Central nervous action of interleukin-1 mediates activation of limbic structures and behavioural depression in response to peripheral administration of bacterial lipopolysaccharide.

Although receptors for the pro-inflammatory cytokine interleukin-1 have long been known to be expressed in the brain, their role in fever and behavioural depression observed during the acute phase response (APR) to tissue infection remains unclear. This may in part be due to the fact that interleukin-1 in the brain is bioactive only several hours after peripheral administration of bacterial lipopolysaccharide (LPS). To study the role of cerebral interleukin-1 action in temperature and behavioural changes, and activation of brain structures during the APR, interleukin-1 receptor antagonist (IL-1ra; 100 microg) was infused into the lateral brain ventricle 4 h after intraperitoneal (i.p.) LPS injection (250 microg/kg) in rats. I.p. LPS administration induced interleukin-1beta (IL-1beta) production in systemic circulation as well as in brain circumventricular organs and the choroid plexus. Intracerebroventricular (i.c.v.) infusion of IL-1ra 4 h after i.p. LPS injection attenuated the reduction in social interaction, a cardinal sign of behavioural depression during sickness, and c-Fos expression in the amygdala and bed nucleus of the stria terminalis. However, LPS-induced fever, rises in plasma corticosterone, body weight loss and c-Fos expression in the hypothalamus and caudal brainstem were not altered by i.c.v. infusion of IL-1ra. These findings, together with our previous observations showing that i.c.v. infused IL-1ra diffuses throughout perivascular spaces, where macrophages express interleukin-1 receptors, can be interpreted to suggest that circulating or locally produced brain IL-1beta acts on these cells to bring about behavioural depression and activation of limbic structures during the APR after peripheral LPS administration.

The role of the vagus nerve in mediating the long-term anorectic effects of leptin.

Leptin, the product of the obese (ob) gene, is mainly known for its regulatory role of energy balance by direct activation of hypothalamic receptors. Recently, its function in the acute control of food intake was additionally attributed to activation of the vagus nerve to regulate meal termination. Whether vagal afferent neurones are involved in longer term effects of leptin on food intake, however, remains undetermined. Using vagotomised (VGX) rats, we sought to clarify the contributions of vagal afferents in mediating the long-lasting effect of leptin on appetite suppression. Intraperitoneal (i.p.) injection of leptin (3.5 mg/kg) attenuated food intake at 4, 6, 8 and 24 h and body weight at 24 h postinjection in SHAM-operated rats; however, this response was not abrogated by vagotomy. In a separate study using immunohistochemistry, we observed leptin-induced Fos expression in the nucleus tractus solitarii, a brain structure where vagal afferent fibres terminate. This signal was not attenuated in VGX animals compared to the SHAM group. Moreover, leptin treatment led to a similar level of nuclear STAT3 translocation, a marker of leptin signalling, in the hypothalami of SHAM and VGX animals. In addition to the effects of leptin, vagotomy surgery itself resulted in a decrease of 24 h food intake. Analyses of brains from saline-treated VGX animals revealed a significant induction of Fos in the nucleus tractus solitarii and changes in agouti-related peptide and pro-opiomelanocortin mRNA expression in the hypothalamus compared to their SHAM counterparts, indicating that the vagotomy surgery itself induced a modification of brain activity in areas involved in regulating appetite. Collectively, our data suggest that vagal afferents do not constitute a major route of mediating the regulatory effect of leptin on food intake over a period of several hours.

The role of cytokines in mediating effects of prenatal infection on the fetus: implications for schizophrenia.

Maternal infections with bacterial or viral agents during pregnancy are associated with an increased incidence of schizophrenia in the offspring at adulthood although little is known about the mechanism by which maternal infection might affect fetal neurodevelopment. Exposure of pregnant rodents to the bacterial endotoxin, lipopolysaccharide (LPS), results in behavioral deficits in the adult offspring that are relevant to schizophrenia. It is however unknown whether these effects are due to the direct action of the inflammatory stimulus on the developing fetus, or due to secondary immune mediators (cytokines) activated at maternal/fetal sites. In this study we sought to elucidate the site of action of LPS, following a single intraperitoneal (i.p.) injection, in pregnant rats at gestation day 18. Animals received 5 muCi of iodinated LPS ((125)I-LPS) and its distribution was assessed in maternal/fetal tissues (1-8 h). In addition, induction of the inflammatory cytokines, TNF-alpha, IL-1beta and IL-6, was measured in maternal/fetal tissues following maternal LPS challenge (0.05 mg/kg, i.p.) (2-8 h). (125)I-LPS was detected in maternal tissues and placenta, but not the fetus. This distribution was accompanied by significant increases in TNF-alpha, IL-1beta and IL-6 in maternal plasma and placenta, but not in fetal liver or brain. A significant increase in IL-1beta was however detected in fetal plasma, possibly due to transfer from the maternal circulation or placenta. Collectively, these data suggest that effects of maternal LPS exposure on the developing fetal brain are not mediated by the direct action of LPS, but via indirect actions at the level of the maternal circulation or placenta.

