Studies on the interaction between phagocytes and tubercle bacilli. I. Observations on the metabolism of guinea pig leucocytes and the influence of phagocytosis

In memoriam: Filippo Rossi (1926-2022)

Description of the scientific life of Filippo Rossi, who died in October 2022.

Antimicrobial actions of reactive oxygen species

Everything should be as simple as it can be, but not simpler.

—Attributed to Albert Einstein (1)

Reactive oxygen species (ROS) are produced by host phagocytes and exert antimicrobial actions against a broad range of pathogens. The observable antimicrobial actions of ROS are highly dependent on experimental conditions. This perspective reviews recent controversies regarding ROS in Salmonella-phagocyte interactions and attempts to reconcile conflicting observations from different laboratories.

A histochemical study of phagocytic and enzymatic functions of rabbit mononuclear and polymorphonuclear exudate cells and alveolar macrophages. I. Survey and quantitation of enzymes, and states of cellular activation

The cytochrome oxidase (CO), aminopeptidase (AMP), succinic dehydrogenase (SD), acid phosphatase, esterase, and alkaline phosphatase of rabbit mononuclear (MN) and polymorphonuclear (PMN) peritoneal exudate cells and pulmonary alveolar macrophages (AM) - air dried on Mylar strips - were characterized by histochemical techniques with respect to stability, activators, inhibitors, and pH optima. A granule count method was established for the quantitation of these enzymes. For the acid phosphatase of MN, in which the most precise results were obtained, time, pH, substrate, and inhibitor curves resembled those commonly obtained biochemically. Five of these enzymes were usually more active in AM than MN, whereas the sixth, alkaline phosphatase, was not present in either cell type. AM also tended to consume more oxygen than MN and to divide more frequently. Since the most active cells in the population would be first involved in the host's defense against microbial agents, a comparison was made of the 10 per cent of the AM and MN with the highest enzymatic activities. No differences were found in the granule counts that were not reflected by the means. However, within a given AM population, cells containing ingested dust particles seemed to have higher enzymatic activities than those without particles. MN had greater acid phosphatase and SD activities than PMN and consumed more oxygen, but the CO, AMP, and esterase activites of both types of cells were of similar magnitude. PMN showed high alkaline phosphatase activity; MN showed none. A survey of the histochemical literature indicates that a positive correlation between the enzymatic and phagocytic activities of both MN and PMN exists in vivo.

The effect of antibody on intracellular parasitism of Salmonella typhimurium in mononuclear phagocytes in vitro: prolonged survival of infected monocytes in presence of antibody

The effect of antibody on the fate of Salmonella typhimurium within mononuclear phagocytes (MN) of rabbits was studied in vitro. Monocytes and bacteria were incubated either in absence or presence of antibody. After 45 minutes during which phagocytosis occurred infected cells were washed to remove extracellular bacilli and free antibody. The cells were then reincubated in a medium without addition of antibody, and the interaction between the MN and bacteria was followed, correlating bacterial viability and the morphology of the mixture. The following results were obtained. The anti-Salmonella antibody was not bactericidal even in presence of complement and did not enhance phagocytosis. Regardless of whether antibody was present or absent during phagocytosis, the bacteria appeared to multiply within the cells. When no antibody was present during phagocytosis the infected cells were severely damaged within a few hours of incubation, and extensive extracellular multiplication was dominating. When antibody was present during phagocytosis MN packed with bacteria persisted for a long time. Little or no extracellular growth occurred. It was possible to demonstrate the presence of the antibody within the infected MN, using the fluorescent antibody technique. The antibody appeared as a coat around the bacteria. Antibody entered the cells only during phagocytosis, presumably attached to the bacteria. The active factor of the immune serum was found in the gamma globulin fraction and reacted specifically with the somatic antigen of Salmonella typhimurium. The antiflagellar portion of the antiserum was not involved in the phenomenon described. It is concluded that this antibody protects monocytes against the effect of intracellularly located Salmonella.

