NPY and VGF Immunoreactivity Increased in the Arcuate Nucleus, but Decreased in the Nucleus of the Tractus Solitarius, of Type-II Diabetic Patients

Ample animal studies demonstrate that neuropeptides NPY and α-MSH expressed in Arcuate Nucleus and Nucleus of the Tractus Solitarius, modulate glucose homeostasis and food intake. In contrast is the absence of data validating these observations for human disease. Here we compare the post mortem immunoreactivity of the metabolic neuropeptides NPY, αMSH and VGF in the infundibular nucleus, and brainstem of 11 type-2 diabetic and 11 non-diabetic individuals. α-MSH, NPY and tyrosine hydroxylase in human brain are localized in the same areas as in rodent brain. The similar distribution of NPY, α-MSH and VGF indicated that these neurons in the human brain may share similar functionality as in the rodent brain. The number of NPY and VGF immuno positive cells was increased in the infundibular nucleus of diabetic subjects in comparison to non-diabetic controls. In contrast, NPY and VGF were down regulated in the Nucleus of the Tractus Solitarius of diabetic patients. These results suggest an activation of NPY producing neurons in the arcuate nucleus, which, according to animal experimental studies, is related to a catabolic state and might be the basis for increased hepatic glucose production in type-2 diabetes.


Introduction
Animal studies have shown that the levels of circulating insulin and glucose are monitored in the Arcuate nucleus (ARC) of the hypothalamus and the caudal brainstem. The ARC is also referred to as Infundibular nucleus in the human brain. Changing concentrations of insulin and glucose modify the activity of several neuronal populations within these areas. These nutrient/hormone sensitive neurons project to pre-autonomic neurons located in the hypothalamus and brainstem and modulate the autonomic output to liver and pancreas. In both areas, neurons were identified that are either activated or inhibited by changes in glucose availability [1,2,3,4]. At present, the ARC is mainly characterized by two antagonistic neuronal groups: the Neuropeptide Y (NPY) and proopiomelanocortin (POMC) expressing cells. In experimental animals, NPY neurons in the ARC are activated by negative metabolic challenges, such as fasting and hypoglycemia [5,6] and NPY signaling is required to stimulate hepatic glucose production in response to a decrease in plasma glucose levels [7,8]. POMC neurons, in contrast, are activated by positive energy conditions and satiety factors, and promote catabolism and energy expenditure. POMC is the precursor of several neuropeptides of which alpha-melanocyte stimulating hormone (a-MSH) is the cleavage product that mediates the effects on metabolism [9]. VGF is a neuropeptide precursor co-expressed in NPY and POMC neurons of the rat arcuate nucleus. The ablation of the VGF gene produces a hyper phagic and hyper metabolic phenotype, and results in increased NPY and diminished POMC expression; typical of prolonged fasting [10].
Notwithstanding the wealth of data present in experimental animals nearly nothing is known about the response and changes of these peptide containing-neurons in the human brain. Except for a study that examined the changes of NPY neurons in the ARC in obese humans [11] no information is available about the possible changes in NPY, a-MSH and VGF neurons in the ARC and NTS in the diabetic condition. We hypothesized that if animal studies have a predictive value for what may happen in the human, the metabolic alterations caused by, or causing diabetic illness may modify the expression of the metabolic neuropeptides in the brain and thus indicate the disturbance of insulin and glucose detection in the brain of the diabetic. Hereto, we investigated the expression of NPY, a-MSH and VGF in the post mortem brain of type-2 diabetic individuals which was compared with their expression in brains obtained from age and sex matched non-diabetic patients. Moreover, we investigated the co-expression of VGF in human NPY and POMC neurons. Finally, because noradrenalin (NA) is proposed to play an important role in glucose homeostasis [3], we verified the relationship of catecholaminergic neurons with NPY neurons in the brainstem of diabetic and nondiabetic individuals.
The present results show that, especially where it concerns NPY neurons in the ARC, a significant increase in the number of neurons expressing NPY can be found in the postmortem human diabetic brain. These changes, when extrapolated to experimental animal studies, suggest that the human ARC in the diabetic state is reflecting a catabolic state and is possibly unable to sense high circulating glucose levels. In addition the results show that the same neuronal phenotype in the NTS area responds quite differently to the diseased diabetic state.

