Introduction
Food consumption is characterized by three aspects: what, when, and how much. The content, the timing, and the size of meals are important factors in dietary effects on health [
1]. Ingesting the same food at different times of the day has different consequences on health, because systemic metabolic efficiency fluctuates over the course of the day [
2]. In humans, food intake at later times in the circadian rhythm was associated with increased adiposity, independent of the content or amount of food intake [
3]. Indeed, among humans enrolled in weight-loss programs, those that ingested more calories earlier in the day lost more weight and showed more improvement in metabolic markers compared to those that ingested more calories late in the day [
4,
5]. Therefore, eating at the correct time is beneficial for health.
High fat diets (HFDs, or “what” you eat) promote obesity. First, HFDs cause excessive energy intake (affecting “how much” you eat); second, HFDs disturb various rhythms, such as feeding (affecting “when” you eat), locomotor activity, and metabolism [
6‐
8]. A lack of coordination in feeding, activity, and metabolism can desynchronize energy intake and expenditure. Correcting the feeding schedule was sufficient to prevent HFD-induced obesity in mice, even without changing caloric intake [
9,
10]. Accumulating evidence has emphasized the importance of feeding patterns (when you eat) in maintaining energy balance and health. To facilitate healthy eating behavior, it is essential to understand the mechanisms that drive ad libitum feeding patterns. However, it remains elusive how feeding patterns are regulated physiologically and how they become disturbed with a HFD.
Some studies have shown that the excessive energy intake associated with HFD ingestion is mediated by activated microglia, which cause hypothalamic inflammation [
11,
12]. HFDs contain saturated long-chain fatty acids (SFAs), such as palmitic acid (C16:0) and stearic acid (C18:0), which accumulate in the hypothalamus [
13]. This accumulation activates toll-like receptor 4 signaling, which induces the transcription factor, NF-κB, to upregulate inflammatory cytokine production (e.g., TNF-α) [
12,
14]. The acute microglial activation induced by HFD feeding is restricted to the arcuate nucleus (ARC) of the hypothalamus [
11,
12], which is the primary center for the homeostatic control of body weight [
15]. The resulting hypothalamic inflammation causes resistance to the central anorexigenic signals of leptin and insulin. This resistance disrupts the homeostatic regulation of feeding and body weight and leads to hyperphagia and weight gain [
12,
16]. However, it remains unclear what role hypothalamic inflammation plays in feeding patterns. Importantly, inflammatory cytokines are not the only pathway for spreading neuroinflammation. Indeed, neuroinflammation mediated by activated microglia is known to spread through two pathways: the inflammatory cytokine pathway (mediated by TNF-α, IL-1β, etc.) and the gap-junction hemichannel pathway [
17]. The hemichannel pathway was shown to play a major causative role in promoting neuronal damage in neurodegenerative diseases [
17].
INI-0602 is a central-acting, pan-connexin inhibitor with a higher affinity for hemichannels than for gap junctions [
18]. INI-0602 blocks only microglial release of small molecules (such as glutamate) through hemichannels, without attenuating acute inflammatory cytokine induction [
18]. We previously reported that inhibiting this pathway with INI-0602 did not affect inflammatory cytokines; nevertheless, it suppressed disease progression in mouse models of amyotrophic lateral sclerosis and Alzheimer’s disease by blocking glutamate release into the extracellular space [
18]. Importantly, INI-0602 was designed to target the central nervous system (CNS); the CNS redox system oxidizes the dihyropyridine moiety of the pro-drug, which results in the drug becoming trapped within the CNS [
19]. Therefore, INI-0602 accumulates in the CNS, including the hypothalamus, although it is rapidly cleared from the circulation and peripheral tissues [
18]. Due to this pharmacodynamic property, INI-0602 demonstrated no systemic side effects in mice, even after chronic administration for 5 months [
18]. Therefore, INI-0602 is an ideal tool for investigating whether the gap junction hemichannel pathway plays a role in HFD-induced hypothalamic inflammation.
