Review Paper:

Sakurai, T. et al. Orexins and orexin receptors: a family of hypothalamic neuropeptides and G protein-coupled Receptors that regulate feeding behavior. Cell 92: 573-585 1998.

The authors of this paper were initially interested in seven-transmembrane G protein coupled cell surface receptors (GPCRs). A wide variety of GPCRs exist, and they respond to a vast array of stimuli. When a ligand binds to a specific GPCR, signal transduction occurs by activation of G proteins. While the functions of many GPCRs are already known, a variety of GPCRs still have unknown ligands. Such GPCRs whose ligands are unknown are termed Òorphan GPCRs.Ó The authors, in testing various orphan GPCRs discovered a family of neuropeptide ligands that, because they were primarily found in the hypothalamus, were hypothesized to be involved in appetite and energy homeostasis. They initially called the receptor HFGAN72 and its neuropeptide ligands Òorexins,Ó derived from the Greek orexis, which means appetite. The hypothesis the authors attempt to prove is that orexin-A and -B act as positive regulators of food consumption in the central nervous system.

The first step taken was to purify the orexins through four steps of HPLC. This is summarized in Fig. 1. Panel A shows the first step of purification. Crude extracts of rat brain were loaded onto a C18 reverse phase HPLC column and 3 ml fractions were collected. A chromatogram was made for these fractions (represented by retention time) at 280 nm, showing the peaks of absorbance corresponding to the fractions potentially containing the proteins of interest. The dotted line corresponds to the acetonitrile gradient. Below the chromatogram are [Ca²⁺]i transient assays. Orexins elicit increases in cytoplasmic Ca²⁺ levels in HEK293 cells transgenic for the orexin receptor HFGAN72. As such, each fraction collected was assayed in the presence of cells transgenic for HFGAN72, and the relative activity was determined by assaying [Ca²⁺]_i. As a control, similar assays for each fraction were performed using cells transgenic for the unrelated orphan receptor HPRAJ70. It was determined that all but three of the fractions induced similar activity responses for HPRAJ70 as they did for HFGAN72, suggesting that these fractions were the result of endogenous components of the host HEK293 cells. Activity peaks contained in both assays were the result of ligands acting with endogenous receptors from the host cells. As a result, the activity peaks labeled A, B, and BÕ were designated as specific for the HFGAN72 receptor and were assumed to contain orexins. In fact, peak A corresponds to orexin-A, peak B corresponds to orexin-B, and peak BÕ corresponds to orexin-B[3-28]. These will be discussed later.

In Figure 1B, the fractions corresponding to peak A from Fig. 1A were pooled and loaded onto a cation exchange HPLC column. Again, a chromatogram was obtained (this time at 220 nm), and fractions were collected. A [Ca²⁺]_i was performed in the same way as in (A) and the fractions with activity were pooled and loaded onto another reverse phase HPLC column. The data obtained from this third step in the HPLC purification are summarized in Figure 1C. The active fractions collected in this step were loaded onto a final reverse phase HPLC column at 40û C. One fraction was determined to have activity, and the chromatogram peak (210 nm) above it corresponds to virtually pure orexin-A. Now that orexin-A was purified, the contents of fractions B and BÕ were purified in the same manner. When this was done, the peptides were sequenced, shown in Figure 2A. Human, rat, mouse, and bovine orexin-A were all found to be identical. Rat and mouse orexin-B were identical, while human orexin-B was slightly different. Orexin-B in rats was determined to be 46% identical to orexin-A. Orexin-A was determined to be a 33-amino acid peptide with a mass of about 3.5 kD, while orexin-B was 28 amino acids long with a mass of approximately 3.0 kD. Orexin-A has two disulfide bonds, indicated by bars above the sequence (Figure 2A). Orexin-B has no such disulfide bonds.

Once the peptide sequences were obtained, a BLAST search was attempted in an effort to find a cDNA corresponding to the orexins. The search detected no similar entries. As a result, a cDNA fragment encoding a portion of orexin-A was produced by RT-PCR of rat brain mRNA. Primers were made based on the amino acid sequence of orexin-A. Using 5Õ-RACE and 3Õ-RACE reactions, the full-length cDNA was obtained. This cDNA encodes prepro-orexin, a 130 residue peptide that is a precursor to both orexin-A and orexin-B and is shown in Figure 2B for human, rat, and mouse. The boxes represent residues of orexin-A and orexin-B; the remaining residues are cleaved to produce one of each of the mature peptides. Equal signs above the peptide sequence represent secretory signal sequences, while prohormone convertase cleavage and amidation sites are labelled with asterisks. Identical residues between the three species are indicated by vertical lines. Rat prepro-orexin is 95% homologous with its mouse counterpart and 83% homologous with human prepro-orexin.