Role of endogenous interleukin-1 receptor antagonist in regulating fever induced by localised inflammation in the rat.

1. Interleukin (IL)-1 is a mediator of host defence responses to inflammation and injury, including fever, but its sites of synthesis and action have not been fully elucidated. The actions of IL-1 are antagonised by IL-1 receptor antagonist (IL-1ra). The present study tested the hypothesis that IL-1 and IL-1ra are produced locally at sites of peripheral inflammation in rats, and that endogenous IL-1ra acts to limit the fever resulting from the inflammation. 2. Injection of lipopolysaccharide (LPS; 100 microg kg-1) into a subcutaneous air pouch (I.PO.) of rats induced a significant increase in body temperature. Virtually all (approximately 85 %) of the injected LPS was recovered from the pouch between 1 and 8 h (when the experiment was terminated) after injection of LPS, but LPS was undetectable (< 50 pg ml-1) in plasma at any time. Concentrations of immunoreactive IL-1alpha and IL-1beta were increased significantly in the pouch at 1, 2, 3, 5 and 8 h after injection of LPS, corresponding with the rise in body temperature and the fever peak. The appearance of IL-1ra was delayed until 2 h. Thereafter, the concentrations of IL-1beta and IL-1ra increased in parallel with the development of fever, while the concentrations of IL-1alpha remained constant. IL-1ra, but not IL-1alpha or IL-1bet, was detected in significant quantities in the plasma of LPS-injected animals. 3. Treatment of rats with an anti-IL-1ra serum (2 ml, I.PO.) at the time of injection of LPS (10 or 100 microg kg-1, I.PO.) abolished the appearance of IL-1ra in the circulation. Although neutralisation of endogenous IL-1ra did not affect the maximum body temperature reached after injection of submaximum (10 microg kg-1, I.PO.) or maximum (100 microg kg-1, I.PO.) doses of LPS, the duration of the fever was significantly prolonged, and was associated with a 3- to 4-fold increase in immunoreactive IL-1beta concentrations in the pouch fluid, but not in the plasma, at the 8 h time point. 4. These data show that effects of local (I.PO.) injection of LPS are not due to its action in the circulation or at distant sites (such as at the blood-brain barrier). These data also show that locally produced IL-1ra, in response to injection (I.PO.) of LPS, inhibits the production and/or action of locally produced IL-1beta. The ability of IL-1ra to limit the duration, rather than the magnitude of the fever, is consistent with its delayed production, relative to IL-IL-1ra, therefore, appears to play a key role in the resolution of fever induced by localised inflammatory responses.

The vagus nerve mediates behavioural depression, but not fever, in response to peripheral immune signals; a functional anatomical analysis.

Cytokines act on the brain to induce fever and behavioural depression after infection. Although several mechanisms of cytokine-to-brain communication have been proposed, their physiological significance is unclear. We propose that behavioural depression is mediated by the vagus nerve activating limbic structures, while fever would primarily be due to humoral mechanisms affecting the preoptic area, including interleukin-6 (IL-6) action on the organum vasculosum of the laminae terminalis (OVLT) and induction of prostaglandins. This study assessed the effects of subdiaphragmatic vagotomy in rats on fever, behavioural depression, as measured by the social interaction test, and Fos expression in the brain. These responses were compared with induction of the prostaglandin-producing enzyme cyclooxygenase-2 and the transcription factor Stat3 that translocates after binding of IL-6. Vagotomy blocked behavioural depression after intraperitoneal injection of recombinant rat IL-1beta (25 microg/kg) or lipopolysaccharide (250 microg/kg; LPS) and prevented Fos expression in limbic structures and ventromedial preoptic area, but not in the OVLT. Fever was not affected by vagotomy, but associated with translocation of Stat3 in the OVLT and cyclooxygenase-2 induction around blood vessels. These results indicate that the recently proposed vagal link between the immune system and the brain activates limbic structures to induce behavioural depression after abdominal inflammation. Although the vagus might play a role in fever in response to low doses of LPS by activating the ventromedial preoptic area, it is likely to be overridden during more severe infection by action of circulating IL-6 on the OVLT or prostaglandins induced along blood vessels of the preoptic area.

