Defective central leptin signalling and impaired leptin entry into the CNS (central nervous system) represent two important aspects of leptin resistance in obesity. In the present study, we tested whether circulating human CRP (C-reactive protein) not only diminishes signalling of leptin within the CNS, but also impedes this adipokine's access to the CNS. Peripheral infusion of human CRP together with co-infused human leptin was associated with significantly decreased leptin content in the CSF of ob/ob mice. Furthermore, following peripheral infusion of human leptin, the CSF (cerebrospinal fluid) concentration of leptin in transgenic mice overexpressing human CRP was sharply lower than that achieved in similarly infused wild-type mice. Administration of LPS (lipopolysaccharide) to human CRP-transgenic mice dramatically elevated the concentrations of human CRP in the CSF. The i.c.v. (intracerebroventricular) delivery of human CRP into the lateral ventricles of ob/ob mice blocked the satiety and weight-reducing actions of human leptin, but not those of mouse leptin. I.c.v. injection of human CRP abolished hypothalamic signalling by human leptin, and ameliorated the effects of leptin on the expression of NPY (neuropeptide Y), AgRP (Agouti-related protein), POMC (pro-opiomelanocortin) and SOCS-3 (suppressor of cytokine signalling 3). Human CRP can impede the access of leptin to the CNS, and elevation of human CRP within the CNS can have a negative impact on the physiological actions of leptin.
Chronic low-grade inflammation is an early event in the pathogenesis of Type 2 diabetes and atherosclerosis, conditions in which the blood level of the inflammatory marker CRP (C-reactive protein) is increased [1–3]. Recent research has also revealed that circulating concentrations of CRP are elevated in overweight and obese individuals and are positively correlated with BMI (body mass index), blood leptin concentrations and central obesity [4–6]. In fact, human CRP is an independent risk factor for obesity, insulin resistance, central fat disposition, hepatic steatosis and cardiovascular disease [1,3,7]. Despite the variety and strength of these statistical associations, biological evidence that CRP influences energy balance and metabolism remains scarce, and our knowledge about how CRP might achieve such an effect is limited.
Leptin resistance is causal of not only obesity, but also some obesity-associated metabolic diseases [7,8], and studies indicate that diminished cellular leptin signalling at the hypothalamus is an important component of leptin resistance [9,10]. Several cellular mechanisms have been proposed for the decreased central leptin signalling, which involve the activities of SOCS-3 (suppressor of cytokine signalling 3) [11,12], PTP-1B (protein tyrosine phosphatase 1B) [13,14] and TCPTP (T-cell protein tyrosine phosphatase) . Additional mechanisms may also contribute to leptin resistance. Recent studies that we have performed suggest that human CRP not only is a marker for obesity-related co-morbidities, but also is, in fact, actively involved in the regulation of adiposity through direct interaction with leptin and subsequent diminishment of leptin's physiological actions . Accordingly, human CRP, when pre-incubated with human leptin, was able to inhibit leptin-induced signalling in cultured cells in vitro . When co-infused peripherally into ob/ob mice, human CRP attenuated the satiety, weight-reducing and anti-diabetic actions of human leptin . Thus human CRP, along with those recently identified cellular mechanisms, may have also contributed to leptin resistance.
Another important aspect of leptin resistance lies in the apparent impairment of access of leptin to the CNS (central nervous system). In this regard, clinical evaluation of leptin concentrations in the CSF (cerebrospinal fluid) compared with plasma pools indicate that leptin is not proportionally transported into the CNS when its blood concentration rises . Consequently, leptin's CSF/plasma concentration ratio shows a negative correlation with BMI as well as with the level of leptin circulating in the periphery . Thus, particularly in the case of obese subjects, the blood concentrations of leptin as well as their adiposity are not appropriately reflected in the CNS. Existing paradigms of leptin resistance do not adequately explain such a defect in leptin's entry into the CNS.
Our recent studies raised the possibility that, in humans, elevated blood CRP might hamper the transport of leptin into the CNS. Clinical studies have reported that human CRP can cross the BBB (blood–brain barrier) and that its CSF concentration is elevated in accordance with the state of inflammation [18,19]. Thus human CRP potentially has two means by which it could regulate the functions of human leptin. Confirming the mode of action of CRP is a critical issue to be resolved since the primary acting site of leptin in regulating energy balance is at the hypothalamus. To this end, we used both physiological and biochemical approaches to test the hypothesis that human CRP may act as an inhibitory factor on both the transport of human leptin into the CNS and on its physiological actions at the hypothalamus.
