The rate of obesity has been rising sharply in industrialized countries, especially in the United States (Hales et al., 2020), and poses a serious threat to already overburdened healthcare systems (Finkelstein et al., 2009; Cawley and Meyerhoefer, 2012). Obese individuals are at higher risk for the development of cardiovascular disease, hypertension, non-insulin-dependent diabetes, certain cancers, and metabolic syndrome (Keller and Lemberg, 2003; Wolin et al., 2010; Kachur et al., 2017). Importantly, the greatest risk factor for most of these diseases is old age (Kaeberlein, 2013), and obese older individuals are at even greater risk than either group alone (Samper-Ternent and Al Snih, 2012). It is estimated that 38% percent of the population of the United States over 60 years of age is obese, and another 40% of obese people aged 40–60 will enter geriatric age in the next 25 years (Flegal et al., 2016), defining a population at high risk for multiple comorbidities (Zhang et al., 2019).

The field of geroscience seeks to understand mechanisms that connect biological aging with age-related pathologies (Kennedy et al., 2014; Sierra, 2016). Several therapeutic approaches that target the molecular hallmarks of biological aging have been shown to increase lifespan and delay age-related pathologies in mice (Kaeberlein, 2017). Among these, inhibition of the mechanistic target of rapamycin (mTOR) has proven particularly robust and reproducible (Johnson et al., 2013; Kennedy and Lamming, 2016). Genetic knockdown of mTOR complex 1 (mTORC1) components or pharmacological inhibition of mTORC1 with rapamycin is sufficient to increase lifespan by up to 30% and delay or reverse functional declines of aging across many tissues and organs in mice including brain, heart (Wilkinson et al., 2012; Quarles et al., 2020), liver (Wilkinson et al., 2012), intestine (Yilmaz et al., 2012), immune system (Chen et al., 2009), tendon (Wilkinson et al., 2012; Zaseck et al., 2016), bone (Ma et al., 2018; An et al., 2020), skeletal muscle (Bitto et al., 2016; Joseph et al., 2019; Tang et al., 2019), auditory system (Altschuler et al., 2021), adipose (Mau et al., 2020), ovaries and endometrium (Wilkinson et al., 2012; Dou et al., 2017; Garcia et al., 2019), adrenal gland (Wilkinson et al., 2012), and the oral cavity (An et al., 2017; An et al., 2020).

C/EBP-β is a basic leucine zipper (bZIP) domain transcription factor highly expressed in liver, kidney, intestine, adipose, pancreatic islets, and innate immune cells. It regulates the expression of several genes involved in development, immune function, regeneration, differentiation, and metabolism (Guo et al., 2015; Pulido-Salgado et al., 2015). C/EBP-β is an intronless gene but it generates three isoforms with distinct yet not fully characterized functions, termed Liver Activating Protein* (LAP*), Liver Activating Protein (LAP), and Liver Inhibitory Protein (LIP) (Calkhoven et al., 2000). LIP, the shortest isoform, contains a bZIP DNA-binding domain but lacks a transactivation domain to recruit the RNA polymerase II complex. LAP and LAP* encode the full-length factor and contain a transactivation domain in addition to the bZIP and are considered functionally identical. Thus, LIP is considered a transcription inhibitor, while LAP and LAP* (hereafter referred to collectively as LAP) are regarded as transcription activators with essentially the same function.

The relative abundance of LAP and the ratio between LAP and LIP are regulated by mTORC1 through differential translation of the isoforms. The full length C/EBP-β mRNA contains a short inhibitory upstream open reading frame (uORF) prior to the LAP initiation site (Calkhoven et al., 2000). When mTORC1 activity is high, translation of the uORF is more likely, which causes the ribosome to skip over the LAP start site and favors translation of LIP. When mTORC1 activity is low, such as under nutrient depletion or in the presence of rapamycin, uORF translation is reduced, and translation of the full-length LAP is favored. Consistent with this model, pro-longevity interventions such as calorie restriction or rapamycin treatment increase LAP and reduce LIP expression in the liver of mice. Interestingly, genetic knockout of the uORF in mice (Cebpb^ΔuORF) ablates expression of LIP, increases female lifespan (Müller et al., 2018), and improves several metabolic parameters including weight, adiposity, adipocyte size, energy expenditure, and adiponectin levels (Zidek et al., 2015).