Metabolic patterns in three types of phagocytizing cells

Some chemical and metabolic characteristics of polymorphonuclear leukocytes and monocytes from peritoneal exudates of the guinea pig, and of alveolar macrophages from the same animal, have been compared. Changes in the metabolic patterns of these three types of cell have been followed during the act of phagocytosis. The effect of conventional inhibitors of metabolism, and of anaerobiosis on the phagocytic ability of each of the three cell types mentioned has also been determined. From these studies it was found that alveolar macrophages depend to a considerable degree upon oxidative phosphorylation to provide energy for phagocytosis. The other two types of cell depend only on glycolysis as the source of metabolic energy for that function. In some experiments aimed at obtaining information on the possible role of complex lipids in the function of the cell membrane, it was noted that phagocytosis stimulated the incorporation of inorganic phosphate-P³² into the phosphatides of both types of cell from peritoneal exudates—whether these were free-swimming or adherent to a surface. This phenomenon has not yet been detected in the case of alveolar macrophages.

Composition and synthesis of lipids in resting and phagocytizing leukocytes

The lipid content of rabbit polymorphonuclear leukocytes, obtained from peritoneal exudates, constituted 8.7 ± 2.9 per cent of the dry weight of these cells; 60 per cent of all lipids were phospholipids, 20 per cent triglycerides, and the remainder cholesterol and cholesterol esters and a small amount of non-esterified fatty acids (2 to 4 per cent). The composition of the fatty acids in leukocytes, as determined by gas-liquid chromatography, was slightly different from rabbit serum and red blood cells, but markedly different from the dietary fat. The synthesis, turnover, and composition of lipids in rabbit leukocytes at rest and during phagocytosis in vitro were compared. Lipid content and composition were not affected by the phagocytic process. However, active phagocytosis resulted in an increase in the rate of turnover of lipids. This stimulation of lipid metabolism was more marked in triglycerides and cholesterol esters than in phospholipids. It is suggested that the increased turnover of lipid during phagocytosis may reflect a general metabolic stimulation accompanying this process, rather than a specific synthesis of phospholipid for the production of new cell membrane.

Functional and metabolic properties of polymorphonuclear leucocytes. I. Observations on the requirements and consequences of particle ingestion

The phagocytosis and intracellular destruction of bacteria by rabbit polymorphonuclear leucocytes has been studied in vitro under defined conditions. The efficient and continuing ingestion of bacteria was dependent upon (a) opsonic factors present in fresh rabbit serum as well as upon, (b) the availability of an adequate supply of glucose in the medium. The effects of selected enzymatic inhibitors on the metabolic and functional activities of the leucocytes was investigated. Cyanide which inhibited oxygen consumption had no effect on the ingestion or inactivation of bacteria, Iodoacetate and arsenite which blocked glycolysis produced a marked inhibition in particle ingestion. 2,4-Dinitrophenol which stimulated both oxygen consumption and glycolysis, depressed phagocytosis after a 1 hour latent period. It was concluded that phagocytosis was an energy-requiring process in which glycolysis served as the most important source of energy. Leucocytes which were ingesting heat-killed bacteria exhibited increases in oxygen consumption, glucose utilization, and lactic acid synthesis. The effect of particle ingestion on glycogen metabolism was characterized by an initial period of glycogenolysis followed by an enhanced rate of glycogen synthesis. Leucocytes which had previously ingested heat-killed bacteria also demonstrated increased rates of phagocytosis.