MATERIALS AND METHODS
Male C57BL/6J ob/ob mice (8-week-old) were obtained from the Jackson Laboratory (Bar Harbor, ME, U.S.A.). hCRP-Tg mice (human CRP-transgenic mice) and the ob/ob mice were fed standard chow (containing 14% fat, 26% protein and 60% carbohydrate by calories, Prolab Isopro RMH 3000, PMI Nutrition International) and water ad libitum and housed individually at 22°C with a 12-h light/12-h dark cycle (light from 07:00 to 19:00 hours). A diet-induced obesity model was established in male C57BL/6J mice fed on a high-fat diet (containing 42% fat, 15% protein and 43% carbohydrate by calories, Harlan Teklad TD88137) starting at 6 weeks of age and continued for 12 weeks. Food intake and body weight were measured daily. All experimental procedures were performed according to a protocol approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh.
Subcutaneous infusion of human CRP and human leptin
Micro-osmotic pumps (Model 1007D, Alzet, Durect) were implanted subcutaneously on the backs of mice. These pumps were filled such that they released human CRP (1 mg/kg per day) (Chemicon International), human leptin (0.2 mg/kg per day) (NHPP-NIDDK) or human CRP plus human leptin (1 and 0.2 mg/kg per day respectively). Control animals received pumps filled with PBS. When hCRP-Tg mice were used, these were implanted with pumps that released human leptin (0.6 mg/kg per day). In each case, serum and CSF were collected at the end of the infusion period and used for measurement of human leptin and human CRP.
Intracerebroventricular cannulation and injection
Mice were anaesthetized with 250 mg/kg Avertin delivered i.p. (intraperitoneally) (Sigma–Aldrich). A 6-mm-long plastic cannula was inserted into the lateral cerebral ventricle: 0.6 mm posterior to the bregma, 1.3 mm lateral of the midline and 2 mm below the calvarium. The exteriorized cannula was secured onto the skull with Super Glue. A 28-gauge stainless steel pinhead was inserted into the cannula to keep it patent. The animals were allowed to recover for 1 week after surgery. Thereafter the cannula served as a port for the direct introduction of human leptin and human CRP into the CNS. Each protein (1 μl volumes) was injected slowly using a 5 μl syringe (Hamilton).
CSF was collected from the cisterna magna using a previously described protocol . Mice were anaesthetized with 250 mg/kg Avertin (i.p.), an incision was made extending from the top of the skull to the dorsum, and the intervening musculature was removed to expose the meninges that overlay the cisterna magna. The animal was then placed on a platform beneath a dissecting microscope (Zeiss), with its head tilted downward at a 45° angle. The tissue above the cisterna magna was excised with care and the surrounding area was gently cleaned with cotton swabs to remove residual blood. A borosilicate glass micropipette (VWR) was prepared using a micropipette puller (Model P-97 Puller, Sutter Instrument Co.), then used to puncture the arachnoid membrane while carefully avoiding the surrounding blood vessels. Routinely, ∼15 μl of CSF filled the pipette as a result of internal luminal pressure, and this was collected for analysis. The CSF was inspected with a microscope and also tested biochemically, and only CSF samples devoid of blood contamination were used for further analysis (see Supplementary Figure S1).
Measurement of human CRP and human leptin
The concentration of human CRP and human leptin in serum and CSF were measured by ELISA (Helica Biosystems and R&D Systems respectively). The minimal detectable doses were 7.8 pg/ml for leptin and 0.022 ng/ml for CRP. For leptin ELISA, we added 5% Tween 20 to each sample and standard, and pre-treated these at 60°C for 10 min before determination of human leptin values, all according to the manufacturer's instructions.
Molecular signalling studies
The hypothalamus was surgically removed immediately after mice were killed, and a portion was subjected to protein extraction. Activation of STAT3 (signal transducer and activator of transcription 3) was evaluated in the protein extracts via Western blot assay, using an antibody against phosphotyrosine STAT3 (Cell Signaling Technology). The membrane was stripped and re-probed with antibody against STAT3 to evaluate the overall level of STAT3 (Cell Signaling Technology).
