ABSTRACT
The immune response to the opportunistic pulmonary pathogen Pneumocystis can have beneficial and harmful effects on the host despite the presence of corticosteroids. We hypothesized that this deleterious hyperinflammatory response is associated with exaggerated cytokine production. The adoptive transfer of at least 107 immune splenocytes reduced the cyst count in rats with corticosteroid-induced pneumocystosis. About 18% of these rats developed clinical illness, an increased lung weight/body weight (LW/BW) ratio, and elevated levels of interleukin 1α (IL-1α), IL-1β, IL-6, tumor necrosis factor alpha, IL-5, IL-10, and gamma interferon in the lungs. This hyperinflammatory reaction was not observed in rats that remained clinically well or in control rats. Thus, in this model, corticosteroids have little effect on the cytokine cascade or other adverse effects of the host immune response to Pneumocystis.
Human immunodeficiency virus (HIV)-infected persons with low CD4+-cell counts and patients who receive corticosteroids (CS) for the treatment of diseases such as cancer and organ transplantation are susceptible to pneumonia caused by the extracellular fungal pathogen Pneumocystis. Yet there is increasing recognition that the host's immune responses to Pneumocystis can have harmful, as well as helpful, effects on the lung. Symptoms of Pneumocystis pneumonia in non-HIV patients often do not begin until after CS have been tapered off (36). Increased levels of neutrophils and IL-8, a potent chemoattractant, in the bronchoalveolar lavage fluid (BALF) of HIV patients with Pneumocystis pneumonia are associated with a worse prognosis (6). HIV patients with Pneumocystis pneumonia experience worsening of respiratory function soon after receiving anti-Pneumocystis drugs, and this can be prevented by CS (1). HIV patients who have recovered from Pneumocystis pneumonia and are started on highly active antiretroviral therapy may develop the “immune reconstitution syndrome,” which is characterized by pulmonary infiltrates leading to respiratory impairment (38).
Rodent models have proven to be useful in studying the dual effects of the host immune response to Pneumocystis and the role of CD4+ and other cells in this response (2, 11, 13, 14, 17, 35, 39). One experimental approach has involved the adoptive transfer of immune splenocytes or purified T-cell populations into athymic (nude) or severe-combined-immunodeficiency (SCID) mice with Pneumocystis pneumonia. The first study to demonstrate this was that of Furuta et al. (9), in which splenocyte transfer to nude Pneumocystis pneumonia mice reduced the Pneumocystis cyst burden but caused intense cellular reactions. Transfer of splenocytes in SCID Pneumocystis pneumonia mouse models also successfully lowered the organism burden with potentially damaging cellular infiltration of the lungs (23, 40, 41). When purified CD4+ T cells were utilized as donor cells to recipient SCID mice, a greater reduction of organism burden occurred, but it was accompanied by a hyperinflammatory reaction (HIR), evidenced by increased lung weight/body weight (LW/BW) ratios that led to premature death (23). CD8+-T-cell adoptive transfer in the SCID mouse Pneumocystis pneumonia model, on the other hand, had no effect on either the Pneumocystis burden or the host lung (23).
A second approach has involved depletion of CD4+ and/or CD8+ cells to induce Pneumocystis pneumonia in normal mice. Anti-CD4+-T-cell antibody (Ab) treatment with GK1.5 rendered mice susceptible to Pneumocystis (27), with an intense inflammatory response that involved CD8+ cells and gamma interferon (IFN-γ) (2), while anti-CD8+-T-cell Ab treatment with YTS169.4 alone had no effect (3, 39). Mice depleted of both CD4+ and CD8+ T cells developed more severe Pneumocystis pneumonia, but inflammatory responses were varied depending on the mouse strain, suggesting that CD8+ cells played an interactive role with CD4+ cells in controlling Pneumocystis infection and lung inflammation (3, 39, 42).
Our laboratory has used a different experimental approach that involves the administration of CS to induce Pneumocystis pneumonia in normal rats. In some respects, this model better mimics HIV than the mouse models described above. CS have broad immunosuppressive effects, but some degree of host immune function is preserved. These agents also have complex (enhancing or suppressing) effects on cytokine function, depending on the conditions of the experiment (7). CS are given throughout an experiment and will have an effect on any immunologic manipulation of the host. Rather than being administered as freshly obtained cells, immune splenocytes are first sensitized to a specific Pneumocystis antigen, the major surface glycoprotein (MSG). Thus, in our model, these cells more closely resemble a form of immunotherapy rather than immune reconstitution.