Human Brains
For this study, 22 human brains were examined, 11 brains from patients (age mean 6 SEM: 60.164.3, 6 males; 5265.1, 5 females) who died of complications of type II-diabetes (DBT group) and 11 brains of non-diabetic individuals (52.264.3, 6 males; 54.265.7, 5 females), who died of conditions not related with a chronic metabolic impairment, considered as controls (CTR group) ( Table 1,2). Diabetic and control cases were age and sex matched, post mortem time due to hospital practice was always between 8 and 12 hours and brains of controls and diabetics were always stained in pairs to avoid staining intensity differences due to procedure.   consequently the medication received was very limited or nil. Patients were admitted to the public hospital when their health was already seriously compromised and moreover most did not receive any long-term drug treatment for their pathologies and their complications. With the exception of three diabetic patients, from whom we don't know how long they suffered type-2 diabetes, the rest of the individuals were diagnosed and affected by the diabetic disease for at least 10 years. The causes of death of the  persons included in both the diabetic and control groups are mostly acute renal failure and severe infections, with some cases of respiratory insufficiency, gastrointestinal hemorrhages and myocardial infarction. Diabetic and control individuals were affected by similar complications elicited by a series of chronic pathologies, such as cardiovascular, renal and hepatic dysfunctions, sometimes associated with pneumonia or cancer.

Tissue Processing and Immunohistochemistry
Brains were removed during routine autopsy always within 8-12 hours after the death. Brains were fixed by immersion in phosphate buffer (0.01 M, ph7.2) 4% paraformaldehyde (Sigma-Aldrich Corp., St. Luis, MO, USA) for two weeks. Dissected hypothalami and brainstems were stored in 4% paraformaldehyde for additional 30 days and then transferred to 30% sucrose-0.04% NaN 3 in Phosphate Buffer Saline (PBS, 0.01 M, ph7.6) at 4uC until sectioning. At 220uC the medio basal hypothalamus and the caudal brainstem were cut in slices of 50 mm such to obtain alternate sections of the ARC and the NTS. Sections were collected in 6 series and then stored in 30% sucrose-0.04% NaN 3 in PBS for immunohistochemical procedures.
For each antibody, one series of 10 sections was used for immunohistochemical analysis, each section 300 mm apart, thus covering an area of 3 mm of the hypothalamus including the ARC or the dorsal vagus complex (Figure 1). Sections of controls and diabetic brains were always incubated together in order to avoid differences in staining intensities. Sections were incubated with either NPY, a-MSH or VGF antibodies. For this, a rabbit anti-NPY 1:4000 (see for details on antibody specificity [12]) and a sheep anti-a-MSH, 1:5000 (Chemicon International, Millipore  Corporation, Billerica, MA, USA) were used. A rabbit antibody was raised against the human VGF C-terminal sequence VGF 609-615 , 1:10000. [13]. A series of sections from the brainstem was incubated also with a Tyrosine Hydroxylase (TH), 1:4000 (Millipore Corporation, Billerica, MA, USA) antibody. Dilution series of all antibodies were made and the optimal dilution was selected in such a way that never a saturation of the staining was obtained.

Staining Quantification
To quantify the immunoreactivity, 10 sections of the ARC and 10 sections of the NTS were photographed at 40X and 63X magnification respectively and analyzed unilaterally for each neuropeptide. Digital pictures were taken using an Axioplan (Zeiss, Jena, Germany) microscope equipped with a digital color photo camera (Olympus DP25, Olympus, Japan). In order of identify the nuclei of interest the several bibliographic sources [14,15,16] were consulted and the areas delineated accordingly (Figure 1). The total optical density in cells and fibers as well as the total number of positive cells for each neuropeptide were counted with the program ImageJ. Data are represented as mean 6 standard error of the mean (S.E.M.), analyzed with a one-way ANOVA analysis results were considered statistically different when they deviated from each other with a p,0.05.

