Issue published May 1, 1996

M utations in the gene defective in Fanconi anemia complementation group C, FAC, are responsible for a subset of Fanconi anemia, a group of autosomal recessive disorders characterized by chromosomal instability, hypersensitivity to cross-linking agents, and cancer susceptibility. Although abnormalities in DNA repair have been suspected, localization of the FAC gene product to the cytoplasm has cast doubt on such a mechanism. Monitoring of interstrand DNA cross-linking shows that the predominant defect in group C cells is in the initial induction of cross-links, not in repair synthesis. Both the cross-linking defect and the enhanced cytotoxicity of cross-linkers on Fanconi anemia group C cells are corrected completely by cytoplasmic isoforms of the FAC protein, but not by an isoform targeted to the nucleus. The ability of FAC to correct these phenotypic abnormalities reaches a maximum threshold despite overexpression leading to higher levels of cytosolic protein. These results demonstrate that cytoplasmic localization is essential for the intracellular activity of the FAC protein. It is proposed that this activity is coupled to a cytoplasmic defense mechanism against a specific class of genotoxic agents.

R ecently, a new human collagenase, collagenase-3 has been identified. Since collagen changes are of particular importance in cartilage degeneration, we investigated if collagenase-3 plays a role in osteoarthritis (OA). Reverse transcriptase-PCR analysis revealed that in articular tissues collagenase-3 was expressed by the chondrocytes but not by the synoviocytes. Northern blot analysis of the chondrocyte mRNA revealed the presence of two major gene transcripts of 3.0 and 2.5 kb, and a third one of 2.2 kb was occasionally present. Compared to normal, OA showed a significantly higher (3.0 kb, P < or = 0.05; 2.5 kb, P < or = 0.03) level of collagenase-3 mRNA expression. Collagenase-3 had a higher catalytic velocity tate (about fivefold) than collagenase-1 on type II collagen. With the use of two specific antibodies, we showed that human chondrocytes had the ability to produce collagenase-3 as a proenzyme and as a glycosylated doublet. The chondrocyte collagenase-3 protein is produced in a significantly higher (P < or = 0.04) level in OA (approximately 9.5-fold) than in normal. The synthesis and expression of this new collagenase could also be modulated by two proinflammatory cytokines, IL-1 beta and TNF-alpha, in a time- and dose-dependent manner. This study provides novel and interesting data on collagenase-3 expression and synthesis in human cartilage cells and suggest its involvement in human OA cartilage patho-physiology.

o b protein is hypothesized to be a circulating feedback signal in the regulation of energy balance. Obese, overfed rats have high levels of ob mRNA expression and suppressed voluntary food intake, indicating the presence of a potent satiety factor. The objectives of this experiment were to determine whether feeding rats their normal daily intake in three meals, compared with ad libitum feeding, increased ob mRNA expression and to determine the degree of obesity required to stimulate expression of ob mRNA. Rats were fed ad libitum, were tube-fed their normal intake in three meals a day, or were tube-fed twice normal intake, ob mRNA was measured by Northern blot analysis after 0, 2, 7, 14, 21, and 32 d of tube-feeding. After only 2 d ob mRNA was threefold higher in tube-fed animals than in ad libitum controls. By day 21 there was a further increase in ob mRNA expression in overfed rats which were at 130% control weight. These results suggest that a metabolic consequence of meal-feeding increases ob mRNA expression in the absence of increased food intake or weight gain. There is a further increase in ob mRNA expression once significant obesity is established.

C lonal expansion of CD4+ T cells is a characteristic finding in patients with RA and is only infrequently found in patients with psoriatic arthritis and healthy controls. Expanded CD4+ clonotypes are present in the blood, infiltrate into the joint, and persist over years. We have not addressed the question of whether the expanded clonotypes have unique functional and phenotypic properties which may explain the preferential in vivo expansion in RA. In contrast to most CD4+ T cells, expanded clonotypes lacked the expression of the CD28 and CD7 cell surface molecules. Accordingly, the subsets of CD4+ CD28- (9.7 vs 1.7, P = 0.00002) and CD4+ CD7- T cells (21.5 vs 12.26, P = 0.018) were increased in RA patients compared with age-matched normal individuals. Despite the lack of CD28 expression, clonally expanded CD4+ T cells were not anergic but proliferated in response to immobilized anti-CD3 and could be maintained in tissue culture. In vivo expanded CD4+ T cells were autoreactive to ubiquitously distributed autoantigens. They responded in an autologous mixed lymphocyte reaction, and T cell clones isolated from selected patients proliferated to autologous peripheral blood adherent cells. These data suggest that in RA patients selected CD4+ T cells which share the CD7- CD28- phenotype escape from peripheral tolerance.