Chemotactic factors released in culture by intact developing and healing skin lesions produced in rabbits by the irritant sulfur mustard

Development, peak and healing lesions were induced in the skin of rabbits by topical applications (on different days) of the chemical irritant sulfur mustard (SM). Immediately after the rabbits were euthanized, the intact lesions were excised and organ-cultured for 17 to 20 hours. The culture fluids from early, peak and healing SM lesions all showed high chemotactic activity for both PMN and MN. This finding suggests that the PMN and MN, seen microscopically in tissue sections of the lesions, were entering continuously, even during the healing process. The chemotaxins identified were the eicosanoid LTB4, the chemokine IL-8, and proteases producing the complement fragment C5a. Other studies from our laboratory showed that the number of cells containing IL-1, IL-8, MCP-1, and GRO mRNAs was increased in SM lesions. Chemotactic activity was released by both live and dead (frozen and thawed) cell suspensions of PMN, MN, and fibroblasts, suggesting that these cells were major sources of the chemotaxins produced by the SM lesion explants. Explants of normal skin produced considerable chemotactic activity for MN, but not for PMN. Chemotactic activity for PMN, and the release of LTB4, IL-8 and proteases cleaving C5 to C5a, occurred only in explants infiltrated by leukocytes.

HYDROLYTIC ENZYMES OF RABBIT MONONUCLEAR EXUDATE CELLS. I. QUANTITATIVE ASSAY AND PROPERTIES OF CERTAIN PROTEASES, NON-SPECIFIC ESTERASES, AND LIPASES OF MONONUCLEAR AND POLYMORPHONUCLEAR CELLS AND ERYTHROCYTES

Oil-induced mononuclear phagocytes (MN) were quantitatively assayed for various hydrolases as unfractionated suspensions of frozen and thawed cells. They apparently contain two proteases. The first, measured with urea- or acid-denatured hemoglobin, was similar to purified Proteinase I of lung with respect to pH optimum (pH 4), stability, hydrolytic and polymerizing activities, and reactions to various inhibitors. The second protease resembled chymotrypsin in its hydrolysis of glycyl-L-phenylalanine amide, acetyl-L-tyrosine ethyl ester and N-benzoyl-DL-phenylalanine-beta-naphthol ester (BPN). With the latter, its pH optimum was between 5.0 and 5.8, and its action was inhibited by diisopropylphosphorofluoridate (DFP) and p-chloromercuribenzoate. When assayed under the above conditions, polymorphonuclear exudate cells (PMN) and red blood corpuscles (RBC) manifested little or no hydrolysis of either hemoglobin or BPN. MN also contained esterases that split methyl butyrate and beta-naphthyl acetate. The pH optimum with the latter was 7.4, and its hydrolysis was partially inhibited by DFP, fluoride, taurocholate, and eserine. PMN had low esterase activity; RBC had little or none. MN, but not PMN or RBC, contained a stable lipase with a pH optimum of 6.1 in maleate buffer. Protamine, NaCl, heat, p-chloromercuribenzoate, ethylenediamine tetraacetate, taurocholate, and DFP were inhibitory, but no appreciable activation occurred in the presence of heparin or serum. Thus it possessed some of the characteristics of Korn's lipoprotein lipase, but not others.

Relationship between pulmonary oxygen consumption, lung inflammation, and calculated venous admixture in patients with acute lung injury

OBJECTIVE

To determine in patients with acute lung injury whether increased pulmonary oxygen consumption (VO2pulm), computed as the difference between oxygen consumption measured by indirect calorimetry (VO2meas) and calculated by the reverse Fick method (VO2Fick), would: (1) correlate with the degree of lung inflammation assessed by bronchoalveolar lavage (BAL); (2) lead to an overestimation of calculated venous admixture (Qva/Qt).

DESIGN

Prospective study.

SETTING

University hospital, medical intensive care unit.

INTERVENTION

None.

