A COMPARISON OF IN VIVO AND IN VITRO STATHMOKINETIC METHODS FOR THE STUDY OF CELL PROLIFERATION IN HAMSTER ORAL EPITHELIA

Mitotic indices (MI) expressed as numbers of metaphase figures per 100 basal cells in the cheek pouch and palatal epithelium of the Syrian hamster following metaphase arrest with vinblastine sulphate (VLB) were compared using in vivo and in vitro techniques. The MI in vivo 4 1/2 hr after intraperitoneal injection of 4 mg VLB/kg body weight was 2·69 ± 0·37 for cheek pouch and 12·08 ± 1·09 for palate. MI in vitro was measured using small tissue explants cultured for 4 hr in medium supplemented with VLB at concentrations ranging from 6-600 μg/ml. The maximum MI for cheek pouch epithelium in vitro (2·7) did not differ significantly from that observed in vivo (P > 0·50) and was obtained in the presence of 12–30 μg VLB/ml, a concentration comparable with that used in vivo. In contrast, the maximum MI for palate epithelium in culture (5·6) was significantly lower than that in vivo (P < 0·001) and was only achieved in the presence of extremely high concentrations of VLB. Possible reasons are discussed for the discrepancy between the MI for palatal epithelium in vivo and in vitro.     

Human colorectal tumours in short-term organ culture A stathmokinetic study

Abstract. Short-term organ culture, using a technique to preserve epithelial/stromal interaction and metabolism, is a useful technique for carrying out kinetic studies on human colorectal carcinoma and adjacent normal mucosa, providing initial perturbations of proliferative indices are allowed to settle. Tumours require 3.0 μg/ml vincristine for complete metaphase arrest compared with mucosa, which needs 0.5 μg/ml, a 6-fold difference. Using a stathmokinetic technique, the birth rate of tumour cells is 10.21 cells/1000 cells per hr, compared with 7.73 cells/1000 cells per hr for mucosa, a statistically significant difference ( P < 0.01).     

STUDIES ON THE REGULATION OF LYMPHOCYTE PRODUCTION IN THE MURINE THYMUS AND SOME EFFECTS OF A CRUDE THYMUS EXTRACT

Cell proliferation in the murine thymus was studied in vivo under normal conditions and from 0 to 24 hr after a single injection of a water-soluble extract from mouse thymus, mouse spleen, and mouse skin. The thymus extract reduced during the first 24 hr the mitotic activity 40%; the spleen extract had a weaker inhibitory effect. The skin extract had no such effect. The thymus extract and spleen extract inhibited the flux of cells into the S phase 0–8 hr after the injection of the extract. Initial labelling index was also reduced in this period. Eight hours after injection of the thymus or spleen extracts the inhibited cells initiated DNA synthesis. The rate of progression of blast cells through the cell cycle was normal 24 hr after the injection of the extracts. It was deduced from the analysis that the thymus extract inhibits processes triggering Go/Gi cells into DNA synthesis, the inhibition of G₂ efflux being of minor importance. Finally a model for the regulation of proliferating thymic blast cells and the emigration of small lymphocytes from the thymus is proposed.     

THE PROLIFERATION KINETICS OF NHIK 1922 CELLS IN VITRO AND IN SOLID TUMOURS IN ATHYMIC MICE

The proliferation kinetics of cells of the line NHIK 1922 grown in vitro and as solid tumours in the athymic mutant nude mouse has been studied. In vitro, growth curves were determined for exponentially growing populations and for populations synchronized by mitotic selection. The phase durations for these populations were determined by flow cytofluorometric measurements of DNA-histograms and pulsed incorporation of [³H]TdR respectively. The generation time and the phase durations for synchronized populations were found to be about equal to those for exponentially growing populations. The duration of the phases G₁, S and G₂+ M was found to be 8·5–9·5, 11·0–12·0 and 6·0–6·5 hr respectively, i.e. the generation time was 26·5–27·0 hr. The proliferation kinetics in vivo were studied by flow cytofluorometry and by the technique of percentage labelled mitoses. The median duration of S-phase and (G₂+ M)-phase in vivo was found to be approximately the same as that observed in vitro, while the median duration of G₁-phase was found to be approximately 5 hr longer in vivo than under the present in vitro growth conditions. The growth fraction in vivo was estimated to be approximately 50%. The non-proliferative compartment of the tumour cells was found to consist mainly of cells with the DNA-content of cells in G₁-phase. It is concluded that the reduced rate of proliferation of NHIK 1922 cells in vivo is correlated with alterations in the duration of G₁-phase and, hence, the proportion of cells in G₁-phase.     

