Mouse model of X-linked chronic granulomatous disease, an inherited defect in phagocyte superoxide production.

Chronic granulomatous disease (CGD) is a recessive disorder characterized by a defective phagocyte respiratory burst oxidase, life-threatening pyogenic infections and inflammatory granulomas. Gene targeting was used to generate mice with a null allele of the gene involved in X-linked CGD, which encodes the 91 kD subunit of the oxidase cytochrome b. Affected hemizygous male mice lacked phagocyte superoxide production, manifested an increased susceptibility to infection with Staphylococcus aureus and Aspergillus fumigatus and had an altered inflammatory response in thioglycollate peritonitis. This animal model should aid in developing new treatments for CGD and in evaluating the role of phagocyte-derived oxidants in inflammation.

Absence of respiratory burst in X-linked chronic granulomatous disease mice leads to abnormalities in both host defense and inflammatory response to Aspergillus fumigatus.

Mice with X-linked chronic granulomatous disease (CGD) generated by targeted disruption of the gp91phox subunit of the NADPH-oxidase complex (X-CGD mice) were examined for their response to respiratory challenge with Aspergillus fumigatus. This opportunistic fungal pathogen causes infection in CGD patients due to the deficient generation of neutrophil respiratory burst oxidants important for damaging A. fumigatus hyphae. Alveolar macrophages from X-CGD mice were found to kill A. fumigatus conidia in vitro as effectively as alveolar macrophages from wild-type mice. Pulmonary disease in X-CGD mice was observed after administration of doses ranging from 10(5) to 48 spores, none of which produced disease in wild-type mice. Higher doses produced a rapidly fatal bronchopneumonia in X-CGD mice, whereas progression of disease was slower at lower doses, with development of chronic inflammatory lesions. Marked differences were also observed in the response of X-CGD mice to the administration of sterilized Aspergillus hyphae into the lung. Within 24 hours of administration, X-CGD mice had significantly higher numbers of alveolar neutrophils and increased expression of the proinflammatory cytokines IL-1 beta and TNF-alpha relative to the responses seen in wild-type mice. By one week after administration, pulmonary inflammation was resolving in wild-type mice, whereas X-CGD mice developed chronic granulomatous lesions that persisted for at least six weeks. This is the first experimental evidence that chronic inflammation in CGD does not always result from persistent infection, and suggests that the clinical manifestations of this disorder reflect both impaired microbial killing as well as other abnormalities in the inflammatory response in the absence of a respiratory burst.

Selectins and neutrophil traffic: margination and Streptococcus pneumoniae-induced emigration in murine lungs.

The roles of selectins in the pulmonary margination and emigration of neutrophils were investigated by using mice genetically deficient in both E- and P-selectins (E/P mutants) and/or by intravenous injections of fucoidin (inhibiting both L- and P-selectins). E/P mutants were neutrophilic (14.7 +/- 4.9 x 10(6) vs. 0.8 +/- 0.1 x 10(6) neutrophils/ml). This neutrophilia was associated with increased margination of neutrophils within pulmonary capillaries (39.7 +/- 9.4 vs. 4.6 +/- 1.1 neutrophil profiles per 100 red blood cell profiles) but no change in margination within noncapillary pulmonary microvessels. After intratracheal instillation of Streptococcus pneumoniae, lungs of E/P mutants displayed increased neutrophil emigration (564 +/- 92 vs. 116 +/- 19 neutrophils per 100 alveolar profiles), edema (5.3 +/- 1.5 vs. 1.5 +/- 0.4 microliter/g body weight), and histologic evidence of lung injury compared with those in wild-type (WT). Fucoidin treatment did not affect neutrophil emigration during streptococcal pneumonia in WT or E/P mice. During pneumonia, the number of white blood cells (WBC) tethered to or spread upon the noncapillary vessel endothelium increased in both WT and E/P lungs. These are the first data demonstrating that neutrophil margination in uninfected pulmonary capillaries does not require E- and P-selectins; that streptococcal pneumonia induces an E- and P-selectin-independent increase in WBC interactions with noncapillary endothelium; and that migration of neutrophils to alveoli can occur despite deficiency or inhibition of all of the known selectins.

Neutrophil emigration in the skin, lungs, and peritoneum: different requirements for CD11/CD18 revealed by CD18-deficient mice.

