17‐a‐estradiol late in life extends lifespan in aging UM‐HET3 male mice; nicotinamide riboside and three other drugs do not affect lifespan in either sex

Abstract In genetically heterogeneous mice produced by the CByB6F1 x C3D2F1 cross, the “non‐feminizing” estrogen, 17‐α‐estradiol (17aE2), extended median male lifespan by 19% (p < 0.0001, log‐rank test) and 11% (p = 0.007) when fed at 14.4 ppm starting at 16 and 20 months, respectively. 90th percentile lifespans were extended 7% (p = 0.004, Wang–Allison test) and 5% (p = 0.17). Body weights were reduced about 20% after starting the 17aE2 diets. Four other interventions were tested in males and females: nicotinamide riboside, candesartan cilexetil, geranylgeranylacetone, and MIF098. Despite some data suggesting that nicotinamide riboside would be effective, neither it nor the other three increased lifespans significantly at the doses tested. The 17aE2 results confirm and extend our original reports, with very similar results when started at 16 months compared with mice started at 10 months of age in a prior study. The consistently large lifespan benefit in males, even when treatment is started late in life, may provide information on sex‐specific aspects of aging.


| INTRODUC TI ON
Each year, the Interventions Testing Program (ITP; http://www.nia. nih.gov/resea rch/dab/inter venti ons-testi ng-progr am-itp) tests the effects of a variety of compounds on lifespan in a genetically heterogeneous mouse model (UM-HET3), the first-generation offspring of the CByB6F1 x C3D2F1 cross, to produce a diverse heterogeneous population that is reproducible across time and place (Roderick, 1963). ITP studies are conducted simultaneously at The University of Texas (UT), University of Michigan (UM), and The Jackson Laboratory (TJL) in Bar Harbor, ME. Details of the ITP design have been published (Harrison et al., 2009;Harrison et al., 2014;Harrison et al., 2019;Macchiarini et al., 2020;Miller et al., 2011Miller et al., , 2014Miller, Harrison, et al., 2020;Strong et al., 2008Strong et al., , 2016Strong et al., , 2020. Interventions that are found, in an initial experiment, to increase lifespan are then subsequently tested for physiological and pathological changes with age. The data in this paper come from the initial survival cohorts, and thus focus on lifespans, which integrate all biological effects that may lead to death. The spectrum of lethal pathology in UM-HET3 mice has been tabulated by Lipman et al. (2004), using groups of 136 and 208 virgin females, 268 multiparous females, plus 117 and 157 virgin males.
The interventions for the present study were chosen for the following reasons: (a) 17α-estradiol (17aE2) is a relatively "non-feminizing" estrogen which shows reduced activation of classical estrogen receptors compared with 17β-estradiol (Anstead et al., 1997). Harrison et al. (2014) reported that in UM-HET3 mice fed 4.8 mg 17aE2/ kg (4.8 ppm) diet from 10 months of age, median male lifespans increased 12% (p = 0.0012, pooled across the three sites), while 17aE2 did not alter female lifespan. Strong et al. (2016) showed that using a threefold higher dose (14.4 ppm) from 10 months of age, pooled median male lifespans increased 19% (p < 0.001); the 90% lifespan increased 12%, but females still did not benefit. Thus, only males were tested in the present study. To determine whether 17aE2 treatment is effective when initiated in older mice, males were treated beginning at 16 or 20 months of age, choosing middle age, and early old age before many natural deaths.
(b) Nicotinamide riboside (NR) is a precursor of nicotinamide adenine dinucleotide (NAD) via the cell's salvage pathway (Trammell, Schmidt, et al. (2016)). Total NAD levels decline with age, in a wide range of species. Importantly, increasing NAD levels benefit a wide variety of tissues in species including mice and human beings. Rajman et al. (2018), for example, suggest that NAD+boosters may "..delay aging and age-related physical decline." Zhang et al. (2016) reported that NR delays senescence of neural SCs and melanocyte SCs and increases mouse life span, even when given in old age (5% increase at 20 months of age). Trammell, Schmidt, et al. (2016) reported that in mice and humans NR is bioactive when given by mouth, unlike most other nicotinamide derivatives. In 2016b, they reported that NR improved liver function and protected against diabetic neuropathy. When fed to C57BL/6 J mice from 10 weeks of age, NR protects against highfat diet (HFD)-induced obesity and promotes oxidative metabolism by increasing the NAD+/NADH ratio in muscle, liver, and brown adipose tissue (Canto´et al., 2012). Ryu et al. (2016) found that increasing NAD+stores with NR supplementation improved muscle function and alleviated heart defects in a mouse model of muscular dystrophy. Mitchell et al. (2018) reported that an NR metabolite, nicotinamide, did not increase lifespan when started at 12 months in C57BL/6 J mice but improved some health outcome measures.
Due to its benefits in a variety of diseases, and reports of benefits in mouse lifespans, NR treatment was proposed to increase lifespan in UM-HET3 mice. c) Candesartan cilexetil (CC) is an angiotensin-receptor blocker, which lowers blood pressure (Ikeda et al., 2006) and improves cardiovascular function and insulin sensitivity in obese, hypertensive patients (Grassi et al., 2003). Importantly, angiotensin-receptor knockout increases lifespan of mice (Benigni et al. 2009). Because CC is effective against age-related diseases, and sensitizes the body to insulin, and because the angiotensin-receptor knockout increases lifespan of mice, treatment with CC was hypothesized to increase lifespan.
(d) To maintain good quality protein in the body, heat shock proteins (HSPs) are vital. Geranylgeranylacetone (GGA) induces heat shock protein (Hsp70) in mammalian tissues and promotes insulin sensitivity in old mice (Silverstein et al., 2015), while van Marion et al. (2020) showed that it increases HSP expression in atrial tissue after heart surgery. Pride et al. (2015) showed that long-lived species, compared with related short-lived species within the same order, have elevated HSP levels in conjunction with better proteostasis. To test whether treatment with an established HSP inducer neither it nor the other three increased lifespans significantly at the doses tested. The 17aE2 results confirm and extend our original reports, with very similar results when started at 16 months compared with mice started at 10 months of age in a prior study.
The consistently large lifespan benefit in males, even when treatment is started late in life, may provide information on sex-specific aspects of aging.