Quantitative analysis of gene expression levels
Hypothalamic RNA was prepared using TRIzol® reagent (Invitrogen), treated with DNase I (Ambion), then reverse-transcribed with a Superscript kit (Invitrogen). Real-time PCRs were carried out on an ABI7900 Taqman machine with the following cycles: one cycle of 50°C for 2 min, one cycle of 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min. Mouse 18S rRNA was used as an internal control for input cDNA quantity. Linearity of each assay was determined based on a co-efficient value of at least 0.99 for the standard curve. The cDNA concentrations of all samples were within the linear range of the standard curve. The probes were labelled on the 5′-end with FAM (6-carboxyfluorescein) and on the 3′-end with BHQ [2,5-di-(t-butyl)-1,4-hydroquinone]. The probe and primer sequences were as follows. Mouse NPY: probe, 5′-FAM-TCCAGCCCTGAGACACTGATTTCAGACC-BHQ-3′; forward primer, 5′-CACCAGACAGAGATATGGCAAGAG-3′; reverse primer, 5′-CGTTTTCTGTGCTTTCCTTCATT-3′. Mouse AgRP (Agouti-related protein): probe, 5′-FAM-CGCGAGTCTCGT-TCTCCGCGTC-BHQ-3′; forward primer, 5′-CGGAGGTGC-TAGATCCACAGA-3′; reverse primer, 5′-GGACTCGTGCAG-CCTTACACA-3′. Mouse POMC (pro-opiomelanocortin): probe, 5′-FAM-CAGTGCCAGGACCTCACCACGGA-BHQ-3′; forward primer, 5′-TGCTTCAGACCTCCATAGATGTGT-3′; reverse primer, 5′-GGATGCAAGCCAGCAGGTT-3′. Mouse SOCS-3: probe, 5′-FAM-ACTGTCAACGGCCACCTGGACT-CCT-BHQ-3′; forward primer, 5′-CCACCCTCCAGCATCT-TTGT-3′; reverse primer, 5′-CAGGCAGCTGGGTCACTTTC-3′. Mouse 18S rRNA: probe, 5′-FAM-TGGACCGG-CGCAAGACGGAC-BHQ-3′; forward primer, 5′-CGCCGCT-AGAGGTGAAATTC-3′; reverse primer, 5′-TTGGCA-AATGCTTTCGCTC-3′. Detection and expression analysis of the MC4R (melanocortin 4 receptor) gene and NPY1R (neuropeptide Y receptor 1) gene were performed using real-time SYBR Green PCR. MC4R forward primer, 5′-TCAT-CTGTAGCCTGGCTGTG-3′; reverse primer, 5′ GGTACTG-GAGCGCGTAAAAG-3′. NPY1R forward primer, 5′-CTGAT-GGACCACTGGGTCTT-3′; reverse primer, 5′-GAAGAAGCC-ACTGCAAGGAC-3′. 18S forward primer, 5′-CTCAACACG-GGAAACCTCAC-3′; reverse primer, 5′-CGCTCCACCA-ACTAAGAACG-3′.
Data are presented as means±S.E.M. Differences between groups were analysed by two-tailed Student's t tests or one-way ANOVA with P<0.05 considered significant.
To understand the impact of circulating human CRP on the ability of human leptin to access the CNS, we installed micro-osmotic pumps into ob/ob mice and infused either leptin alone (0.2 mg/kg per day) or leptin plus human CRP (leptin at 0.2 mg/kg per day, human CRP at 1 mg/kg per day). At the end of a 4-day infusion period co-infusion of human CRP completely neutralized the satiety-inducing effect of human leptin (Figure 1A). On day 4 of infusion, the serum concentration of human CRP reached 0.61±0.07 μg/ml in mice infused with CRP alone or CRP plus leptin. This amount of serum CRP is in the normal baseline range for healthy humans . Serum concentrations of human leptin were comparable for mice infused with human leptin and mice infused with human leptin plus human CRP (Figure 1B, left-hand panel), but the CSF concentration of human leptin was significantly lower in mice infused with human leptin plus CRP (Figure 1B, right-hand panel).