Metaphase-Arrest Technique Applied to Human Cervical Epithelium. Ii. Cell Production Rates In Normal and Pathological Cervical Epithelium

An application of the metaphase-arrest technique to human cervical epithelium, in vivo, was utilized to obtain cell birth rate data for seventy-six patients with normal and pathological cervical epithelium. Mean cell production rates for basal and parabasal layers of normal epithelium were 0.91 and 0.92 cells/1000 cells/hr respectively. Histologically normal epithelium adjacent to CIN (cervical intra-epithelial neoplasia) had a significantly higher rate for the parabasal layer compared to the ‘normal’ group (P < 0.05). Values for the lower two-thirds of CIN III were 8–10 times higher than for normal epithelium, with microinvasive carcinoma having the highest rates of all. Values for wart-affected cervical epithelium (NCWVI) were intermediate between normal and CIN, but there was activity in the superficial layer, possibly reflecting activity of the papilloma virus. Large variation in birth rates between individuals in the same histological category was noted for each group, this being particularly the case in six patients with early invasive carcinoma. The data is used to attempt to elucidate some of the uncertainties surrounding the aetiology and biological behaviour of cancer of the cervix and its precursors. Sources of inaccuracy are emphasized and practical difficulties discussed.     

THYMIDINE SUICIDE IN VIVO AND IN VITRO OF SPLEEN COLONY FORMING AND AGAR COLONY FORMING CELLS OF MOUSE BONE MARROW

The ‘thymidine suicide’technique for indicating differences in the proliferation rate of early haemopoietic progenitor cells (spleen colony forming and agar colony forming cells) in C57BL mice has been evaluated. Special care was taken to use the same bone marrow cell suspension for the two progenitor cell assays. Both the in vivo and the in vitro techniques were employed. Following ³H-TdR in vivo, about 20% of both types of progenitor cell are killed in normal mice; however, after incubation in vitro with ³H-TdR, 35% of agar colony forming cells but only 4% of spleen colony forming cells are killed. Reasons for the difference between the in vivo and the in vitro results are discussed. With bone marrow from continuously irradiated animals, the thymidine suicide for both agar colony forming and spleen colony forming cells is in the range 42–50%, and there is no difference between in vivo and in vitro suicide. The in vivo results support the conclusion, based on the effect of proliferation dependent cytotoxic agents, that in C57BL mice agar colony forming and spleen colony forming cells are proliferating at the same rate in normal animals, and are speeded up to the same extent by continuous γ-irradiation. It is considered that in normal C57BL mice the in vitro method does not give a correct estimate of the proliferation rate of these progenitor cells. It would seem that the similarity in the proliferation rate of agar colony forming and spleen colony forming cells in C57BL mice is not true for other strains of mice: indeed using normal CBA and in vivo suicide, we have shown a significantly greater thymidine suicide for agar colony forming cells compared to spleen colony forming cells.     

THE EFFECT OF CYTOSINE ARABINOSIDE IN VITRO ON AGAR COLONY FORMING CELLS AND SPLEEN COLONY FORMING CELLS OF C57BL MOUSE BONE MARROW

This study reports the effect of cytosine arabinoside in culture on two classes of bone marrow progenitor cells in C57BL mice, agar colony forming cells (ACU) and spleen colony forming cells (CFU). Both normal cells and rapidly proliferating cells were studied. The results show that in normal mice, 23 % of ACU but only 7 % of CFU are killed following 1 hr incubation with the drug. With longer periods of incubation, the survival of ACU in the controls is poor, and the results for the drug-treated cultures suggest that the cells are held up in cycle. In continuously irradiated mice, the proportion of ACU and CFU killed after 1 hr incubation with drug is increased to 43–54%, confirming previous results that these cells are proliferating more rapidly than in normal mice. In mice treated with myerlan, 54 % of ACU are killed by 1 hr in vitro exposure to cytosine arabinoside, again confirming that ACU are rapidly proliferating. However, the proportion of CFU killed is lower (23 %). These results are compared with other studies of the effect of cytosine arabinoside in vivo and also with thymidine suicide in the same strain of mice. The results show that cytosine arabinoside has the same effect as tritiated thymidine, and also that the proportion of CFU killed by these agents in vitro is lower than when the agents are injected in vivo. It is suggested that the conditions in culture have an adverse effect on CFU, which cease DNA synthesis, and are protected from the killing effect of cytosine arabinoside and tritiated thymidine. Since cytosine arabinoside in vitro has an effect similar to tritiated thymidine in vitro on bone marrow progenitor cells in C57BL mice, in vitro incubation with cytosine arabinoside could be an alternative method to thymidine suicide for measuring differences in cell proliferation rate.     

