Basis Of Cell Size Research Articles

Cytoarchitectonic atlas of the cochlear nucleus of the chinchilla, Chinchilla laniger

A detailed cytoarchitectonic atlas of the chinchilla cochlear nucleus complex was prepared in the transverse plane with the Nissl method. Subdivisions of the cochlear nucleus were defined on the basis of cell size, cell packing density, and, in some cases, on cytological features of cell types. In general, the chinchilla cochlear nucleus has an organizational plan similar to that described for other mammalian species. As in other rodents, the chinchilla has a large and well-developed dorsal cochlear nucleus consisting of three distinct layers. The ventral cochlear nucleus consists of two distinct nuclear masses, a posterior nuclear group and an anterior nuclear group, each composed of several subdivisions, which are qualitatively similar to those described for other mammals. Thus it is now possible to compare detailed observations, such as tonotopic maps, in the chinchilla with findings from the analogous cell populations in other mammals, such as the cat, with considerable precision. In the chinchilla, three cell groups, previously undescribed in mammals, have been defined and their counterparts in the cat identified.

GABA A/benzodiazepine receptor-like immunoreactivity in rat and monkey cerebellum

The cellular and subcellular localization of GABA A/benzodiazepine receptor-like immunoreactivity in the rat and monkey cerebellum has been studied with a monoclonal antibody (E9) directed against the α-subunit of purified GABA A/benzodiazepine receptors. At both the light and electron microscopic level E9 immunoreactivity is located in all 3 layers of the cerebellar cortex and within the deep cerebellar nuclei. The reaction product accumulates within the cytoplasm of neurons and their dendrites but axons are not immunoreactive. Glial cells in the white matter and the cortical layers are also unlabeled, although in some instances Bergmann glia do contain reaction product. The overall distribution and cellular and subcellular localization of E9 immunoreactivity is identical for both monkey and rat cerebellum. On the basis of cell size, morphology,and locatization it is evident that E9 immunoreactivity occurs in examples of all 5 neuronal types in the cerebellar cortex: Purkinje cells, Golgi type II cells, granule cells, and stellate and basket cells. However, the distribution of the reaction product within the cells is more selective. For example, electron microscopy demonstrates that axonal processes and terminals are not E9 immunoreactive with the single exception of the mossy fiber terminals in the granular layer. Also, examples of unlabeled axon terminals resembling those derived from Golgi type II cells, basket cells, and stellate cells form synapses with immunoreactive dendrites and cell bodies in the cortical layers. Finally, in the deep cerebellar nuclei unreactive axon terminals make symmetric synapses with immunostained neurons and dendrites. These results show that E9 monoclonal antibodies label neurons and portions of their processes which are postsynaptic in GABA-mediated inhibitory circuits, and demonstrates that this antiserum can be used a morphological marker for cells which make GABA A/benzodiazepine receptors.

Endogenous lipoxygenase metabolites mediate A23187 induced macrophage cytostasis towards P815 tumor cellsin vitro

The calcium ionophore A23187 stimulated cytostatic activity of peritoneal macrophages towards P815 tumor cells in coculture served as a model for macrophage activation. A macrophage enriched preparation, separated on the basis of cell size in a discontinuous fetal calf serum gradient column, expressed cytostatic activity when stimulated by A23187. This was inhibited dose-dependently, by AA-861 but not by nordihydroguaiaretic acid (NDGA). AA-861 inhibited the 5-lipoxygenase specifically, NDGA inhibited both lipoxygenase- and cyclooxygenase activity: The ratio cyclooxygenase/lipoxygenase products increased with AA-861 but not with NDGA. These results show that lipoxygenase products are necessary for expression of cytostatic activity of these arachidonic acid metabolite producing macrophages and that the ratio of cyclooxygenase/lipoxygenase metabolites plays an important role in macrophage activation.

Specific lipoxygenase inhibition reverses macrophage cytotasis towards P815 tumor cells in vitro induced by the calcium lonophore A23187

A23187-stimulated cytostatic activity of peritoneal macrophages towards P815 tumor cells served as a model for macrophage activation: A macrophage enriched preparation, separated on the basis of cell size in a discontinuous FCS gradient column, expressed cytostatic activity when stimulated by A23187. This was inhibited dose-dependently, by AA-861 but not by nordihydroguaiaretic acid (NDGA). AA-861 inhibited 5-lipoxygenase specifically, NDGA inhibited both 5-lipoxygenase- and cyclooxygenase activity. The ratio cyclooxygenase/lipoxygenase products increased with AA-861 but not with NDGA. These results show that lipoxygenase products are necessary for expression of cytostatic activity of these arachidonic acid metabolite-producing macrophages and that the ratio cyclooxygenase/lipoxygenase metabolites plays an important role in macrophage activation.