To determine the role of CD11/CD18 complexes in neutrophil emigration, inflammation was induced in the skin, lungs, or peritoneum of mutant mice deficient in CD18 (CD18-/- mutants). Peripheral blood of CD18-/- mutants contained 11-fold more neutrophils than did blood of wild-type (WT) mice. During irritant dermatitis induced by topical application of croton oil, the number of emigrated neutrophils in histological sections of dermis was 98% less in CD18-/- mutants than in WT mice. During Streptococcus pneumoniae pneumonia, neutrophil emigration in CD18-/- mutants was not reduced. These data are consistent with expectations based on studies using blocking antibodies to inhibit CD11/CD18 complexes, and on observations of humans lacking CD11/CD18 complexes. The number of emigrated neutrophils in lung sections during Escherichia coli pneumonia, or in peritoneal lavage fluid after 4 h of S. pneumoniae peritonitis, was not reduced in CD18-/- mutants, but rather was greater than the WT values (240 +/- 30 and 220 +/- 30% WT, respectively). Also, there was no inhibition of neutrophil emigration during sterile peritonitis induced by intraperitoneal injection of thioglycollate (90 +/- 20% WT). These data contrast with expectations. Whereas CD11/CD18 complexes are essential to the dermal emigration of neutrophils during acute dermatitis, CD18-/- mutant mice demonstrate surprising alternative pathways for neutrophil emigration during pneumonia or peritonitis.

Infectious susceptibility and severe deficiency of leukocyte rolling and recruitment in E-selectin and P-selectin double mutant mice.

During the initial phase of the inflammatory response, leukocytes marginate and roll along the endothelial surface, a process mediated largely by the selectins and their ligands. Mice with mutations in individual selectins show no spontaneous disease and have mild or negligible deficiencies of inflammatory responses. In contrast, we find that mice with null mutations in both endothelial selectins (P and E) develop a phenotype of leukocyte adhesion deficiency characterized by mucocutaneous infections, plasma cell proliferation, hypergammaglobulinemia, severe deficiencies of leukocyte rolling in cremaster venules with or without addition of TNF-alpha, and an absence of neutrophil emigration at 4 h in response to intraperitoneal Streptococcus pneumoniae peritonitis. These mice provide strong evidence for the functional importance of selectins in vivo.

The cytoplasmic domain of L-selectin interacts with cytoskeletal proteins via alpha-actinin: receptor positioning in microvilli does not require interaction with alpha-actinin.

The leukocyte adhesion molecule L-selectin mediates binding to lymph node high endothelial venules (HEV) and contributes to leukocyte rolling on endothelium at sites of inflammation. Previously, it was shown that truncation of the L-selectin cytoplasmic tail by 11 amino acids abolished binding to lymph node HEV and leukocyte rolling in vivo, but the molecular basis for that observation was not determined. This study examined potential interactions between L-selectin and cytoskeletal proteins. We found that the cytoplasmic domain of L-selectin interacts directly with the cytoplasmic actin-binding protein alpha-actinin and forms a complex with vinculin and possibly talin. Solid phase binding assays using the full-length L-selectin cytoplasmic domain bound to microtiter wells demonstrated direct, specific, and saturable binding of purified alpha-actinin to L-selectin (Kd = 550 nM), but no direct binding of purified talin or vinculin. Interestingly, talin potentiated binding of alpha-actinin to the L-selectin cytoplasmic domain peptide despite the fact that direct binding of talin to L-selectin could not be measured. Vinculin binding to the L-selectin cytoplasmic domain peptide was detectable only in the presence of alpha-actinin. L-selectin coprecipitated with a complex of cytoskeletal proteins including alpha-actinin and vinculin from cells transfected with L-selectin, consistent with the possibility that alpha-actinin binds directly to L-selectin and that vinculin associates by binding to alpha-actinin in vivo to link actin filaments to the L-selectin cytoplasmic domain. In contrast, a deletion mutant of L-selectin lacking the COOH-terminal 11 amino acids of the cytoplasmic domain failed to coprecipitate with alpha-actinin or vinculin. Surprisingly, this mutant L-selectin localized normally to the microvillar projections on the cell surface. These data suggest that the COOH-terminal 11 amino acids of the L-selectin cytoplasmic domain are required for mediating interactions with the actin cytoskeleton via a complex of alpha-actinin and vinculin, but that this portion of the cytoplasmic domain is not necessary for proper localization of L-selectin on the cell surface. Correct L-selectin receptor positioning is therefore insufficient for leukocyte adhesion mediated by L-selectin, suggesting that this adhesion may also require direct interactions with the cytoskeleton.

