INTRODUCTION

Rheumatoid arthritis (RA) is an autoimmune and chronic inflammatory disease that mainly affects the synovium but also induces systemic manifestations causing pain, swelling, stiffness, unsteadiness, and deformity. Of note, RA is frequently associated with fatigue, weakness, fever, and weight loss (1-3). The mechanisms underlying metabolic complications in RA are not well understood, but a serious catabolic status, driven predominantly by proinflammatory cytokines, might be responsible for body cell mass loss, a common feature of RA (4, 5). In fact, there is increasing evidence about the contribution of the dysregulation of adipose tissue to RA, in particular dysregulated secretion of adipokines (4-7).

Obesity can play a dual role in RA, both by exacerbating its development and as a result of the disease progression, in part due to the patient's inability to carry out physical exercise and thereby inducing weight gain, something that may also be boosted by certain drugs used in the management of the illness (1, 8-10). There is a general consensus that obesity prevention is important in patients with RA since it improves pain perception and metabolic and cardiovascular risk, as well as favoring a better response to treatments, such as anti–tumor necrosis factor (TNF) (11) and anti–interleukin-6 (anti–IL-6) receptor antibody (12).

A link between RA and altered levels of energy balance modulators acting at the central level, such as ghrelin, leptin, and other adipokines, has also been described (1, 6, 7, 13). Moreover, a substantial amount of data highlighted a close relationship between alterations in different neuronal populations, some of them hypothalamic, and experimental arthritis (EA) (14-18). However, whether a dysregulation of these hypothalamic mechanisms is a cause or a consequence of the disease or, more importantly, whether targeting these hypothalamic nuclei may have a positive impact on the development of EA, remains unclear. Here, we aimed to investigate whether the hypothalamus may be involved in the physiopathology of EA. We focused on AMP-activated protein kinase (AMPK) in a specific set of neurons located in the ventromedial nucleus of the hypothalamus (VMH), which recently emerged as a critical canonical mechanism controlling energy homeostasis (19, 20). In this sense, current evidence has shown that inhibition of AMPK in steroidogenic factor 1 cells of the VMH leads to sympathetic nervous system–mediated activation of the brown adipose tissue (BAT) thermogenesis, leading to increased energy expenditure and feeding-independent weight loss (19-24). Notably, this mechanism mediates the actions of key thermogenic factors, such as thyroid hormones, bone morphogenetic protein 8B, estradiol, liraglutide receptor agonism, and nicotine (19-24).

Therefore, our aim was to investigate whether this central pathway might be involved in the metabolic alterations induced in an experimental model of arthritis. We used a model of EA induced by intradermal injection of Freund's complete adjuvant (CFA), which does not reflect every aspect of human RA but is a routinely used model and resembles some of the articular and extraarticular features of the disease (13, 15, 16).

DISCUSSION

The relationship between obesity and several inflammatory and autoimmune diseases, such as RA, has been broadly studied over the last decades. However, the underlying mechanism is still under debate. There is a general consensus that both diseases are associated with an imbalance between proinflammatory and antiinflammatory cytokines contributing to the onset and progression of RA and obesity (4, 5). Therefore, in this study we aimed to clarify whether obesity could influence RA and to uncover the molecular mechanism responsible for RA-induced altered energy balance. With this in mind, we induced EA by CFA inoculation (13, 15, 16) in a rat model (those fed a control diet [SD] versus those fed an HFD) and also assessed its impact on peripheral and central mechanisms regulating energy balance. We focused specifically on BAT thermogenesis, since it is known that induction of EA by CFA is characterized, in some cases, by increased energy expenditure (28, 29). Furthermore, it is known that RA is characterized by weight loss and wasting, a state known as rheumatoid cachexia (RC), but the mechanism by which some RA patients lose weight is not well defined and may be multifactorial (1-3). A similar situation is present in cancer-induced cachexia, where activation of brown fat thermogenesis has been described (30-33).

The phenotype observed in our preclinical model is consistent with the definition of pre-cachexia, since it fulfils the features required to be present in patients with an underlying disease: chronic and systemic inflammation, hypophagia, and weight loss (34). However, it should be acknowledged that there are some clear differences between cachexia induced by other diseases, such as cancer, and RC. In classic cachexia, loss of body weight, due to muscle and fat loss, is a common feature. These outcomes are consistent with data showing increased resting energy expenditure induced by BAT activation (30-33) or WAT browning (35-37), in both rodent models of cachexia and in patients with cachexia. In contrast, in RC, for which a consensus diagnostic criterion does not exist, the loss of body weight and adiposity rarely occurs (38, 39).