A lpha-Melanocyte-stimulating hormone (MSH) is a potent anti-inflammatory agent in many models of inflammation, suggesting that it inhibits a critical step common to different forms of inflammation. We showed previously that alpha-MSH inhibits nitric oxide (NO) production in cultured macro-phages. To determine how alpha-MSH acts in vivo, we induced acute hepatic inflammation by administering endotoxin (LPS) to mice pretreated with Corynebacterium parvum, alpha-MSH prevented liver inflammation even when given 30 min after LPS administration. To determine the mechanisms of action of alpha-MSH, we tested its influence on NO, infiltrating inflammatory cells, cytokines, and chemokines. Alpha-MSH inhibited systemic NO production, hepatic neutrophil infiltration, and increased hepatic mRNA abundance for TNF alpha, and the neutrophil and monocyte chemokines (KC/IL-8 and MCP-1). We conclude that alpha-MSH prevents LPS-induced hepatic inflammation by inhibiting production of chemoattractant chemokines which then modulate infiltration of inflammatory cells. Thus, alpha-MSH has an effect very early in the inflammatory cascade.

A rginine vasopressin (AVP) acts synergistically with corticotropin-releasing hormone (CRH) to stimulate ACTH release from the anterior pituitary. In a previous study of bilateral simultaneous inferior petrosal sinus (IPS) sampling in healthy human subjects, we observed lateralized ACTH secretion, suggesting lateralized secretion of an ACTH-regulating hypothalamic factor. To investigate this possibility, we measured ACTH, CRH, AVP, and oxytocin (OT) levels in the IPS and the peripheral circulation in nine normal volunteers, before and after 1 microgram/kg i.v. bolus ovine CRH (oCRH). At baseline, ACTH, AVP, and OT exhibited a significant (P < 0.05) two to threefold intersinus gradient (ISG), indicating the existence of a dominant petrosal sinus. Endogenous CRH was undetectable in all samples. Despite similar exogenous oCRH levels in both petrosal sinuses, oCRH caused a significant increase (P < 0.001) in the ACTH ISG (15.8 +/- 5.6, mean +/- SEM), suggesting increased responsiveness of one dominant side of the anterior pituitary. This was associated with an ipsilateral CRH-induced AVP release and a significant increase (P < 0.01) in the AVP ISG (8.6 +/- 2.3), suggesting lateralized AVP secretion by the hypothalamus. Furthermore, the increased AVP ISG after oCRH correlated strongly with the ACTH ISG (r = 0.92, P < 0.01). oCRH administration did not affect OT. These findings suggest that there is a dominant petrosal sinus in healthy volunteers that appears to reflect a dominant side of the adenohypophysis, characterized by increased functional activity and/or responsiveness of the pituitary corticotrophs. This may reflect lateralized hypothalamic and/or suprahypothalamic function resulting in CRH-responsive lateralized secretion of AVP from parvocellular and/or magnocellular axons in the median eminence and the posterior pituitary. Although the functional and teleologic significance of these findings remains to be investigated, our data suggest a novel mechanism for CRH-mediated ACTH release, namely CRH-induced release of AVP which then enhances CRH action on the corticotrophs. Furthermore, our data represent the first direct evidence for the concept of brain lateralization with respect to neuroendocrine secretion.

I nterleukin 1 and nitric oxide (NO) from infiltrating macrophages and activated mesangial cells may act in concert to sustain and promote glomerular damage. To evaluate if such synergy occurs, we evaluated the effect if IL-1 beta and NO on the formation of prostaglandin (PG)E2 and cyclooxygenase (COX) expression. The NO donors, sodium nitroprusside and S-nitroso-N-acetylpenicillamine, alone did not increase basal PGE2 formation. However, these compounds amplified IL-1 beta-induced PGE2 production. Similarly, sodium nitroprusside and S-nitroso-N-acetylpenicillamine by themselves did not induce mRNA and protein for COX-2, the inducible isoform of COX; however, they both potentiated IL-1 beta-induced mRNA and protein expression of COX-2. The stimulatory effect of NO is likely to be mediated by cGMP since (a) an inhibitor of the soluble guanylate cyclase, methylene blue, reversed the stimulatory effect of NO donors on COX-2 mRNA expression; (b) the membrane-permeable cGMP analogue, 8-Br-cGMP, mimicked the stimulatory effect of NO donors on COX-2-mRNA expression; and (c) atrial natriuretic peptide, which increases cellular cGMP by activating the membrane-bound guanylate cyclase, also amplified IL-1 beta-induced COX-2 mRNA expression. These data indicate a novel interaction between NO and COX pathways.