IN VITRO PRODUCTION OF COLONY-STIMULATING ACTIVITY. I. EXPOSURE OF MOUSE PERITONEAL CELLS TO ENDOTOXIN

Cell suspensions, obtained from bone marrow, spleen, thymus, lung, liver, and from peritoneal washings, were incubated in vitro with low concentrations of endo-toxin and the supernatant media assayed for colony-stimulating activity (CSA). Peritoneal cells were markedly responsive. The kinetics of CSA production in vitro by peritoneal cells were not remarkably different from that seen in vivo following intravenous administration of endotoxin. The activities of CSA prepared from peritoneal cells and serum were compared following serial dilution; both gave a similar linear relationship when plotted as a function of log-concentration. The bulk of the CSA was produced by adherent peritoneal cells. Separation of peritoneal cells by velocity sedimentation showed that the CSA-producing cell had a sedimentation velocity of 7 mm/hr. Cells with this sedimentation velocity were found to be large mononuclear cells which demonstrated adherence and phagocytosis.     

COMPARATIVE PROLIFERATIVE KINETICS OF CELLS FROM NORMAL HUMAN EPIDERMIS AND BENIGN EPIDERMAL HYPERPLASIA (PSORIASIS) IN VITRO

Proliferation kinetics of epidermal cells from normal human skin and lesions of psoriasis (benign epidermal hyperplasia) were studied in vitro. Epithelial out-growths were obtained from skin explants and the cell cycle studied using the conventional method of following two successive curves of labeled mitoses after an initial pulse with ³H thymidine. The length of T_c was 59 hr and 53.5 hr respectively for normal and psoriatic cells. The shorter T_c for psoriasis was due to a shorter duration of S. The growth fraction was 66% and 74% for normal and psoriatic cells respectively as determined by continuous labeling with ³H thymidine. Under the conditions of the present experiments, therefore, normal and psoriatic epidermal cells showed no significant difference in proliferative capacity.     

Analysis of Cell Flow and Cell Loss Following X-Irradiation Using Sequential Investigation of the Total Number of Cells In the Various Parts of the Cell Cycle

The cell flow and cell loss of an in vivo growing Ehrlich ascites tumour were calculated by sequential estimation of changes in the total number of cells in the cell cycle compartments. Normal growth was compared with the grossly disturbed cell flow evident after a 5 Gy X-irradiation. The doubling time of normal, exponentially growing cells was 24 hr. the generation time was 21 hr based on double-isotope labelling studies and the potential doubling time was 21 hr. Thus, the growth fraction was 1.0 and the cell loss rate about 0.5%/hr. Following irradiation, a transiently increased relative outflow rate from all cell cycle compartments was found at about 3 and 40 hr, and from S phase at 24 hr after irradiation. Minimum flow rates from all compartments were found up to 20 hr. Cell loss as calculated from the cell flow was compared with non-viable cells determined by Percoll density separation. Increase in cell loss as well as non-viable cells was observed at 24 hr after irradiation at the time of release of the irradiation-induced G₂ blockage. Up to 50 hr, about 70% of the initial total number of cells were lost. the experiments show the applicability and limitations of cell flow and cell loss calculations by sequential analysis of the total number of cells in the various parts of the cell cycle.     

The effects of cortisol on the primary response of mouse spleen cell cultures to heterologous erythrocytes

Cell viability and the production of direct PFC were studied in mouse spleen cell cultures after cortisol treatment in vivo or in vitro at various times relative to primary stimulation with SRBC in vitro.Cortisol treatment in vivo reduced spleen cell numbers by 88% after 48 hr, but cultures of the remaining cells produced as many PFC in vitro as did cultures of equal numbers of normal spleen cells.In normal spleen cell cultures incubated with cortisol for 4 hr prior to the addition of antigen, peak responses of PFC/culture and PFC/10⁶ cells occurred 24 hr later than in controls and averaged, respectively, 27% and 141% of control values. Minimum viable cell numbers were observed in cortisol-treated cultures after 3 days; thereafter cell numbers gradually increased. These results were not significantly altered when cultures were treated simultaneously with cortisol and antigen.The response was not suppressed if the addition of antigen preceded that of cortisol by more than 4 hr. Suppression was also considerably reduced if fetal calf serum was used when preparing cells for culture.     