Pronephric leucocytes of Cyprinus carpio: Isolation, separation and characterization

Since the teleost pronephros is an important source of diverse immunocytes, suspensions of pronephric cells from young adult carp have been characterized. In freshly prepared suspensions, adherent, spreading cells (macrophages?) constituted <3% of the total population. Granulocytes and lymphocytes were co-dominant (<80%) leucocyte types. Continuous Percoll density gradient centrifugation yielded discrete subpopulations with these ϱ values and cytological characteristics: Fraction I & II ϱ = 1.055 − 1.070 thrombocytes, monocytes, macrophages, and lymphocytes. Fraction III ϱ = 1.080 − 1.090 granulocytes, type 1. Fraction IV ϱ = 1.105 − 1.110 erythrocytes and granulocytes, type 2. Fraction V ϱ = 1.118 − 1.125 granulocytes, type 3. Fraction VI ϱ = 1.140 − 1.150 granulocytes, type 4. Granulocyte motility increased markedly over the first 24 hr in vitro, and was enhanced by components washed from intact yeast. The subtypes of granulocytes were distinguishable by not only the ϱ values, but also on the basis of cell size, ultra-structure of the granules, and their histochemical and phagocytic characteristics. After simultaneous in vivo injection of Bacillus megaterium (Gram + ve), Aeromonas hydrophila(Gram − ve) and Saccharomyces cerevisiae (yeast), individual pronephric leucocytes were found capable of phagocytosing all three types of particle. Granulocytes which had phagocytosed B. megaterium were slower than macrophages in their ability to kill the bacteria. Encounter with B. megaterium or S. cerevisiae in vitro elicited a clumping reaction which involved mostly the larger leucocytes [granulocytes]. Both adherent cells and non-adherent cells were phagocytic in vitro.

Antibodies reactive with class II antigens encoded for by the major histocompatibility complex inhibit human B cell activation.

Although class II antigens encoded by genes in the major histocompatibility complex (MHC) are important as recognition structures for immunoregulatory cell interactions, the precise functional role of these molecules in the biological responses of B lymphocytes is unknown. In the studies described here, we have examined the effects of six monoclonal antibodies reactive with human class II MHC antigens on B cell activation and proliferation. Peripheral blood IgM+ B cells purified by fluorescence-activated cell sorter (FACS) techniques were stimulated with anti-mu antibodies, protein A-bearing Staphylococcus aureus (SAC), or in T cell-dependent activation cultures. The B cell proliferative responses induced by these stimuli were inhibited 68 to 90% by low concentrations (1 to 5 micrograms/ml) of antibodies reactive with class II MHC antigens. Antibodies specific for DR and DQ antigens were both effective inhibitors of B cell proliferation. This inhibition was not due to the binding of antibody to B cell Fc-IgG receptors, because IgM and IgG anti-class II antibodies were equally potent as inhibitors. When responses of B cells fractionated on the basis of cell size by forward angle light scatter were analyzed, anti-DR and anti-DQ antibodies inhibited the proliferation of small, resting IgM+ cells induced by T-independent as well as T-dependent stimuli. Activation-dependent increases in B cell size and RNA synthesis were similarly inhibited. In contrast, the responses of large B cells (that had been preactivated in vivo) to T cell-derived B cell growth factors were not affected by anti-class II antibodies. These data suggest that class II MHC molecules do not serve merely as cellular interaction structures but also directly participate in early events of the B cell activation cascade that precede cell enlargement or increased RNA synthesis. After activation and expression of receptors for growth factors, however, B cell class II MHC antigens no longer mediate signals required for mitogenesis.

Correlation between the presence of intracellular OK-432 and antitumor activity of peritoneal macrophages

The mechanism of the generation of tumoricidal macrophages by intraperitoneal administration of OK-432 was investigated. An association between the presence of intracellular FITC-conjugated OK-432 and the increase of macrophage cytostatic activity was observed after i.p. administration of OK-432. When peritoneal exudate cells were fractionated on the basis of cell size by velocity sedimentation, a higher proportion of macrophages and a strong cytostatic activity were shown in the large cell fractions. Furthermore, in these fractions, increased numbers of macrophages with intracellular FITC conjugated and 3H-acetate labelled OK-432 were observed. Thus, the more cytocidal macrophages can be identified by the presence of intracellular microorganisms. A role for OK-432 in the direct activation of macrophages is suggested.

