesterolemia, hypertriglyceridemia, dyslipidemia and gonadal dysfunction are also seen in humans (Kaplan et al., 2014;MacDonald, 2001). This would be expected to promote atherosclerosis and these side effects are often seen as too severe to seriously consider rapamycin treatment in humans (Bitto et al., 2016). Rapamycin-induced insulin resistance was reported to be mediated by mTORC2 disruption (and therefore uncoupled from longevity effects) (Lamming et al., 2012), and more specific mTORC1 inhibitors are seen by some as a possible avenue to reduce some of these side effects (Mahoney et al., 2018;Schreiber et al., 2019). Unexpectedly, six weeks of treatment with the rapamycin analogue RAD001 was reported to improve adaptive immune function in elderly people (Mannick et al., 2014). However, an oral, selective mTORC1 inhibitor, RTB101, failed to meet the primary goal in a clinical trial in older patients, increasing rather than decreasing risk of respiratory illness (Kaeberlein, 2020), as would be expected from an immunosuppressant. The drug later disappeared from the pipeline of the company that acquired its developer, resTORbio. Like in mice, mTOR inhibition interferes in human spermatogenesis and causes infertility, which can be reversed by treatment withdrawal (Deutsch et al., 2007;Oliveira et al., 2017), even though not in all cases (Zuber et al., 2008).

Caloric restriction mimetics in humans -NAD supplementation
Raising NAD+ concentration is the most recent intervention to receive widespread interest as a potential treatment to increase life or health span. In humans, it is still too early to tell if raising NAD+ levels might have any long-term beneficial effects, whether through weight loss or otherwise. Nevertheless, short term studies are available, and increasing circulating NAD+ levels using various supplements seems to be save and well tolerated, with no significant effects on a variety of physiological parameters, including weight (Conze et al., 2019;Dollerup et al., 2018;Martens et al., 2018). Caution however, is appropriate: a combination of nicotinamide riboside (NR, a NAD+ precursor, 500 mg per day) and pterostilbene (PT, a resveratrol analogue and SIRT1 activator, 100 mg per day), increased total and LDL-cholesterol ("bad" cholesterol) in both normal and overweight individuals by a highly significant 0.48 mmol/L (p < 0.001) (Dellinger et al., 2017). Lowering LDL-cholesterol by 1.0 mmol/L reduces the risk of major vascular events (fatal and non-fatal myocardial infarctions, strokes, etc.) by approximately 20% (Cholesterol Treatment Trialists et al., 2012), so this purportedly safe and effective way to increase NAD+ levels (Dellinger et al., 2017) would be expected to increase major vascular events by about 10% at the above dose. As about one in four people die of heart disease in the US (Kochanek et al., 2011), long term consumption of this supplement would be expected to kill about 1 in 40 of long-term consumers at the above dose, and 1 in 80 at its recommended dose. Since the increase in LDL-cholesterol is probably attributable not to NR but to pterostilbene (Brenner and Boileau, 2019), this potential public health nightmare is avoidable and not tied to increasing NAD+ levels.