THE EFFECT OF VINCRISTINE ON MOUSE JEJUNAL CRYPT CELLS OF DIFFERING CELL AGE: DOUBLE LABELLING AUTORADIOGRAPHIC STUDIES USING 3H- AND 14C-TdR

The mechanism of action of the alkaloid vincristine (VCR) has been investigated in vitro on HeLa cells in culture and in vivo on jejunal crypt cells of the mouse. The in vitro experiments with HeLa cells show that VCR affects not only mitotic but also interphase cells. The VCR-affected cells first continue their passage through the cell cycle undisturbed but after reaching mitosis they are arrested in metaphase. This agrees well with the results obtained by Madoc-Jones & Mauro (1968) and Madoc-Jones (1973) on synchronized cell cultures. Until now there has been no investigation of the mechanism of action of VCR in vivo. This is due to the absence of a suitable technique for synchronization in vivo. The present study is based on a method which permits the assessment of the VCR sensitivity as a function of the cell age without synchronization in the usual sense. The jejunal crypt epithelium of the normal mouse was double labelled with ³H- and ¹⁴C-thymidine (TdR) in such a way as to produce a narrow subpopulation of crypt cells with a maximum age difference of 1 hr. On autoradiographs these cells can be distinguished by their characteristic labelling from other cells. As this ‘pseudo’-synchronized subpopulation passes through the cycle the effect of VCR can be studied, i.e. one can analyse the effect in well-defined time intervals of the cycle. The results show that the effect of VCR is the same in vivo as in vitro. The crypt cells which are affected by VCR in interphase continue their passage through the cycle, but upon entering mitosis they are arrested in metaphase. VCR has, at the concentration used in the present study, no effect on the duration of the S and G₂ phases. The necrotic cells seen after VCR application are formed from arrested metaphases.     

EFFECT OF AN INHIBITING FACTOR ISOLATED FROM RAT LIVER ON DNA POLYMERASES IN REGENERATING RAT LIVER

We partially purified an inhibitory factor (LIF), isolated from 105,000 g supernatant of a saline adult rat liver homogenate. LIF stopped in vitro cell multiplication by blocking the G₁—S transition, and reduced in vivo [³H]thymidine incorporation into liver DNA in two-thirds hepatectomized rats. This reduction in DNA synthesis was observed at 24 hr after hepatectomy, even when the LIF was injected before the beginning of the S phase, 10 hr after hepatectomy, i.e. when DNA polymerase activity had not yet increased. Under these experimental conditions, LIF in vivo treatment prevented α DNA polymerase activity from increasing after partial hepatectomy, so that enzyme activity at 24 hr in LIF-treated rats decreased compared to the controls. No direct inhibitory effect of LIF on α DNA polymerase was detected. LIF did not affect β DNA polymerase. These results suggest that LIF plays a part in controlling liver growth.     

STUDY OF CELL DEATH IN FRIEND LEUKAEMIA

Cell death of splenic Friend leukaemic cells has been studied in vivo, using ¹²⁵I-UdR and ³H-TdR pulse labelling. The evolution of the splenic specific activity has been measured by autoradiography and external counting during 40 hr after injection of the labelled precursor. These two techniques show the existence of a large reutilization of ³H-TdR (50%), which is measurable as soon as 7 hr after the injection. The DNA turnover rate is rapid, 83-8 % of the splenic cellular DNA being renewed per day. These results confirm that most of the cells produced in the Friend leukaemic spleen are rapidly lost; they also demonstrate that this cell loss is mainly due to a massive death, which occurs in proerythroblastic and erythroblastic compartments after one or two cell divisions. Friend leukaemic cells, which are characterized by a limited capacity of proliferation and a short lifespan, do not appear to be malignant.     

