Physiological Roles for Butyrylcholinesterase: A BChE-Ghrelin Axis

BChE is a ghrelin hydrolase

Our own views of BChE’s catalytic interactions with ghrelin changed very recently, after accidentally observing that mice with high blood levels of BChE generated by viral gene transfer showed dose-dependent reductions in plasma ghrelin. Thus, after a moderate viral dose (e.g., 10^10 particles per mouse), there was a marginal decline (~ 10%). A 10-fold higher dose caused a ~ 50% decline, and a 100-fold higher dose reduced ghrelin levels 90% or more. These effects arose within ten days and lasted for as long as high BChE levels were sustained (months to years). In fact, ghrelin is an unusual peptide hormone with a biological activity that requires octanoylation of its Ser3 residue [5, 7]. BChE, an esterase with wide substrate specificity, does cleave this ester bond and transforms ghrelin into desacyl-ghrelin, which is inactive at GHSR1a although it may have other physiological targets [8, 9]. Because gene transfer of irrelevant luciferase did not alter plasma ghrelin, it seemed likely that the effects of high BChE were specific. To test that conclusion we gave vector-treated mice an i.p. injection of the selective BChE inhibitor, iso-OMPA. This treatment returned ghrelin levels to near normal for 12–24 hr in mice with excess BChE [10]. Furthermore we observed that plasma ghrelin levels in untreated, BChE-null C57BL/6 mice were consistently 40% higher than in wild type animals of the same strain [10].

The summarized findings left little doubt that BChE is a regulator of circulating ghrelin, but did not explain how the enzyme can deacylate such a large substrate. We addressed that issue by molecular dynamics simulations of ghrelin and BChE. These showed most of the peptide backbone excluded from the active site, but the crucial ester bond was in an orientation conducive to hydrolysis. In fact, the consensus structure [10] had ghrelin’s carbonyl ester carbon positioned closer to the catalytic serine hydroxyl than was previously seen with the equivalent atoms in BChE-cocaine simulations [11].

Further studies using commercial immunoassays for ghrelin and desacyl-ghrelin yielded identical rates for the disappearance of intact peptide and the appearance of its breakdown product, confirming the original findings of ghrelin hydrolysis by BChE. Similar results were obtained by quantitative mass spectrometry in collaboration with Schopfer et al, [12], and again with a new assay developed to track ³H-ghrelin deacylation by quantitative transfer of free octanoic acid from aqueous media into toluene [13]. Although BChE hydrolyzes ghrelin less efficiently than smaller substrates, with a K_m of 3.6 ± 0.3 μM and k_cat of 2.2 ± 0.1 min⁻¹ it can exert a strong influence on circulating levels of ghrelin. BChE is highly concentrated in the amygdala, hypothalamus, and pituitary gland [14, 15], where ghrelin and its receptor are prominently localized [16]. This spatial arrangement may facilitate a central BChE-ghrelin regulatory circuit that impacts emotional behaviors. In short, past and present observations provide good grounds for concluding that BChE has evolved in part to serve as a natural ghrelin hydrolase. The next logical issue is to determine fully how it affects ghrelin-driven events in live organisms.

Physiological role of BChE

It is now reasonable to ask this question: does BChE serve as an important ghrelin modulator or is its hydrolytic activity a mere curiosity? Existing data support the idea that BChE does impact ghrelin-driven events. In 2008, Lockridge’s group reported that BChE^−/− S129 mice showed excessive weight gain on a high fat diet under parameters that did not promote obesity in mice with normal BChE expression [17]. That same year we found excess weight gain in rats given daily exposures to low doses of chlorpyrifos [18]. This irreversible organophosphate inhibitor impacts BChE more readily than acetylcholinesterase [19]. We have now repeated Lockridge’s finding in a second BChE^−/−mouse strain (C57BL/6) and also have shown that a BChE gene transfer can restore resistance to obesity in genetically BChE-deficient mice (unpublished data). These observations are in line with the idea that ghrelin arouses hunger and stimulates eating, and that BChE’s ability to deacylate ghrelin contributes to the regulation of these phenomena. Oddly, however, neither we nor Lockridge saw increased food consumption in BChE-null animals. Other apparent paradoxes have come to light as well. For example, plasma ghrelin levels in obese BChE^−/− mice and humans are actually lower than in lean individuals [20]. One possible reason is that ghrelin as a hormone is influenced by many factors, only one of which is diet. We hypothesize that it may affect fat deposition more directly via actions within the adipocyte, yet to be discovered. In fact, although ghrelin is a documented regulator of food intake and energy metabolism, its essential function is still obscure. For example, eliminating the gene encoding the ghrelin peptide or its activating enzyme, ghrelin octanoyl acyl transferase (GOAT), does not seem to produce any severe phenotype in mice, such as altered body weight, length, and composition, or even a change in daily food intake [21–24]. Thus, the BChE-ghrelin axis now appears far more complex than was originally envisioned, and the intertwined roles of BChE and ghrelin in obesity progression clearly deserve intensive further investigation.

Emotional behaviors represent another complex arena in which BChE can exert a major impact by deacylating ghrelin. Our awareness of this link was prompted by cage-side observations revealing that group-housed male Balb/c mice accumulated fewer bite wounds over time if they and their cage mates had ever received BChE gene transfer. As time continued, the BChE vector recipients outlived the untreated or luciferase vector treated controls (Fig 1). These observations were not made under conditions that met standards for unbiased analysis of mouse longevity, and they were certainly influenced by ongoing cage-fighting and unequal housing conditions. Thus they have no claim to be definitive, but they strongly indicate that administration of BChE vector can alleviate psychosocial stress.