Although many investigators have extensively studied the importance of inflammatory cytokine signaling in the pathogenesis of obesity, the role of the gap junction hemichannel pathway in HFD-induced obesity has not been addressed. Therefore, in this study, we investigated whether treating HFD-fed mice with INI-0602 would be sufficient for preventing HFD-induced obesity and the associated feeding pattern disturbances. By monitoring feeding patterns in detail, we found that INI-0602 prevented HFD-induced feeding pattern disturbances, characterized by excessive feeding during the light cycle. This effect was associated with the prevention of the HFD-induced obesity in mice.
Discussion
In this work, we used a central-acting reagent, INI-0602, which blocks the gap junction hemichannel pathway, but not the inflammatory cytokine pathway, during neuroinflammation. We analyzed its effects on feeding and locomotor activity patterns in detail, within the context of HFD-induced obesity. We addressed two major questions; first, how does HFD disturb feeding and locomotor activity patterns at the behavioral levels? Second, does the gap junction hemichannel pathway play any role in the pathogenesis of HFD-induced behavioral pattern disturbances and obesity (independently of the inflammatory cytokine signaling pathway). We found that HFD caused acute hyperphagia mainly by increasing light cycle feeding; moreover, the feeding pattern disturbances worsened with increasing proportions of long-chain SFAs in the HFD. These results suggested that the long-chain SFAs in HFDs were responsible for disturbing feeding rhythms. We found that INI-0602 prevented HFD-induced obesity and HFD-induced feeding pattern disturbances, but not HFD-induced activity pattern disturbances. However, INI-0602 failed to restore an existing HFD-induced feeding pattern disturbance, when it was given after HFD feeding had been initiated. These results suggested that the gap junction hemichannel pathway mediated the effect of HFD feeding on the feeding pattern only at the beginning phase of the disturbance. Importantly, preventing the initial step was sufficient to attenuate the degree of HFD-induced obesity in mice, despite HFD-induced locomotor activity pattern disturbances.
Long-chain SFAs are known to activate microglia within the ARC and induce inflammatory cytokines [
12]. The initial hypothalamic inflammation is associated with the acute activation of microglia within 3 days followed by the activation of astrocytes in 7 days [
11]. HFD-responsive microgliosis occurs specifically in the ARC and not in the adjacent ventromedial nucleus of the hypothalamus [
11,
23]. The activation of the ARC residential microglia induces the subsequent recruitment of peripheral myeloid cells to the ARC by 4 weeks of HFD ingestion, further promoting ARC microgliosis [
23]. These microglial inflammation further triggers astrocyte inflammation, which also contribute to the establishment of the diet-induced obesity [
24]. Therefore, the NF-κB-dependent microglial activation is critical for the entry of bone-marrow-derived myeloid cells to the ARC, propagation of the hypothalamic inflammation, and the subsequent development of the HFD-induced obesity. The instant disruption of feeding rhythm upon HFD ingestion suggested the involvement of the acute hypothalamic ARC inflammation, which is mostly mediated by microglia.
In addition to the induction of the NF-κB-dependent inflammatory cytokine signaling, neuroinflammation activates the gap junction hemichannel pathway [
17]. This hemichannel pathway is blocked by INI-0602, without affecting the acute induction of inflammatory cytokines. Here, we demonstrated that the INI-0602 block prevented HFD-induced feeding pattern disturbances. Therefore, we speculated that HFD-induced obesity had two components; the gap junction hemichannel pathway, which affected “when you eat”, and the inflammatory cytokine signaling pathway, which affected “how much you eat.”