K E Y W O R D S
17α-estradiol, candesartan cilexetil, geranylgeranylacetone, heterogeneous mice, lifespan, macrophage migration inhibitory factor, nicotinamide riboside can increase lifespan in a mammalian model, UM-HET3 mice were treated with GGA.
This may include the chronic inflammation that increases with age, as suggested by the finding of Harper et al. (2010) that MIFknockout mice live significantly longer than controls. Because it is orally bioavailable and shows MIF inhibitory activity in mouse models of hyperoxic lung injury, as well as in other diseases (Sauler et al., 2015), treatment with MIF098 was proposed to increase lifespan by decreasing chronic inflammation and disease (Poulsen et al., 2019).
Our new data show that Nicotinamide riboside (NR) failed to increase lifespan. Only 17aE2 increased lifespan, and benefits in males occurred even when the drug was not fed until late middle or early old age (16 and 20 months of age, respectively). The range of ages for which treatment is effective suggests that benefits from 17aE2 do not depend on effects earlier in life, such as growth alteration.
Interventions that are effective when started at a late age have considerable translational potential.

| RE SULTS
Compared to controls, UM-HET3 male mice fed a diet with 14.4 ppm 17aE2 starting at 16 or 20 months of age had, respectively, 19% (p < 0.0001, log-rank test) and 11% (p = 0.007) increases in median lifespans (data pooled across three sites). Lifespans at the 90th percentile were significantly increased 7% when treatment was started at 16 months (p = 0.004, Wang-Allison statistic); however, when treatment was started at 20 months, the 5% increase in the 90th percentile lifespan was not significant (p = 0.17) (Table 1; Figure 1).
These data hint that benefits from 17aE2 diminish when treatment is started at 20 rather than at 16 months; however, when comparing the median lifespan results between the two treatment groups, no significant difference was detected (p = 0.24, log-rank test).
Mice were weighed every 6 months beginning at 6 months of age. Prior to treatment, at 6 and 12 months of age, the mean weight GGA mice were fed 600 ppm geranylgeranylacetone starting at 9 months of age.
N = number of mice tested; data were pooled, with about 1/3 of the mice from each testing site; see Table 2 caption for details.
Under Median lifespan: Days-median ages; % change calculated with respect to controls. p-value = probability that lifespans are the same as the controls using two-tailed log-rank test on pooled data stratified by sites; "removed" mice were included as censored (see Experimental Procedures).
Under lifespan at 90th percentile: Days = age at 90th percentile, % change from control.
Wang-Allison p-value = probability that the proportion of live mice is the same in treated as in the control group at the 90th percentile age, evaluated by the procedure of Wang et al. (2004).