I mmunization with cardiac myosin induces T cell-mediated myocarditis in genetically predisposed mice and serves as a model for autoimmune heart disease. This study was undertaken to identify pathogenic epitopes on the myosin molecule. Our approach was based on the comparison of the pathogenicity between cardiac (alpha-)myosin and soleus muscle (beta-)myosin. We show that alpha-myosin is the immunodominant isoform and induces myocarditis at high severity and prevalence whereas beta-myosin induces little disease. Therefore the immunodominant epitopes of alpha-myosin must reside in regions of different amino acid sequence between alpha- and beta-myosin isoforms. Cardiac myosin peptides corresponding to these regions of difference were synthesized and tested for their ability to induce inflammatory heart disease. Three pathogenic peptides were identified. One peptide that is located in the head portion of the molecule induced severe myocarditis, whereas two others that reside in the rod portion possessed only minor pathogenicity. The identification of pathogenic epitopes on the cardiac myosin molecule will allow detailed studies on the recognition of this antigen by the immune system and might be used to downmodulate ongoing heart disease.

W e investigated the role of the costimulatory molecules, CD40 and its ligand CD40L, in the pathogenesis of human SLE. In comparison to normal subjects or patients in remission, PBMC from active lupus patients had a 21-fold increase in the frequency of CD40L-expressing, CD4+T cells. However, the expression of CD40L induced in either lupus or normal T cells by mitogenic stimulation could be down-regulated equally well by CD40 molecules on autologous B cells. Active lupus patients also had a 22-fold increase in percentage of CD8+ T cells expressing CD40L, consistent with their unusual helper activity in SLE. Surprisingly, patients with active lupus had a 20.5-fold increase in B cells that spontaneously expressed high levels of CD40L, as strongly as their T cells. Although lupus patients in remission had low levels of CD40L+ cells in the range of normal subjects, mitogen-induced upregulation of CD40L expression in the T and B cells was markedly greater than normal, suggesting an intrinsic defect. A mAb to CD40L blocked significantly the ability of lymphocytes from lupus patients with active and established disease to produce the pathogenic variety of antinuclear autoantibodies in vitro, bolstering the possibility of anti-CD40L immunotherapy for lupus. Future studies on the hyperexpression of CD40L could elucidate a regulatory defect in the pathogenic T and B cells of lupus.

5 1p1 is an immunoglobulin VH gene that is frequently expressed in B cell chronic lymphocytic leukemia and early in B cell ontogeny. The 51p1 gene locus is highly polymorphic, consisting of 13 alleles that can be classified as being either 51p1-related or hy1263-related, based on distinctive sequence motifs in the second complementarity determining region. Two of the 51p1-related genes usually occur as a linked pair on the same haopltype, resulting from gene duplication. Consequently, a person can have a total of zero to four copies of 51p1-related genes. These genes are detectable in genomic DNA by sequence-specific RFLP analysis using oligonucleotide probes. Ig encoded by nonmutated 51p1-related genes can be detected by G6, a murine antiidiotypic mAb. We have now studied lymphocytes from 35 human tonsils to examine the relation between the number of 51p1-related germlime gene copies and the proportion of IgD-bearing tonsillar B cells that react with G6. All subjects who had zero copies of 51p1-related genes lacked any G6-reactive B cells, whereas those with four copies of 51p1-related genes had the highest proportions of G6-positive IgD B cells, up to 11.4%. Subjects with intermediate gene doses had intermediate proportions of G6-reactive B cells. Over the entire data set, the percentage of IgD-bearing B cells that reacted with G6 was proportional to the 51p1-related gene copy number (r = 0.92, p < 0.001), with each copy accounting for 2.4-4.0% of the IgD-bearing B cells. We conclude that 51p1-related genes are expressed by a relatively large percentage of IgD+ tonsillar B cells and this percentage is proportional to the germline copy number of 51p1-related genes.