MEASUREMENTS AND RESULTS

In nine mechanically ventilated patients with acute lung injury (Apache II 12 +/- 5, lung injury score 2 +/- 0.6, mean +/- SD), whole-body VO2 (VO2wb) was determined simultaneously by indirect calorimetry and the reverse Fick technique, after which BAL was immediately performed. VO2meas was significantly higher than VO2Fick (128 +/- 24 and 102 +/- 18 ml/min per m2, respectively, p < 0.001). Median VO2pulm was 25.3 ml/min per m2 (range 1.98-51.5), thus representing 19 +/- 11% of VO2wb. Total BAL cellularity was increased in all patients (median 47, range 24-200 x 10(4)/ml), as was the total polymorphonuclear (PMN) count (median 78 range 5-93 x 10(4)/ml). Macrophage counts were in the normal range. There were raised BAL levels of interleukin-6 (IL-6) (median 945, range 23-1800 ng/ml) and elastase (median 391, range 5-949 ng/ml). Median protein levels were 270 micrograms/ml (range 50-505). There was no correlation between VO2pulm and BAL cellularity, PMNs, elastase, IL-6, or protein. Qva/Qt was 31.7 +/- 8%. Qva/Qt, corrected for the presence of VO2pulm, (Qva/Qtcorr), was 30.3 +/- 8% (P < 0.01 vs Qva/Qt), a 4.2% overestimation due to VO2plum. There was no correlation between Qva/Qt or Qva/Qtcorr and VO2pulm.

CONCLUSIONS

In mechanically ventilated patients with acute lung injury, VO2pulm was increased and led to a 19% underestimation of VO2wb determined by the reverse Fick method, as well as to a 4.2% overestimation of calculated Qva/Qt. Lung inflammatory activity was increased, as assessed by BAL cellularity, IL-6 and elastase levels. However, there was no correlation between VO2pulm and the intensity of pulmonary inflammation.

New concepts in the pathophysiology of oxygen metabolism during sepsis

Sepsis is an intriguing pathological condition associated with many complex metabolic and physiological alterations. In this review a novel hypothesis in the pathophysiology of oxygen metabolism during sepsis is explored. It is proposed that the hypermetabolic response to sepsis results from enhanced reactive oxygen generation by phagocytes. Reactive oxygen detoxification by host enzyme systems subsequently leads to alterations in oxidative metabolism. The similarities between the metabolic consequences of reactive oxygen metabolism and the metabolic changes observed during sepsis are outlined. A unified concept is presented to help explain the pathophysiological changes in oxygen metabolism during sepsis.

The superoxide-forming enzymatic system of phagocytes

The formation of oxygen-derived free radicals by the phagocytes (neutrophils, eosinophils, monocytes and macrophages) is catalysed by a membrane-bound NADPH oxidase which is dormant in resting cells and becomes activated during phagocytosis or following interaction of the cells with suitable soluble stimulants. This enzyme is under investigation in many laboratories but its molecular structure remains to be clarified. Possible components such as flavoproteins, cytochrome b558, and quinones have been proposed on the basis of enzyme purification studies, effects of inhibitors, kinetic properties and analysis of genetic defects of the oxidase. An extensive discussion of the evidence for the participation of these constituents is reported. On the basis of the available information on the structure and the catalytic properties of the NADPH oxidase, a series of possible models of the electron-transport chain from NADPH to O2 is presented. Finally, the triggering mechanism of the respiratory burst is discussed, with particular reference to the stimulus-response coupling and the final modification(s) of the oxidase (phosphorylation, assembly, change of lipid environment, etc.) which are involved in its activation.

Fluoride-mediated activation of guinea pig neutrophils

In guinea pig peritoneal neutrophils NaF at a concentration of above 5 mM elicited a dose-dependent, delayed and sustained activation of NADPH oxidase. Unlike in human neutrophils, in guinea pig cells, this response was independent of extracellular calcium. Fura2 fluorescence measurements indicated also a fluoride-mediated moderate elevation in the level of cytosolic calcium concentration. Pretreatment of neutrophils with pertussis toxin, blocked fluoride-promoted activation of NADPH oxidase, indicating that NaF stimulation was mediated by a G protein which is a pertussis toxin substrate. NaF-elicited calcium elevation was insensitive to the toxin. Upon transfer of NaF-stimulated cells to a fluoride-free medium, superoxide release declined and calcium levels diminished. The response of the deactivated, fluoride-prestimulated guinea pig neutrophils to a secondary stimulation with phorbol myristate acetate (PMA) or fMet-Leu-Phe, was either unaffected by the previous challenge with NaF (PMA) or augmented by it (the chemotactic peptide). In parallel to the activation of NADPH oxidase, NaF also induced translocation of protein kinase C to cell membranes. This effect was also abolished by a pretreatment with pertussis toxin.