Subcutaneous co-infusion of human CRP and human leptin impedes access of human leptin to the CNS in ob/ob mice
The findings shown in Figure 1(B) are consistent with human CRP in the periphery impairing entry of human leptin into the CNS. To explore this possibility further, we took advantage of a transgenic mouse model that expresses human CRP in the liver (hCRP-Tg mice) [22,23]. Previous studies have established that expression of human CRP in hCRP-Tg mice is regulated in a fashion very similar to that in humans [22–24]. We reasoned that, if circulating human CRP has a negative effect on the ability of peripherally administered human leptin to access the CNS, then the amount of human leptin entering the CSF of hCRP-Tg mice should be reduced compared with their wild-type littermates. The average serum concentration of human CRP in the male hCRP-Tg animals used for these tests was 15.62±3.73 μg/ml, a level that would be indicative of mild inflammation in humans. At the end of a 2-day infusion period (human leptin 0.6 mg/kg per day), the resultant level of human leptin in the sera of hCRP-Tg mice and their wild-type littermates did not differ (Figure 2A, left-hand panel). However, the CSF concentration of human leptin in the hCRP-Tg mice was less than half that in the wild-type littermates (Figure 2A, right-hand panel). Importantly, the CSF concentrations of leptin achieved in the wild-type and the hCRP-Tg mice were approximately 0.9% and 0.4% of the associated serum concentrations respectively (Figure 2B). Such ratios are comparable with those (∼0.5–1%) observed in human CSF and plasma samples [17,25–27].
Subcutaneously infused human leptin has impaired access to the CNS in hCRP-Tg mice
There have been reports of increased CRP in the CSF of humans with severe inflammation . To corroborate these observations, we challenged hCRP-Tg mice with intraperitoneally administered endotoxin [LPS (lipopolysaccharide)], and measured the human CRP concentration in both serum and CSF 24 h later (Figure 3). Importantly, these hCRP-Tg mice reportedly do not express human CRP in the brain . As expected, the serum level of CRP was elevated in the LPS-injected group (Figure 3, left-hand panel). In the same mice the CRP concentration in the CSF was also elevated (Figure 3, right-hand panel), although the net increase in the CSF was less than that in the serum. The concentrations of CRP detected in the CSF of hCRP-Tg mice were within the ranges reported in previous clinical studies of human CSF samples [18,29]. The combined data provide indirect evidence that circulating CRP can cross the BBB and direct evidence that its CSF concentration is reflective of the inflammatory state.
Human CRP is elevated in the CSF in association with an inflammatory response
We further investigated whether human CRP can attenuate leptin resistance in DIO (diet-induced obesity) mice. By installing micro-osmotic pumps into DIO mice and infusing either leptin alone (0.2 mg/kg per day) or leptin plus human CRP (0.2 and 1 mg/kg per day respectively), we recapitulated similar results on food intake and body weight as we did in ob/ob mice (Figures 4A and 4B). On day 4 of infusion, the serum concentration of human CRP reached 0.52±0.04 μg/ml in mice infused with CRP alone or CRP plus leptin. Serum concentrations of human leptin were comparable for mice infused with human leptin and mice infused with human leptin plus human CRP (Figure 4C, left-hand panel); the CSF concentration of human leptin was significantly lower in mice infused with human leptin plus CRP (Figure 4C, right-hand panel).
Subcutaneous co-infusion of human CRP and human leptin impairs the metabolic action of human leptin in DIO mice
We next evaluated the effects of human CRP on the physiological actions of human leptin within the CNS. To do so, we installed cannulas in the lateral ventricles of 8-week-old male ob/ob mice. After allowing the mice sufficient post-operative recovery (food-intake reached at least 90% of pre-operative level), we injected human CRP directly into the lateral ventricles (low dose: 1.2 μg per injection; high dose: 2.3 μg per injection) using the cannula as an access port. This i.c.v. (intracerebroventricular) injection of human CRP was followed 24 h later by injection of human leptin (100 ng/injection). In a parallel control experiment, we performed i.c.v. injection of PBS followed by human leptin. Food was returned to animals 1 h after each injection, and the animals’ food intake and body weight were measured 24 h later. Animals injected with PBS consumed ∼6.5 g of food per day (Figure 5A, white bars), and this rate was not affected by injection of human CRP (Figure 5A, diagonally striped bars). As expected, injection of human leptin (Figure 5A, black bars) or mouse leptin (Figure 5A, light grey bars) after treatment with PBS both led to a significant decrease in food-intake (Figure 5A). Importantly, injection of mice with 1.2 μg of human CRP blunted the satiety-inducing action of subsequently administered human leptin, and a higher dose of CRP (2.3 μg) blocked it completely (Figure 5A). I.c.v. injection of human CRP did not impair the satiety-inducing effect of mouse leptin, consistent with our previous observation that human CRP does not block the signalling of mouse leptin . As predicted on the basis of their food intake, central injection of human leptin or mouse leptin following PBS delivery led to significant weight reduction (Figure 5B). Also pre-injection of human CRP neutralized the weight-reducing function of human leptin, but not that of mouse leptin (Figure 5B). These data suggest that, when introduced directly into the central nervous system, human CRP can attenuate the physiological functions of human leptin in a dose-dependent manner.