Our data showed that both lean and obese rats displayed a similar increase in paw volume after EA induction. No correlation was found between body weight and body weight loss. Remarkably, although both SD-fed and HFD-fed animals with CFA-induced EA displayed initial hypophagia, they restored their food intake after the peak of the illness; however, while SD-fed rats with CFA-induced EA were able to show a progressive body weight recovery, HFD-fed rats with CFA-induced EA failed to recuperate their body weight and continued to lose body weight and fat masses. To further improve our understanding of EA-induced alterations in energy balance, we assessed the effect of CFA-induced EA on BAT thermogenesis. We found a marked activation of BAT in all the stages of EA, in both SD-fed and HFD-fed animals, as shown by increased BAT temperature and/or increased levels of UCP-1 in brown fat, as well as browning of WAT.

Several mechanisms could explain this increased thermogenic tone in our experimental EA model, acting centrally or directly on brown and white adipocytes. For example, it is known that cancer cachexia–induced browning is dependent on IL-6 (35). However, considering that the proinflammatory milieu represses the thermogenic activity of brown and beige fat via cytokines that inhibit noradrenergic signaling (25), central effects might be more important than direct peripheral actions on adipose cells. Given that AMPK in different hypothalamic neuronal populations regulates whole-body energy homeostasis, from feeding to BAT thermogenesis and browning of WAT (19, 20, 23, 40), we next investigated the effect of EA in this pathway. Our data revealed that VMH AMPK is decreased in EA. Next, we investigated whether this effect was mechanistically associated with EA-induced actions on energy balance. Thus, we targeted AMPKα1 in the VMH, a nucleus where this catalytic subunit has been involved in both the modulation of feeding and BAT thermogenesis (19, 20, 27).

Our data showed that specific VMH AMPK activation using virogenetic strategies was enough to ameliorate the negative energy balance included by CFA-induced EA. Remarkably, besides body weight gain, restored feeding, and diminished BAT and browning tones, AMPKα1 activation in the VMH decreased the circulating levels of inflammatory cytokines, as well as improving the physical appearance of the animals. These later effects are quite relevant since it is assumed that proinflammatory cytokines are at the root of some of the most serious consequences of RA (4, 5). In this sense, the mechanisms underlying metabolic complications in RA are unclear, although proinflammatory cytokines might also be responsible for the loss of body cell mass (4, 5). In addition, the link between the hypothalamic AMPK axis and the inflammatory status raises very interesting pathophysiologic as well as physiologic questions. It is known that inflammation of tissues is under neural control, involving the neuroendocrine, sympathetic, and central nervous systems (18, 41). Data from the 1990s had already demonstrated an association between sympathetic ganglia and the pathogenesis of EA (42, 43). Of note, CFA-induced EA in Lewis rats has been linked to changes in the sympathetic nerves in the spleen and is also responsible for the activation of immune cells in the red pulp of that organ (44, 45).

Remarkably, the spinal BAT sympathetic preganglionic neurons in the intermediolateral nucleus of the thoracolumbar spinal cord are in the same area as those innervating the spleen (46, 47). Thus, activation of the same centers may promote both BAT thermogenesis and immune activation in the spleen. This connection is functionally supported by our data and a recent report showing that propranolol (a nonselective beta blocker) promotes, in addition to antiarrhythmic effects, a systemic antiinflammatory action in a model of collagen-induced arthritis in Lewis rats (48). Overall, this evidence seems to indicate that reduced sympathetic tone ameliorates EA symptoms, offering a possible alternative mechanism to the antiinflammatory effect of AMPKα1 adenoviral treatment in the VMH.

To our knowledge, this is the first study linking the canonical hypothalamic AMPK–BAT/WAT axis to the development of the symptoms of a systemic disease, such as RA. This is relevant because targeting hypothalamic AMPK, which has been proposed as a potential therapy for obesity (19), may also be a possible strategy to ameliorate the negative energy balance and to improve the inflammatory state associated with RA. In this sense, recent and profuse evidence has shown that metformin, a drug administered for the treatment of type 2 diabetes mellitus that activates AMPK, promotes metabolic improvement in RA patients and in animal models of pharmacologically induced and autoimmune arthritis (49, 50).

In summary, our data show that negative energy balance caused by CFA-induced EA is independent of initial body weight, and it is associated with VMH AMPK–mediated activation of BAT thermogenesis and browning. Notably, activation of AMPK in the VMH not only ameliorates the metabolic outcome in CFA-induced EA but also improves the inflammatory status of the animals. Taken together, these findings provide new mechanistic insight into the pathophysiology of RA and suggest new therapeutic strategies for its possible clinical management and treatment.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. López, who is the lead author, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design

Seoane-Collazo, Rial-Pensado, Estévez-Salguero, Nogueiras, Diéguez, Gualillo, López.

Acquisition of data

Seoane-Collazo, Rial-Pensado, Estévez-Salguero, Milbank, García-Caballero, Ríos, Liñares-Pose, Scotece, Gallego, Gualillo, López.

Analysis and interpretation of data

Seoane-Collazo, Rial-Pensado, Estévez-Salguero, Milbank, Ríos, Fernández-Real, Nogueiras, Diéguez, Gualillo, López.