Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

BUOYANT DENSITY ANALYSIS OF MYELOID COLONY-FORMING CELLS IN GERMFREE AND CONVENTIONAL MICE

BUOYANT DENSITY ANALYSIS OF MYELOID COLONY-FORMING CELLS IN GERMFREE AND CONVENTIONAL MICE v. B Y R N E ,H I L D E G A R D E I T A N D H w. H E I T Department o Internal Medicine, Hemaiology Division, and Department f of Clinical Physiology, University o Ulm, Federal Republic Germany f (Received 24 August 1978; revision accepted 30 March 1979) ABSTRACT Granulocyte-macrophage colony-forming cells (CFUc), in the bone marrow of germfree and conventional CBA mice, were compared quantitatively and qualitatively. Cells were separated on the basis of their buoyant density by equilibrium centrifugation in continuous albumin density gradients. C F U c in the density subpopulations were detected by culture in agar containing three different types of colony stimulating factor (CSF). The sources of the CSF were postendotoxin mouse serum (CSF,,), mouse lung conditioned medium (CSF MLCM) and human urine (CSF,,). Mice were removed from the germfree environment and the buoyant density status of their C F U c was examined 1,4 and 8 weeks later. No difference was found between germfree and conventional mice in the number of nucleated cells per femur or in their modal density. Neither was the number of C F U c per femur different. The cell cycle status of CFUc, as determined by the thymidine suicide technique was not significantly different. Functional heterogeneity was found among the density subpopulations for both groups of mice. This depended on the type of CSF. The density distribution of C F U c was significantly different in germfree mice. There were proportionately more low density CFUc. The mean modal density of C F U c under CSF,, stimulation was less by 0.0045 g/cm3 in germfree mice. The removal of mice from the germfree environment resulted in a shift of the distribution to higher densities. The trend was towards the conventional situation. The significance of the buoyant density status of C F U c is discussed. Germfree (axenic) mice, because of the absence of a bacterial flora, offer an interesting model for studying the effect of the microbial environment on the granulocyte-macrophage pathway. Correspondence: Dr Peter Byrne, Albert Einstein College of Medicine, Yeshiva University (Microbiology and Immunology), F426, 1300 Morris Park Avenue, Bronx, N.Y.10461, U.S.A. 3 0008-8730/79/1100-0635$02.00 $ 1979 Blackwell Scientific Publications 635 P. V. Byme, H . Heit and W. Heit Comparative studies on the kinetics of this system have been carried out with conventional and germfree mice. Boggs et al. (1967) investigated the total number of neutrophil precursors in the bone marrow of germfree mice. They found no significant differences from conventional mice. However, the blood neutrophil concentration was found to be slightly lower in germfree mice. Fliedner, Fache & Adolphi (1966) found no difference in the turnover rate of blood granulocytes. The maturation time for neutrophils was found to be longer in germfree mice. Germfree mice were more resistant to whole-body X-irradiation (Wilson, 1963; Heit, Fliedner & Fache, 1972). The recovery of myeloid progenitor cells (CFUc) after irradiation was delayed in germfree mice (Kubanek et a[., 1975). The regeneration could be influenced by endotoxin or by hypertransfusion. The level of colony-stimulating factor (CSF), which is necessary for in vitro colony formation and may play a physiological role in uivo, has been examined in germfree mouse serum. Metcalf, Foster & Pollard (1967b) found low but detectable serum levels in normal germfree mice. In contrast to conventional mice, there was no elevation of serum CSF levels in germfree mice after X-irradiation. This suggested that the host bacterial flora plays a role in determining CSF level (Morley et af., 1972). After injection with endotoxin, both conventional and germfree mice responded with an elevation of the serum CSF level (Quesenberry et al., 1974). This suggests that bacterial products from the host’s flora may influence the serum CSF level and indirectly act on the myelopoietic system. Cell separation methods have been used to characterize progenitor cells of granulocytes and/or macrophages (CFUc). Separation of cells on the basis of their sedimentation rate by velocity sedimentation (Metcalf & MacDonald, 1975) or buoyant density by continuous albumin density gradients (Byrne, Heit & Kubanek, 1977), showed that CFUc are a heterogeneous population with respect to the type of colony they produce and their responsiveness to different types of CSF. Preliminary studies indicated that CFUc from germfree mice had a lower density. Further work was undertaken using different types of CSF with different actions on CFUc. The findings are reported here and establish that the density distribution of CFUc in germfree mice is significantly different from that in conventional animals and reflects a different physiological state. MATERIALS AND METHODS Mice Both conventional and germfree mice were CBA/Ca aged between 10 and 16 weeks. Germfree mice were housed in flexible film isolators (Trexler & Reynolds, 1957). When mice were removed from the isolators they were placed in cages adjacent to the conventional mice and fed the same diet. Cell suspension All femoral marrow suspensions were prepared in a HEPES-buffered balanced salt solution (pH 7.2) and of mouse serum osmolarity (308 mOs equivalent to 0.168 M NaCl), with 10% foetal calf serum (Shortman, Williams & Adams, 1972). The osmolarity of foetal calf serum was found to be approximately the same as that of mouse serum. All suspensions were cleared of large and small debris as described elsewhere (Shortman et al., 1972). CFUc buoyant density in germfree mice Osmometry The osmolarity of media was measured with a Knauer Vapour Pressure Osmometer. Mannitol was used as the basic standard (Williams, Kraft & Shortman, 1972). Density gradient separation Bone marrow cells were separated on the basis of their buoyant density in a continuous density gradient according to the method of Shortman (1968). Bovine Serum Albumin (BSA), Fraction V (Armour, Chicago) was used as the gradient medium (pH 5.1) and its osmolarity was strictly controlled at 308 