Does circulating leptin have the ability to cross the blood-brain barrier and target neurons directly?

It is in general agreement that proteins with high molecular weights do not cross the blood-brain barrier (BBB), because of the tight junctions between endothelial cells that characterize the cerebral microvasculature. The tight junctions are indeed essential features of the cerebrovascular system, which filtrates water-soluble molecules and protects parenchymal elements of the brain against potentially toxic substances. In addition to be the critical defense wall for neurons, the cerebral endothelium acts as a secretory membrane. There are, however, profound differences in the permeability of the blood vessels within specific regions of the brain. This is actually the case for the circumventricular organs (CVOs), in which the numerous vascular loops are relatively permeable to soluble proteins, pathogens, and even cells of the immune system (1). The vascular system of the brain is composed of penetrating arterioles, small capillaries, and venules that, all together with the blood, constitute approximately 8 –10% of the total volume of the brain and spinal cord. The capillary density is also highly variable between regions and nuclei. For example, white matter contains 260 mm blood vessels/mm of total tissue, whereas the paraventricular and supraoptic nuclei of the hypothalamus are composed of 2000 mm/mm (2). These hypothalamic nuclei are among the most vascularized regions of the organism, explaining their crucial role in the control of most homeostatic systems. Ordinarily, small capillaries are not permeable due the tight junctions and robust astrocytic foot processes, but penetrating arterioles and venules exhibit high degrees of variability that may give access to proteins, such as leptin. This hormone is produced mainly by the white adipose tissue and circulates into the bloodstream as a 167-amino acid protein with a relative molecular mass of 16 kDa (3). Such high molecular mass indicates that the protein has very limited access to the cerebral tissue through the regions that are more permeable or through selective penetrating arterioles or venules. The main target region of leptin is the ventrobasal hypothalamus just above the median eminence, the region where all neurosecretagogues are released into the hypophyseal-portal system (4). Median eminence is also a CVO equipped with a rich fenestrated vascular network extending to various subcellular regions of the arcuate nucleus. Therefore, direct interaction between leptin and its long-form receptor (Ob-Rb) expressed in neurons just adjacent to the median eminence remains a likely mechanism. However, leptin is also capable of activating neurons that are not adjacent to the median eminence, and the exact mechanisms explaining this phenomenon are still not clear at this point. In a recent paper published in this journal, Flier and colleagues (5) provided evidence supporting the existence of an active transport system via the leptin short-form receptor (ObRa). They found high levels of both Ob-Ra and Ob-Rc transcripts in isolated rat cerebral microvessels and a decrease in brain uptake of leptin in mice lacking leptin receptor. The cerebral uptake of leptin was also reduced in diet-induced obese mice and in an animal model responding to central, but not peripheral, leptin (5). In this issue of Endocrinology, Hosoi et al. (6) further support this concept and report that circulating leptin has the ability to trigger signal transduction events not only in the hypothalamus, but in different populations of neurons in the brainstem. Although it is not a cytokine, leptin receptor contains docking sites for Janus kinases (JAK) that phosphorylate specific members of the signal transducers and activators of transcription (STAT) family (7). Once activated, STAT proteins may activate different genes in combining their Src homology 2 domains and forming homodimers. Activation of JAK2 and STAT3 takes place in response to the binding of leptin with its transmembrane Ob-Rb. Both Western blot and immunocytochemistry techniques confirmed the presence of phosphorylated (p) STAT3 in the hypothalamus and brainstem, but not in the hippocampus, cortex, and cerebellum, of animals challenged with a single bolus of leptin (6). More specifically, positive pSTAT3 cells were found in the commissural and lateral part of the nucleus of the solitary tract (NTS) and in the superior and external lateral parts of the parabrachial nucleus and periaqueductal gray. Previous experiments using c-fos as a marker of neuronal activation reported similar data in these brainstem regions, and the double-labeling approach provided anatomical evidence that cholecystokinin neurons in the superior parabrachial nucleus and glucagon-like peptide-1 neurons in the NTS were activated by leptin (8). Although the nuclear protein Fos was detected in Ob-Rb-expressing cells in different hypothalamic regions, low Ob-Rb mRNA levels were found in the brainstem, and most Fos-immunoreactive nuclei failed to depict positive signal for the gene encoding the Ob-Rb (8). These data either suggest that leptin does not act directly onto these neurons, or in situ hybridization is not sensitive enough to detect Ob-Rb transcript in regions in which mRNA levels are less abundant. In this Abbreviations: BBB, Blood-brain barrier; CVO, circumventricular organ; JAk, Janus kinase; NTS, nucleus of the solitary tract; Ob-Ra, leptin short-form receptor; Ob-Rb, leptin long-form receptor; p, phosphorylated; SOCS-3, suppressor of cytokine signaling 3; STAT, signal transducer and activator of transcription. 0013-7227/02/$15.00/0 Endocrinology 143(9):3211–3213 Printed in U.S.A. Copyright © 2002 by The Endocrine Society doi: 10.1210/en.2002-220655