Direct administration of human CRP and human leptin into the CNS impairs the metabolic action of human leptin in ob/ob mice
We also explored the temporal effects of i.c.v.-injected human CRP on the physiological actions of human leptin. After an initial injection of PBS or human CRP (2.3 μg/injection) 24 h earlier, human leptin (100 ng) was injected daily for 2 days. This treatment regimen led to the expected gradual decrease in food intake and body weight (Figures 6A and 6B respectively, left-hand panels). In comparison, administration of human CRP 24 h before the first injection of leptin delayed the satiety-inducing and weight-reducing actions of the adipokine and blunted its action (Figures 6A and 6B, right-hand panels). The data suggest that, once inside the CNS, human CRP has long-lasting effects on leptin action.
Time course of human CRP's effect on human leptin's metabolic actions
To determine whether the presence of human CRP in the CNS dampens signalling mediated by leptin, we assessed the ability of injected leptin to stimulate tyrosine phosphorylation of STAT3 in hypothalamic tissues. In these experiments, the hypothalamus was harvested approximately 2 h after i.c.v. delivery of leptin. Injection of human leptin 24 h after PBS administration induced a sharp increase in the levels of phospho-STAT3 (Figure 7A), which was dramatically diminished by pre-injection of 2.3 μg of human CRP (Figure 7A). Consistent with the metabolic observations shown above, pre-injection of human CRP did not block mouse leptin-stimulated STAT3 phosphorylation (Figure 7A). Like its effect on leptin induced STAT3 phosphorylation, introduction of human CRP into the CNS of ob/ob mice largely ameliorated leptin-suppressed hypothalamic expression of NPY (neuropeptide Y) mRNA (Figure 7B) and AgRP mRNA (Figure 7C). Meanwhile, the central presence of human CRP also attenuated the stimulatory effect of human leptin on POMC and SOCS-3 mRNA in the hypothalamus (Figures 7D and 7E). Further determination of the expression of MC4R and NPY1R, the predominant receptors for the anorexic or orexigenic actions of POMC or NPY respectively, found no significant changes of their expression in the same i.c.v. treatment schemes, suggesting that i.c.v. delivery of CRP, an inflammatory protein, did not influence these receptors (Figures 7F and 7G).
The presence of human CRP in the CNS ameliorates signalling and gene expression mediated by human leptin
In a separate group of ob/ob mice, we found that CSF samples extracted 24 h after injection of 2.3 μg of human CRP contained 47±8 ng of human CRP per ml (n=8), and CSF extracted 2 h after administration of 100 ng of human leptin contained 461±64 pg of human leptin per ml (n=4). Previous studies in humans have reported that baseline concentrations of CRP in human CSF were between 30 to 200 ng/ml [18,29], and that the concentrations of leptin in human CSF were in the range 100–500 pg/ml depending on the level of adiposity of the subjects [17,27]. Although these reported concentrations were measured using different methods and thus might contain systematic differences, it appears that the CSF concentrations of human CRP and human leptin achieved in the ob/ob mice we injected are well within the ranges reported for humans.
Leptin is an adipocyte-derived hormone that acts on hypothalamic regions of the brain to regulate energy intake and expenditure, fertility, and glucose and lipid metabolism [30–32]. Consequently, deficiencies in leptin or its receptors can lead to metabolic anomalies such as hyperphagia, gross obesity and insulin resistance. Paradoxically, the great majority of overweight and obese individuals have elevated levels of leptin rather than depressed ones . In a randomized clinical trial, it was shown that very high doses of leptin are required to reduce the body weight of obese people , a situation now known as ‘leptin resistance’. Leptin resistance, however, can be the consequence of leptin signalling deficiency at the hypothalamus of the CNS and/or impaired access of leptin to the CNS [17,27].