mOs (Williams & Van den Engh, 1975). A linear isotonic density gradient was generated at 4OC by mixing 17% (w/w) BSA with 30% (w/w) BSA containing approximately 1 x lo8 cells in a mixing chamber. The cells were brought to equilibrium in the gradient by centrifuging for 45 min at 4000 g and 4OC. The gradient was fractionated into fifteen to twenty fractions of equal volume by upward displacement with bromobenzene. The density of each fraction was accurately determined on a Kerosine-bromobenzene gradient. Each fraction was diluted with isotonic buffer and the cells recovered by centrifugation. The cells were counted and plated in agar. A computer program was written to process the results and plot the normalized density profiles. An additional program was used to a plot a composite profile from replicate gradients. Agar culture The culture system was a modification of the method of Bradley & Metcalf (1966). Bone marrow cells were cultured in 0.3% agar in McCoy’s 5A medium supplemented with 20% serum and a source of CSF. Each Petri dish contained approximately 50,000 cells and in addition 2-mercaptoethanol at a concentration of 1 x M (Byrne et al., 1977) in a volume of 1.1 ml. The serum was a mixture of horse and foetal calf serum in the proportions that gave maximum colony numbers. The cultures were incubated at 37OC in a fully humidified atmosphere of 3% CO, in air in sealed air-tight glass jars (Firket, 1969). Colonies, that is aggregates of more than fifty cells, were scored with a dissecting microscope after 6 days of incubation. The morphology of the colonies at day 6 was examined by transferring colonies with a fine Pasteur pipette on to glass slides and staining with 0.6% Orcein in 60% acetic acid. The colonies were typed as pure granulocytic, pure macrophage or mixed by criteria described elsewhere (Metcalf, Bradley & Robinson, 1967a). Where possible, a sample of at least fifty sequential colonies was taken. Colony-stimulatingfactor Post-endotoxin mouse serum (CSF,,). Male NMRI mice weighing approximately 30 g were injected intravenously with 50 pg lipopolysaccaride B, S . abortus Equi (Difco, Detroit). The mice were bled after 3 hr (Metcalf, 1971) and the serum was heat treated-for 20 min at 56OC. Some serum samples showed high dose inhibition and were appropriately diluted to ensure maximum colony numbers when 0.1 ml was added to each plate. Mouse lung conditioned medium (CSFMLcM). lungs from the same mice were excised The after 3 hr and conditioned medium was prepared by a modification of the method of Sheridan & Metcalf (1973). Each lung was incubated in 5 ml of serum-free McCoy’s 5A supplemented medium for 48 hr at 37°C in a fully humidified atmosphere of 3% CO, in air. The medium P. V. Byme, H. Heit and W. Heit was harvested, heat treated for 30 min at 56OC and centrifuged at 12,000 g for 15 min at 4OC. It was then dialysed for 3 days with 5 x 10 volume changes of distilled water and again centrifuged. This was followed by batch chromatography with calcium phosphate gel. One ml of the gel preparation (Stanley et al., 1972; Senn, Messner & Stanley, 1974) was added per 5 mg medium protein. The adsorbed C S F was block eluted with phosphate buffer p H 7.4 between 0.01 M and 0.05 M. The C S F containing eluant was Millipore filtered (0.45 p) and stored frozen. Human urinary colony-stimulating factor (CSF,,). This was prepared according to Stanley et al. (1972). Dialysed urine underwent batch chromatography with calcium phosphate gel. C S F was eluted with 0.03 M phosphate buffer p H 7.4. The C S F containing eluant was Millipore filtered and stored frozen. Different batches of the three types of C S F were used in the present experiments which were carried out over a period of 3 years. Dose-response curves indicated n o high dose or CSF,, and with each batch, 0-1 ml of the undiluted inhibition with either CSF,,,, preparation ensured plateau levels of CSF. Tritiated thymidine suicide Unfractionated bone marrow cells were incubated with high specific activity tritiated thymidine ([3H]TdR) to indicate the proportion of C F U c in the S phase of the cell cycle. The method was basically as described elsewhere (Iscove, Till & McCulloch, 1970; Williams & Moore, 1973). 2.5 x lo6 cells per ml were incubated with either 50 pCi/ml [)H]TdR (21 Ci/mM) or 100 pg/ml 'cold' TdR at 37OC for 20 min with continuous agitation. The suspension was in HEPES-buffered balanced salt solution p H 7.2 and 308 mOs. Afterwards, the cells were washed with ice-cold buffer containing 10% foetal calf serum and 100 pg/ml cold TdR and centrifuged at 400 g for 10 min at 4°C. This was repeated twice. The cells were plated in agar at two cell doses, 2.5 and 5 x lo4 cells/plate. The culture medium contained in addition 50 pglml cold TdR. RESULTS Femoral granulocyte-macrophage progenitor cells (CFUc) Conventional and germfree mice were compared for the incidence and total content per femur of CFUc. The cultures were stimulated with three different types of CSF. Table 1 summarizes the results. TABLE Colony incidence and total C F U c per femur (mean k s.e.m.) I. Conventional Germfree CSF No. of experiments* Colonies per platet 69 k 4.3 53 k 3 . 9 62 i 5 . 7 Femur (x No. of experiments Colonies per plate 55 f 7.6 61 & 7.1 53 ? 7.9 Femur (x 10-3) ES MLCM HU 17.8 ? 1 . 4 13.9 _+ 1.3 16.3 ? 1.7 17.5 t 4.2 18.0 k 3.5 14.4 ? 4 . 0 * Each experiment was carried out with pooled marrow cells from at least two mice. t Each experiment had at least three plates, each containing approx. 