Preferential radioprotection to DNA of normal tissues by ferulic acid under ex vivo and in vivo conditions in tumor bearing mice

Our previous study showed that ferulic acid (FA) offered good radioprotection under in vitro and in vivo conditions to DNA and enhanced the DNA repair process in the peripheral blood leucocytes of mice in vivo. This study concerns radioprotection of normal versus tumor cells. Administration of FA (50 mg/kg body weight) to mice bearing fibrosarcoma tumor, 1 h prior to/ or immediately after radiation exposure (4 Gy) showed preferential radioprotection to normal cells i.e. peripheral blood leucocytes and bone marrow cells in comparison to tumor cells. This preferential protection under in vivo conditions could be attributed to poor vasculature in the tumor or peculiar characteristics of the tumor cells either to restrict its entry inside the cells or metabolize or inactivate the drug. To resolve these ex vivo study was carried out using bone marrow and tumor cells. It was found that under ex vivo condition also only bone marrow cells were protected by FA. Thus the studies revealed that FA showed preferential protection to normal cells under both in vivo and ex vivo conditions. (Mol Cell Biochem xxx: 1–10, 2005)     

Liposomal-benzocaine gel formulation: correlation between in vitro assays and in vivo topical anesthesia in volunteers

The aim of the present study was to characterize a liposome-based benzocaine (BZC) formulation designed for topical use on the oral mucosa and to evaluate its in vitro retention and permeation using the Franz-type diffusion cells through pig esophagus mucosa. To predict the effectiveness of new designed formulations during preclinical studies, a correlation between in vitro assays and in vivo efficacy was performed. Liposomal BZC was characterized in terms of membrane/water partition coefficient, encapsulation efficiency, size, polydispersity, zeta potential, and morphology. Liposomal BZC (BL10) was incorporated into gel formulation and its performances were compared to plain BZC gel (B10) and the commercially available BZC gel (B20). BL10 and B10 presented higher flux and retention on pig esophagus mucosa with a shorter lag time, when compared to B20. BZC flux was strongly correlated with in vivo anesthetic efficacy, but not with topical anesthesia duration. The retention studies did not correlate with any of the in vivo efficacy parameters. Thus, in vitro permeation study can be useful to predict anesthetic efficacy during preclinical tests, because a correlation between flux and anesthetic efficacy was observed. Therefore, in vitro assays, followed by in vivo efficacy, are necessary to confirm anesthetic performance.     

Isolating and maintaining highly polarized primary epithelial cells from normal human duodenum for growth as spheroid-like vesicles

Summary A method is described for the three-dimensional (3-D) in vitro culture of nontransformed gastrointestinal epithelial cells from the human duodenal mucosa. Biopsies obtained from human duodenum were finely minced. The tissue fragments were suspended in culture medium supplemented with 5% fetal calf serum and the appropriate antibiotics. The suspended mucosal fragments generated spheroid-like multicellular vesicles consisting of highly prismatic absorptive and goblet cells retaining most of the histological features of the tissue in vivo. We performed immunocytochemical studies to determine the origin of the vesicles using monoclonal antibodies against EP4. The histochemistry of the vesicles showed alkaline phosphatase activity. Ultrastructural studies revealed that these cells exhibit characteristics of normal duodenal cells in vivo: apical microvilli, glycocalyx, tight junctions and desmosomes, lateral membrane interdigitations, mucous droplets, and a well-developed Golgi apparatus. An overgrowth of the vesicles by fibroblasts was not seen during cultivation. In contrast with the two-dimensional cell cultures grown on artificial supports, the vesicle cells show organization similar to that of natural epithelia. The polarization and cytoarchitecture of normal gastrointestinal epithelial cells cultured as 3-D vesicles are comparable to those known for the native tissue. This study was undertaken to provide a morphological baseline for subsequent infection experiments.     

The N-Cadherin/Catenin Complex in Colon Fibroblasts and Myofibroblasts

Fibroblasts and myofibroblasts were isolated respectively from normal colon mucosa and from colon cancers. Immunostaining with an antibody against α-smooth muscle actin (α-SMA) of the tissues of origin and of early passage cultures showed equal proportions of α-SMA positive myofibroblasts in vivo as in vitro. Immunocytochemistry, immunoprecipitation of metabolically labelled cells followed by Western blotting and RT-PCR of RNA isolates demonstrated the presence of a N-cadherin/catenin complex in both fibroblasts and myofibroblasts. This complex was found preferentially at the cell-cell boundaries. Immunocytochemistry and, to a lesser extent, co-immunoprecipitation indicated partial colocalisation of catenins and α-SMA. Transforming growth factor β1 (TGF-β1) greatly enhanced the expression of α-SMA, but left the N-cadherin/catenin complex unaltered. We speculate that the N-cadherin/catenin complex may have different functions in myofibroblasts than in fibroblasts because of its interaction with α-SMA.     