Activation of the respiratory burst of guinea pig neutrophils by dicyclohexylcarbodiimide

DCCD activates the respiratory burst in guinea pig peritoneal neutrophils. The onset of the superoxide producing activity is preceeded by a lag, inversely proportional to the dose of the stimulant and to the temperature. Initial rates of superoxide formation exhibit different dependencies on the concentrations of DCCD and on temperature. Activation of NAD(P)H oxidase is inhibited by preincubation of neutrophils with 2-deoxyglucose and does not require the presence of extra cellular Ca2+.

The isolation and characterisation of guinea-pig polymorphonuclear leucocyte actin and myosin

The contractile proteins actin and myosin have been isolated from the soluble phase of guinea-pig polymorphonuclear leucocytes and partially characterised. Two forms of actin have been identified, designated 'Mg-actin' and 'KCl-actin'. They have different polymerising properties but their propensity to form synthetic homologous and heterologous actomyosins and to inhibit DNAase-1 does not significantly differ. Both show beta and gamma isoelectric forms in focusing gels and the Mg-actin accounts for about 5% of the soluble-phase protein and te KCl-actin around 2%. Leucocyte myosin has been isolated by affinity chromatography on N6-ADP-Sepharose with a good enrichment of both Ca2+-ATPase and the ATPase activity measured in the absence of Ca2+ or Mg2+ and in the presence of EDTA. This protein, too, has the capacity to form synthetic homologous and hybrid actomyosins with enhancement of the basal Mg2+-ATPase activity. The ratio of actin to myosin in the leucocyte calculated on a molar basis is well in excess of 100, a figure consistent with the findings from other non-muscle cells.

Superoxide generation by digitonin-stimulated guinea pig granulocytes. A basis for a continuous assay for monitoring superoxide production and for the study of the activation of the generating system

Stimulation of guinea pig granolocytes by digitonin results in superoxide (O-2) generation. A continuous assay shows that there is a lag between the addition of digitonin and the onset of O-2 production. The rate of activation of the O-2 generating system is dependent upon the concentration of digitonin and the temperature. The final linear rate of O-2 production is affected by the concentration of digitonin, temperature, pH, and the presence of exogenous reduced pyridine nucleotides. Thus, factors which alter either the activation process or the activity of the O-2 generating system can affect O-2 production by stimulated granolocytes.

Superoxide production by digitonin-stimulated guinea pig granulocytes. The effects of N-ethyl maleimide, divalent cations; and glycolytic and mitochondrial inhibitors on the activation of the superoxide generating system

N-ethylmaleimide, divalent cations, ethylene glycol bis (beta aminoethyl ether) N,N,N',N',-tetraacetate, 2-deoxyglucose, cyanide, and dinitrophenol were examined for their effect on the ability of guinea pig granulocytes to generate superoxide (O(2) (-)) when stimulated by digitonin. N-ethylmaleimide (1 mM) inhibits only when added before complete activation of the O(2) (-) generating system, and at lower concentrations (0.05-0.2 mM) slows the activation process. Ca(++) is required for maximum O(2) (-) generation, and Mg(++) decreases the amount of Ca(++) required. Ethylene glycol bis (beta aminoethyl ether) N,N,N',N',-tetraacetate (10 mM) inhibits only if added before complete activation. Incubation of cells in 2-DOG causes a time- and concentration-dependent inhibition of O(2) (-) generation. It also increases the time required for activation of this system. Cyanide and dinitrophenol increase the rate of O(2) (-) production. However, when these compounds are added to cells whose O(2) (-) production is partially inhibited by incubation in 2-deoxyglucose, complete inhibition results. If cyanide or dinitrophenol is added after activation of 2-deoxyglucose-treated cells, no further inhibition occurs. On the basis of the above results, we conclude that the activation of the O(2) (-) generating system is N-ethylmaleimide sensitive, Ca(++) dependent, and energy requiring, but that the activity of the enzyme system in the cell is not.