50,000 bone marrow cells. CFUc buoyant density in germfree mice There were no significant difference with either CSF,,,, or CSF,, stimulation. The colony incidence was slightly lower in germfree mice under CSF,, stimulation. This was probably significant by the Student's t-test (P< 0-05). The total number of nucleated cells per femur was not significantly different. There were and 12.3 2 0.9 ( x nucleated cells per femur (+s.e.m.) in 12.8 5 0.6 ( x conventional and germfree mice respectively. The were no significant differences between germfree and conventional mice in terms of total CFUc per femur with each type of CSF. Tritiated thymidine suicide Conventional and germfree mice were compared in terms of the percentage of CFUc in the S phase of the cell cycle. Table 2 presents the findings which indicate the percentage in S phase. TABLE Tritiated thymidine suicide of CFUc from conventional and germtree mice 2. Y Reduction u in CFUc (mean 2 s.e.m.)t CSFH, No. of Mice Conventional Germ free experiments* CSFES 45.1 2 10.1 36.6 k 4.2 ~ ~ CSFMLCM 16.7 & 2.5 22.5 f 4.9 ~~~ ~~ ~~ t Percentage reduction was determined from the mean colony count at two cell doses each with three plates. * Pooled femur cells from three mice in each experiment. With conventional mice, it was found that the suicide index, i.e. the percentage of CFUc in S phase with CSFMLcM was significantly lower than with CSF,, (P < 0.01) and probably significantly lower than CSF,, (P < 0-05). There was no significant difference between CSF,, and CSF,,. With germfree mice, the percentage of cells in S phase with CSFMLcM found to be was significantly lower than with CSF,,. However, no other significant differences were found. When the two groups of mice were compared, no significant differences were found. Density distribution of nucleated cells With conventional mice the modal density was in the range 1.075-1.079 g/cm3 with a mean of 1.0775 g/cm3. Figure 1 shows the composite distribution profile from a number of gradients. The mean recovery of nucleated cells was 63.9% and 77.1% for conventional and germfree mice respectively. The denser fractions contained predominately granulocytes. The modal density for germfree mice lay in the range 1-073-1.078 g/cm3 with a mean of 1.0755 g/cm3. The profiles showed considerable variation as demonstrated by the error bars in Fig. 1. The difference in modal density between the two groups was not significant. There were relatively more cells in the low density regions in germfree mice. P . V. Byme, H . Heit and W. Heit 60:. J/ Density (g/cm’) 1-06 1-08 FIG. 1. Composite nucleated cell density profile derived from ten gradients with conventional mice (V) and four gradients with germfree mice (. Each composite was generated by computer. Each . ) point represents the total number of nucleated cells per fraction (per density increment) expressed as a percentage of the maximum. Each point is the mean of interpolated values at density increments of 0.0010 g/cm’. Vertical lines represent the standard error of the mean. Density distribution o CFUc f Figure 2 shows the density profiles from a typical gradient for conventional mice and for germfree mice. There is an obvious difference between the two groups, with the germfree distribution lying at lower density. There were more CFUc of low density in germfree mice and this was reflected in the relationship between profiles with the CSF,,,, and CSF,,. CSFMLcM stimulated predominately low density CFUc and CSF,, predominately high density ones. The shift in modal density was more striking with CSF,, which stimulated both high and low density CFUc. The overall picture of the density relationship between conventional and germfree CFUc is presented in Fig. 3. Each density profile is a composite produced by the computer from a number of gradients. The recovery of CFUc in each gradient was approximately the same as that for nucleated cells indicating that there was no selective loss of CFUc or concentration of cells (which might inhibit colony formation in certain fractions). With conventional mice, the mean CFUc recovery was 64.6, 66.8 and 57.7% when stimulated with CSF,,, CSF,,,, and CSF,, respectively. With germfree mice the mean recoveries were 73.4, 72.4 and 76.1% respectively. All profiles with germfree mice are shifted towards the low density end of the gradient relative to conventional mice. Table 3 summarizes the differences in distribution in terms of the modal density. CFUc buoyant density in germfree mice h:; I f E .x u- 1-06 Density (g/cm’) 1-09 FIG.2. The density distribution of CFUc in conventional (a) and germfree (b) mice. Cultures were stimulated by CSF,, (+), CSF,,,, (0) CF, (0). point represent the total number and S, Each of colonies per fraction per density increment expressed as a percentage of the maximum. With CSF,,, the profde was highly reproducible and the modal density had a standard deviation of 0.0006 g/cm3 (equivalent to 0.25-0.5 a density fraction). The difference in the variation modal density was 0.0045 g/cm3 and was highly significant. With CSF,,,,, was greater but the difference (04040g/cm3) was less significant. With CSF,, the difference was only 0.0025 g/cm3, but was significant. P . V.Byme, H . Heit and W. Heit Density (g/cm3) CFUc buoyant density in germfree mice TABLE Mean modal density of CFUc (ks.d.) 3. Conventional CSF ES MLCM HU Germfree No. of gradients g/cm3 1.0733 k 0.0006 1.0704 f 0.003 1 1.0737 f 0.0012 No. of gradients g/cm’ 1.0688 f 0.0006 1.0671 f 0.0018 1.0712 0.0005 Student’s r-test P < 0.001 P < 0.05 P < 0.01 Colony morphology The morphology of colony cells derived from different density fractions was examined. Figure 4 shows the relative proportions of colonies stimulated by CSF,,. Pure granulocytic and mixed colonies predominated in the low density regions for both types of mice. Pure macrophage colonies tended to predominate in higher density regions, CFUc responding to CSF,,,, which produced mainly granulocytic colonies had a low density. Responders to CSF,,, which produced predominately macrophage colonies, had a high density Removal of mice from the germfree environment Germfree mice were removed from isolators and the bone marrow cells were investigated 1, 4 and 8 weeks afterwards. The density distribution of nucleated cells at each time was not noticeably different from the germfree or conventional state. An intestinal flora almost identical to that in conventional mice had been established by 1 week. The density distribution of CFUc was markedly altered from the germ-free state, This is illustrated in Fig. 5. There was a predominance of high density CFUc. There was a clear distinction between the profiles with CSFMLcM CSF,,. After 1 week, the profiles spread into higher density and regions. The maximum was found with CSF,,,,. After 4 weeks, the pattern was radically altered and resembled that in conventional mice as regards the modal density. The maximum was found with CSF,,. After 8 weeks, the modal density with CSF,, had become even higher. DISCUSSION There appear to be at