Preservation of complement-induced lung injury in mice with deficiency of NADPH oxidase.

Mice with chronic granulomatous disease (X-CGD mice) generated by mutating the X-linked gene for a subunit of NADPH oxidase have been analyzed for their ability to respond to intravenous injection of purified cobra venom factor (CVF). This agent in wild-type mice produces a neutrophil-dependent and catalase-sensitive form of lung injury. Lung injury was evaluated by measuring the accumulation of extravascular albumin. Quite unexpectedly, the lungs of X-CGD mice showed no difference in the increased accumulation of extravascular albumin after injection of CVF when compared to wild-type mice. In both X-CGD and wild-type mice, full development of injury required neutrophils. While catalase was highly protective in wild-type mice, its protective effects were completely lost in the X-CGD mice. Furthermore, a competitive antagonist of L-arginine, N(G)-methyl-L-arginine, was protective in X-CGD mice but not in wild-type mice. Allopurinol was protective in both types of mice. Both the basal and the CVF-inducible lung mRNA for inducible nitric oxide synthase and IL-1beta was similar in X-CGD and wild-type mice. These data indicate that oxygen radical production and lung injury in response to injection of CVF occurs through alternative pathways in mice with genetic deletion of NADPH oxidase.

Neutrophil margination, sequestration, and emigration in the lungs of L-selectin-deficient mice.

These studies tested the hypothesis that L-selectin plays a role in neutrophil traffic in the lungs, particularly in neutrophil margination, sequestration, and emigration, using L-selectin-deficient mice. No defect in neutrophil margination within either capillaries or arterioles and venules was observed in uninflamed lungs of L-selectin-deficient mice. The initial rapid sequestration of neutrophils within the pulmonary capillaries 1 min after intravascular injection of complement fragments was not prevented. In contrast, L-selectin did contribute to the prolonged neutrophil sequestration (> or = 5 min). Interestingly, neutrophil accumulation within noncapillary microvessels required L-selectin at both 1 and 5 min after complement injection. During bacterial pneumonias, L-selectin played a role in neutrophil accumulation within noncapillary microvessels in response to either Escherichia coli or Streptococcus pneumoniae and within capillaries in response to E. coli but not S. pneumoniae. However, L-selectin was not required for emigration of neutrophils or edema in response to either organism. These studies demonstrate a role for L-selectin in the prolonged sequestration of neutrophils in response to intravascular complement fragments, in the intracapillary accumulation of neutrophils during E. coli-induced pneumonia, and in the accumulation of neutrophils within noncapillary microvessels when induced by either intravascular complement fragments or

Role of the intercellular adhesion molecule-1(ICAM-1) in endotoxin-induced pneumonia evaluated using ICAM-1 antisense oligonucleotides, anti-ICAM-1 monoclonal antibodies, and ICAM-1 mutant mice.

This study examined the effectiveness of antisense oligonucleotides targeted to intercellular adhesion molecule-1 (ICAM-1) to inhibit endotoxin-induced upregulation of ICAM-1 and neutrophil emigration and compared the apparent role of ICAM-1 when examined using antisense oligonucleotides, anti-ICAM-1 antibodies, and ICAM-1 mutant mice. Antisense oligonucleotides inhibited upregulation of ICAM-1 mRNA at 4 and 24 h after instillation of endotoxin in a dose-dependent manner. Neutrophil emigration into the alveolar spaces at 24 h was inhibited by 59%, similar to inhibition using the anti-ICAM-1 antibodies 3E2 (58%) and YN1/1 (75%). No inhibition was observed in the ICAM-1 mutant compared to wild-type mice. These data show that antisense oligonucleotides targeted to ICAM-1 inhibit the endotoxin-induced upregulation of ICAM-1 in the lung and are as effective as anti-ICAM-1 antibodies in preventing neutrophil emigration. The incomplete inhibition by either antisense oligonucleotides or antibodies suggests that alternative adhesion pathways that do not require ICAM-1 are important in neutrophil emigration in the lungs. The disparity in the role of ICAM-1 when evaluated using antisense or antibodies compared to mutant mice suggests that either these inhibitors are exerting additional effects on endothelial cells other than blockade of ICAM-1 or mutant mice have upregulated the ICAM-1-independent pathways to compensate for the long-term loss of ICAM-1.