Studies of leukocyte exudates in cyanate-treated animals

Cyanate (NCO), which impedes sickling by increasing the oxygen affinity of sickle haemoglobin, may react with many proteins in the body and potentially interfere with many functional systems. Leukocyte glycogen was studies because elevated liver glycogen has been noted after high dose NCO treatment in rats. Leukocyte function was also studied in peritoneal exudates. In animals given NCO alone, blood leukocyte glycogen was 2.8 plus or minus 0.3 (SE) mg per 10+9 WBC while control values were 3.3 plus or minus 0.6, a difference not statistically significant. In casein-induced peritoneal exudates, more WBC were recovered from NCO-treated animals (265 x 10+6 WBC vs. 214 x 10+6 WBC; P = 0.0009). Glycogen in blood leukocytes of casein-stimulated animals was not significantly different from NCO-treated animals. Leukocyte glycogen in peritoneal exudates was markedly increased over blood leukocyte glycogen in both controls (19.4 plus or minus 0.6 vs. 2.7 plus or minus 0.3 mg per 10+9/WBC; P less than 0.0001) and NCO-treated animals (17.5 plus or minus 0.7 vs. 3.6 plus or minus 0.5 mg per 10+9 WBC; P less than 0.001), although levels in exudates from control and NCO-treated animals did not significantly differ from each other. 14-C-glucose incorporation into glycogen was significantly increased (P = 0.020) in exudate leukocytes of NCO-treated animals. Initial phagocytic rates were equal in exudate leukocytes of control and NCO-treated animals. Although NCO treatment has demonstrable effects on ome aspects of leukocyte glycogen metabolism, exudation of leukocytes and phagocytic function are not impaired.

The effect of differing ambient oxygen tensions on wound infection

Wound infections were studied in rabbits using two standard inocula (approximately equal to 10-4 and approximately equal to 10-6) of Pseudomonas aeruginosa injected into subcutaneous wound dead space made by implantation of standard wire mesh cylinders. The inoculation was done on the fourth day after implantation of the cylinders in animals kept from the day of implantation in atmospheres of 12%, 21%, or 45% oxygen content. Samples of wound fluid (0.2 ml) were removed for quantitative culture just before inoculation and 3, 7, 14, and 21 days later. No positive cultures resulted from samples taken before inoculation. One uninoculated wound served as a control in each animal. None of these control wounds became infected. Culture counts were significantly highest in the anoxic group and lowest in the hyperoxic group. Established infections were significantly lowest in the hyperoxics and highest in the hypoxics. The percent of wounds showing a significant culture count showed a similar trend. The mechanisms of this effect is not known, but a possible mechanism lies in the relative inability of leucocytes to kill this bacterium under hypoxic conditions.

Iodinating ability of various leukocytes and their bactericidal activity

A rapid method that employs monolayers of different phagocytic cells, primarily from guinea pigs and mice, has allowed a kinetic determination of (a) ingestion by these cells of labeled particles, (b) fixation of (131)I and (c) microbicidal activity in the cells after periods as short as 5' of exposure of bacteria to phagocytes. Phagocytes so examined included polymorphonuclear leukocytes (PMN) elicited into the peritoneal cavity, elicited peritoneal mononuclear cells (monocytes) (MN), and peritoneal macrophages (MAC) obtained simply by lavage. Circulating PMN from normal human subjects and from children afflicted with chronic granulomatous disease were also studied. The potential for generation of H(2)O(2) (a key component of the iodinating system) of all the normal cells studied, gauged by their content of cyanide-insensitive NADH oxidase, seemed comparable. Peroxidase levels varied widely, and were highest in PMN and almost undetectable in MAC. Catalase was at negligible levels in all the cell types obtained from mice. The fixation of (131)I by phagocytes ingesting (14)C-labeled dead tubercle bacilli appeared to be primarily a function of the cellular peroxidase content. Thus, mouse macrophages, with virtually no peroxidase, displayed no fixation of iodide. PMN proved far more able to fix (131)I during phagocytosis than did MN. In experiments comparing PMN from normal human subjects and from children with chronic granulomatous disease (CGD), a sex-linked condition characterized by a deficiency of H(2)O(2) production during phagocytosis and low microbicidal activity, the iodination ratio of CGD cells was dramatically less than that of normal PMN (by about two orders of magnitude). Capacity for iodination was correlated with bactericidal activity toward E. coli. At low bacterial loads (ca. 5:1), phagocytes killed efficiently, and little discrepancy in ability among cell types was apparent. Under the stress of higher loads of (14)C-labeled E. coli (ca. 100:1), differences in bactericidal activity were exaggerated, and a substantial disparity between MN and PMN was observed in favor of the latter. The hierarchy for killing efficiencies therefore agreed with that for iodination, with one notable exception: mouse MAC were consistently competent in their killing activity, more so than MN, even though they virtually lack peroxidase and the ability to iodinate ingested bacteria.