least two distinct subclasses of CSF. One subclass stimulates granulocytic colony formation with subsequent concentration dependent macrophage formation. This has been purified from mouse lung conditioned medium (Burgess, Camakaris & Metcalf, 1977). The other subclass only stimulates macrophage colony formation and has been purified from human urine and conditioned medium (Stanley, Chen & Lin, 1978) These two subclasses may act wholly or partly on different progenitor cells. This is shown by the fact that the cells responding to the different types of CSF have different FIG. 3. Composite density profiles of femoral CFUc of conventional (v)and germfree mice (0). Cultures were stimulated with CSF,, (a). CSF,,,, (b) and CSF,, (c). Each point represents the mean of interpolated values at density increments of 0.0010 g/cmJ from a number of different gradients. Vertical bars represent the standard error of the mean. The number of gradients with conventional mice was in (a) eight, (b) five and (c) five. The number of gradients with germfree mice was in (a) four, (b) four and (c) three. P . V. Byrne, H . Heit and W . Heit @\.-a MIX v1 MIX 0 ._ I*06 Density (g/cm3) I SO8 1-09 FIG. 4. The proportional distributions of colony types derived from density subpopulations of conventional (a) and germfree (b) mice. Cultures were stimulated with CSF,,. Colonies were typed as pure granulocytic (G), pure macrophage (M) or mixed (MIX) and expressed as a percentage of the total sample. buoyant densities and proportions in S phase of the cell cycle in both conventional and germfree mice. Low density CFUc, detectable with CSF,,,,, form predominately granulocytic colonies, but macrophage differentiation occurs with longer incubation (Byrne et al., 1977) and lower CSF concentrations (Metcalf 8c MacDonald, 1975; Burgess et al., 1977). High density CFUc detectable with CSF,,, show preferential differentiation to macrophages (Byrne et al., 1977). An important consideration is do low and high density CFUc represent a differentiation sequence? The low density CFUc appear to be bipotent and give rise to both granulocytes FIG. 5 . The density distribution of C F U c of mice removed 1 (a), 4 (b) and 8 weeks (c) previously (0)and from the germfree environment. Cultures were stimulated with CSF,, (+), CSF,,,, CSF,, (0).Each point represents the total number of colonies per fraction per density increment expressed as a percentage of the maximum. Results from single gradients. CFUc buoyant density in germfree mice 90; 30: 20: .Al 1-09 P. V. Byme, H. Heit and W. Heit and macrophages. The high density CFUc, on the other hand, appear to be mainly restricted to macrophage differentiation. This raises the possibility that high density C F U c are more mature and may be the progeny of low density CFUc. Commitment of C F U c to macrophage formation may occur by development of specific cell receptors for the macrophage type of C S F which is antigenically distinct from that in mouse lung conditioned medium (Stanley, 1978). This transition may also be associated with an increase in C F U c buoyant density. There was no quantitative difference between germfree and conventional mice in the number of C F U c per femur. Qualitatively, there was a significant difference. There was a predominance of low density C F U c in germfree mice. The nature of this density difference remains open to speculation but it may reflect the relationship between nuclear and cytoplasmic size. If the buoyant density status of C F U c is characteristic of the degree of stress in the system and the demand for mature granulocytes and macrophages then we would expect this status to alter when mice are removed from the germfree environment, as was found. At each time studied after the mice were removed from the germfree environment there was an increase in high density CFUc. These changes can be interpreted as being due to bacterial association giving a changing demand for mature phagocytic cells. This has been reported to occur within hours of removal from the germfree environment (Boggs et al., 1967). These investigators also found that the post-mitotic maturation pool of the marrow was depleted and the mitotic pool was enlarged relative to both conventional and germfree mice after removal from isolators. The system was not normal after 2 months but had returned to normal by 3 months. In the present study, the mice were found to have established an almost normal gastrointestinal flora after 1 week. The modal density of the C F U c of germfree mice is intermediate between that of conventional adult marrow and that of mouse foetal liver which is of lower density (Moore, McNeill & Haskill, 1970). We found the modal density of CFUc, detected with CSF,,, in 16 day foetal liver to be 1.0644 g/cm3 under conditions of mouse osmolarity (unpublished). It has been demonstrated that C F U c have a capacity for self-renewal (Sumner et al., 1972). C F U c density may serve as an index of ‘generation age’ with the C F U c of germfree mice being ‘younger’ in the sense that they have undergone fewer self-renewals. Foetal liver C F U c were found to increase in density with increasing gestation age (Moore el al., 1970). The density distribution of pluripotent stem cells (CFUs) is lighter but partially overlaps that of C F U c (Worton, McCulloch & Till, 1969; Haskill, McNeill & Moore, 1970). This has also been found in foetal liver at different stages of development (Moore et al., 1970). This ‘out of phase’ relationship suggests that the two populations are closely related. If this relationship persists in germfree mice then we would expect C F U s in germfree mice to have a lower density than in conventional mice perhaps reflecting a different functional state. This aspect should be considered in relation to the improved survival of germfree mice after irradiation. ACKNOWLEDGMENT Research was supported by the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 1 12 (Zellsystemphysiologie). We are indebted to Mrs Sigrid Schropfer for her excellent technical assistance. CFUc buoyant density in germfree mice http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Cell Proliferation Wiley

BUOYANT DENSITY ANALYSIS OF MYELOID COLONY-FORMING CELLS IN GERMFREE AND CONVENTIONAL MICE

Loading next page...