P-selectin/ICAM-1 double mutant mice: acute emigration of neutrophils into the peritoneum is completely absent but is normal into pulmonary alveoli.

Neutrophil emigration during an inflammatory response is mediated through interactions between adhesion molecules on endothelial cells and neutrophils. P-Selectin mediates rolling or slowing of neutrophils, while intercellular adhesion molecule-1 (ICAM-1) contributes to the firm adhesion and emigration of neutrophils. Removing the function of either molecule partially prevents neutrophil emigration. To analyze further the role of P-selectin and ICAM-1, we have generated a line of mice with mutations in both of these molecules. While mice with either mutation alone show a 60-70% reduction in acute neutrophil emigration into the peritoneum during Streptococcus pneumoniae-induced peritonitis, double mutant mice show a complete loss of neutrophil emigration. In contrast, neutrophil emigration into the alveolar spaces during acute S. pneumoniae-induced pneumonia is normal in double mutant mice. These data demonstrate organ-specific differences, since emigration into the peritoneum requires both adhesion molecules while emigration into the lung requires neither. In the peritoneum, P-selectin-independent and ICAM-1-independent adhesive mechanisms permit reduced emigration when one of these molecules is deficient, but P-selectin-independent mechanisms cannot lead to ICAM-1-independent firm adhesion and emigration.

Endotoxin-induced mortality is related to increased oxidative stress and end-organ dysfunction, not refractory hypotension, in heme oxygenase-1-deficient mice.

BACKGROUND: Heme oxygenase (HO)-1 is an enzyme that degrades heme to generate CO (a vasodilatory gas), iron, and the potent antioxidant bilirubin. A disease process characterized by decreases in vascular tone and increases in oxidative stress is endotoxic shock. Moreover, HO-1 is markedly induced in multiple organs after the administration of endotoxin (lipopolysaccharide [LPS]) to mice. METHODS AND RESULTS: To determine the role of HO-1 in endotoxemia, we administered LPS to mice that were wild-type (+/+), heterozygous (+/-), or homozygous null (-/-) for targeted disruption of HO-1. LPS produced a similar induction of HO-1 mRNA and protein in HO-1(+/+) and HO-1(+/-) mice, whereas HO-1(-/-) mice showed no HO-1 expression. Four hours after LPS, systolic blood pressure (SBP) decreased in all the groups. However, SBP was significantly higher in HO-1(-/-) mice (121+/-5 mm Hg) after 24 hours, compared with HO-1(+/+) (96+/-7 mm Hg) and HO-1(+/-) (89+/-13 mm Hg) mice. A sustained increase in endothelin-1 contributed to this SBP response. Even though SBP was higher, mortality was increased in HO-1(-/-) mice, and they exhibited hepatic and renal dysfunction that was not present in HO-1(+/+) and HO-1(+/-) mice. The end-organ damage and death in HO-1(-/-) mice was related to increased oxidative stress. CONCLUSIONS: These data suggest that the increased mortality during endotoxemia in HO-1(-/-) mice is related to increased oxidative stress and end-organ (renal and hepatic) damage, not to refractory hypotension.

Combinatorial requirements for adhesion molecules in mediating neutrophil emigration during bacterial peritonitis in mice.

To investigate the requirements for adhesion molecules in neutrophil emigration during peritonitis, mice received intraperitoneal injections of Streptococcus pneumoniae while the functions of multiple adhesion molecules were blocked. Emigration after 4 h was compromised by antibodies against ICAM-1 or genetic deficiency of ICAM-1. Anti-CD11a/CD18 antibodies decreased emigration in ICAM-1 mutant mice, suggesting that ICAM-1 independent emigration requires CD11/CD18 complexes. In contrast, mice mutant in ICAM-1 plus E-selectin showed no defect in emigration, suggesting that E-selectin commits neutrophils to an ICAM-1-dependent pathway during streptococcal peritonitis. However, in mutant mice lacking the three endothelial adhesion molecules E-selectin, P-selectin, and ICAM-1, emigration after 4 h was significantly compromised. Thus, P-selectin is essential to ICAM-1- and E-selectin-independent acute peritoneal inflammation. After 24 h of peritonitis, there were no differences between WT and E-selectin/P-selectin/ICAM-1 mutant mice, demonstrating that these endothelial adhesion molecules are not essential to neutrophil emigration during later stages of peritonitis.

Platelet distribution in rabbits following infusion of liposomes.