Fluorescent Immunohistochemistry
To investigate the co-localization of VGF with NPY and a-MSH, sections of the ARC and the NTS from at least three controls and three form the diabetic group were incubated for 48 h with rabbit anti-VGF antibody with either sheep anti-NPY or sheep anti-a-MSH (from Santa Cruz Biotechnology, Santa Cruz, CA, USA and Chemicon International, Millipore Corporation, Billerica, MA, USA, respectively). Sections from the brainstem of both groups were also incubated with rabbit anti-TH and sheep anti-NPY. Sections were then incubated with a mix of donkey anti-rabbit and donkey anti-sheep antibodies conjugated with fluorescent dyes (Cy3 TM -anc Cy2 TM -conjugated, respectively), all provided by Jackson Immunoresearch, West Grove, PO, USA. At the end of the reaction, sections were processed with 10% solution of Sudan Black (Sigma-Aldrich Corp., St. Luis, MO, USA) in ethanol, for 109, to prevent autofluorescence. Finally, sections were mounted on gelatinized slides, cover slipped with 30% glycerol in PBS and analyzed with the LSM 5 Pascal confocal microscope (Zeiss, Jena, Germany).

Ante Mortem Data
Unfortunately little or no ante mortem data such as glucose levels were available; for a part because 6 controls and one diabetic died from cardiovascular problems before they could reach the hospital. At the other hand blood glucose levels at admission were of special interest because they were less influenced by medication. In the controls glucose levels were in the range of 76-145 mg/dl with a mean of 100 mg/dl. The diabetics showed much more variation 72-1058 mg/dl with a mean of 228.1 mg/dl of which the low values were obtained after treatment with oral hypoglycemic drugs before hospitalization. Because of the missing values no statistical analysis could be done. Unfortunately also no data on body weight were available since such measurement is not routinely performed during hospitalization in this hospital.

NPY, a-MSH and VGF in ARC and NTS of the Human Brain
The expression of NPY and POMC mRNAs, as well as the presence of NPY, a-MSH and TH peptides, in the human brain has been extensively described, and our results essentially concur with these descriptions [11,17,18,19]. NPY cellular elements can be encountered throughout the whole diencephalon, but the ARC is one of the areas with the highest number of NPY-expressing neurons. In the ARC of both diabetic and non-diabetic patients, NPY positive neurons were scattered in the whole area of the nucleus (Figure 2-A). The distribution of NPY and TH neurons in the medulla oblongata also corresponds to what previously has been reported for the human brain [14,23,8,20,21]. In the brainstem, the majority of NPY and TH immuno reactive cells are found in the NTS and the ventro lateral medulla, whereby NPY is often co-expressed in catecholaminergic neurons (Figures 3,  4, 5, 6, and 7). In the hypothalamus, POMC expression has been reported to be restricted to the medio-lateral ARC [22,23,24]. However, our results indicate that a-MSH neurons, in addition to the ARC, are also present in the human ventromedial hypothalamus and in the adjacent ventro lateral area (Figure 8).
a-MSH, immunoreactivity was hardly detectable in the NTS of both diabetic and control groups. Indeed, no references about POMC expression in human brainstem are available in literature.
In humans, VGF mRNA has been observed in neural and endocrine cells [25]. We observed VGF-containing neurons in both ARC and NTS, as well as in the catecholaminergic areas of the brainstem (Figure 2-B, 5, 9). In our present study following the NPY, a-MSH and VGF immunohistochemical procedure, we have never seen any indication for a staining in glial cells, and as is also clear from the figures, only cells clearly identifiable as neurons were stained.

NPY and VGF Show Higher Immunoreactivity in the ARC of type-2 Diabetic Subjects
The ARC of diabetic patients displayed a significant increase in the number and intensity of immuno reactive NPY cells as well as in the total NPY optical density (cells and fibers) (p,0.03 and p,0.02, respectively) and for VGF (p,0.003 for both parameters), when compared with non-diabetic controls (Figure 2, 10). For a-MSH, in contrast, no-significant changes were found. Both diabetics and non-diabetics showed a large variation in the number of positive cells (p = 0.430589), as well as in the optical density (p = 0.412148) (Figure 8, 10A).
As can be seen in figure 2 in the hypothalamic target areas of the arcuate nucleus also an increased staining for NPY can be observed reflecting the higher amount of peptide present in the fibers. In contrast to this increase in ARC staining is the observation in the NTS where a diminishment of NPY and VGF staining was observed. This result shows that the changes in NPY and VGF staining are not due to a general change in fixation or procedural artifact but is due to local peptide changes.