Figure 2C shows the amino acid sequences for the HFGAN72 receptor in human and rat. Transmembrane domains across the neuron membrane are labeled and indicated with equal signs above the peptide sequence. The receptor is now called the OX1 receptor (OX₁R), and the authors determined that it is similar to many other neuropeptide receptors. They say that this is consistent with the hypothesis that OX₁R is the receptor for orexins, but there is still not sufficient proof. To further prove this hypothesis, they performed in vitro functional assays, summarized in Figure 3. Panel A demonstrates that radioactive [125I]orexin-A binding to OX₁R is inhibited by very small concentrations of non-labeled orexin-A in CHO cells transfected with an OX₁R expression vector. The same conditions hold true of orexin-B inhibiting the binding of radiolabelled orexin-B. However, it is apparent that orexin-B binds to OX₁R with significantly less affinity than orexin-A. Panel C follows up the data from (A); it shows that as the ligandsÕ (orexin-A or -B) concentration is increased, so is the cytoplasmic Ca2+ concentration. Again, orexin-A has a greater effect than orexin-B. This led to two major conclusions. First, the authors felt that the cytoplasmic calcium concentration is mobilized by receptor:ligand binding because the ligand orexin-A is activated in the binding process. Such calcium mobilization was not seen in non-transfected CHO cells but was similarly seen in transfected HEK293 cells, although the authors do not show this data. Nonetheless, these results confirm that orexin-A is the ligand, or an agonist, for OX₁R . Second, the lower affinity of orexin-B prompted the authors to hypothesize that another orexin receptor might exist that is more specific for orexin-B. They did a BLAST search and discovered another GPCR, which they called OX₂R. OX₂Rwas determined to 64% homologous with OX₁R. Panels C and D repeat panels A and C, except the receptor used was OX₂R. The results show that the oxerin-A and -B ligands bind with equal affinities to OX₂R; otherwise, OX₂Rbehaves in much the same way as OX₁R . The ultimate conclusion of Figure 3 is that OX₁R is a specific receptor for oxerin-A, while OX₂Ris nonspecific and will bind to either ligand with equal affinity. These results make sense, but the argument would have been stronger if the authors had included (and not just briefly mentioned) the results from the non-transfected CHO cells as a negative control.

The next step in the research was to determine which tissues express prepro-orexin (and thus both orexin-A and -B), OX₁R, and OX₂R. To accomplish this, a Northern hybridization was done, using probes for prepro-orexin and both receptors. Figure 4 shows that in rats, the orexins are primarily produced in the brain and some in the testis. OX₁R and OX₂R, according to the authors, only seem to be present in the brain, but I can see bands indicating presence elsewhere, especially for OX₂R in the kidneys. Actin mRNA was used as a positive control to ensure that cellular mRNA was loaded into each lane. The authors conclude that the results are consistent with the hypothesis that orexins are neuropeptides, but they do not address any possibilities for why the ligands or receptors are present in tissues other than the brain.

Since the authors had shown that the brain is the primary location of orexin expression, they did in situ hybridization and immunohistochemical analyses in rat brains to determine precisely which regions of the brain are producing prepro-orexin mRNA and ultimately orexin-A and -B. Using a prepro-orexin cRNA probe, the results showed that neurons containing orexins are primarily found in hypothalamic and subthalamic regions of adult rat brains. This data is shown in Figure 5. Both immunofluorescence and in situ hybridization gave these same results. Thalamic and hypothalamic regions have repeatedly been shown to influence food consumption, so the results prompted the idea that orexin-A and -B may play a role in controlling appetite.

To test whether orexin-A and -B had any effects on appetite, the neuropeptides were administered to rats via a catheter directly into the left lateral ventricle of the brain. The results are summarized in Figure 6. We see that both orexin-A and -B administration increased the ratsÕ food consumption compared to the vehicle control. In the panel on the left, it is clear that higher doses (30 nmol) of orexin-A enhanced food consumption more than lower doses (3 nmol). Both dosages, however, increased food consumption several fold, and the effects were apparent after just one hour. Asterisks indicate that food consumption was significantly different than the vehicle control, while crosses show significant differences between the two dosages. For orexin-A, the food consumption is greatly enhanced in both dosages, and higher doses seem to further enhance consumption. Orexin-B administration had similar effects; feeding significantly increased after one hour, but the increase did not seem to last as long. Whereas orexin-A persisted for 4 solid hours, the effects of orexin-B began to decline after 2 hours, especially in the lower dose. As an explanation to the shorter action of orexin-B, the authors suggest two possibilities. First, because orexin-A has disulfide bonds and posttranslational modifications at both termini, it may be more resistant to degradation than orexin-B. To test this hypothesis, they could have done a time dependent Northern hybridization for radiolabelled orexin-A and -B to see if orexin-A is, in fact, present in cells longer than -B. Second, there is a distinct possibility that orexin-A and orexin-B each have distinct mechanisms by which they work to influence appetite and food consumption. What is clear is that compared to vehicle controls, rats administered orexins have an increased appetite due to some mechanism that takes place in and around the hypothalamus.