 
/lp/wiley/buoyant-density-analysis-of-myeloid-colony-forming-cells-in-germfree-TW7pBadKs0

References (29)

Publisher
Wiley
Copyright
1979 Blackwell Publishing Ltd
ISSN
0960-7722
eISSN
1365-2184
DOI
10.1111/j.1365-2184.1979.tb00182.x
Publisher site
See Article on Publisher Site

Abstract

v. B Y R N E ,H I L D E G A R D E I T A N D H w. H E I T Department o Internal Medicine, Hemaiology Division, and Department f of Clinical Physiology, University o Ulm, Federal Republic Germany f (Received 24 August 1978; revision accepted 30 March 1979) ABSTRACT Granulocyte-macrophage colony-forming cells (CFUc), in the bone marrow of germfree and conventional CBA mice, were compared quantitatively and qualitatively. Cells were separated on the basis of their buoyant density by equilibrium centrifugation in continuous albumin density gradients. C F U c in the density subpopulations were detected by culture in agar containing three different types of colony stimulating factor (CSF). The sources of the CSF were postendotoxin mouse serum (CSF,,), mouse lung conditioned medium (CSF MLCM) and human urine (CSF,,). Mice were removed from the germfree environment and the buoyant density status of their C F U c was examined 1,4 and 8 weeks later. No difference was found between germfree and conventional mice in the number of nucleated cells per femur or in their modal density. Neither was the number of C F U c per femur different. The cell cycle status of CFUc, as determined by the thymidine suicide technique was not significantly different. Functional heterogeneity was found among the density subpopulations for both groups of mice. This depended on the type of CSF. The density distribution of C F U c was significantly different in germfree mice. There were proportionately more low density CFUc. The mean modal density of C F U c under CSF,, stimulation was less by 0.0045 g/cm3 in germfree mice. The removal of mice from the germfree environment resulted in a shift of the distribution to higher densities. The trend was towards the conventional situation. The significance of the buoyant density status of C F U c is discussed. Germfree (axenic) mice, because of the absence of a bacterial flora, offer an interesting model for studying the effect of the microbial environment on the granulocyte-macrophage pathway. Correspondence: Dr Peter Byrne, Albert Einstein College of Medicine, Yeshiva University (Microbiology and Immunology), F426, 1300 Morris Park Avenue, Bronx, N.Y.10461, U.S.A. 3 0008-8730/79/1100-0635$02.00 $ 1979 Blackwell Scientific Publications 635 P. V. Byme, H . Heit and W. Heit Comparative studies on the kinetics of this system have been carried out with conventional and germfree mice. Boggs et al. (1967) investigated the total number of neutrophil precursors in the bone marrow of germfree mice. They found no significant differences from conventional mice. However, the blood neutrophil concentration was found to be slightly lower in germfree mice. Fliedner, Fache & Adolphi (1966) found no difference in the turnover rate of blood granulocytes. The maturation time for neutrophils was found to be longer in germfree mice. Germfree mice were more resistant to whole-body X-irradiation (Wilson, 1963; Heit, Fliedner & Fache, 1972). The recovery of myeloid progenitor cells (CFUc) after irradiation was delayed in germfree mice (Kubanek et a[., 1975). The regeneration could be influenced by endotoxin or by hypertransfusion. The level of colony-stimulating factor (CSF), which is necessary for in vitro colony formation and may play a physiological role in uivo, has been examined in germfree mouse serum. Metcalf, Foster & Pollard (1967b) found low but detectable serum levels in normal germfree mice. In contrast to conventional mice, there was no elevation of serum CSF levels in germfree mice after X-irradiation. This suggested that the host bacterial flora plays a role in determining CSF level (Morley et af., 1972). After injection with endotoxin, both conventional and germfree mice responded with an elevation of the serum CSF level (Quesenberry et al., 1974). This suggests that bacterial products from the host’s flora may influence the serum CSF level and indirectly act on the myelopoietic system. Cell separation methods have been used to characterize progenitor cells of granulocytes and/or macrophages (CFUc). Separation of cells on the basis of their sedimentation rate by velocity sedimentation (Metcalf & MacDonald, 1975) or buoyant density by continuous albumin density gradients (Byrne, Heit & Kubanek, 1977), showed that CFUc are a heterogeneous population with respect to the type of colony they produce and their responsiveness to different types of CSF. Preliminary studies indicated that CFUc from germfree mice had a lower density. Further work was undertaken using different types of CSF with different actions on CFUc. The findings are reported here and establish that the density distribution of CFUc in germfree mice is significantly different from that in conventional animals and reflects a different physiological state. MATERIALS AND METHODS Mice Both conventional and germfree mice were CBA/Ca aged between 10 and 16 weeks. Germfree mice were housed in flexible film isolators (Trexler & Reynolds, 1957). When mice were removed from the isolators they were placed in cages adjacent to the conventional mice and fed the same diet. Cell suspension All femoral marrow suspensions were prepared in a HEPES-buffered balanced salt solution (pH 7.2) and of mouse serum osmolarity (308 mOs equivalent to 0.168 M NaCl), with 10% foetal calf serum (Shortman, Williams & Adams, 1972). The osmolarity of foetal calf serum was found to be approximately the same as that of mouse serum. All suspensions were cleared of large and small debris as described elsewhere (Shortman et al., 1972). CFUc buoyant density in germfree mice Osmometry The osmolarity of media was measured with a Knauer Vapour Pressure Osmometer. Mannitol was used as the basic standard (Williams, Kraft & Shortman, 1972). Density gradient separation Bone marrow cells were separated on the basis of their buoyant density in a continuous density gradient according to the method of Shortman (1968). Bovine Serum Albumin (BSA), Fraction V (Armour, Chicago) was used as the gradient