The University of Arizona Health Sciences Center Tucson, Arizona 85724
Address correspondence and requests for reprints to: Seymour Reichlin, Department of Medicine, The University of Arizona Health Sciences Center, 1501 N. Campbell Avenue, P.O. Box 25099, AHSC Room 7338, Tucson, Arizona 85724-5099.
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() |
---|
Little attention has been paid over the years since the publications of Harris and the Scharrers to the possibility that the brain might secrete other physiologically important substances into the peripheral blood. In this issue of JCE&M, Wiesner et al.(4)(see pages 22702274) propose that leptin is one such brain-derived secretion. They found that women and obese men have higher concentrations of immunoassayable leptin concentration in the internal jugular vein than in arterial blood. These workers hypothesize that the brain secretes leptin into the blood and that the higher peripheral blood leptin levels observed in women could be due in part to the sexually dimorphic brain contribution of leptin. Their demonstration of positive arterial/venous gradients across the forearm in some individuals (though not significant by statistical test) is compatible with the well established finding that leptin is secreted by fat tissue (5); their demonstration of a negative arterial/venous gradient across the renal vein is compatible with other studies showing that the kidney is the main site of excretion/destruction of circulating leptin (6).
Based on current knowledge of the tissue distribution of leptin, the claim that the brain might contribute to the circulating peripheral leptin pool is completely unanticipated. In the initial report, in which leptin was identified in mouse adipose tissue by cloning methods, survey of other potential tissue sites of synthesis found that leptin messenger RNA (mRNA) was not expressed in brain (7). Similar negative findings were reported for extracts of sheep brain (8). However, a relatively small amount of leptin mRNA synthesized by a highly restricted pool of neurons could well escape detection when diluted with a large amount of whole brain mRNA; no studies of leptin mRNA of human brain have been published. Binding, immunohistochemical and in situ mRNA hybridization has convincingly shown the specific distribution of leptin receptors in choroid plexus and in several groups of neurons (5, 9, 10, 11), but there appear to be no published reports of leptin-immunopositive structures in brain. In an unpublished study, neurons containing leptin could not be identified in the rat (J. Elmquist, personal communication). However, negative immunohistochemical findings are not conclusive because there may be technical problems in antigen detection, and some neuronal (or glial) products may be contained in a very small storage pool that is turning over rapidly.
Another bit of evidence suggesting that leptin is not synthesized in the brain is the demonstration that leptin concentration in the cerebrospinal fluid is much lower than in peripheral blood. For example, in nonobese humans, cerebrospinal fluid to serum leptin concentration ratios ranged beween 0.023 and 0.0068) (12, 13, 14, 15) and in obese rats lacking the brain leptin receptor (which is postulated to be the leptin transporter from blood to brain), ratios of cerebrospinal fluid to serum leptin were less than one tenth of those of normal animals (0.0029 vs. 0.031) (16). Our group found that concentrations of leptin in the superior sagittal sinus (the principal site of venous drainage from the brain) in rats were no greater than in blood obtained from the aorta (Reichlin S, G Chen, M Nicolson, unpublished). This finding corresponds to the results obtained by Wiesner et al. in normal weight males, but is not relevant to the groups in which the jugular/arterial gradient was elevated. It thus remains to be shown by direct determination that any brain structure, especially in the human, is capable of synthesizing leptin. The paper by Wiesner et al. should encourage such studies.
If an intrinsic brain source of leptin cannot be identified, how then can we explain the findings of Wiesner et al.? One possibility is that the antibody used in the radioimmunoassay detected a cross-reacting substance, with the (admittedly) unlikely possibility that it was present only in obese men and in women. Another possibility considered by Wiesner et al. is that the sampling procedure could have included blood from other fat-containing parts of the head. They sought to exclude this artifact by positioning the catheter high up toward the base of the skull. Nevertheless, it is possible that brain-derived blood in the internal jugular vein was mixed with blood refluxed from orbital and facial fat. This could have been the case as the venous catheterization procedure was carried out in the supine position, and the elevated gradients were observed in obese men and in women who had higher peripheral (presumably adipocyte-derived) blood levels of leptin.