A new experimental approach was used to determine whether a eucaloric, low fat, high carbohydrate diet increases fatty acid synthesis. Normally volunteers consumed low fat liquid formula diets (10% of calories as fat and 75% as glucose polymers, n = 7) or high fat diets (40% of calories as fat and 45% as glucose polymers, n = 3) for 25 d. The fatty acid composition of each diet was matched to the composition of each subject's adipose tissue and compared with the composition of VLDL triglyceride. By day 10, VLDL triglyceride was markedly enriched in palmitate and deficient in linoleate in all subjects on the low fat diet. Newly synthesized fatty acids accounted for 44 +/- 10% of the VLDL triglyceride. Mass isotopomer distribution analysis of palmitate labeled with intravenously infused 13C-acetate confirmed that increased palmitate synthesis was the likely cause for the accumulation of triglyceride palmitate and "dilution" of linoleate. In contrast, there was minimal fatty acid synthesis on the high diet. Thus, the dietary substitution of carbohydrate for fat stimulated fatty acid synthesis and the plasma accumulation of palmitate-enriched, linoleate-deficient triglyceride. Such changes could have adverse effects on the cardiovascular system.

E rythropoietin (rHuEPO) has proven to be effective in the treatment of anemia of chronic renal failure (CRF). Despite improving the quality of life, peak oxygen uptake after rHuEPO therapy is not improved as much as the increase in hemoglobin concentration ([Hb)] would predict. We hypothesized that this discrepancy is due to failure of O2 transport rates to rise in a manner proportional to [Hb]. To test this, eight patients with CRF undergoing regular hemodialysis were studied pre- and post-rHuEPO ([Hb] = 7.5 +/- 1.0 vs. 12.5 +/- 1.0 g x dl-1) using a standard incremental cycle exercise protocol. A group of 12 healthy sedentary subjects of similar age and anthropometric characteristics served as controls. Arterial and femoral venous blood gas data were obtained and coupled with simultaneous measurements of femoral venous blood flow (Qleg) by thermodilution to obtain O2 delivery and oxygen uptake (VO2). Despite a 68% increase in [Hb], peak VO2 increased by only 33%. This could be explained largely by reduced peak leg blood flow, limiting the gain in O2 delivery to 37%. At peak VO2, after rHuEPO, O2 supply limitation of maximal VO2 was found to occur, permitting the calculation of a value for muscle O2 conductance from capillary to mitochondria (DO2). While DO2 was slightly improved after rHuEPO, it was only 67% of that of sedentary control subjects. This kept maximal oxygen extraction at only 70%. Two important conclusions can be reached from this study. First, the increase in [Hb] produced by rHuEPO is accompanied by a significant reduction in peak blood flow to exercising muscle, which limits the gain in oxygen transport. Second, even after restoration of [Hb], O2 conductance from the muscle capillary to the mitochondria remains considerably below normal.

A fter erythropoietin (rHuEPO) therapy, patients with chronic renal failure (CRF) do not improve peak O2 uptake (VO2 peak) as much as expected from the rise in hemoglobin concentration ([Hb]). In a companion study, we explain this phenomenon by the concurrent effects of fall in muscle blood flow after rHuEPO and abnormal capillary O2 conductance observed in CRF patients. The latter is likely associated with a poor muscle microcirculatory network and capillary-myofiber dissociation due to uremic myopathy. Herein, cellular bioenergetics and its relationships with muscle O2 transport, before and after rHuEPO therapy, were examined in eight CRF patients (27 +/- 7.3 [SD] yr) studied pre- and post-rHuEPO ([Hb] = 7.8 +/- 0.7 vs. 11.7 +/- 0.7 g x dl-1) during an incremental cycling exercise protocol. Eight healthy sedentary subjects (26 +/- 3.1 yr) served as controls. We hypothesize that uremic myopathy provokes a cytosolic dysfunction but mitochondrial oxidative capacity is not abnormal. 31P-nuclear magnetic resonance spectra (31P-MRS) from the vastus medialis were obtained throughout the exercise protocol consisting of periods of 2 min exercise (at 1.67 Hz) at increasing work-loads interspersed by resting periods of 2.5 min. On a different day, after an identical exercise protocol, arterial and femoral venous blood gas data were obtained together with simultaneous measurements of femoral venous blood flow (Qleg) to calculate O2 delivery (QO2leg) and O2 uptake (VO2leg). Baseline resting [phosphocreatine] to [inorganic phosphate] ratio ([PCr]/[Pi]) did not change after rHuEPO (8.9 +/- 1.2 vs. 8.8 +/- 1.2, respectively), but it was significantly lower than in controls (10.9 +/- 1.5) (P = 0.01 each). At a given submaximal or peak VO2leg, no effects of rHuEPO were seen on cellular bioenergetics ([PCr]/[Pi] ratio, %[PCr] consumption halftime of [PCr] recovery after exercise), nor in intracellular pH (pHi). The post-rHuEPO bioenergetic status and pHi, at a given VO2leg, were below those observed in the control group. However, at a given pHi, no differences in 31P-MRS data were detected between post-rHuEPO and controls. After rHuEPO, at peak VO2, Qleg fell 20% (P < 0.04), limiting the change in QO2leg to 17%, a value that did not reach statistical significance. The corresponding O2 extraction ratio decreased from 73 +/- 4% to 68 +/- 8.2% (P < 0.03). These changes indicate that maximal O2 flow from microcirculation to mitochondria did not increase despite the 50% increase in [Hb] and explain how peak VO2leg and cellular bioenergetics (31P-MRS) did not change after rHuEPO. Differences in pHi, possibly due to lactate differences, between post-rHeEPO and controls appear to be a key factor in the abnormal muscle cell bioenergetics during exercise observed in CRF patients.