TA B L E 1 Effects of interventions on
lifespan; data pooled from the 3 sites of each group of mice designated for treatment was comparable to the weight of the group designated as the untreated control group.
Weights of male mice started on 17aE2 at 16 months decreased about 20% by 18 months, and weights of mice started on 17aE2 at 20 months decreased about 20% by 24 months ( Figure 2B). At 24 months, the weights of both of the 17aE2 treatment groups were similar, and both were about 8.5 g lighter than controls (p < 0.001 in both cases).
Site-specific effects on lifespans were also not significant, although the site-specific data sets had far less statistical power (Table 2). NR led to a marginally significant increase in female lifespan at one site, but this was balanced by a marginally significant decline in female lifespan at another (as well as a marginally significant decline in lifespans of males at a single site) ( Table 2). The ITP primary endpoint is always the pooled data set, and NR had no significant effect, positive or negative, in the pooled data sets for either sex. Figure 2 gives effects of these agents on body weights. In females, weights of mice fed GGA were similar to weights in controls, while females fed NR, CC, and MIF098 were about 5 g lighter than controls. In males, weights of mice fed NR, CC, GG, and MIF098 were similar to weights in controls. In a separate study, NR and its metabolites were tested in liver, brain cortex, plasma, gastrocnemius muscle, heart, kidney, iWAT (inguinal white adipose tissue), gWAT (perigonadal white adipose tissue), and leg muscle from 6-to 10-month-old UM-HET3 mice fed control or 1,000 ppm NR diet for 6-6.5 weeks, with 3 males and 4 females per group. The most interesting positive finding was lower levels (p = 0.004) of nicotinamide in cortex (though not liver) of NRtreated mice. NR-treated mice also had lower NR levels in cortex (but not in liver), compared with controls (p = 0.03). The ratio of NAD to NADH in the cortex was lower in NR-treated males (p = 0.012).
These results are preliminary, because numbers of mice tested were low, and the diets were only fed to young mice for 6 weeks.  Figure 1) and site-specific data (  Table 1. Days = median survival in days, % change from control, and the log-rank P that the group differs from the controls (Cont_16). The rightmost column shows the average (mean) of the changes in median lifespan across the three sites.

F I G U R E 3 Lifespan curves of controls, and the 4 interventions without effects/.
Panel A shows females, and Panel B shows males. Data are from the same mice whose lifespans are shown in Table 1 sex hormones regulate lifespan. In UM-HET3 mice, diet restriction (DR) increases male and female lifespans to a similar degree (Flurkey et al., 2010). This is not the case for 17aE2. Both male and female mice given 14.4 ppm 17aE2 diets starting at 10 months of age were 14%-18% lighter than controls at 12, 18, and 24 months yet lifespans only increased in males (Strong et al., 2016, Figure 1, Figure S1), implying that the male-specific effects are not mediated by DR.
Rapamycin, like 17aE2, increases lifespan when treatment is initiated at 20 months of age; however, unlike 17aE2, rapamycin benefits both male and female mice (Harrison et al., 2009Miller et al., 2011Miller et al., , 2014. Thus, the male-specific lifespan benefit of 17aE2 contrasts with other treatments that produce comparable benefits in both sexes, making this model valuable for investigation of sexually dimorphic, and potentially novel, mechanisms of aging. found changes in liver amino acids, and urea cycling, in untargeted metabolomic analyses; these changes also were specific to intact males. If testicular hormones play an important role in lifespan extension when 17aE2 is given at 10-20 months of age, perhaps the reduction in benefits observed between 16 and 20 months is partly caused by a decline in androgen with age in male mice (Coquelin & Desjardins, 1982). Only males (Garratt et al., 2018) had elevated levels of two conjugated estriols when given 17aE2. This suggests that males, but not females, metabolize 17aE2 into one or more estriol derivatives and that the sex-specific beneficial effects may be due to an estriol derivative rather than to 17aE2 per se. The ITP is testing this hypothesis by comparing lifespans in male and female mice given diets containing estriol. Strong et al. (2016) noted that 17aE2 at the standard 14.4 ppm dose reduced body weights in females and altered uterine weights in ovariectomized mice. Thus, 17aE2 is partially active in females, so the absence of an effect on female lifespan is not because bioavailability is entirely absent. These observations support the idea that testicular hormones play an important role in lifespan extension by 17aE2. Interestingly, the wide range of cancers that cause most deaths in UM-HET3 mice Lipman et al., 2004;Miller & Chrisp, 2002)  in 20-month-old male C57BL/6 mice. The 17aE2 had no effect on protein-to-DNA synthesis rates compared with controls in any tissue, while 18% DR produced the large changes in protein-to-DNA synthesis rates expected with lifelong 40% DR. Using RNA-seq, Tyshkovskiy et al. (2019) found feminizing effects with most of the anti-aging candidate drugs they tested, increasing RNAs characteristic of females and decreasing RNAs characteristic of males.
Interestingly, the exception was 17aE2. It might be informative to feed 17aE2 to young or middle-aged mice and see how quickly the RNA profile is modified to resemble that seen in mice given 17aE2 for most of their lives.
None of the other four drugs tested in the 2015 cohort, that is, NR, CC, GGA, and MIF098, led to lifespan improvement in either sex. As reviewed in the introduction, in many species, total NAD levels decline with age, increasing NAD levels benefits many physiological systems in mice and men, and a small increase in mouse lifespan was reported. Our finding that NR has no effect on lifespan in the genetically variable mice that best model the human population is thus a surprise. In fact, the two points of most interest in this paper are that 17aE2 is effective when given later in life, and that NR has no effect on lifespan.
Each drug was detectable in food pellets, and metabolites of NR and CC were detected in plasma of treated mice. While it is possible that one or more of these drugs might have led to health benefits if used at a different dose, or in another stock of mice, or if started or stopped at a different age, the most plausible interpretation is that none of the drugs slows aging or prevents disease in a genetically heterogeneous mouse population.
Positive results are important, as showing here that 17aE2 is beneficial when started later in life. But negative results are also important, as showing here that NR failed to increase lifespan.