Caloric restriction mimetics in humansmetformin
Metformin has been in use for diabetes for several decades (Clarke and Duncan, 1968), so despite being a more recent addition to the catalogue of CR mimetics, its long history of human use puts it in the forefront of candidates for trials in humans. Billions of prescriptions have produced irrefutable evidence of its beneficial effects in type 2 diabetes patients, where it produces durable weight loss (Stumvoll et al., 1995;UKPDSGroup, 1998). The molecular mechanism of metformin action is still a matter of debate (Kalender et al., 2010), but is most often attributed to reduced hepatic glucose production caused by its inhibition of complex I of the mitochondrial respiratory-chain, which activates AMPK (Viollet et al., 2012;Zhou et al., 2001). The weight loss effects of metformin are well known (Lachin et al., 2007), and metformin is an effective weight-loss drug also in non-diabetic humans (Knowler et al., 2002;Lexis et al., 2015;Pedersen, 1965;Preiss et al., 2014;Seifarth et al., 2013). The percentage of weight loss was independent of age and body mass index. Heavier patients lost more weight and insulin-resistant patients lost more weight than insulin-sensitive patients (Seifarth et al., 2013). Decreased food intake with metformin has been reported (Lee and Morley, 1998;Malin and Kashyap, 2014), and recently, anti-diabetic effects of metformin and its ability to lower body weight have been attributed to suppressing food intake by elevating the anorectic peptide hormone GDF15 (Coll et al., 2020;Kleinert and Müller, 2019), as in mice.
Targeting Aging with Metformin (TAME) is a clinical trial to test whether metformin can delay the onset of several age-related diseases, including cancer and cardiovascular disease (Barzilai et al., 2016). Given that this trial is held in a population with a high prevalence of obesity, improvements in variety of biomarkers of aging, resembling the changes seen with caloric restriction in mice or fasting in humans, along with weight loss, are expected. The TAME study was developed to obtain a new FDA indication to target aging and to allow industry to justify the development of next-generation drugs to target aging (Barzilai, 2017). It is highly questionable whether opening the door to weight-loss drugs pretending to extend health span is a good idea. The gastrointestinal side effects of metformin (Flory et al., 2019) might well outweigh the benefit, even if there were any in healthy weight humans. On the other hand, the median BMI for middle aged humans in the U.S. is approaching 30 (Fryar et al., 2018), so most Americans would benefit from weight loss. The risk of developing a range of tumors is increased in type 2 diabetes (López-Suárez, 2019), and metformin treatment reduces this risk (Evans et al., 2005), especially for colon and pancreatic cancers (Currie et al., 2009). Whether through weight loss or other effects, metformin can prevent diabetes (Knowler et al., 2002) and might very well reduce cancer in a non-diabetic but overweight population as well (Renehan et al., 2008).