Circulating interleukin-6 mediates the febrile response to localised inflammation in rats.

Interleukin (IL)-6 is an important mediator of the host response to disease and has been proposed, largely based upon circumstantial evidence, as the principal endogenous circulating pyrogen responsible for activating CNS mechanisms in fever during infection and inflammation. In the present investigation, we studied the role of peripheral IL-6 in fever and its relationship with IL-1, itself an important endogenous pyrogen and a potent stimulus of IL-6 production. Injection of lipopolysaccharide (LPS) into a sterile, subcutaneous air pouch (i.po.) in rats evoked an increase in body temperature which peaked at 3 h, and which was abolished in animals pretreated (intraperitoneally) with IL-6 antiserum. The increase in body temperature was accompanied by a significant elevation in concentrations of (immunoreactive) IL-1 and IL-6 at the site of inflammation (pouch), but only IL-6 in the circulation and cerebrospinal fluids. We propose that much of the circulating IL-6 originates at the site of inflammation, since injection of human recombinant (hr)IL-6 (i.po.) was detected (10 min after the injection) in the plasma using an ELISA specific for human IL-6. However, despite the relatively high concentration of IL-6 injected (25 microg kg-1, i.po.), this cytokine had no effect on body temperature when injected alone, but did induce fever when co-injected with a non-pyrogenic dose (when given alone) of IL-1beta, and exacerbated the fever to a pyrogenic dose of IL-1beta. The results from the present study demonstrate that IL-6 is a circulating endogenous pyrogen during LPS-induced fever, which acts in concert with IL-1beta at the local site of inflammation, before entering the circulation. Circulating IL-6 can then activate CNS mechanisms resulting in the development of the febrile response during disease.

Central administration of rat IL-6 induces HPA activation and fever but not sickness behavior in rats.

Interleukin (IL)-6 has been proposed to mediate several sickness responses, including brain-mediated neuroendocrine, temperature, and behavioral changes. However, the exact mechanisms and sites of action of IL-6 are still poorly understood. In the present study, we describe the effects of central administration of species-homologous recombinant rat IL-6 (rrIL-6) on the induction of hypothalamic-pituitary-adrenal (HPA) activity, fever, social investigatory behavior, and immobility. After intracerebroventricular administration of rrIL-6 (50 or 100 ng/rat), rats demonstrated HPA and febrile responses. In contrast, rrIL-6 alone did not induce changes in social investigatory and locomotor behavior at doses of up to 400 ng/rat. Coadministration of rrIL-6 (100 ng/rat) and rrIL-1beta (40 ng/rat), which alone did not affect the behavioral responses, reduced social investigatory behavior and increased the duration of immobility. Compared with rhIL-6, intracerebroventricular administration of rrIL-6 (100 ng/rat) induced higher HPA responses and early-phase febrile responses. This is consistent with a higher potency of rrIL-6, compared with rhIL-6, in the murine B9 bioassay. We conclude that species-homologous rrIL-6 alone can act in the brain to induce HPA and febrile responses, whereas it only reduces social investigatory behavior and locomotor activity in the presence of IL-1beta.

Interleukin-1 mediates a rapid inflammatory response after injection of adenoviral vectors into the brain.