The Nucleoside-analog Reverse Transcriptase Inhibitor (NRTI) adefovir dipivoxil has also been identified as a modulator of the LAP/LIP in a in vitro screening for compounds that reduce LIP translation (Zaini et al., 2017).

Since obesity and aging are overlapping risk factors for multiple pathologies, and rates of obesity increase with age (at least until about age 65), we set out to test whether the geroprotective intervention rapamycin could impact diet-induced obesity in adult mice. Here we report that mice treated with rapamycin are strongly protected against weight gain and fat accumulation associated with an obesogenic diet. Rapamycin-treated mice also show activation of C/EBP-β in the liver. Further, a small-molecule activator of C/EBP-β, adefovir dipivoxil, is sufficient to phenocopy the anti-obesogenic effects of rapamycin and to up-regulate enzymes involved in fatty acid metabolism in the liver. These observations support the model that mTORC1-inhibition prevents diet induced obesity by activation of C/EBP-β and identify adefovir dipivoxil as a potential therapeutic for obesity and aging.

In this study, we provide evidence that inhibiting mTORC1 protects against diet-induced obesity by activating fatty acid catabolism through C/EBP-β. Furthermore, we show that adefovir dipivoxil is able to activate C/EBP-β in vivo and function as a potent anti-obesogenic agent. The effects of both rapamycin and adefovir dipivoxil cannot be attributed to reduced food consumption, but instead appear to be mediated, at least in part, through increased fatty acid oxidation. Taken together with prior studies, these observations suggest a model (Figure 8) whereby many of the metabolic effects of mTORC1 inhibition are likely mediated through activation of C/EBP-β, and that both mTOR and C/EBP-β are druggable targets to induce positive effects in the context of diet-induced obesity.

This work builds on prior studies which have explored targeting the mTOR pathway to prevent or reduce obesity in rodents, with mixed results. One study found that 14 ppm encapsulated rapamycin in the food exacerbated the effects of an obesogenic high-fat diet in 2 month-old male mice (Liu et al., 2014), while the same dose and route of administration showed beneficial effects in 2 month-old-male mice on a high-fat, high-sucrose diet (den Hartigh et al., 2018). This suggests that diet composition may impact the effects of mTORC1 inhibition on obesity. The diet used in this study was not high in sucrose and, interestingly, reduced mTORC1 signaling relative to the normal diet control (Supplementary Figure S2). Thus, it is not the case that carbohydrate composition alone accounts for these differences. Route of administration, bioavailability, and age likely all play significant roles in the anti-obesogenic effects of rapamycin.

Sex also appears to influence the efficacy of rapamycin as an anti-obesogenic treatment in mice. For example, 2 mg/kg/week intraperitoneal rapamycin was sufficient to decrease body weight in obese male mice (Chang et al., 2009), though oral gavage with 1 mg/ml rapamycin was only effective in female but not male 9 month-old mice on the same high fat diet used in our study. Though the small sample size likely prevented us from detecting a significant difference in weight and fat mass change in rapamycin-treated male mice on a HFD, we observed a clear trend (Supplementary Figure S1). There are numerous reports that female mice are more sensitive to rapamycin treatment and show higher blood levels of the drug at a given delivery concentrations (Harrison et al., 2009; Miller et al., 2011; Halloran et al., 2012; Wilkinson et al., 2012; Fok et al., 2014; Miller et al., 2014; Zhang et al., 2014; Fischer et al., 2015). Female mice also show consistently greater lifespan extension from rapamycin treatment at lower concentrations of the drug, but at higher concentrations male mice catch up to female mice and eventually show larger lifespan extension at the highest concentrations tested to date (Miller et al., 2014; Bitto et al., 2016). Thus, we speculate that rapamycin may have greater anti-obesogenic effects in male mice at other doses, and future studies should more carefully explore this question.