Antigen presentation by splenic B cells: resting B cells are ineffective, whereas activated B cells are effective accessory cells for T cell responses.

In this study, we have investigated the ability of splenic B cells to act as antigen-presenting cells. Previous data had established that lipopolysaccharide (LPS)-activated B cells were effective antigen-presenting cells; however, the relative capacity of resting B cells to carry out this function remains controversial. Splenic B cells from naive BALB/c mice were depleted of macrophages, dendritic cells, and T cells, and were fractionated on the basis of cell density by using Percoll gradient centrifugation. Fractions were collected from the 50/60, 60/65, and 65/72% interfaces and from greater than 72% (pellet). Cytofluorograph analysis of the fractionated B cells showed that the two lower density fractions (50/60 and 60/65) contained a number of cells which, by cell size determination, appeared to be activated B cells, whereas the two higher density fractions (65/72 and greater than 72) appeared to contain predominantly small resting B cells contaminated by many fewer activated B cells. Functionally, the capacity of fractionated B cells to act as accessory cells for a concanavalin A response or present the antigens chicken ovalbumin (OVA) or OVA-tryptic digest gave similar results, which indicated a striking hierarchy of accessory cell function in the different Percoll fractions. When normalized to the most active low-density fraction (50/60%), the activity of the other fractions were: 60/65 = 78%; 65/72 = 25%; and greater than 72 = 4%. The differences in the functional capacity between the various Percoll fractions did not appear to be due to differences in Ia expression. Although the expression of Ia varied approximately 12-fold within any one fraction, there was little difference in the mean amount of Ia on cells obtained from the various fractions. Kinetic studies showed that activation of B cells with LPS and dextran sulfate resulted in the expression of two stages of functional development. The first stage was an increased efficiency of accessory cell function that was abrogated by irradiation with 4000 rad followed by a second stage, which was characterized by the acquisition of resistance to treatment with 4000 rad. When nonfractionated B cells that had been stimulated with LPS and DexSO4 were sorted on the basis of cell size into a small B cell fraction and a large B cell fraction, only the large B cells were able to present antigen. Taken together, these data suggest that much of the accessory cell function associated with splenic B cells can be accounted for by the relatively small percentage of activated B cells present in the spleen.(ABSTRACT TRUNCATED AT 400 WORDS)

G0 B cells activated by anti-mu acquire the ability to proliferate in response to B cell-activating factors independently from entry into G1 phase.

In the present studies we investigated the early steps in B cell activation and determined that activation could be separated from entry into cell cycle. Purified B cells from BALB/c nu/nu mice, activated under controlled conditions by affinity-purified rabbit IgG anti-mu or F(ab')2 fragments prepared from the same antibodies, were sorted according to size by flow cytometry. Approximately 80% of the B cells were small and were shown to be in Go state by quantitative RNA analysis and by [3H]uridine and [3H]thymidine incorporation studies. The sorted Go B cells, when again incubated in serum-free medium with the appropriate anti-mu preparations, remained in Go. However, Go B cells in the presence of anti-mu have undergone significant change(s), in as much as they were now able to proliferate in response to soluble mediators. The ability to develop a proliferative response to B cell-activating factor(s), acquired independently from entry into cell cycle, characterizes another important step in early B cell activation and also indicates that Go B cells characterized on the basis of cell size, density, and/or RNA content may be in an activated state.

Immunohistochemical studies of substance P, cholecystokinin-octapeptide and somatostatin in dorsal root ganglia of the rat