| Mouse production, maintenance, and estimation of lifespan
UM-HET3 mice were produced at each of the three test sites as previously described (Harrison et al., 2009;Miller et al., 2011;Strong et al., 2008), where environmental conditions are presented in detail. The dams of the test mice were CByB6F1/J, TJL stock #100009 (dams, BALB/cByJ; sires, C57BL/6 J). The sires of the test mice were C3D2F1/J, TJL stock #100004 (dams, C3H/HeJ; sires, DBA/2 J). In each site, breeding mice were fed LabDiet ® 5008 mouse chow (PMI Nutritional International, Bentwood, MO). As soon as mice were weaned, they were fed LabDiet ® 5LG6 distributed to the test sites from the same batch. Males were initially housed 3 per cage, while females were housed 4 per cage; numbers per cage declined as mice died. Figure 4 shows that lifespans for both female and male controls were similar at all 3 sites.
Details of the methods used for health monitoring were provided previously (Harrison et al., 2009;Miller et al., 2011;Strong et al., 2008). In brief, each of the three colonies was evaluated four to twelve times each year for infectious agents. All such surveillance tests were negative for pathogens at all three sites throughout the entire study period.

| Removal of mice from the longevity population
Mice were removed from the study because of fighting or accidental death (e.g., during chip implantation) or chip failure, or because they were used for another experimental purpose. For log-rank survival analyses, all such mice were "censored," that is, treated as alive at the date of their removal from the protocol and lost to follow-up thereafter. These mice were not included in calculations of median longevity. For the mice in Table 1, % males removed: control, 9%; 17aE2-16, 3%; 17aE2-20, 10%; NR, 7%; CC, 12%; GGA, 13%; and MIF098, 6%. Most of these males were removed because of fighting. Fewer females were removed: control, 0.3%; NR, 1.5%; CC, 3%; GGA, 0; and MIF098, 0.
Thus, mice removed from the longevity population were included in log-rank test calculations. The question as to whether removal is random does not apply, since no mice were removed for other purposes.
The statistical approach used is the usual Kaplan-Meier calculation, based on the log-rank test, in which mice that do not die a natural death (typically because of removal for humane reasons) are treated as known to be alive at the date of removal, with unknown date of death.

| Estimation of age at death (lifespan)
At UM and UT, mice were examined daily for signs of ill health from the time they were set up in the experiment. At TJL, mice over 500 days of age were examined daily and twice a day once they were marked as ill. Mice were euthanized for humane reasons if so severely moribund that they were considered, by an experienced technician, unlikely to survive for more than an additional 48 hours.
TJL's definitive endpoint criterion is the non-responsiveness of a mouse to being touched, and which is usually accompanied by one or more of the following: slow respiration, feeling cold to the touch, a hunched-up appearance with matted fur, signs of sudden weight loss, failure to eat and drink, prominent appearing ribs and spine, and sunken hips. The age at which a moribund mouse was euthanized was taken as the best available estimate of its natural lifespan.
Mice found dead were also noted at each daily inspection, giving the lifespan. The 2015 cohort of UM-HET3 mice was also used to test the effects of canagliflozin, an inhibitor of SGLT2 . The raw data used here are available at the Mouse Phenome

| Statistical methods
Database https://pheno me.jax.org/proje cts/ITP1, The Jackson Laboratory, Bar Harbor, ME. This is supported by the Mouse Phenome Project and is a public data repository that provides the authoritative source for the raw and summary data from the ITP, along with visualizations for exploration of lifespan and related phenotype data.

ACK N OWLED G M ENTS
This work was funded in part by NIA grants AG022308 (

CO N FLI C T O F I NTE R E S T
None.