medium (pH 5.1) and its osmolarity was strictly controlled at 308 mOs (Williams & Van den Engh, 1975). A linear isotonic density gradient was generated at 4OC by mixing 17% (w/w) BSA with 30% (w/w) BSA containing approximately 1 x lo8 cells in a mixing chamber. The cells were brought to equilibrium in the gradient by centrifuging for 45 min at 4000 g and 4OC. The gradient was fractionated into fifteen to twenty fractions of equal volume by upward displacement with bromobenzene. The density of each fraction was accurately determined on a Kerosine-bromobenzene gradient. Each fraction was diluted with isotonic buffer and the cells recovered by centrifugation. The cells were counted and plated in agar. A computer program was written to process the results and plot the normalized density profiles. An additional program was used to a plot a composite profile from replicate gradients. Agar culture The culture system was a modification of the method of Bradley & Metcalf (1966). Bone marrow cells were cultured in 0.3% agar in McCoy’s 5A medium supplemented with 20% serum and a source of CSF. Each Petri dish contained approximately 50,000 cells and in addition 2-mercaptoethanol at a concentration of 1 x M (Byrne et al., 1977) in a volume of 1.1 ml. The serum was a mixture of horse and foetal calf serum in the proportions that gave maximum colony numbers. The cultures were incubated at 37OC in a fully humidified atmosphere of 3% CO, in air in sealed air-tight glass jars (Firket, 1969). Colonies, that is aggregates of more than fifty cells, were scored with a dissecting microscope after 6 days of incubation. The morphology of the colonies at day 6 was examined by transferring colonies with a fine Pasteur pipette on to glass slides and staining with 0.6% Orcein in 60% acetic acid. The colonies were typed as pure granulocytic, pure macrophage or mixed by criteria described elsewhere (Metcalf, Bradley & Robinson, 1967a). Where possible, a sample of at least fifty sequential colonies was taken. Colony-stimulatingfactor Post-endotoxin mouse serum (CSF,,). Male NMRI mice weighing approximately 30 g were injected intravenously with 50 pg lipopolysaccaride B, S . abortus Equi (Difco, Detroit). The mice were bled after 3 hr (Metcalf, 1971) and the serum was heat treated-for 20 min at 56OC. Some serum samples showed high dose inhibition and were appropriately diluted to ensure maximum colony numbers when 0.1 ml was added to each plate. Mouse lung conditioned medium (CSFMLcM). lungs from the same mice were excised The after 3 hr and conditioned medium was prepared by a modification of the method of Sheridan & Metcalf (1973). Each lung was incubated in 5 ml of serum-free McCoy’s 5A supplemented medium for 48 hr at 37°C in a fully humidified atmosphere of 3% CO, in air. The medium P. V. Byme, H. Heit and W. Heit was harvested, heat treated for 30 min at 56OC and centrifuged at 12,000 g for 15 min at 4OC. It was then dialysed for 3 days with 5 x 10 volume changes of distilled water and again centrifuged. This was followed by batch chromatography with calcium phosphate gel. One ml of the gel preparation (Stanley et al., 1972; Senn, Messner & Stanley, 1974) was added per 5 mg medium protein. The adsorbed C S F was block eluted with phosphate buffer p H 7.4 between 0.01 M and 0.05 M. The C S F containing eluant was Millipore filtered (0.45 p) and stored frozen. Human urinary colony-stimulating factor (CSF,,). This was prepared according to Stanley et al. (1972). Dialysed urine underwent batch chromatography with calcium phosphate gel. C S F was eluted with 0.03 M phosphate buffer p H 7.4. The C S F containing eluant was Millipore filtered and stored frozen. Different batches of the three types of C S F were used in the present experiments which were carried out over a period of 3 years. Dose-response curves indicated n o high dose or CSF,, and with each batch, 0-1 ml of the undiluted inhibition with either CSF,,,, preparation ensured plateau levels of CSF. Tritiated thymidine suicide Unfractionated bone marrow cells were incubated with high specific activity tritiated thymidine ([3H]TdR) to indicate the proportion of C F U c in the S phase of the cell cycle. The method was basically as described elsewhere (Iscove, Till & McCulloch, 1970; Williams & Moore, 1973). 2.5 x lo6 cells per ml were incubated with either 50 pCi/ml [)H]TdR (21 Ci/mM) or 100 pg/ml 'cold' TdR at 37OC for 20 min with continuous agitation. The suspension was in HEPES-buffered balanced salt solution p H 7.2 and 308 mOs. Afterwards, the cells were washed with ice-cold buffer containing 10% foetal calf serum and 100 pg/ml cold TdR and centrifuged at 400 g for 10 min at 4°C. This was repeated twice. The cells were plated in agar at two cell doses, 2.5 and 5 x lo4 cells/plate. The culture medium contained in addition 50 pglml cold TdR. RESULTS Femoral granulocyte-macrophage progenitor cells (CFUc) Conventional and germfree mice were compared for the incidence and total content per femur of CFUc. The cultures were stimulated with three different types of CSF. Table 1 summarizes the results. TABLE Colony incidence and total C F U c per femur (mean k s.e.m.) I. Conventional Germfree CSF No. of experiments* Colonies per platet 69 k 4.3 53 k 3 . 9 62 i 5 . 7 Femur (x No. of experiments Colonies per plate 55 f 7.6 61 & 7.1 53 ? 7.9 Femur (x 10-3) ES MLCM HU 17.8 ? 1 . 4 13.9 _+ 1.3 16.3 ? 1.7 17.5 t 4.2 18.0 k 3.5 14.4 ? 4 . 0 * Each experiment was carried out with pooled marrow cells from at least two mice. t Each experiment had at least three plates, each containing approx. 