Discussion
A 2019 anniversary review of aging research lists seven interventions to increase health span and/or lifespan (Campisi et al., 2019). Five of them (CR, rapamycin, NAD precursors, sirtuin-activating compounds and metformin) might be little more than rodent diet aids. Exercise is the sixth, and certainly an intervention with almost no downside. Exercise undoubtedly improves health and survival (Nocon et al., 2008;Vina et al., 2012), but there seems to be little evidence that, once corrected for the associated lower body weight, it has any significant impact on the rate of aging (Holloszy, 2000). Lack of exercise contributes to chronic diseases (Booth et al., 2011), but there seems to be no enrichment of athletes or physical laborers in centenarians (Centenarians, 1997). Exercise had no effect on the fundamental rate of aging in mice (Garcia--Valles et al., 2013) or rats (Holloszy et al., 1985), despite lowering body weight. The beneficial effects of food restriction and exercise on survival are not additive or synergistic in rats (Holloszy, 1997). Most researchers justifiably roll their eyes upon the suggestion to prevent aging with more exercise.
Doubts about CR have certainly been raised previously. CR is questioned at length (though not dismissed) by Austad (Austad, 1997). According to Austad and Kristan (Austad and Kristan, 2003), several books seem to have speculated that the effects of CR on rodent aging might be artefacts of overfeeding under captive conditions (Cutler, 1982). But even though they show that laboratory mice are extremely obese compared to wild mice, which have 8-9 times less body fat than C57BL/6 J, but up to three times higher metabolic rates per g body weight (Austad and Kristan, 2003), they conclude that, because ad libitum laboratory mice eat no more than wild mice after correcting for body mass, CR restricts energy consumption beyond that of mice in nature (Austad and Kristan, 2003). As if a sedentary 200 kg person eating 4 pizzas is more calorically restricted than a 55 kg marathon runner eating one. Minority opinions like that the correct interpretation of caloric restriction experiments in mice may be that overfeeding reduces longevity (Hayflick, 2010;Le Bourg, 2018) have been dismissed with distractions about yeast and worms, and the fantastic sleight of hand of relabeling healthy nutrient intake as CR (Fontana et al., 2010).
Dismissing CR mimetics as mere diet aids does not mean that they have no effect besides weight loss. Testing the above treatments against an at least moderately calorie-restricted control group might clarify whether any of the above compounds can delay aging in healthy subjects through pathways unrelated to growth retardation and preventing preventable weight gain. Interestingly, the 2020 cohort of the NIA Interventions Testing Program includes 2,4-dinitrophenol, a mitochondrial uncoupler and famous weight loss agent with significant acute toxicity and risk of death (Grundlingh et al., 2011). What should be clear though is that weight loss-mediated "reversal of aging" (when measured with biomarkers), however lifesaving (in case of bariatric surgery) or healthy (in case of dieting, fasting, or other weight loss in overweight individuals), is ignorant of the mechanism of aging in healthy weight organisms, and should not be expected to inform us about how to prevent or delay healthy human aging.
Why is CR so well accepted as an age-retarding intervention? The beauty of CR and most of its appeal as a fundamental method to retard aging lies in its ability to delay most phenotypes of aging in such a way that it is appropriate to state that CR delays aging in general (Masoro, 1985(Masoro, , 1988(Masoro, , 2005van den Boogaard et al., 2020;Yu et al., 1985), not only one or a few specific aspects of aging. CR really does delay aging against an obese control group. Why is obesity so bad? I think looking at why the alleviation of obesity, or the endocrine retardation of growth (Bartke, 2019;Bartke et al., 2013) slows aging tells us something about the fundamental mechanism of aging. The underlying aspect both obesity and growth have in common is that they accelerate the pace of cell division: CR decreases proliferation, while IGF-1 reverses it back to normal (Dunn et al., 1997). Disrupted GH/IGH-1 signaling is associated with small size and increased longevity in many vertebrate species (Berryman et al., 2008). Many long-lived dwarf mice are relatively obese (Berryman et al., 2014), and genetically obese mice on a CR diet, despite having almost 50% body fat, live longer than normal, less obese mice given unlimited access to food, showing that fat deposits by themselves are not an important determinant of longevity (Harrison et al., 1984). Elevated IGF-1 is a risk factor for several human cancers (Sonntag et al., 1999), as is obesity (Renehan et al., 2008). Effects of CR are generally the opposite of IGF-1, with dramatic reductions in many types of murine cancers (Speakman and Mitchell, 2011), reduced pathologies and increased life span (Sonntag et al., 1999). Severe stunting, low IGF-1, but increased longevity with protein restriction despite compensatory overeating (Miller et al., 2005;Solon-Biet et al., 2014) supports anabolic growth and proliferation, but not calorie intake or adipose tissue mass, as the key determinants of longevity, even in humans (Levine et al., 2014). The obvious explanation why lowered intake of particular proteins and specific amino acids rather than overall calories regulate aging (Fontana and Partridge, 2015) is because these nutrients regulate proliferative signaling through mTOR (Laplante and Sabatini, 2012). Cell proliferation is necessary to build up body mass during increased nutrient availability, and the resulting increase in fitness and resilience is key to achieving higher reproductive success. In the following article, I argue that somatic mutations accumulating mostly during cell division are the fundamental driver of aging, and that the pace and the genetic precision with which cell division is accomplished mostly determines the rate of aging in higher organisms. Malignant growth is a fundamental problem for large and long-lived mammals, and tumor suppression mechanisms had to evolve together with the preservation of genomic integrity, an observation known as Peto's paradox. The tumor suppression theory of aging (Wolf, 2021) (originally part of this article (Wolf, 2021a) and briefly mentioned in (Wolf, 2021b)) proposes that only oncogenic mutations driving clonal expansion are relevant for aging, and that most aging phenotypes are the consequence of tumor-suppressive replicative exhaustion and cell loss due to apoptosis, differentiation and senescence.