Immunofluorescence histochemistry was used to determine the distribution of substance P, somatostatin and cholecystokinin-octapeptide-immunoreactive perikarya in C6, T6, T10, L2 and S1 dorsal root ganglia of rat. Five different categories of immunoreactive primary afferent neurons were distinguished on the basis of cell size, cytology and peptide immunoreactivities. The population of small cells (diameter <20 μm) included three groups which were identified as containing somatostatin, substance P, or substance P + cholecystokinin-octapeptide. Two groups of cells were identified in an intermediate size range (diameter 21–43 μm) as containing cholecystokinin-octapeptide or cholecystokininoctapeptide + substance P. These categories may reflect four distinct populations of primary afferent neurons. The relative abundance of dorsal root ganglion cells containing substance P, cholecystokininoctapeptide or somatostatin immunoreactivities was significantly different within segmental levels. More neurons were immunoreactive for cholecystokinin-octapeptide than substance P in ganglia C6, T6 and T10. Somatostatin-containing cells were fewest in number regardless of level. The number of immunoreactive cells also varied among spinal ganglia. L2 contained the greatest number of immunoreactive cells; S1 contained the fewest. These studies are relevant to our understanding of dorsal root ganglia in two ways. Firstly, the data document significant variation in the distribution of peptide-containing neurons among spinal ganglia associated with various cord levels. The variation in peptide-containing cell populations among spinal ganglia may reflect differences in populations of modality-specific primary afferent fibers as well as in populations of somatic and visceral primary afferent fibers at each level. Furthermore, the data indicate that the relative abundance of a population of peptide-containing primary afferent neurons cannot be extrapolated from the examination of spinal ganglia from a single level. Secondly, substance P and cholecystokinin-octapeptide did not co-exist in all spinal ganglion cells as previously reported. In conjunction with immunostaining characteristics and cell size, the differential distribution of the two peptides defined four cell types, raising the possibility that each cell type may mediate a different modality.

A biometrical analysis of morphological variation within a section of genusEuplotes (ciliata, hypotrichida), with special reference to theE. vannus complex of sibling species

Some 50 clones belonging to a rather homogeneous, presumably monophyletic, section of genus Euplotes proved to be distributed between two clear-cut subgroups on the basis of cell size and relative peristome length. These subgroups are referred to as vannus and minuta morphotypes.

Effectiveness of AMSA alone or in combination with radiation on murine fibrosarcoma pulmonary nodules

The cytotoxic effects in vivo of 4'-(9-acridinylamino) methanesulfon-m-anisidide (AMSA), radiation or both modalities in combination on murine fibrosarcoma (FSa) cells grown as pulmonary tumors were determined. Fourteen days following the i.v. injection of viable FSa cells, recipient mice developed between 100 and 150 visible pulmonary nodules. At that time, tumor-bearing animals were exposed to either single or combined modality treatments, as well as single and fractionated dose regimens. Animals were sacrificed 1 hour after the last treatment. Tumor nodules were excised, made into a single cell suspension and separated on the basis of cell size by centrifugal elutriation. Flow microfluorometry (FMF) was used to determine the cell-cycle parameters and the relative synchrony of the separated populations, as well as the percentage contamination by normal diploid cells in each of the tumor cell populations. Known numbers of viable cells from each elutriator fraction were injected into recipient mice to determine their colony-forming efficiency (CFE). Surviving fractions were determined by comparing the CFEs of treated FSa cells from each of the separated elutriator fractions with those of appropriate untreated controls. Following a single i.v. dose of AMSA (30 mg/kg), populations of cells enriched in S phase were the most sensitive. When a single dose of AMSA was combined with a single dose of radiation (100 rad), there was a marked schedule dependence with the more effective sequence, especially if a 12 hour interval was chosen between doses, being AMSA followed by irradiation. No schedule dependence was observed if both modalities were combined and administered under a fractionated protocol of four fractions of AMSA (5 mg/kg per fraction) and four fractions of radiation (300 rad per fraction). Under these conditions the greatest reduction in CFE was in cell subpopulations most enriched in S and G2 + M phase cells.

Survival of purified motor neurones in vitro: Effects of skeletal muscle-conditioned medium

Spinal motor neurones in the adult mouse were labelled retrogradely with both True Blue and horseradish peroxidase (HRP). Using the cell sorter, motor neurones were separated on the basis of cell size and the intensity of True Blue fluorescence. Cultures of the sorted cells were then prepared and the motor neurones were identified by their HRP labelling. There was found to be a 4.8-fold increase in motor neurones over the unsorted population such that 40% of the cells present in culture was labelled. Medium conditioned over skeletal muscle was shown to enhance the survival of motor neurones in these cultures.

Cell Types in the Mesencephalic Trigeminal Nucleus of Mauremys caspica

AbstractA classification of the neurons of the mesencephalic trigeminal nucleus (MTN) has been carried out in the adult turtle Mauremys caspica on the basis of cell size and the shape and arrangement of the Nissl granules, following Sheinin's (1930) criteria for the dog MTN. The authors describe the nucleus as being divided in two cell populations with different neuronal composition and tectal location and probably distinct embryonic sites of origin. The morphological variation observed in the trigeminal neurons might be an expression of functional specialization.