50,000 bone marrow cells. CFUc buoyant density in germfree mice There were no significant difference with either CSF,,,, or CSF,, stimulation. The colony incidence was slightly lower in germfree mice under CSF,, stimulation. This was probably significant by the Student's t-test (P< 0-05). The total number of nucleated cells per femur was not significantly different. There were and 12.3 2 0.9 ( x nucleated cells per femur (+s.e.m.) in 12.8 5 0.6 ( x conventional and germfree mice respectively. The were no significant differences between germfree and conventional mice in terms of total CFUc per femur with each type of CSF. Tritiated thymidine suicide Conventional and germfree mice were compared in terms of the percentage of CFUc in the S phase of the cell cycle. Table 2 presents the findings which indicate the percentage in S phase. TABLE Tritiated thymidine suicide of CFUc from conventional and germtree mice 2. Y Reduction u in CFUc (mean 2 s.e.m.)t CSFH, No. of Mice Conventional Germ free experiments* CSFES 45.1 2 10.1 36.6 k 4.2 ~ ~ CSFMLCM 16.7 & 2.5 22.5 f 4.9 ~~~ ~~ ~~ t Percentage reduction was determined from the mean colony count at two cell doses each with three plates. * Pooled femur cells from three mice in each experiment. With conventional mice, it was found that the suicide index, i.e. the percentage of CFUc in S phase with CSFMLcM was significantly lower than with CSF,, (P < 0.01) and probably significantly lower than CSF,, (P < 0-05). There was no significant difference between CSF,, and CSF,,. With germfree mice, the percentage of cells in S phase with CSFMLcM found to be was significantly lower than with CSF,,. However, no other significant differences were found. When the two groups of mice were compared, no significant differences were found. Density distribution of nucleated cells With conventional mice the modal density was in the range 1.075-1.079 g/cm3 with a mean of 1.0775 g/cm3. Figure 1 shows the composite distribution profile from a number of gradients. The mean recovery of nucleated cells was 63.9% and 77.1% for conventional and germfree mice respectively. The denser fractions contained predominately granulocytes. The modal density for germfree mice lay in the range 1-073-1.078 g/cm3 with a mean of 1.0755 g/cm3. The profiles showed considerable variation as demonstrated by the error bars in Fig. 1. The difference in modal density between the two groups was not significant. There were relatively more cells in the low density regions in germfree mice. P . V. Byme, H . Heit and W. Heit 60:. J/ Density (g/cm’) 1-06 1-08 FIG. 1. Composite nucleated cell density profile derived from ten gradients with conventional mice (V) and four gradients with germfree mice (. Each composite was generated by computer. Each . ) point represents the total number of nucleated cells per fraction (per density increment) expressed as a percentage of the maximum. Each point is the mean of interpolated values at density increments of 0.0010 g/cm’. Vertical lines represent the standard error of the mean. Density distribution o CFUc f Figure 2 shows the density profiles from a typical gradient for conventional mice and for germfree mice. There is an obvious difference between the two groups, with the germfree distribution lying at lower density. There were more CFUc of low density in germfree mice and this was reflected in the relationship between profiles with the CSF,,,, and CSF,,. CSFMLcM stimulated predominately low density CFUc and CSF,, predominately high density ones. The shift in modal density was more striking with CSF,, which stimulated both high and low density CFUc. The overall picture of the density relationship between conventional and germfree CFUc is presented in Fig. 3. Each density profile is a composite produced by the computer from a number of gradients. The recovery of CFUc in each gradient was approximately the same as that for nucleated cells indicating that there was no selective loss of CFUc or concentration of cells (which might inhibit colony formation in certain fractions). With conventional mice, the mean CFUc recovery was 64.6, 66.8 and 57.7% when stimulated with CSF,,, CSF,,,, and CSF,, respectively. With germfree mice the mean recoveries were 73.4, 72.4 and 76.1% respectively. All profiles with germfree mice are shifted towards the low density end of the gradient relative to conventional mice. Table 3 summarizes the differences in distribution in terms of the modal density. CFUc buoyant density in germfree mice h:; I f E .x u- 1-06 Density (g/cm’) 1-09 FIG.2. The density distribution of CFUc in conventional (a) and germfree (b) mice. Cultures were stimulated by CSF,, (+), CSF,,,, (0) CF, (0). point represent the total number and S, Each of colonies per fraction per density increment expressed as a percentage of the maximum. With CSF,,, the profde was highly reproducible and the modal density had a standard deviation of 0.0006 g/cm3 (equivalent to 0.25-0.5 a density fraction). The difference in the variation modal density was 0.0045 g/cm3 and was highly significant. With CSF,,,,, was greater but the difference (04040g/cm3) was less significant. With CSF,, the difference was only 0.0025 g/cm3, but was significant. P . V.Byme, H . Heit and W. Heit Density (g/cm3) CFUc buoyant density in germfree mice TABLE Mean modal density of CFUc (ks.d.) 3. Conventional CSF ES MLCM HU Germfree No. of gradients g/cm3 1.0733 k 0.0006 1.0704 f 0.003 1 1.0737 f 0.0012 No. of gradients g/cm’ 1.0688 f 0.0006 1.0671 f 0.0018 1.0712 0.0005 Student’s r-test P < 0.001 P < 0.05 P < 0.01 Colony morphology The morphology of colony cells derived from different density fractions was examined. Figure 4 shows the relative proportions of colonies stimulated by CSF,,. Pure granulocytic and mixed colonies predominated in the low density regions for both types of mice. Pure macrophage colonies tended to predominate in higher density regions, CFUc responding to CSF,,,, which produced mainly granulocytic colonies had a low density. Responders to CSF,,, which produced predominately macrophage colonies, had a high density Removal of mice from the germfree environment Germfree mice were removed from isolators and the bone marrow cells were investigated 1, 4 and 8 weeks afterwards. The density distribution of nucleated cells at each time was not noticeably different from the germfree or conventional state. An intestinal flora almost identical to that in conventional mice had been established by 1 week. The density distribution of CFUc was markedly altered from the germ-free state, This is illustrated in Fig. 5. There was a predominance of high density CFUc. There was a clear distinction between the profiles with CSFMLcM CSF,,. After 1 week, the profiles spread into higher density and regions. The maximum was found with CSF,,,,. After 4 weeks, the pattern was radically altered and resembled that in conventional mice as regards the modal density. The maximum was found with CSF,,. After 8 weeks, the