This investigation determined the organ distribution of liposomes containing egg phosphatidylcholine and cholesterol with egg phosphatidylglycerol (PG liposomes) or without (PC liposomes) and the effect of each liposome on platelet distribution in rabbits. Eight minutes after 51chromium-labelled platelets were given intravenously, either saline (n = 7), iodinated PG liposomes (n = 5) or iodinated PC liposomes (n = 5) were infused. Two minutes later the organ distribution of 51Cr-platelets and 125I-liposomes were compared. The PG liposomes produced a 41 +/- 5% reduction in circulating platelet counts while PC liposomes did not. The PG liposomes decreased circulating 51Cr-platelets by a factor of 2 and increased platelet recoveries in the liver and lungs. The increased platelet recovery in the liver was associated with a greater PG liposome recovery. When animals receiving PG liposomes were studied over 60 minutes, both the labelled and unlabelled platelet counts returned to control values by 30 minutes and the 51Cr-platelet distribution between organs was similar to control values. These data indicate that platelets and PG liposomes initially sequester together and that this platelet-liposome interaction is specific for PG liposomes. However, the platelet sequestration is transient and by 60 minutes the platelets were released and circulating.

Pulmonary blood volume and edema in postpneumonectomy lung growth in rats.

After pneumonectomy in young animals, the contralateral lung undergoes compensatory growth and generally attains the same weight and air space volume as both lungs in age-matched controls. In this study, we determined the contribution of lung edema and increased blood volume to the weight gain in rats. Three weeks after pneumonectomy (n = 18) or sham pneumonectomy (n = 17), the pulmonary blood volume and the extravascular water and albumin were evaluated by use of 51Cr-labeled erythrocytes and 125I-labeled albumin. The air space volume, blood-free lung weights, and DNA and protein content were also compared. The data show that the total pulmonary blood volumes and the blood volume per gram of blood-free dry lung were similar in pneumonectomized and age-matched sham controls. The total extravascular albumin and the extravascular albumin per gram of blood-free dry lung were also similar as well as the extravascular lung water, wet-to-dry weight ratios, DNA and protein content, and air space volumes. These data indicate that the increased weight of the postpneumonectomy lung was due to cellular and stromal proliferation. The blood volume and interstitial fluid increased in proportion to the increase in lung parenchyma. Neither vascular congestion nor increased extravascular protein and water contributed to the observed weight gain.

Effect of zymosan-activated plasma on the deformability of rabbit polymorphonuclear leukocytes.

Intravascular infusion of inflammatory mediators causes a sudden neutropenia due to the sequestration of polymorphonuclear leukocytes (PMN) within the microvasculature of the lung and other organs. This sequestration could be due to a decrease in the ability of PMN to deform and pass through the narrow capillary bed. The purpose of this study was to determine if the complement fragments present in zymosan-activated plasma (ZAP) caused a rapid stiffening of PMN. The PMN deformability was determined by measuring the pressure required to pass PMN through a polycarbonate filter containing 5-micron pores at a constant flow rate as well as the extraction of PMN compared with red blood cells and 125I-labeled albumin by the filter. The role of the cytoskeleton in PMN deformation was examined in studies where F-actin formation was inhibited using cytochalasin B or microtubule assembly was inhibited using colchicine. The results showed that treatment with ZAP induced a rapid decrease in PMN deformability. Inhibiting the formation of F-actin made the unstimulated PMN more deformable and reduced the stiffening induced by ZAP. In contrast, inhibition of microtubule reassembly did not alter either normal deformability or the ZAP-induced decrease in deformability. In vivo, colchicine increased normal PMN margination but did not inhibit the rapid sequestration of PMN induced by infusion of ZAP. These studies indicate that ZAP induces a rapid decrease in PMN deformability that is mediated through the cytoskeleton. They suggest that this decrease is due to the polymerization of F-actin.

Computational modeling of RBC and neutrophil transit through the pulmonary capillaries.

A computational model of the pulmonary microcirculation is developed and used to examine blood flow from arteriole to venule through a realistically complex alveolar capillary bed. Distributions of flow, hematocrit, and pressure are presented, showing the existence of preferential pathways through the system and of large segment-to-segment differences in all parameters, confirming and extending previous work. Red blood cell (RBC) and neutrophil transit are also analyzed, the latter drawing from previous studies of leukocyte aspiration into micropipettes. Transit time distributions are in good agreement with in vivo experiments, in particular showing that neutrophils are dramatically slowed relative to the flow of RBCs because of the need to contract and elongate to fit through narrower capillaries. Predicted neutrophil transit times depend on how the effective capillary diameter is defined. Transient blockage by a neutrophil can increase the local pressure drop across a segment by 100--300%, leading to temporal variations in flow and pressure as seen by videomicroscopy. All of these effects are modulated by changes in transpulmonary pressure and arteriolar pressure, although RBCs, neutrophils, and rigid microspheres all behave differently.

