to the editor: In a recent letter to the editor, Kreier and Buijs (5a) attempted to clear up doubts about the existence of significant parasympathetic/vagal innervation of white adipose tissue in the rat. Using pseudorabies transneuronal retrograde tracing, these investigators had originally concluded that several fat depots were significantly innervated by the efferent vagus nerve (6, 7). However, these observations were seriously questioned when similar studies in Siberian hamsters resulted in very limited retrograde labeling in the vagal motor nuclei and in complete absence of nerve fibers in white adipose tissue containing vagus-specific markers (5). This prompted us to discuss the potential problems and pitfalls that might explain the opposing findings and lead us to conclude that the data presented, to date, do not provide convincing evidence for significant vagal innervation of white adipose tissue (1). Unfortunately, in their letter to the editor, Kreier and Buijs do not offer any new angle or insight to the discussion, leaving our conclusion completely unchanged.
As we extensively discussed, retrograde tracing from peripheral organs has been inherently plagued by problems of leakage and cross-contamination, leading to false-positive results (4), and the use of pseudorabies virus as a retrograde tracer does not make this problem any easier. Unless the proper controls and alternative strategies are employed, the positive results of Kreier et al. (6, 7) have to be put into this category. Among the essential controls is the identification of pseudorabies virus (PRV) in and around the peripheral injection site. Demonstrating complete absence of virus in abdominal organs adjacent to the injected fat pad would go a long way in ruling out leakage over time. Again, simply putting virus into the peritoneal cavity is not a sufficient control, as it rapidly disperses and inactivates. Among the alternative strategies is positive proof of labeled fibers in white adipose tissue following anterograde labeling of vagal preganglionic motor neurons and/or identification of postganglionic neurons. Innervation of other abdominal and thoracic organs by vagal efferents was established beyond doubt because of anterograde tracing and identification of the postganglionic neurons, following the motto, seeing is believing.
The most puzzling finding in the Kreier et al. study (6) is the dense bilateral labeling of the dorsal motor nucleus after unilateral virus injection into so-called sympathetically denervated retroperitoneal fat depots. For one, the explanation that sympathetic denervation forces more virus into the intact vagal innervation is pure speculation and requires rigorous testing. In no way do we agree with this rather bizarre interpretation. Further, tissue trauma caused by the manipulation of abdominal tissues required to achieve a complete mechanical sympathetic denervation of the fat pads, as described by Kreier et al. (6), would have facilitated virus uptake by nerve terminals supplying tissues adjacent to the fat pads (4). Consistent with this interpretation, the use of chemical sympathectomy by Giordano et al. (5), which would not have introduced any abdominal tissue trauma was associated with very limited viral labeling of vagal preganglionic neurons.
In addition, vagal efferent innervation is clearly lateralized as seen by the selective innervation of the ventral esophagus and stomach wall by the ventral subdiaphragmatic trunk and the dorsal esophagus and stomach wall by the dorsal trunk, selectively originating from the left and right dorsal motor nucleus, respectively (3, 8). Thus, equally dense labeling of both sides of the dorsal motor nucleus with unilateral fat pad injection is highly suspicious and suggests a leakage problem. In a similar vein, the PRV labeling of the nucleus ambiguus that occurred in parallel with staining of the dorsal motor nucleus in the Kreier et al. (6) study is diagnostic of virus redistribution from the injection site (4). To counter this suggestion of ours, Kreier and Buijs (5a) cite two studies that demonstrated innervation of other abdominal organs by the nucleus ambiguus. However, these studies also suffered from lack of controls for tracer or virus spread, and the existence of the neural pathways from the nucleus ambiguus to the pancreas or colon inferred is inconsistent with the known functions and projections of the nucleus ambiguus. Therefore, these most likely represent false-positive staining artifacts as previously evaluated (4).
Finally, the complete absence in white adipose tissue of nerve fibers containing vesicular acetylcholine transporter (VAChT), nitric oxide synthase, and vasoactive intestinal peptide in the Giordano et al. (5) paper cannot easily be disposed of by technical arguments. It is simply not true that evidence for these markers has been difficult to obtain in other organs. With regard to VAChT, it is very prominently present throughout the gastrointestinal tract and pancreas, but very rare in the liver. It thus perfectly parallels the strong vagal efferent innervation of the gastrointestinal tract and pancreas but weak innervation of the liver (2). The study cited by Kreier and Buijs (5a) in favor of a significant vagal innervation of the liver is seriously flawed because of the well-known lack of specificity of the acetylcholine esterase staining method for cholinergic nerve fibers. If Kreier and Buijs speculate that the neurochemical marker for vagal innervation of white adipose tissue is pituitary adenylate cyclase-activating polypeptide, leucine-enkephalin, or peptide histidine methionine, why do they not test this possibility, again following the motto, seeing is believing.
Therefore, currently available data suggest that white adipose tissue is not innervated by vagal efferents to any significant extent. In our opinion, only identification of a significant number of nerve fibers with clearly vagal origin in white fat tissue can change this assessment.
- Copyright © 2007 the American Physiological Society