Another possibility considered by Wiesner et al. is that the jugular vein/arterial gradient reflects the release of leptin from brain to blood that has been taken up by the brain, assuming that the conditions under which the studies have been done (after an overnight fast and a 350 kCal light breakfast) induce release. The occurrence of the phenomenon in obese men and in women may reflect the fact that these are the states associated with the highest blood leptin level and hence the highest brain receptor loading. Leptin enters the brain by a specific transport system (5, 17, 18) postulated to be the truncated form of the leptin receptor located in high concentrations in the choroid plexus (5) and on capillary endothelia throughout the brain (18, 19). Once in the brain, leptin readily exits by the mechanism of bulk transport, at the rate with which cerebrospinal fluid is both formed and secreted into the blood (17, 20).
However, it is unlikely (as inferred from pharmacokinetic studies in rodents) that enough leptin can be sequestered in the brain to account for the magnitude of release over time calculated by Wiesner et al. to contribute as much as 40% to the circulating body pool in some individuals. In the mouse, the highest concentration of radioiodinated leptin in the brain relative to blood is at 20 min after iv bolus injection; at that time, the concentration of the dose per gram of brain is 0.171% (17). If one assumes mouse brain weight to be 0.445 gm, only 0.076% of the injected dose is present in brain at that time (W.A. Banks, personal communication).
Although earlier studies indicate that the synthesis of leptin by brain would not be anticipated, it would not be surprising if it was found by further examination. A number of nonadipocyte tissues have been shown to synthesize and secrete leptin. These include gastric mucosa (21), mammary epithelial cells (22), myocytes (23), rat hepatic stellate cells (24), and placenta (25, 26). In brain, leptin acts as a regulator of secretion of anorexogenic peptides (5, 7, 9, 10), GnRH (27), and TRH (28), but of its action in nonadipose sites, little is known. In view of its cytokine-like properties, it is likely that leptin secreted locally outside of fat tissue is a paracrine regulator.
Finally, one may consider whether other substances synthesized in brain can be considered to be secretions of the brain. Brain biogenic amine neurotransmitters and their metabolites spill over into the human internal jugular vein (29), but their peripheral function is not fully established. In male songbirds circulating blood estrogens arise from aromatization of precursors carried out in the brain (30). Brain natriuretic peptide appears in the circulating blood of patients suffering from subarachnoid hemorrhage in concentrations high enough to explain the salt wasting observed in this disorder (31).
The brain may also secrete bacterial toxins and proinflammatory
cytokines into peripheral blood after central infection or inflammation
(32, 33, 34, 35, 36). Elevation of peripheral cytokines after central brain damage
likely accounts for the endocrine and metabolic changes associated with
the acute phase reaction that is characteristic of several kinds of
brain damage in humans: closed head injury (37), stroke (38),
meningitis (39), and brain death (40). Although TNF-, IL-1, and IL-6
may be expressed by subpopulations of neurons, it is more likely that
the bulk of proinflammatory cytokines that enter the blood from the
brain originate from brain vascular endothelia, meninges, glia, and
choroid plexus (32, 33, 41), all of which secrete these substances when
exposed to bacterial endotoxin. Thus, secretions of the brain may
derive from brain supportive structures as well as from neurons.
Leptin (17, 20), cytokines such as IL-6 and TNF-, and endotoxin all
leave the brain as soluble constitutents of cerebrospinal fluid by bulk
transport as shown by pharmacokinetic kinetic studies of
radioiodine-labeled tracers (32, 33). They exit by way of both
fenestrated endothelia in the arachnoid villi (42) and perineuronal
channels that communicate between the subarachnoid space and
interstitial spaces in the retroperitoneum and the nasopharyngeal
submucosa (43). In contrast to the tightly controlled and restricted
entry of substances into the brain determined by the blood-brain
barrier, the pathways of exit are unrestricted by molecular size and
are not regulated by specific transport systems. Thus, leptin, a
peptide approximately 16 kD in size, and other molecules, if
synthesized in brain whether by neurons or by other supportive tissues,
can readily become secretions of the brain if they gain entry to the
cerebrospinal fluid.
![]() |
Note Added In Proof |
---|
![]() ![]() ![]() ![]() |
---|
Received April 19, 1999.
Accepted May 7, 1999.
![]() |
References |
---|
![]() ![]() ![]() ![]() |
---|