Adenovirus-mediated gene transfer into the brain is associated with significant inflammation and activation of anti-vector and anti-transgene immune responses that curtail the gene delivery of adenoviruses and therapeutic efficacy. Elucidating the molecular mediators of inflammatory and immune responses to adenoviruses injected into the brain should allow us to inhibit their inflammatory actions, thereby reducing vector clearance and enhance adenoviral-mediated gene transfer into the CNS. Cytokines are primary mediators of the immune response and are released during inflammation. Here we report for the first time that injection of replication-deficient adenovirus vectors into the cerebral ventricles of rats causes a rapid increase in body temperature. This fever response precedes any vector-encoded transgene expression and occurs with vectors encoding no transgene, as well as with vectors encoding a therapeutic transgene i.e., HSV1-thymidine kinase. No fever is detected after infection of the striatum, an important brain target in studies on neurodegeneration. After infection of the brain ventricles, CSF levels of immunoreactive tumor necrosis factor (TNF)-alpha and interleukin (IL)-1beta increase significantly (up to 300-fold). In the hypothalamus, the locus of thermoregulation in the brain, only IL-1beta and IL-6 are significantly elevated. A neutralizing TNF-alpha antibody has no effect on adenovirus-induced fever. However, pretreatment with either the IL-1 receptor antagonist or the cyclooxygenase inhibitor flurbiprofen completely abolishes adenovirus-induced fever, suggesting that IL-1 and prostaglandins are direct mediators of this response. These results are the first to demonstrate that IL-1, but not TNF-alpha, is the main mediator of a very early inflammatory response to adenovirus in the brain.

Fever and production of cytokines in response to repeated injections of muramyl dipeptide in guinea-pigs.

Fever and systemic plasma levels of the cytokines tumour necrosis factor alpha (TNF) and interleukin-6 (IL-6) were measured in guinea-pigs in response to single or repeated intramuscular injections of 100 micrograms/kg muramyl-dipeptide (MDP). In a pilot study (experiment 1), MDP-induced fever was monitored for 8 h. The first fever phase 90-360 min after injection of MDP was followed by the second phase which continued beyond the duration of this experiment. High circulating levels of TNF and IL-6 were detected just before body core temperature started to rise. Within the next 90 min TNF declined again by more than 90% while IL-6 remained elevated. In experiment 2, the effects of repeated injections of MDP (5 times at intervals of 3 days) on the same parameters were investigated. In this paradigm, the febrile response started earlier (60 min after injection) and the first phase of fever remained manifest until 360 min after injection, while the late phase, measured 360-720 min after injection, was attenuated. Circulating, bioactive TNF and IL-6, measured 60 and 180 min after MDP was administered, were the same in response to the first, third, and fifth injection. In experiment 3, the influence of five repeated MDP injections on the abdominal temperature was measured for 22 h, and circulating cytokines were analysed before (360 min after injection) and during (480 min after injection) the late phase of MDP-induced fever. The late phase of MDP-induced fever 7-22 h after injection was attenuated in response to the second and further administrations of this pyrogen. At 6 h after the first, third, and fifth administration of MDP, only traces of TNF alpha were measured, 2 h later no bioactive TNF was detected at all. At these times also IL-6 declined again, compared with the activity of this cytokine measured during the early phase of MDP fever, but was still present in elevated amounts. Compared with the values measured in response to the third and fifth injections of MDP, circulating IL-6 was higher 360 min and 480 min after the first injection. It remains speculative whether the longer duration of elevated IL-6 in plasma is related to the development of the long-lasting, late phase of MDP-induced fever, which was only observed after the first of five repeated injections of MDP at intervals of 3 days.

Sites of action of IL-1 in the development of fever and cytokine responses to tissue inflammation in the rat.