This study also expands on the relationship between mTOR signaling and regulation of metabolism by C/EBP-β. Prior work has shown that reduced mTORC1 signaling inhibits translation of C/EBP-β-LIP in mouse liver and adipose tissue, with subsequent activation of fatty acid catabolism and increased energy expenditure (Zidek et al., 2015). Our results are consistent with these findings, although under obesogenic conditions, rapamycin increased expression of both LAP and LIP isoforms as well as the expression of downstream effectors of C/EBP-β (Figures 3, 5). In addition to the potential differential regulation of C/EBP-β by mTORC1 under HFD conditions, it is also worth noting that the dose of rapamycin used in this study (8 mg/kg/day) is sufficient to inhibit both mTORC1 and mTORC2 (Martin-Perez et al., 2020). Inhibition of mTORC2 by chronic rapamycin treatment has been proposed to cause changes in glucose metabolism and insulin sensitivity in mice fed both normal and obesogenic diets (Liu et al., 2014), although the role of C/EBP-β has not been explored. It will be of interest to determine whether effects of rapamycin on C/EBP-β LAP are solely mediated by mTORC1, or if mTORC2 also plays a role, and how this influences the relationship between diet composition, adiposity, and fat metabolism. Notably, HFD alters AKT phosphorylation at serine 473 which could reflect changes in mTORC2 signaling in our study (Supplementary Figure S2). We did not detect any large changes in glucose homeostasis in response to rapamycin treatment (Figure 1), although we did not formally test glucose and insulin tolerance in these animals, nor can we completely rule out potential effects of either rapamycin or adefovir dipivoxil on alternative targets.

Like rapamycin, the effects of adefovir dipivoxil on LAP and LIP expression in vivo appear to be more complex than has been previously appreciated from in vitro studies. Adefovir dipivoxil was identified as a C/EBP-β activator from a cell-based screen for inhibitors of LIP translation that could mimic the effects of rapamycin (Zaini et al., 2017). Interestingly, we found that the drug decreased total abundance of both LIP and LAP without altering the ratio at a timepoint 24 h after the most recent treatment (Figure 5), but it significantly increased the expression of LAP (Figure 7) and the LAP/LIP ratio 3–5 h after treatment (Figure 7). Correspondingly, at both timepoints we observed increased expression of the C/EBP-β target gene EDIL3, and several additional C/EBP-β targets involved in fatty acid metabolism were induced at the 3–5 h timepoint . The mechanism by which adefovir dipivoxil is impacting C/EBP-β activity and fat metabolism remains unclear. Based on its nucleoside analogue structure, it is tempting to speculate that the drug could impair mitochondrial DNA replication and induce a mitohormetic response (Ristow and Zarse, 2010; Merry and Ristow, 2016); however, we currently have no direct evidence to support this hypothesis over alternative models.

Taken together, the various effects of rapamycin and adefovir dipivoxil observed here suggest that modulating the expression of both C/EBP-β isoforms, LAP and LIP, is likely critical for nutrient homeostasis. Mice lacking both isoforms are resistant to diet-induced obesity, but either die of severe hypoglycemia shortly after birth or live with severe metabolic imbalances (Croniger et al., 2001; Millward et al., 2007). Moreover, reintroduction of LIP in Cebp ^−/− animals restores adiposity and lipid homeostasis, including plasma leptin, free fatty acids, and triglycerides levels (Bégay et al., 2018), thus demonstrating that LIP has more than purely inhibitory function, as initially suggested by the lack of a transactivation domain (Calkhoven et al., 2000). C/EBP-β is highly expressed in liver, kidney, intestine, adipose, pancreatic islets, and innate immune cells and regulates the expression of several genes presiding over development, immune function, regeneration, differentiation, and metabolism (Guo et al., 2015; Pulido-Salgado et al., 2015). Tissue-specific transgenic models will be required to better understand the multiple effects of C/EBP-β isoforms on metabolism and nutrient homeostasis.

Beginning from a geroscience framework, we showed that reduced mTORC1 signaling prevents diet-induced obesity by increasing the activity of C/EBP-β. Furthermore, we identified adefovir dipivoxil as a potent activator of C/EBP-β in vivo and a potential anti-obesogenic therapeutic. Mice engineered to express only C/EBP-β LAP (Cebp ^ΔuORF ) show a marked reduction of age-related impairment and increased lifespan in female animals (Müller et al., 2018). It will be interesting to determine whether treatment with adefovir dipivoxil recapitulates the longevity phenotypes of Cebp ^ΔuORF animals. Nucleoside-analog reverse transcriptase inhibitors such as adefovir dipivoxil are well known for their mitochondrial toxicity and may be involved in HIV-associated neurodegeneration and other accelerated aging phenotypes (Payne et al., 2015); however, it is possible that at appropriate doses adefovir dipivoxil could be an effective drug for preventing obesity or delaying aging.