For quantitative determinations of aggressive behaviors we turned to Klaus Miczek’s rigorous “resident-intruder paradigm.” This model actively provokes aggression in a host male (residing in a cage where he had mated with a female), when another male is introduced [25]. Bites and fights were videotaped and scored by a blinded observer. Under this procedure a distinctly different behavior pattern emerged between mice that had undergone BChE gene transfer versus those that had not [10]. The chief difference was in the rate at which aggressive acts per session rose during repeated daily encounters. The rise in aggression was fastest in untreated mice or mice given a different vector that did not decrease circulating levels of ghrelin (Fig 2A), i.e., vector encoding luciferase or a mutated version of BChE (CocH-6 ΔT) that readily hydrolyzes cocaine but has poor catalytic efficiency with ghrelin [10]. The impact was all the more surprising because ghrelin had not been recognized as a factor in aggression although widely known to affect fear and anxiety [26]. Therefore we used the same model to test the effect of sustaining modestly elevated ghrelin levels by means of double gene transfer with cDNA encoding a) ghrelin peptide and b) its activator enzyme, GOAT. This combined treatment did raise aggression scores, but not when BChE gene transfer was added to prevent the rise in plasma ghrelin (Fig 2B).

In these aggression experiments untreated mice could all be induced to fight, although the rate of onset and intensity of that outcome differed between the two tested strains (Balb/c > C57BL/6). Inbred mice show large strain differences in aggressive behaviors and stress responses. This fact must be considered carefully when interpreting outcomes, especially in purely observational studies. Shortly after our aggression data were published, however, Chatonnet’s group released a study focused on cognitive function in BChE-knockout and wild type mice [27]. Their report included a statement that they were “unable to confirm” our observations on mouse aggression. This wording suggested a divergent view. However, in private exchanges the senior author clarified the intended meaning, which was simply that their experiments had not been designed to assess such behavior. Thus there is no conflict. Meanwhile we acknowledge that experiments have yet to be performed to test the impact of high BChE levels on aggression in females, an important remaining issue.

Complex impact of ghrelin and BChE in emotional behaviors

Clearly, emotional behaviors in rodents must be evaluated with care. The mechanism by which ghrelin affects mood-related behaviors has not yet been fully elucidated. There is general agreement that ghrelin levels rise in response to various stressors, but two opposite models for ghrelin actions in stress responses are under debate. One is based on the concept that ghrelin is in itself a stressor, while the other views ghrelin as calming under stressful conditions. In any event there is fair agreement that ghrelin strongly affects “perceived stress,” which can then influence a wide range of responses. The complexity of the underlying neural circuitry is further enhanced by special features of GHSR1a. GHSR1a is a G protein-coupled receptor mainly expressed in the hypothalamus and pituitary [28]. It has unusually high constitutive activity that contributes to downstream signaling and physiological processes [29]. Thus, even in the total absence of ligand, GHSR1a receptors show significant signaling, substantial inactivation, and unusually high levels of internalization [30]. This curious behavior has contributed to controversy over the true impact of ghrelin in regard to emotion. Prominent neurobehavioral scientists have reached diametrically opposite conclusions from studies of anxiety and fear in rodents. In 2006, Donahue et al reported that growth hormone, which is released in response to ghrelin receptor activation, rises after exposure to stress [31]. That finding linked ghrelin to stress, but it did not indicate whether the hormonal response was a conveyer or a reliever of stress. Soon afterward, mouse experiments from Eric Nestler’s group [32] and rat experiments from an Australian team [33] linked ghrelin to reduced behavioral signs of chronic and acute stress. On the other hand, in rat models of post-traumatic stress disorder, K.A. Goosens and coworkers found that ghrelin administration enhanced fear and anxiety [34].

A plausible explanation for these divergent outcomes can be constructed on the premise that ghrelin-induced release of growth hormone is an effective anti-stress response, but acute pulses of ghrelin and long-term elevations generate opposite effects. Thus, a sudden rise in ghrelin will stimulate functionally sensitized GHSR1a to release growth hormone, but chronically high ghrelin levels will weaken the response of GSHR1a receptors to subsequent pulses of the same peptide and, in turn, reduce growth hormone release. This seemingly contradictory behavior could be explained by the unusually high sensitivity of GSHR1a to desensitization and its unique propensity to self-activate in the absence of ligand [35, 36]. Preliminary results from our lab, designed to examine growth hormone release in response to a synthetic GHSR1a agonist, appear to support this concept (not shown).

Other factors add even further complexity to the task of gaining a comprehensive understanding of ghrelin’s role in modulating aggression and other emotional behaviors. One such factor is the des-acylated peptide, lacking the octanoyl ester. There is a consensus that desacyl-ghrelin (DAG) does not interact productively with GHSR1a, but the literature suggests that DAG has one or more target receptors that remain to be clearly identified and linked to emotional states and behavioral outcomes [37]. Preliminary results from our ongoing studies are consistent with that interpretation and, specifically, they do not exclude the possibility that high levels of circulating DAG generated by BChE’s action on the parent peptide might exert anti-aggressive effects through a separate receptor.