Conclusion
Obesity is common and predicted to increase further with economic development (Ward et al., 2019). Obesity contributes to hypertension and related cardiovascular, renal and metabolic disorders (Hall et al., 2015). Weight loss drugs are very attractive "anti-aging" treatments in an overweight population, and several CR mimetics might be approved as drugs to treat obesity rather than aging. Manipulation of nutrient absorption (Holman et al., 2017;Yang et al., 2014) and satiety (Frias et al., 2018;Husain et al., 2019;O'Neil et al., 2018) are also attractive avenues to achieve rejuvenating weight loss (Kushner, 2014). Finding out why even slight overweight seems to increase risk for so many apparently unrelated ailments (Gallagher and LeRoith, 2015) is very valuable. Increased proliferation might explain accelerated aging, but is probably only one of many factors leading to increased mortality in obese patients. Progeria model mice show impressive extension of life span under caloric restriction (Vermeij et al., 2016), suggesting that in the unusual condition of increased mutation accumulation per cell division, curbing replication as far as possible with a diet that might be stunting for normal children might be beneficial in children with progeria. In such patients (and possibly even in normal aging), controlled suppression of proliferation through, for example, inhibition of mTOR, might yield benefits that outweigh the costs.
What is clear however is that excessive growth suppression has overwhelmingly negative consequences on fitness, and is thoroughly unattractive as a candidate pathway for life extension in otherwise healthy individuals (Churgin et al., 2017;Guevara-Aguirre et al., 2011;Hur et al., 2020;Jenkins et al., 2004;Podshivalova et al., 2017). Rapamycin suppresses crucial cell replication and, for this reason, has probably the harshest side effects among the CR mimetics (Bitto et al., 2016;Kaplan et al., 2014). On the other hand, enhancing growth usually increases fitness. Endocrine messengers like growth hormone (Rudman et al., 1990), young blood (Conboy et al., 2013;Rando and Chang, 2012;Villeda et al., 2014;Wyss-Coray, 2016) or anabolic steroids (Franke and Berendonk, 1997) do so and might rejuvenate epigenetic (Fahy et al., 2019;Mendelsohn and Larrick, 2019), and other (Srinivas-Shankar et al., 2010) biomarkers of aging as well. But like obesity, they will, through the stimulation of cell division, probably accelerate aging rather than delay it (Anisimov and Bartke, 2013;Bartke, 2019;Bartke et al., 2013), illustrating the double edged nature of anabolic stimulation (Ratajczak et al., 2017). Nevertheless, age-appropriate nutrition in old adult humans is very different from what one would conclude when looking at longevity in mice (Levine et al., 2014). Suppressing proliferation through pharmacology or diet might slow aging, but in close to optimal weight humans, potentially positive aspects of doing so seem currently not separable from a heavy loss of fitness and quality of life. Weight loss is not known to be beneficial and therefore not recommended to normal weight humans (Initiative, 1998).
The lack of success and current sorry state of aging research is due to the repeated failure to see caloric restriction as what it is: nothing more than the prevention of obesity. Especially the failure to correctly interpret and accept the outcome of the monkey CR studies (Mattison et al., 2017) is striking. Of course CR improves health and survival of rhesus monkeys (Mattison et al., 2017) if you mean >obese< rhesus monkeys. Of course, CR mechanisms are likely translatable to human health (Mattison et al., 2017) if you mean >obese< humans. Early 20th century anti-aging interventions are now ridiculed as quackery, but at that time were state of the art and endorsed by many leading scientists (Hansson et al., 2020;Kahn, 2005;Rubin, 1923;Sengoopta, 2003). Maybe we don't have to age (Sinclair and LaPlante, 2019), but mistaking the manipulation of obesity for the successful deceleration of aging has produced a fallacious optimism (Campisi et al., 2019;de Cabo et al., 2014;López-Otín et al., 2016;Madeo et al., 2019b;Partridge et al., 2020) and commercial entities selling the hope that, as a consequence, pharmacologic control of human aging is within reach (De Magalhães et al., 2017) or already there and we just have to open our mouths and wallets (Dellinger et al., 2017). But slower aging is a highly valuable trait (Gurven et al., 2006;Lahdenperä et al., 2004;Shanley et al., 2007) and it is unlikely that something that requires little investment of energy like increasing a metabolite or blocking a protein, can improve life span by a measurable amount (Kirkwood, 2005). Viewing aging not as inevitable but as a treatable condition is to be welcomed, but while investors salivate over the projected sales figures of a treatment for a disease that affects 100% of the population, progress in aging research is prevented by the continued investigation of rodent weight loss compounds. The aging process might seem much more malleable than we used to think (Kirkwood, 2017), but, at least in my opinion, remains as malleable as a block of tungsten carbide. The tumor suppression theory of aging described in the following article (Wolf, 2021), if correct, would conclude that slowing aging is the same as preventing cancer. Both is, for now, essentially impossible.

Author contributions
A.M.W. wrote the paper.

Declaration of Competing Interest
The author declares no competing interests.

Acknowledgements
Support was provided by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (16K01736 and 19K11560 to Alexander M. Wolf). Funding sources were not involved in study design or collection, analysis and interpretation of data. Alexander M. Wolf thanks the Government of Japan for 2 years of parental leave, during which the concepts in the paper were developed.

Appendix A. Supplementary data
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.mad.2021.111584.