A method for size separation of murine spleen cells using counterflow centrifugation

A method for the rapid separation of murine spleen cells into subpopulations on the basis of their size has been developed using counterflow centrifugation. Upon separation of normal spleen cells with a mean cell volume of 125.5 +/- 6.0, 5 fractions of cells were obtained with mean cell volumes which ranged from 107.8 +/- 3.2 microns3 in fraction 1 to 152.7 +/- 4.9 microns3 in fraction 5. The cells in these 5 fractions were characterized by analysis on a fluorescence activated cell sorter (FACS) after staining with fluorescein conjugated anti-mu, delta, Ia, or Thy 1.2 antibodies, and by assaying for the presence of non-specific esterase activity. Surface Ig+, Ia+ B lymphocytes and Thy 1.2 T lymphocytes were present in all 5 fractions. However, while these T and B lymphocytes accounted for virtually all of the cells in the first 3 fractions, non-T, non-B cells were found in fractions 4 and 5, and represented 30% of the total population in the 5th fraction. Comparison of the intensity of anti-mu, delta or Ia staining of the B cells in fractions 1-5 revealed differences which suggested that B cell size correlated with different activation states of these cells. Increases in the intensity of the staining of T lymphocytes by anti-theta antibodies were also noted in the various fractions. The capacity of the B and T cells in each fraction to proliferate to B or T cell mitogens, respectively, was proportional to their frequency within the fractions. By contrast, the fractions containing larger cells were enriched in cells which proliferated in vitro in the absence of added mitogen. Furthermore, only fractions containing larger cells had the capacity to stimulate allogeneic T cell proliferation in a mixed lymphocyte reaction. Our data suggest that this technique provides a useful method for separating splenocytes on the basis of cell size. Use of this methodology provides a way to correlate cell size with phenotypic surface markers and functional abilities.

Treatment of artificially-induced pulmonary metastases with fractionated doses of vincristine and/or radiation therapy.

The cytotoxic effects in vivo of vincristine (VC), radiation, or both modalities in combination on murine fibrosarcoma (FSa) cells grown as pulmonary tumors were determined. Fourteen days following the i.v. injection of viable FSa cells, recipient mice developed between 100 and 150 visible pulmonary nodules. At that time, tumor-bearing animals were exposed to either single or combined modality treatments, as well as single and fractionated dose regimens. Animals were sacrificed 1 hour after the last treatment. Tumor nodules were excised and made into a single cell suspension and separated on the basis of cell size by centrifugal elutriation. Flow microfluorometry (FMF) was used to determine the cell-cycle parameters and the relative synchrony of the separated populations, as well as the percentage contamination by normal diploid cells in each of the tumor cell populations. Known numbers of viable cells from each elutriator fraction were injected into recipient mice to determine their colony-forming efficiency (CFE). Surviving fractions were determined by comparing the CFEs of treated FSa cells from each of the separated elutriator fractions with those of appropriate untreated controls. Following a single dose of VC (1 mg/kg), populations of cells enriched in late S and G2 + M were the most sensitive. However, following the administration of five doses (0.25 mg/kg each) over a 24 hour period, populations of cells most enriched in G1 cells exhibited the lowest percentage of survivors. A single dose of radiation (1000 rad) was most effective in killing S-phase cells. When administered in five fractions (250 rad per fraction), each separated by a 6 hour interval, no phase-specific sensitivity was observed. The combination of both modalities exhibited a marked schedule dependence, with a single dose (1000 rad) of radiation followed by five doses of VC (0.25 mg/kg) being the most effective treatment protocol observed.

Effect of elicitation on peritoneal macrophage subpopulations: size distributions, ectoenzyme phenotypes and antitumor activity.