In addition to a direct correlation between circulating leptin and CRP levels, mounting molecular studies have shown that human CRP may have contributed to leptin resistance. Accordingly, leptin is able to stimulate CRP expression not only in human hepatocytes , but also in human coronary artery endothelial cells . Secondly, CRP is able to modulate leptin action. Besides our earlier study , Procopio et al.  demonstrated that the effects of human leptin on endothelial nitric oxide synthase activation were blunted by direct interaction with human recombinant CRP in human aortic endothelial cells and aorta from C57BL6J mice. Thirdly, genetic links have been established between the human leptin receptor locus, human CRP production and chronic inflammatory diseases [36,37].
By peripheral co-infusion of purified human CRP and recombinant human leptin into ob/ob mice, and peripheral infusion of human leptin into hCRP-Tg mice and diet-induced obesity mice, we generated direct evidence that human CRP in the periphery can reduce the amount of human leptin that gains access to the CNS. Leptin may gain access to the CNS by a number of mechanisms. It enters most of the brain and reaches its target neurons via transport across the BBB by combining with the short isoform of leptin receptor (OB-Ra) which has proven to be a saturable transport system [9,17]. The arcuate nucleus of hypothalamus is located very close to the ME (median eminence) with fenestrated capillaries, therefore CRP and leptin may both reach the hypothalamus by direct diffusion. Although the structural mechanisms remain to be elucidated further, we reason that the interaction of human leptin with human CRP, which has a large pentameric structure, blocks the binding of leptin to its receptors, which in turn hinders leptin receptor-mediated transport across the BBB. Such interaction is also predicted to block leptin's binding to its receptors in the hypothalamus, thus dampening its actions. In fact, evidence from our previous study  and another study  strongly supports such a concept that human CRP blocks human leptin binding to its receptor. This CRP–leptin interaction scenario is also supported by the observations we made in our animal studies, but whether it occurs in humans too remains to be established. Nevertheless, our animal models approximate the situation in humans very well as both the serum and CSF concentrations of human CRP and human leptin in the mice tested were comparable with those reported in human samples [17,18,27,29]. Furthermore, the ratios of CSF leptin to serum leptin were also very similar [17,25–27]. The blood concentration of human CRP is elevated in low-grade chronic inflammation, such as in obesity , thus our results fit an attractive theory which needs to be investigated further, i.e. that circulating CRP represents one of the primary factors that can block the access of human leptin to the CNS. The CRP–leptin interaction may indeed be a major reason for leptin resistance in obese subjects.
Our earlier study has shown the species-specific nature of CRP interacting with leptin , which is intriguing considering that the overall primary structure of leptin protein is strongly conserved from mice to humans (83% homology, GenBank® accession numbers ADM72802.1 and AAH69452.1 respectively; see the alignment in Supplementary Figure S2). Nonetheless, further dissection of their amino acid sequences revealed at least two stretches of ten amino acids or more (amino acids 1–10 or 121–132) where the homology falls significantly (50% for the stretch 1–10, and 42% for the stretch 121–132). We speculate that the divergence in these two regions dictates the species-specific nature of the CRP–leptin interaction, and that the corresponding synthetic peptides might potentially be applied as the agents to liberate the inhibitory effect of CRP on leptin's functions. In studying the permeability of BBB to mouse CRP in mouse models, Hsuchou et al.  did not find that mouse CRP crossed the BBB, and that the permeability of mouse leptin in the hypothalamus was not modulated by co-injection of mouse CRP . This is most likely to be due to the fact that, unlike human CRP and rat CRP, mouse CRP circulates at very low levels in blood , and is not equivalent to human CRP as the acute-phase reactant . Rather, mouse SAP (serum amyloid protein) functionally resembles human CRP . Thus hCRP-Tg mice become an ideal model for the studies of human CRP in inflammation, atherosclerosis and metabolism , particularly in the light of the interaction with human leptin.
In the present study, we have also demonstrated that human CRP, when directly administered into the CNS of ob/ob mice, inhibits the satiety-inducing and weight-reducing actions of human leptin, but not those of mouse leptin, in a dose-dependent manner. In keeping with these physiological data, the hypothalamic signalling of human leptin, such as STAT3 phosphorylation, and NPY and AgRP expression, was ameliorated, whereas POMC and SOCS-3 expression were attenuated by the introduction of human CRP into the CNS. Importantly, we confirmed that the CSF concentrations of human leptin achieved in the i.c.v.-injected mice were within the range reported previously in human CSF samples [17,27]. Likewise, we showed that the CSF concentrations of human CRP achieved in i.c.v.-injected mice were also within the range reported for CRP concentrations in human CSF [18,29].