Gap junctions consist of two apposed hemichannels, each contributed by one cell. They are used for intercellular diffusion of second messengers smaller than 1 kDa, such as Ca
2+, IP3, and nucleotides [
25]. Under physiological conditions, astrocytic gap junctions contribute to the stability of neuronal networks; in contrast, resting microglia express low to undetectable levels of connexins (components of gap junction hemichannels). However, under inflammatory conditions, astrocytic gap junctions are shut down by classical inflammatory mediators, and the expression of connexin is induced in activated microglia, and morphological changes induced in microglia upon activation result in the detachment of gap junctions and cell adhesions and form unapposed hemichannels [
25,
26]. Through these unapposed hemichannels, activated microglia release massive amounts of pro-inflammatory factors, such as ATP and glutamate, into the extracellular space. This milieu causes neuronal damage and secondary glial inflammation [
17,
26]. Our findings suggested that SFAs activated this gap junction hemichannel pathway and led to feeding pattern disruptions.
These observations have raised the question: what are the identities of the small molecules that mediate HFD-induced feeding rhythm disturbances? Within the context of neurodegenerative disease models, in which extracellular glutamate plays pivotal roles, we previously showed that INI-0602 prevented lipopolysaccharide-induced microglial glutamate release in vivo and in vitro, and prevented disease progression [
18]. However, the identities of the pathogenic small molecules involved in HFD-induced feeding pattern disturbances remain unknown, because doing microdialysis in the ARC is challenging even in rats and no previous report exists in mice. Furthermore, the microdialysis of the ARC in HFD-fed rat has not been reported. Therefore, in this study, we could not directly determine whether INI-0602 blocked the release of small molecules in the hypothalamus that were relevant to HFD-induced feeding disturbances. This issue must be addressed in future studies.
Interestingly, INI-0602 had differential effects on HFD-induced disturbances in feeding and locomotor activity. The suprachiasmatic nucleus (SCN) of the hypothalamus serves as the master clock; this clock systemically coordinates various biological rhythms by regulating slave clocks located within the brain (such as “feeding clock”, “activity clock”, etc.) and the periphery. This control is demonstrated in the rhythmic expression patterns of clock genes, such as
period1 [
27‐
29]. Both feeding and ambient light serve as zeitgebers (timing cues from the external environmental). It is generally believed that HFDs disturb clocks in the gastrointestinal tract (including the liver), and feeding zeitgebers indirectly reset the phase of the SCN clock through a feedback loop that includes humoral and neural pathways. However, our finding that INI-0602 had differential effects on HFD-induced rhythm disturbances suggested that INI-0602 did not act on the master clock in the SCN to rescue HFD-induced behavioral disturbances. Moreover, a previous study showed that a 12-week HFD feeding period had no effect on the molecular oscillations in the SCN [
30]. Therefore, it is unlikely that INI-0602 restored the rhythmicity of the feeding clock by blocking only the specific influence of the SCN rhythm on the feeding clock (and not its influence on the activity clock). Instead, we hypothesized that the HFD might have disturbed both the feeding clock and the activity clock through two independent mechanisms; thus, INI-0602 treatment could have blocked only the former and not the latter. However, the precise location of the feeding clock in the brain remains an issue of debate. Proposed candidates include the ARC [
31,
32], NPY neurons [
33,
34], and the dopamine system [
35], among others. Moreover, INI-0602 action is not specific for any particular location in the CNS. Therefore, the identity of the feeding clock remains to be uncovered in future investigations.
The findings of this study have shed some light to how the intrinsic feeding rhythm is disturbed by HFDs, and how these disturbances might be prevented. In mice fed ad libitum, the feeding pattern is presumably controlled by the intrinsic feeding drive. In humans, the timing of meals is determined by an intrinsic feeding drive (hunger) combined with the social environment and other factors [
1]. A disruption in a meal schedule (e.g., breakfast, lunch, dinner) could induce hunger-related stress and lead to in-between-meal snacking. Snacks are often rich in fat, which would exacerbate the disturbance in the feeding rhythm. Therefore, we propose that normalizing the intrinsic rhythm of the feeding drive would be beneficial in humans by facilitating a regular feeding schedule and reducing snacking. Identification of the pathogenic small molecule(s) responsible for HFD-induced feeding rhythm disturbances would help improve our strategies for improving the feeding pattern to promote better health.