R esistance to insulin-mediated glucose disposal is a common finding in patients with non-insulin-dependent diabetes mellitus (NIDDM), as well as in nondiabetic individuals with hypertension. In an effort to identify the generic loci responsible for variations in blood pressure in individuals at increased risk of insulin resistance, we studied the distribution of blood pressure in 48 Taiwanese families with NIDDM and conducted quantitative sib-pair linkage analysis with candidate loci for insulin resistance, lipid metabolism, and blood pressure control. We found no evidence for linkage of the angiotensin converting enzyme locus on chromosome 17, nor the angiotensinogen and renin loci on chromosome 1, with either systolic or diastolic blood pressures. In contrast, we obtained significant evidence for linkage or systolic blood pressure, but not diastolic blood pressure, to a genetic region at or near the lipoprotein lipase (LPL) locus on the short arm of chromosome 8 (P = 0.002, n = 125 sib-pairs, for the haplotype generated from two simple sequence repeat markers within the LPL gene). Further strengthening this linkage observation, two flanking marker loci for LPL locus, D8S261 (9 cM telomeric to LPL locus) and D8S282 (3 cM centromeric to LPL locus), also showed evidence for linkage with systolic blood pressure (P = 0.02 and 0.0002 for D8S261 and D8S282, respectively). Two additional centromeric markers (D8S133, 5 cM from LPL locus, and NEFL, 11 cM from LPL locus) yielded significant P values of 0.01 and 0.001, respectively. Allelic variation around the LPL gene locus accounted for as much as 52-73% of the total interindividual variation in systolic blood pressure levels in this data set. Thus, we have identified a genetic locus at or near the LPL gene locus which contributes to the variation of systolic blood pressure levels in nondiabetic family members at high risk for insulin resistance and NIDDM.

B ased upon existing methods of isolating fetal porcine islet tissue, a simple, reliable procedure was developed for the preparation of porcine neonatal islet cell aggregates with a reproducible and defined cellular composition. After 9 d of in vitro culture, tissue from one neonatal pig pancreas yielded approximately 50,000 islet cell aggregates, consisting of primarily epithelial cells (57%) and pancreatic endocrine cells (35%). During the culture period, the total beta cell mass decreased initially, but subsequently increased 1.5-fold between days 3 and 9. Transplantation of grafts consisting of 3 x 10(5) beta cells (1,000 aggregated) under the kidney capsule of alloxan-diabetic nude mice corrected hyperglycemia in 75% (10/13) of the animals, whereas, 100% (20/20) of recipients implanted with 6 x 10(5) beta cells (2,000 aggregates) achieved euglycemia within 8 wk posttransplantation. Nephrectomy of the graft bearing kidney at 14 wk posttransplantation resulted in hyperglycemia in all recipients, and examination of the grafts revealed the presence of numerous well-granulated insulin- and glucagon-containing cells. The cellular insulin content of these grafts was 20 to 30-fold higher than at the time of transplantation. These results indicate that the neonatal porcine pancrease can be used as a source of large numbers of viable islet cells, which have the potential for growth both in vitro and in vivo, and exhibit the metabolic capacity to correct diabetes in nude mice.