Mechanisms of antibody formation: I. Early in vivo proliferation of mouse spleen T cells as an initial step in antibody formation

Spleen cultures prepared from mice injected 24 hr earlier with 2 × 10⁶?2 × 10⁸ sRBC and challenged in vitro with sRBC produced 10 times more anti-sRBC IgM PFC than cultures prepared from uninjected mice. The effect was specific for the particular species of foreign RBC injected in vivo. In vitro responses to TNP were also increased in spleen cultures prepared from animals injected 24 or 12 hr earlier with carrier RBC alone, directly implicating carrier-specific T cells in this process. Similar enhancements of PFC formation occurred in cultures prepared from mice which had been injected with sRBC 24 and 48 hr earlier, but which were exposed to lethal irradiation at 1 hr after injection of antigen, if their spleens were shielded extracorporeally during irradiation. This finding indicated that in vivo recruitment of antigen-reactive extrasplenic X-ray-sensitive cells from the circulating lymphocyte pool by the spleen could not account for the observed enhancement.Proliferation in the spleen of antigen-reactive T cells, commencing 12–20 hr after the administration of antigen, was demonstrated by the tritiated thymidine pulse technique. An 8-hr hot-pulse given to spleen cell cultures from normal animals at 20 hr after in vitro challenge with antigen did not affect the rate of generation of IgM-producing cells; however, administration of a similar pulse to cultures which were initiated at 12 or at 20 hr after the in vivo injection of sRBC eliminated the enhanced generation of PFC and delayed the in vitro response to sRBC by 24 hr.Spleen cell cultures were prepared from mice which had been injected in vivo with sRBC at 12, 20, and 70 hr earlier, and 8- to 10-hr hot pulses were given immediately after initiation of the cultures. The cultures were then challenged with sRBC-TNP; antibody responses to TNP were greatly reduced in hot-pulsed cultures prepared from mice injected in vivo with carrier RBC at 12 or 20 hr prior to initiation of the cultures. In contrast, antibody responses to TNP observed in hot-pulsed cultures prepared from mice which had been injected with carrier RBC at 70 hr prior to initiation of the cultures were generally similar to those of nonpulsed 70 hr control cultures. This result suggests that the onset of T helper cell proliferation begins within 12–20 hr after injection of antigen, but subsides in vivo within 70 hr. By that time, the antigen-reactive T cells have already differentiated to perform their helper function.In spite of the triggering of T-cell proliferation during the first 24 hr after injection of antigen, spleen cell cultures prepared from mice which had been injected 24 hr earlier in vivo with 2 × 10⁸ sRBC produced only minimal numbers of anti-sRBC PFC if no antigen was added to the cultures. The presence of unprocessed antigen thus appears to be a requirement for B-cell proliferation in vitro, even after T-cell division has been triggered. This finding is consistent with earlier suggestions that the function of “helper” T cells may not be limited to passive transport of antigenic determinants to B cells. Evidence is also presented to support the contention that the antigen-reactive T cell involved in this process may have to undergo cell division in order to develop “helper” capacity.     

Sequential Involvement of NK Cells and CD8+ T Cells in Granuloma Formation of Rhodococcus aurantiacus-Infected Mice

We investigated the effect of in vivo administration of antibodies against T-cell subsets and natural killer (NK) cells on endogenous gamma interferon (IFN-γ) production and granuloma formation in Rhodococcus aurantiacus-infected mice. High titers of endogenous IFN-γ were detected in the extracts of the livers and spleens during 24 hr of the infection, reaching the peak at 8 hr, and the IFN-γ production was reduced by in vivo administration of anti-NK 1.1 monoclonal antibody (MAb) or antibody against asialo GM1⁺ cells. Endogenous IFN-γ declined until 2 days of the infection, then reappeared from 1 week and peaked at 3 weeks. Endogenous IFN-γ at 1 and 3 weeks was reduced by in vivo administration of anti-CD8 MAb, but not by anti-CD4 MAb or anti-NK 1.1 MAb. Granulomatous lesions in the livers and spleens began to appear from 1 week of the infection and developed in 3 weeks. In vivo administration of rat anti-IFN-γ MAb reduced the development of granulomas. In addition, granuloma formation was reduced by depletion of NK cells prior to the infection or depletion of CD8⁺ T cells at 1 week of the infection. Based on these findings, it is presumed that the biphasic production of IFN-γ is attributable to NK cells in the early phase of the infection and CD8⁺ T cells in the phase of granuloma formation, and that granuloma formation is regulated by NK cells and CD8⁺ T cells through the secretion of endogenous IFN-γ.