Glycogen in leukocytes from bovine blood and milk

Glycogen content was determined quantitatively by the Anthrone reagent method in leukocytes obtained from blood and milk of five cows. Distribution of glycogen in leukocytes was studied by microscopic examination of slides stained by Periodic acid-Schiff (PAS) reaction. Blood glucose concentrations were investigated in these animals by standard procedures. In two of five cows both blood glucose levels and blood leukocyte glycogen levels on the same day were determined for six consecutive days. One hundred and two blood leukocyte samples from five cows had a mean glycogen content of 1.32 +/- 0.04 (S.E.) mg/10(9) WBC, and 6.11 +/- 0.17 (S.E.) mg/10(9) PMNs. Leukocyte preparations from 80 samples of milk comprising 97 to 98% PMNs contained 3.81 +/- 0.18 (S.E.) mg glycogen/10(9) milk leukocytes. In PAS preparations of blood and milk leukocytes glycogen was found almost exclusively in PMNs. Glycogen granules, present frequently in PMNs and occasionally in monocytes and large lymphocytes from blood, were not observed in those from milk. The glycogen level in milk leukocytes was significantly lower (P = <0.01) than that of the blood PMNs in every cow, and the overall mean difference between levels for milk leukocytes and blood PMNs was highly significant (P = <0.001). Mean blood glucose concentration in the five cows was 44.46 +/- 0.66 (S.E.) mg%. There was no significant relationship between blood glucose and blood leukocyte glycogen levels in the five corresponding cows; nor between blood glucose and blood PMN glycogen levels on the same day in either of two cows investigated. Leukocyte preparations from milk samples obtained on the second day following intramammary infusion of endotoxin consistently contained markedly less glycogen than the leukocyte preparations from first day post-infusion samples. These tended to level off and became intermediate between first and second day levels. It is postulated that the poor phagocytic competence of leukocytes from bovine mammary glands compared to their counterparts in blood observed by various workers may be due partially to low energy reserves in these cells.

Serum-mediated immune cellular responses to Brucella melitensis. IV. Infection of macrophages under anaerobic conditions

Immune mechanisms active against Brucella were studied under conditions of oxygen deficiency. B. melitensis grew in rabbit serum-Tyrode medium flooded with N(2) and CO(2) gas mixtures. Immune sera from rabbits injected with B. melitensis strain Rev I possessed growth-inhibitory activity that operated in anaerobic environments against Rev I and virulent strain 6015. When mixed with macrophages, immune sera mediated even greater inhibition of bacterial growth and slowed the spread of infection throughout the tissue culture. Although under anaerobic conditions the rate of phagocytosis was reduced, the macrophages in immune serum killed significant percentages of Brucella, suggesting that an antibacterial mechanism had been activated. Sonic extracts of macrophages prepared and tested under anaerobic conditions depressed the growth rate of strain Rev I. The extracts, however, exhibited no immediate killing capacity when tested in Tyrode solution. A factor from serum was required for depression of the growth rate.