NPY and VGF Show Lower Immunoreactivity in the NTS of Type-2 Diabetic Subjects
The NTS of the diabetic group shows a reduction in both the number of immuno reactive cells and as well as in the optical density of NPY (p,0.005 and p,0.001, respectively) and VGF (p,0.005 and p,0.001, respectively) (Figure3, 9, 10-B). Although    there was a considerable variation of these two parameters in the diabetic group, the difference with the controls was highly significant. In contrast to NPY, TH showed a tendency to decrease in the NTS of diabetic individuals, but the difference with the controls did not reach significance (p = 0.06 and p = 0.51 for the difference in cell number and optical density, respectively). (Figure 4, 10-B).

VGF co-localizes with NPY and a-MSH in the Human ARC
We examined the co-localization of VGF with both NPY and a-MSH in the human ARC and examined possible differences between controls and diabetics. VGF and NPY co-localize both in soma and fibers, especially in the medial (ventral and periventricular) zone of the nucleus (Figure 11). In the controls, VGF and a-MSH co-localize only sporadically in fibers and cell bodies within the ARC. However, in diabetics, (almost) all a-MSH neurons along the ventro lateral portion of the nucleus also contain VGF (Figure 12). VGF and a-MSH did not co-localize in the immuno reactive neurons clustered in the most ventro lateral part of the ARC, neither in controls nor in diabetics.

NPY Co-localizes with VGF and TH in the Human Brainstem
The distribution of NPY, VGF and TH in the brainstem at the level of the dorso vagal complex was quite similar: neurons and fibers stained for these neuropeptides were found not only within the NTS and the surrounding area, but also along the trajectory of the Vagus nerve and the A1/C1 area ( Figure 5). Because the TH cells showed a tendency to be diminished in diabetic brains we investigated whether there could be a difference in co-localization of TH with VGF or NPY. In the NTS, NPY co-localizes with VGF and TH (Figure 6 and 13). NPY was found to be present in the majority of the TH cell bodies in the A1 area (Figure 7). No differences were observed in the co-localization pattern of NPY, VGF and TH in diabetic brain as compared to the control brain.

NPY Changes in the ARC, Related to Diabetes?
The increase in NPY immunoreactivity in the ARC of diabetic subjects is comparable to the increase of NPY under negative energy status as observed in rodent studies [5][6][7][8]. This suggests that in diabetic persons the NPY neurons react as if glucose levels are lower than normal. In rodents it has been shown that a variation in glucose and insulin signaling to NPY neurons of the ARC results in the generation of physiological responses that culminate in the regulation of hepatic glucose production and lipid metabolism via the autonomic system [8,22,26,27]. Specifically, an increase in plasma glucose and insulin levels is followed by suppression of sympathetic activity to the liver resulting in lower glucose production [27]. Type-2 diabetes patients, instead, display inappropriately high hepatic glucose production, in spite of their hyperglycemic, hyperinsulinemic condition. This high hepatic glucose production is shown to be caused by increased sympathetic tone, which stimulates both glycogenolysis and gluconeogenesis [28,29]. All in all, the present data suggest that the arcuate nucleus of diabetic patients perceives a negative metabolic state which may result in enhanced hepatic glucose production mediated by the sympathetic system.

Technical Considerations
Animal experimental studies have indicated the arcuate nucleus as a major site for the control of metabolism and have suggested its importance for the development of diabetes. However one of the limitations of animal studies is the question whether these findings can be of relevance for human disease. The aim of the present study was to answer at least in part that question. However, also the present analysis of human brain material has several limitations that need to be taken into consideration for the interpretation of the data. First the conditions before and during death vary considerably and cannot be controlled as in animal experiments, in addition no information about body weight was available. Second, medication between the individuals, especially in the last phase of life may have varied enormously. Third, all measurements are done with immunohistochemical techniques and are therefore semi-quantitative and only reflect relative differences between the groups.
Yet, in spite of all these differences and putative problems, the diabetic group reacted quite distinct with respect to staining intensity and number of NPY and VGF neurons in the ARC as well as in the NTS. Moreover, this is not due to a general change in NPY of VGF immunoreactivity or change in neuron number, as can be concluded from the fact that in the NTS the opposite changes were noted as compared to the ARC. Self evidently more research will be needed to substantiate these findings but they indicate that indeed also in the human diabetic brain similar changes can be observed as in the diabetic rodent brain.