T cell cytokines are known to play a major role in determining protection and pathology in infectious disease. It has recently become clear that IL-12 is a key inducer of the type 1 T cell cytokine pattern characterized by production of IFN-gamma. Conversely, IL-10 down-regulates IL-12 production and type 1 cytokine responses. We have investigated whether IL-12 and IL-10 might be involved in a chronic inflammatory reaction, atherosclerosis. In atherosclerotic plaques, we found strong expression of IFN-gamma but not IL-4 mRNAs as compared to normal arteries. IL-12 p40 mRNA and IL-12 p70 protein were also found to be abundant in atherosclerotic plaques. IL-12 was induced in monocytes in vitro in response to highly oxidized LDL but not minimally modified LDL. The cross-regulatory role of IL-10 was indicated by the expression of IL-10 in some atherosclerotic lesions, and the demonstration that exogenous rIL-10 inhibited LDL-induced IL-12 release. These data suggest that the balance between IL-12 and IL-10 production contributes to the level of immune-mediated tissue injury in atherosclerotsis.

L ymphocytes activate adhesion to intracellular adhesion mlecule 1 (ICAM-1) via leukocyte function associated antigen 1 (LFA-1), their major beta 2 integrin, in response to PMA (phorbol 12-myristate 13-acetate) without an increase in the number of receptors expressed. The molecular details of the mechanism are unknown. To determine the effect of PMA activation on LFA-1 movement within the plasma membrane, we used the single particle tracking technique to measure the diffusion rate of LFA-1 molecules on EBV-transformed B cells before and after PMA activation. Diffusion of LFA-1 on unactivated cells was restricted compared to CR1 (CD35), another transmembrane protein of equivalent size. PMA caused a 10-fold increase in the diffusion rate of LFA-1 without any effect on CD35. The increased LFA-1 motion induced by PMA was random, not directed, indicating that it was due to a release of constraints rather than the application of forces. The diffusion rates of LFA-1 are consistent with cytoskeletal attachment before and free diffusion after PMA. Cytochalasin D led to an equivalent increase in mobility and, at low doses, stimulated adhesion, implying that the nonadhesive state of LFA-1 is actively maintained by the lymphocyte cytoskeleton.

T his study compared the therapeutic potential of recombinant, native versus pegylated megakaryocyte growth and development factor (rMGDF and PEG-rMGDF, respectively), as well as that of the combined administration of PEG-rMGDF and r-methionyl human granulocyte colony-stimulating factor (r-metHuG-CSF) on hematopoietic reconstitution after 700 cGy, 60Co gamma, total body irradiation in nonhuman primates. After total body irradiation, animals received either rMGDF, PEG-rMGDF, r-metHuG-CSF, PEG-rMGDF and r-metHuG-CSF or HSA. Cytokines in all MGDF protocols were administered for 21-23 d. Either rMGDF, PEG-rMGDF, or PEG-rMGDF and r-metHuG-CSF administration significantly diminished the thrombocytopenic duration (platelet count (PLT) < 20,000 per microliter)to o.25, 0, 0.5 d, respectively, and the severity of the PLT nadir (28,000, 43,000, and 30,000 per microliter, respectively) as compared with the controls (12.2 d duration, nadir 4,000 per microliter), and elicited an earlier PLT recovery. Neutrophil regeneration was augmented in all cytokine protocols and combined PEG-rMGDF and r-metHuG-CSF further decreased the duration of neutropenia compared with r-metHuG-CSF alone.These data demonstrated that the administration of PEG-rMGDF significantly induced bone marrow regeneration versus rMGDF, and when combined with r-metHuG-CSF significantly enhanced multilineage hematopoietic recovery with no evidence of lineage competition.

T he expression of leptin, the ob gene product, is increased in adipose tissue in response to feeding and energy repletion, while leptin decreases food intake. Because adipose tissue gene expression is regulated by cytokines induced during infection and because infection is associated with anorexia, we tested whether induction of leptin might occur during the host response to infection. Administration of endotoxin (LPS), a model for gram negative infections, induces profound anorexia and weight loss in hamsters. In fasted adipose tissue to levels similar to fed control animals. There is a strong inverse correlation between mRNA levels of leptin and subsequent food intake. TNF and IL-1, mediators of the host response to LPS, also induced anorexia and increased levels of leptin in mRNA in adipose tissue. As assessed by immuknoprecipitation and Western blotting, circulating leptin protein is regulated by LPS and cytokines in parallel to regulation of adipose tissue leptin mRNA. Induction of leptin during the host response to infection may contribute to the anorexia of infection.