Previous clinical studies have reported both the presence of CRP in CSF and its elevated concentrations there in association with inflammation [22,29,40]. These observations were replicated in this animal study, i.e. the CSF concentrations of human CRP were sharply increased in hCRP-Tg mice following an injection of LPS, a condition that generates high blood concentrations of CRP. By extension, the CSF levels of CRP in humans should also be increased in obese individuals due to their mild inflammatory status and their elevated CRP concentrations in circulation. We emphasize that future systematic studies based on a large clinical sample size are needed to validate this concept, but we are encouraged to see from a recent analysis of a limited number of samples (n=4), that the CSF levels of human CRP in the overweight and obese subjects are sharply higher than those in lean individuals (Zhao, A.Z., unpublished work).
Although peripheral tissues can play modulating roles [41–44], the hypothalamus is the primary site of action for leptin to regulate energy metabolism [45,46]. On the basis of existing clinical evidence and the results of the present study, we tentatively propose the following concept to explain the contribution of CRP to leptin resistance in human obesity. In overweight and obese individuals, an increase in body fat leads to an elevated concentration of leptin and other pro-inflammatory cytokines [such as IL-6 (interleukin 6)]. These in turn have been shown to stimulate the production of human CRP by hepatocytes in vitro  and in humans in vivo . Blood CRP then dampens the satiety-inducing and weight-reducing actions of human leptin at two different levels. First, in the periphery, circulating CRP interacts with human leptin and impedes its transport into the CNS via the BBB and ME. Secondly, once inside the CNS, CRP can attenuate the physiological function of human leptin. To overcome CRP-induced leptin resistance, fat stores are increased to produce and secrete more leptin, and the cycle repeats until a new energy balance is reached. This scenario might explain why recombinant human leptin failed to effectively reduce the body weight of obese people who often have high blood concentrations of CRP. We emphasize that the capacity of human CRP to inhibit human leptin's metabolic actions is likely to be restricted to periods of mild rather than acute inflammation. In the case of acute inflammation, the interpretation of energy balance would be complicated by the sharp rise in other cytokines, such as IL-6 and TNFα (tumour necrosis factor α), which are also greatly elevated in the CSF. These cytokines can exert potent anorexic effects via the hypothalamus, independent of the satiety actions of leptin [48–50].
In summary, the results of the present study suggest two potential novel molecular mechanisms contributing to leptin resistance in the CNS, i.e. peripheral impediment of leptin's access to the CNS and central inhibition of leptin signalling within the CNS. Such actions of human CRP, along with the other cellular mechanisms identified [11–14,32], probably contribute to the diminution of leptin's actions and thus promote human obesity. Disruption of the inhibitory effects of human CRP on the functions of leptin and/or reduction of hepatic CRP output in patients with mild inflammation may be a useful therapeutic approach in treating obesity.
Alexander Szalai, Fanghong Li and Allan Zhao conceived and designed the experiments. Jie Li, Dong Wei, Gangyi Yang, Ling Li and Mark McCrory performed the experiments. Jie Li, Dong Wei, Fanghong Li and Allan Zhao analysed the data. Jie Li, Alexander Szalai, Fanghong Li and Allan Zhao wrote the paper. All authors reviewed and commented on the paper before submission.
This work was supported by the National Program on Key Basic Research Project of China (973 Program) [grant number 2013CB945202 (to A.Z.Z. and F.L.)], the National Natural Science Foundation of China (NSFC) [grant numbers 81170780 (to A.Z.Z.) and 81372798 (to F.L.)], the Ph.D. Programs Foundation of the Ministry of Education of China [grant number 20113234110005 (to A.Z.Z.)], Scientific Support Program of Jiangsu Province [grant number BE2012756 (to A.Z.Z.)], Natural Science Foundation of Jiangsu Province of China (JSNFC) [grant number BK20130059 (to A.Z.Z.)], the High-Level Innovative Talents Reward from Jiangsu Province (to F.L.) and a Postdoctoral Fellowship Award from the American Heart Association [grant number 0725418U (to J.L.)].
body mass index
central nervous system
- hCRP-Tg mice
human CRP transgenic mice
melanocortin 4 receptor
neuropeptide Y receptor 1
suppressor of cytokine signalling 3
signal transducer and activator of transcription 3
These authors contributed equally to this work.