1. The objective of the present study was to determine the sites of action of the cytokine, interleukin-1 (IL-1), in the febrile response to local inflammation in the rat, by comparing the importance of IL-1 in the local tissues, the circulation and the brain. This was achieved by injecting lipopolysaccharide (LPS, 100 micrograms kg-1) into a subcutaneous air pouch and testing the effects of blocking IL-1 action with the human recombinant interleukin-1 receptor antagonist (IL-1ra) injected either into the air pouch, intraperitoneally (1 mg kg-1, 0 + 1 h, i.p.), or intracerebroventricularly (200 micrograms/rat, 0 + 1 h, i.c.v.). 2. To investigate the effect of IL-1ra on fever and the induction of local and circulating cytokines (IL-1 and IL-6), separate experiments were performed in which groups of animals were killed 1.5, 3 or 5 h after LPS injection. Plasma and pouch fluid samples were collected for bioassay of IL-1 and IL-6. 3. Injection of LPS into the air pouch significantly increased (1.5 degrees C) body temperature, local (air pouch) concentrations of bioactive IL-1 and IL-6, and circulating bioactive IL-6, compared to saline-treated controls. 4. Injection of IL-1ra into the pouch significantly attenuated LPS fever (P < 0.001). This decrease in body temperature was associated with significant inhibition of local IL-1 bioactivity 1.5 (96%), 3 (84%) and 5 h (72%), and in bioactive IL-6 in pouch lavage fluid 1.5 (45%) and 5 h (35%), after LPS injection. The concentration of bioactive IL-6 in the plasma was significantly reduced (39%) at 3 h, when temperature was approaching the maximal value. 5. Both systemic (i.p.) and central (i.c.v.) administration of IL-1ra significantly attentuated LPS fever (P < 0.05). However, it had no effect on either local concentrations of bioactive IL-1 or IL-6, or circulating IL-6, at any of the sample points. 6. These data suggest that IL-1 is released locally, at the site of tissue inflammation and that it is an important mediator of the febrile response to local inflammation. The results also indicate that IL-1 produced locally may contribute to the production of IL-6 which is released into the circulation, and that IL-1 has important actions in the generation of fever at other sites, including the brain.

Local cytokine induction by LPS in the rat air pouch and its relationship to the febrile response.

Peripheral induction of cytokines is a critical event in the induction of febrile responses. The sequence of induction and site of action of these cytokines, however, remain unclear. The objective of the present study was to investigate the kinetics of local and systemic production of interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF), with the aim of identifying the relationship between these cytokines and the febrile response induced by injection of lipopolysaccharide (LPS) into a subcutaneous air pouch in the rat. Intrapouch injection of LPS induced dose-dependent fevers and increases in the concentration of bioactive IL-6 in the plasma. Further studies using 100 microg/kg LPS demonstrated significant increases in local (air pouch) concentrations of bioactive IL-1, TNF and IL-6, and circulating IL-6. No significant increases in TNF or IL-1 were detected in the plasma of the same animals. Local TNF was induced rapidly and peaked 1 h after LPS injection. The kinetics of local IL-1 and IL-6 induction were similar and both peaked after 3 h. The rise in local IL-6 preceded that of plasma IL-6 and reached a peak concentration that was 25-fold higher than that observed in the plasma. The data indicate that IL-1 and TNF act locally at the site of inflammation and that locally induced IL-6 is the important systemic mediator of the response.

Febrile response to tissue inflammation involves both peripheral and brain IL-1 and TNF-alpha in the rat.

We investigated the role and interaction between tumor necrosis factor (TNF)-alpha, interleukin (IL)-1, and IL-6 in the development of fever and their involvement in brain and systemic pathways in response to localized tissue inflammation caused by injection of turpentine (TPS) in the rat. Intramuscular injection of 10 microl TPS caused significant increases in body temperature, of up to 2 degrees C, compared with saline-treated animals. Fevers were maximal 7-8 h after injection and were preceded by significant increases in plasma bioactive IL-6. No changes in circulating bioactive IL-1 or TNF-alpha were detected. Systemic injection of IL-1 receptor antagonist (IL-1ra, 2 mg/kg i.p.) or anti-TNF-alpha antiserum (0.4 ml i.v.) almost completely abolished the febrile responses to TPS over 8 h and markedly inhibited the rise in plasma IL-6 bioactivity measured 6 h after TPS. To test the involvement of brain cytokines, anti-TNF-alpha antiserum and IL-1ra were injected intracerebroventricularly. Injections of anti-TNF-alpha antiserum (3 microl/rat i.c.v.) or IL-1ra (400 microg/kg i.c.v.) significantly (P < 0.01 and P < 0.05, respectively) inhibited fever induced by TPS. These data suggest that both localized peripheral and brain IL-1 and TNF-alpha are involved directly in the pyrogenic response to inflammation. The results indicate that, in the periphery, IL-1 and TNF-alpha cause increased production of IL-6, the most likely candidate as a circulating endogenous pyrogen.

The role of TNF-alpha in fever: opposing actions of human and murine TNF-alpha and interactions with IL-beta in the rat.