Neutrophil kinetics in rabbits during infusion of zymosan-activated plasma.

Complement activation in vivo produces neutropenia and pulmonary sequestration of neutrophils (PMNs) whereas in vitro activation increases PMN adherence and decreases PMN deformability. The present study examined PMN kinetics in vivo to determine if this sequestration was specific to the lung. Venous or arterial injections of radiolabeled PMNs were given to animals receiving infusions of zymosan-activated plasma (ZAP) or saline, and the PMN distribution was evaluated 10 min later. In control animals, the relative size of the marginated and circulating PMN pools was similar after venous or arterial injection and regional PMN retention increased as blood velocity slowed. ZAP infusion produced threefold increases in PMNs within pulmonary capillaries after venous injection and PMN retention was independent of blood velocity. After arterial injection, ZAP infusion produced PMN sequestration in all organs. Rigid (glutaraldehyde-fixed) PMNs injected into control rabbits showed increased lung recoveries similar to those of fresh PMNs injected into ZAP-treated rabbits. We conclude that activation of the complement system causes PMN sequestration in both the pulmonary and the systemic microvasculature and that the decrease in PMN deformability that occurs with activation of the PMN may be important in the genesis of PMN sequestration.

Neutrophil retention during a single transit through the pulmonary circulation.

Our laboratory has previously reported that 70-80% of polymorphonuclear cells (PMNs) are delayed with respect to erythrocytes (RBCs) in a single pass through the lungs of dogs, whereas only 5-15% of PMNs are delayed in a single pass through human lungs. Because these results were obtained using a direct blood sampling method in animals and an indirect gamma camera method in humans, the reported differences could be related to differences in measurement technique. The present study was designed to settle this question by comparing both techniques in a single species. The results show that the gamma camera technique previously used in humans underestimates the retention of PMNs with respect to RBCs during a single pass through the lung. They also show that this problem can be corrected by modifying the analysis of the data obtained using the gamma camera. We conclude that the pulmonary circulation retains PMNs with respect to RBCs to a comparable degree in animals and humans.

Neutrophil kinetics during active cigarette smoking in rabbits.

Polymorphonuclear leukocyte (PMN) transit through the pulmonary vasculature is slowed during inhalation of cigarette smoke in humans. This study was undertaken to determine the localization of the delayed PMN and whether they release granule-bound enzymes during smoke exposure. Anesthetized New Zealand White rabbits were exposed to cigarette smoke (n = 5) or sham (n = 5) for 10 min while they breathed spontaneously. The cardiac output, pulmonary blood volume and flow, and PMN retention were measured in each of five gravity-defined slices of lung. In three smoke-exposed and three sham animals the lungs were prepared for autoradiography, and the distribution of the radiolabeled PMN was determined. Plasma was assayed for myeloperoxidase in 10 animals. We found that smoke exposure caused increased PMN retention in the top two slices of the lungs without changing hemodynamics. The PMN were randomly distributed in the lobule, and plasma myeloperoxidase was elevated at the beginning of the exposure. We conclude that cigarette smoke may damage the lung by activating PMN in the pulmonary capillary bed.

Effects of anatomic variability on blood flow and pressure gradients in the pulmonary capillaries.

A theoretical model is developed to simulate the flow of blood through the capillary network in a single alveolar septum. The objective is to study the influence of random variability in capillary dimension and compliance on flow patterns and pressures within the network. The capillary bed is represented as an interconnected rectangular grid of capillary segments and junctions; blood flow is produced by applying a pressure gradient across the network. Preferred flow channels are shown to be a natural consequence of random anatomic variability, the effect of which is accentuated at low transcapillary pressures. The distribution of pressure drops across single capillary segments widens with increasing network variability and decreasing capillary transmural pressure. Blockage of one capillary segment causes the pressure drop across that segment to increase by 60%, but the increase falls to < 10% at a distance of three segments. The factors that cause nonuniform capillary blood flow through the capillary network are discussed.