modal density with CSF,, had become even higher. DISCUSSION There appear to be at least two distinct subclasses of CSF. One subclass stimulates granulocytic colony formation with subsequent concentration dependent macrophage formation. This has been purified from mouse lung conditioned medium (Burgess, Camakaris & Metcalf, 1977). The other subclass only stimulates macrophage colony formation and has been purified from human urine and conditioned medium (Stanley, Chen & Lin, 1978) These two subclasses may act wholly or partly on different progenitor cells. This is shown by the fact that the cells responding to the different types of CSF have different FIG. 3. Composite density profiles of femoral CFUc of conventional (v)and germfree mice (0). Cultures were stimulated with CSF,, (a). CSF,,,, (b) and CSF,, (c). Each point represents the mean of interpolated values at density increments of 0.0010 g/cmJ from a number of different gradients. Vertical bars represent the standard error of the mean. The number of gradients with conventional mice was in (a) eight, (b) five and (c) five. The number of gradients with germfree mice was in (a) four, (b) four and (c) three. P . V. Byrne, H . Heit and W . Heit @\.-a MIX v1 MIX 0 ._ I*06 Density (g/cm3) I SO8 1-09 FIG. 4. The proportional distributions of colony types derived from density subpopulations of conventional (a) and germfree (b) mice. Cultures were stimulated with CSF,,. Colonies were typed as pure granulocytic (G), pure macrophage (M) or mixed (MIX) and expressed as a percentage of the total sample. buoyant densities and proportions in S phase of the cell cycle in both conventional and germfree mice. Low density CFUc, detectable with CSF,,,,, form predominately granulocytic colonies, but macrophage differentiation occurs with longer incubation (Byrne et al., 1977) and lower CSF concentrations (Metcalf 8c MacDonald, 1975; Burgess et al., 1977). High density CFUc detectable with CSF,,, show preferential differentiation to macrophages (Byrne et al., 1977). An important consideration is do low and high density CFUc represent a differentiation sequence? The low density CFUc appear to be bipotent and give rise to both granulocytes FIG. 5 . The density distribution of C F U c of mice removed 1 (a), 4 (b) and 8 weeks (c) previously (0)and from the germfree environment. Cultures were stimulated with CSF,, (+), CSF,,,, CSF,, (0).Each point represents the total number of colonies per fraction per density increment expressed as a percentage of the maximum. Results from single gradients. CFUc buoyant density in germfree mice 90; 30: 20: .Al 1-09 P. V. Byme, H. Heit and W. Heit and macrophages. The high density CFUc, on the other hand, appear to be mainly restricted to macrophage differentiation. This raises the possibility that high density C F U c are more mature and may be the progeny of low density CFUc. Commitment of C F U c to macrophage formation may occur by development of specific cell receptors for the macrophage type of C S F which is antigenically distinct from that in mouse lung conditioned medium (Stanley, 1978). This transition may also be associated with an increase in C F U c buoyant density. There was no quantitative difference between germfree and conventional mice in the number of C F U c per femur. Qualitatively, there was a significant difference. There was a predominance of low density C F U c in germfree mice. The nature of this density difference remains open to speculation but it may reflect the relationship between nuclear and cytoplasmic size. If the buoyant density status of C F U c is characteristic of the degree of stress in the system and the demand for mature granulocytes and macrophages then we would expect this status to alter when mice are removed from the germfree environment, as was found. At each time studied after the mice were removed from the germfree environment there was an increase in high density CFUc. These changes can be interpreted as being due to bacterial association giving a changing demand for mature phagocytic cells. This has been reported to occur within hours of removal from the germfree environment (Boggs et al., 1967). These investigators also found that the post-mitotic maturation pool of the marrow was depleted and the mitotic pool was enlarged relative to both conventional and germfree mice after removal from isolators. The system was not normal after 2 months but had returned to normal by 3 months. In the present study, the mice were found to have established an almost normal gastrointestinal flora after 1 week. The modal density of the C F U c of germfree mice is intermediate between that of conventional adult marrow and that of mouse foetal liver which is of lower density (Moore, McNeill & Haskill, 1970). We found the modal density of CFUc, detected with CSF,,, in 16 day foetal liver to be 1.0644 g/cm3 under conditions of mouse osmolarity (unpublished). It has been demonstrated that C F U c have a capacity for self-renewal (Sumner et al., 1972). C F U c density may serve as an index of ‘generation age’ with the C F U c of germfree mice being ‘younger’ in the sense that they have undergone fewer self-renewals. Foetal liver C F U c were found to increase in density with increasing gestation age (Moore el al., 1970). The density distribution of pluripotent stem cells (CFUs) is lighter but partially overlaps that of C F U c (Worton, McCulloch & Till, 1969; Haskill, McNeill & Moore, 1970). This has also been found in foetal liver at different stages of development (Moore et al., 1970). This ‘out of phase’ relationship suggests that the two populations are closely related. If this relationship persists in germfree mice then we would expect C F U s in germfree mice to have a lower density than in conventional mice perhaps reflecting a different functional state. This aspect should be considered in relation to the improved survival of germfree mice after irradiation. ACKNOWLEDGMENT Research was supported by the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 1 12 (Zellsystemphysiologie). We are indebted to Mrs Sigrid Schropfer for her excellent technical assistance. CFUc buoyant density in germfree mice

Journal

Cell ProliferationWiley

Published: Nov 1, 1979

There are no references for this article.