Resident (R), Brewer's thioglycollate broth (TG) and C. parvum (CP)-elicited murine peritoneal cell populations were separated into subpopulations by centrifugation on discontinuous Ficoll gradients consisting of layers of 4, 6, 8 and 10% Ficoll. The resulting subpopulations were shown to be distinct on the basis of cell size, ectoenzyme phenotypes and antitumor activity. The cell size distributions were analyzed by means of a Coulter channelyzer and software developed for a computer. The 4-6% Ficoll interface fraction comprised the smallest macrophages, with most cells in this subpopulation appearing to range in cell volume from 150-275 microns 3. The largest sized macrophages were found in the 10%-pellet fraction, with most cells appearing to range in cell volume from 300-600 microns 3. The ectoenzyme phenotypes of the R, TG and CP unseparated macrophage populations were significantly different. Moreover, the ectoenzyme phenotypes of the smallest, (4-6% Ficoll interface) subpopulation in the R and CP macrophages differed from the other subpopulations. Reduction in alkaline phosphodiesterase 1 (APD-1) ectoenzyme activity (as compared with unseparated R macrophages) appeared to be a marker for acquisition of antitumor activity. The small CP macrophages (4-6% Ficoll interface) showed no antitumor activity while the unseparated CP macrophages and all other CP macrophages subpopulations exhibited antitumor activity. The CP macrophage unseparated population and the subpopulations with antitumor activity expressed reduced ADP-1 activity. Conversely, the R or TG macrophage unseparated populations and their subpopulations, along with the CP macrophage 4-6% subpopulation, lacked antitumor activity and failed to show a change in ADP-1 ectoenzyme activity.

Cholesterol and cholesteryl ester concentration in different size classes of cultured human fibroblasts: Effect of in vitro aging

Early and late passage WI-38 fibroblasts were fractionated on the basis of cell size by gravity sedimentation, and free and esterified cholesterol concentrations were determined in each fraction. The cholesteryl ester concentration in all size classes of late passage cells was greater than that of all size classes of early passage cells; the average of all fractions of late passage cells was 2.5 times greater (pg/μm 3) and 1.7 times greater (μg/mg protein) than that of all fractions of early passage cells ( p< 0.001). The average free cholesterol concentration (μg/mg protein) in late passage cell fractions exhibited a consistent, but not statistically significant, increase over that in early passage cells. These results indicate that the increase in cholesteryl ester concentration in late passage WI-38 fibroblasts is not solely attributable to the large, non-dividing cells which accumulate in senescing cultures.

Cytoarchitectonic heterogeneity of the primate neostriatum: subdivision into Island and Matrix cellular compartments.

The cytoarchitecture of the caudate nucleus was examined in Nissl-stained sections from rhesus monkeys, in some of which the corticostriatal terminals had also been labeled by anterograde transport of tritiated amino acids injected into prefrontal cortex. The cytoarchitectonic analysis revealed the existence of two cellular compartments that could be distinguished on the basis of cell size, density, orientation, and tinctorial properties: (1) cell islands consisting of approximately 1,500 to 15,000 densely packed neurons that form aggregates of variable shapes and sizes embedded in (2) a matrix compartment of slightly larger and more loosely packed neurons that comprise the remaining and greater part of the caudate nucleus. In coronal sections, cellular islands appear mostly as round or elliptically shaped areas, 300-600 micrometer in diameter, but can assume more elongated and complex forms particularly in the sagittal and horizontal planes. They are encircled by fibers arranged in a thin, cell-sparse capsule that sets them apart from the matrix compartment. Analysis of cellular organization and corticostriatal connections in counterstained autoradiograms indicates that the prefrontal cortex projects only to the matrix zone and not to the territory occupied by island cells. Therefore, according to present observations the neostriatum in primates should be viewed as a cytoarchitectonically heterogeneous structure composed of at least two distinct cellular compartments with specific connectivity. These compartments may be related to the histochemical and functional diversity of the neostriatum.

Phase-specific cytotoxicity in vivo of hydroxyurea on murine fibrosarcoma pulmonary nodules

The cytotoxic effects in vivo of hydroxyurea (HU) on murine fibrosarcoma (FSa) cells grown as pulmonary tumours were determined. Tumour cells from 13-day-old nodules were made into suspension and separated on the basis of cell size by centrifugal elutriation. Flow microfluorometry (FMF) was used to determine the cell-cycle parameters and the relative synchrony of the separated populations, as well as the degree of contamination by normal diploid cells in each of the tumour-cell populations. HU cytotoxicity was tested by administering both a single 1 mg/g i.p. dose into mice that had been injected i.v. 20 min earlier with known numbers of synchronized viable FSa cells, and i.p. doses of 1 mg/g each into mice bearing 13-day-old pulmonary nodules. In the latter experiments, animals were killed 1 h after the last dose, and the tumour nodules were excised and made into a single-cell suspension and elutriated. Known numbers of cells from each fraction were injected into recipient mice to determine survival. In both sets of experiments, cell killing by HU correlated with the percentage of S-phase cells. The treatment of 13-day-old pulmonary nodules with 3 doses of HU also depleted the (G2+M) phase tumour cells and increased the heterogeneity between tumour subpopulations, as determined by FMF analysis.