The final data collected in the paper (Figure 7) was to determine whether fasting rats had up-regulation of orexin and NPY. NPY has already been shown to be a positive regulator of food consumption. Upregulation of orexin in fasting rats would suggest that they have a role in the nervous systemÕs control of feeding behavior. Ten total rats were either fasted for 48 hours or fed as much as they wanted. After the 2 days, the rats were sacrificed and total RNA was extracted from thalamic and hypothalamic brain tissue. A Northern hybridization was then conducted (Panel A), using probes for prepro-orexin mRNA, NPY, and actin. By using actin as a positive control for the amount of RNA added in each lane, they could quantify the ratios of orexin mRNA and NPY mRNA to the actin mRNA, which should be constant. Panel B shows the results of these ratios. Fasted rats had relative orexin mRNA expression upregulated more than 2-fold over rats that had been fed. NPY mRNA was also upregulated, although to a lesser extent, but this had been previously shown. The conclusion is that orexins do appear to play a role in the central nervous systemÕs regulation of food consumption, since they are present at varying amounts depending on certain situations. In sum, the authors have shown that orexins play a physiological role in the feedback mechanism of the hypothalamus that regulates feeding patterns.

The results produced in this paper are very believable, and besides the few omissions in Figures 3 and 4, the data was convincing. They should have included the non-transfected cells as a control in Figure 3 to demonstrate that their results were not due to other factors. They neglected to address bands that were present in Figure 4, presumably because the presence of these bands did not fully support their hypothesis. As a follow-up to Figure 6, they could have included an assay to determine whether orexin-B is degraded more quickly than orexin-A. Otherwise, the results are believable and will certainly lead to future research. However, the simple discovery of two ligands that seem to influence food consumption and their receptors in the hypothalamus does little to explain how the mechanism works in vivo. The authors suggest further investigation into the function of orexin-B[3-28], which is truncated at the N-terminus and may be physiologically relevant and could help explain the mechanism by which the neuropeptides work. They also suggest further investigating the relationships between positive regulators of feeding (e.g. NPY and orexins) and negative regulators of feeding (e.g. leptin and melanocortins). Study of the interrelationships of these regulators could lead to a better understanding of how the body maintains energy homeostasis.

One possible way to analyze homeostasis would be to do an assay like the one in Figure 7, but with rats fed under varying conditions. Rats could be fasted and fed varying amounts up to an unlimited food supply. Then, after 48 hours, the rats could be sacrificed and total RNA could be extracted from the thalamus/hypothalamus tissue. A Northern hybridization would then be conducted, but this time one could probe for mRNAs of both positive and negative feeding regulators. Using actin as a control, we could see how these regulators of feeding behavior respond to varying levels of food consumption. This would help explain how the levels of production of each neuropeptide changes in response to food consumption in hopes of maintaining energy homeostasis in the rat.

Another experiment that could be conducted would be to see the levels of production of various positive and negative feeding regulators in obese rats compared to wild type. This would be conducted by taking brain tissue from both obese and wild type rats, isolating total RNA, and proceeding in the method described above. By doing this, one could determine whether obese mice had an excess of positive regulators, a deficiency of negative regulators, or both. Once this was determined, the next logical course of action would be to administer negative feeding regulators and/or to somehow diminish the effects of the positive regulators and see how this influences obese ratsÕ feeding behavior and ultimately their weight. By doing so, it would be possible to determine whether neuropeptides could be manipulated to restore obese mice to more normal feeding patterns.

The ultimate goal for this research would be to find the mechanisms by which orexin-A and -B work, and how the two receptors play a role in these mechanisms. This is a very complicated matter, and it will take more than just one paper to determine these mechanisms. Questions that might be answered are: How do orexins influence neuronal firing in the hypothalamus?; Is simple binding of the neuropeptide to its receptor sufficient to change feeding behavior?; Are orexins necessary for survival, or will other positive regulators pick up the slack if they are absent? Some of these questions are relatively simple to answer and others are more difficult. Hopefully, by answering such questions as these, the mechanism by which these neuropeptides influence feeding behavior will be better understood.

Send comments, questions, and suggestions to: tobeadle@davidson.edu