Brain-liver Interplay
In rodents, inactivation of the insulin receptor, specifically in the periventricular area of the ARC causes an increase of NPY expression of about 50% and, at the same time, produces insulin resistance in the liver and increased hepatic glucose production, even in presence of high levels of insulin in plasma [30,31].
According to the present data, the increase of NPY in the ARC in the human diabetic brain, suggests in addition to the peripheral insulin resistance as defined by the diabetic state, the presence hypothalamic insensivity to circulating insulin and/or glucose. At the other hand changes in leptin or other adipokines as occur in diabetes may also play a role in the observed changes in the ARC.

NPY in the NTS
In the NTS of diabetic patients, NPY was decreased indicating a different mechanism of regulation between brainstem and hypothalamus. A comparable observation was made by Goncharuk et al, in 2007 [32] who reported that CRH neurons are activated in the hypothalamus of hypertensive patients, but not in the brainstem. The present observation suggests that in the NTS of diabetic subjects the proper inhibition of NPY neurons may take place by glucose and insulin suggesting that the gluco-sensing elements within the brainstem of diabetic subjects are functioning properly.
At the other hand our present results may indicate that the brainstem NPY neurons have a different role than those of the ARC. It seems logical to assume that both the NTS and ARC are able to sense and process the same information, i.e. glucose and insulin in the circulation, but will control glucose homeostasis using different mechanisms. Thus, the different NPY signaling as we observed in the NTS of diabetic patients might be the basis for a disruption in one of the multiple components of the homeostatic systems upstream of the NPY neurons.

The POMC System
The contribution of POMC neurons in the control of glucose homeostasis has been a matter of controversy since a long time [33][34][35][36][37]. Thus the wide variation in a-MSH content observed in the ARC of both diabetic and control groups may be due to other factors, as mentioned in technical considerations above. These other effects may influence a-MSH synthesis, making it difficult to distinguish from the changes eventually triggered by hyperglycemia and/or hyperinsulinemia. At the other hand the present observation may suggest that in the human brain the POMC neurons may be less important for the inhibition of glucose production than in the rodent.

VGF -Comparison with NPY and a-MSH
VGF, like NPY, is increased in the ARC of diabetic patients. This result agrees with the observation in mice that lack vgf, suggesting an anabolic role for VGF. VGF KO mice are lean and hyper metabolic, show increased energy consumption, and are resistant to several forms of obesity, hyperglycemia and hyperinsulinemia [10,38]. Animal studies, suggest that both NPY and VGF are upregulated by negative metabolic state and are involved in sympathetic control of hepatic glucose production [10]. The present study also shows that the increase in VGF immunoreactivity coincides with an increase in NPY.
Remarkably a-MSH neurons in the ARC of diabetic patients increase their VGF content. Also in animals, VGF has been shown to be present in a-MSH neurons [39]. The increased colocalization of VGF and a-MSH indicate that, contrarily to what a-MSH immunoreactivity suggests, at least a portion of the POMC neurons is affected in the diabetic group.

VGF in the NTS
For the first time, we report here not only the presence of VGF in the human brain but also that VGF and NPY are co-localized in a subset of neurons within the NTS. In addition, as in the ARC, both peptides follow a similar, though opposite pattern of immunoreactivity in the NTS.
The wide distribution of VGF in the brain, the large number of vgf-inducing factors, the complexity of the vgf ko phenotype and the co-expression of vgf in neuronal populations with opposite effects on metabolism do not allow a conclusion about the involvement of VGF in diabetes. However, we demonstrated that VGF, just as in rodents, is also present in the human brain in neurons that control energy balance and react to metabolic changes similar as NPY.

Concluding Remarks
In the present study we demonstrated that type-2 diabetic patients, in spite of the hyperglycemia that characterizes this pathology, show an increased NPY content in the ARC, suggesting that the diabetic brain sensed a negative metabolic state. In view of the fact that many features of the NPY cells in the human ARC and NTS show similarities to the rodent NPY cells in these areas, we would like to propose that the changed perception of the metabolic condition in the ARC of diabetic persons might be the basis for these changes in NPY neurons. Hence, increased NPY signaling to second order neurons that control the activity of the sympathetic system may contribute to the anomalous, unbalanced hepatic glucose production.