1. The role of tumour necrosis factor-alpha (TNF-alpha) in fever is controversial. Some studies have indicated that TNF-alpha acts as a cryogen to inhibit fever, while others suggest that TNF-alpha is an endogenous pyrogen which mediates fever. The majority of studies in experimental animals supporting a cryogenic action have been conducted using human (h)TNF-alpha, which has been shown to bind only to one (p55) of the two TNF-alpha receptors in rodents. 2. The aim of the present investigation was to study the role of TNF-alpha in fever by comparing effects of hTNF-alpha, which binds only to the p55 receptor, with those of murine (m) TNF-alpha, which binds to both p55 and p75 TNF-alpha receptors, and to investigate the relationship between TNF-alpha and interleukin-1 (IL-1), an important endogenous pyrogen. 3. Injection of hTNF-alpha (0.3-10 micrograms kg-1, i.p.) had no effect on core temperature in conscious rats (measured by remote radiotelemetry), whereas mTNF-alpha (3 micrograms kg-1) induced fever which was maximal 1 h after the injection (38.2 +/- 0.2 degrees C compared to 37.3 +/- 0.1 degrees C in controls). Intracerebroventricular (i.c.v.) administration of either form of TNF-alpha elicited dose-dependent fever at doses higher than 0.12 microgram kg-1. 4. Peripheral injection of hIL-1 beta (1 microgram kg-1) resulted in fever (38.3 +/- 0.2 degree C compared to 37.2 +/- 0.1 degrees C in controls at 2 h), which was significantly attenuated (P < 0.01) by co-administration of a sub-pyrogenic dose of hTNF-alpha (1 microgram kg-1), but was unaffected by co-administration of mTNF-alpha (0.1 or 0.3 microgram kg-1, i.p.). In contrast, intracerebroventricular (i.c.v.) co-administration of a sub-pyrogenic dose (0.12 microgram kg-1) of hTNF-alpha did not attenuate fever induced by intraperitoneal (i.p.) injection of IL-1 beta, and sub-pyrogenic dose (0.12 microgram kg-1, i.c.v.) of mTNF-alpha significantly prolonged the febrile response to IL-1 beta. Pretreatment of animals with anti-TNF-alpha antiserum (i.c.v.) did not affect the febrile response to systemic IL-1 beta. 5. Animals injected i.p. with a pyrogenic dose of mTNF-alpha developed fever (38.2 +/- 0.2 degrees C compared to 37.3 +/- 0.1 degrees C in controls 2 h after the injection) that was completely abolished by peripheral administration of IL-1ra (2 mg kg-1, P < 0.001), while i.c.v. administration of IL-1ra (400 micrograms/rat) did not affect mTNF-alpha-induced fever. 6. These data indicate that endogenous TNF-alpha is probably a pyrogen and that previous results suggesting cryogenic actions of TNF-alpha resulted from the use of a heterologous protein in the rat. The markedly contrasting effects of mTNF-alpha and TNF-alpha could result from different interactions with the two TNF-alpha receptor subtypes. The data also suggest that fever induced by exogenous TNF-alpha is mediated via release of IL-1 beta in peripheral tissues, but not in the brain.

Tumor necrosis factor-alpha and fever after peripheral inflammation in the rat.

The involvement of endogenous tumor necrosis factor-alpha (TNF-alpha) in the pyrogenic [i.e., rise in colonic temperature (Tc)] and thermogenic [increase in oxygen consumption (VO2)] responses to inflammation was investigated in rats subjected to an intramuscular injection of turpentine. Turpentine administration caused a rise in Tc and VO2 within 2 h (0.9 +/- 0.1 degrees C, 27 +/- 2%, respectively). Eighteen to twenty hours after turpentine, the magnitude of these responses had increased (2.3 degrees C fever and a 28% increase in metabolic rate compared with control animals) and was associated with marked inflammation in the injected limb. A rapid (by 4 h) and sustained rise in the plasma concentration of the endogenous pyrogen IL-6, but not TNF-alpha, was also observed. Intravenous pretreatment with a TNF-alpha antiserum attenuated the rise in Tc observed 2, 8, and 18 h after turpentine injection and almost abolished the hypermetabolic response observed at 18 h. In addition, the TNF-alpha antiserum inhibited the peak rise (8 h) in plasma IL-6 by 76%. These findings indicate that endogenous TNF-alpha is involved in fever and hypermetabolism during inflammation and that it may exert these effects by inducing the release of IL-6 into circulation.