Previous studies showed that the adoptive transfer of MSG-sensitized donor splenocytes to recipient rats resulted in reduced Pneumocystis burden (30, 34). The reduction of the Pneumocystis burden was improved with a purified sensitized CD4+-T-cell population, but some of the recipient rats perished due to a HIR, evidenced by increased LW/BW ratios. This adverse effect was ameliorated when CD8 cells were administered in conjunction with CD4+ cells (34). These data imply that donor cells can contribute to both Pneumocystis clearance and the development of inflammation. Due to the fact that some host immune function remains following immunosuppression, the donor cells may activate and be assisted by the recipient cells during organism clearance. We hypothesized that since CS did not prevent clinical illness and HIR following adoptive transfer, these drugs would not prevent the cytokine overproduction that has been associated with these adverse events in other animal models (43). We undertook the present study to test this hypothesis.
MATERIALS AND METHODS
Animals and experimental design.Male Lewis rats were acquired from Charles River Laboratories (Hollister, Calif.). All of the rats were 6 to 8 weeks of age and weighed 125 to 150 g at the beginning of the experiments. The animals were housed in microisolator cages in a bioBubble (bioBubble, Fort Collins, Colo.) to control aerosol contamination and were nourished with autoclaved food and water as described previously (32). Ampicillin (1 mg/ml; Clonmel Healthcare Ltd., Clonmel, Ireland), amoxicillin (1 mg/ml; Clonmel), or cephalexin (1 mg/ml; TEVA Pharmaceuticals, Sellersville, Pa.) was given in the water to control secondary bacterial infections. All animals were handled according to institutionally recommended guidelines. The rats were exposed to Pneumocystis by being housed with CS-treated rats with active Pneumocystis pneumonia, as shown previously (32). Pneumocystis pneumonia was induced in the rats by subcutaneous injections of 4.0 mg of methylprednisolone acetate (Depo-Medrol; Pharmacia and Upjohn Co., Kalamazoo, Mich.)/0.2 ml/week.
The experimental design was similar to that used in previous reports (30, 34) with some modifications. The recipient rats were administered CS on Thursdays for 10 weeks. Adoptive transfer of immune cells was performed on the Monday of the seventh week. Therefore, donor splenocyte transfer occurred 3 days before the seventh CS treatment. In previous studies, the rats were sacrificed at a single time (30, 34). Here, the animals were sacrificed at different times (on days 1, 2, 3, 7, 10, 14, and 21 and at times created to accommodate the sacrifice of rats that appeared clinically ill) following adoptive transfer. In preliminary experiments, we found that we could not obtain enough purified CD4 cells for these sequential studies. Therefore, we explored different numbers of splenocytes that would result in a reduction in organism burden as well as in the HIR. Another difference in experimental design was that we examined the effects of adoptive transfer not only on Pneumocystis cysts but also on other developmental stages.
Pneumocystis antigen purification.Enriched samples of Pneumocystis MSG were obtained from the processed lungs of CS-treated rats with Pneumocystis pneumonia as previously described (30), with modifications. The lungs were minced in 10.0 ml of phosphate-buffered saline (PBS), homogenized in a Stomacher 80 (Tekmar Inc., Cincinnati, Ohio), and filtered through gauze. These preparations were then treated with 0.85 M ammonium chloride to lyse red blood cells, washed, and suspended in PBS. Lung homogenates were pooled and digested with Zymolyase 100 T (ICN Biomedicals Inc., Costa Mesa, Calif.), 1.0 mM phenylmethylsulfonyl fluoride, and 5.6 × 10−2 M β-mercaptoethanol (Sigma, St. Louis, Mo.) for 30 min at room temperature. The digest was then spun at 47,900 × g using a J2-MI centrifuge with a JA-18 rotor (Beckman Coulter, Inc., Fullerton, Calif.), and the supernatant was dialyzed in 10% PBS using a Spectra/Por CE (cellulose ester) membrane with a molecular mass cutoff of 100,000 Da (Spectrum Laboratories, Inc., Rancho Dominguez, Calif.). Unused aliquots of this enriched-MSG (eMSG) preparation were stored at −20°C.
Donor cell preparation.Untreated donor rats exposed to Pneumocystis pneumonia rats were sacrificed by an overdose of CO2 (30), and the spleens were excised. Lymphocytes were obtained from the spleens by homogenization using a Tenbroeck tissue grinder (Wheaton, Millvill, N.J.) and filtered through a 0.4-mm-pore-size cell strainer, and the red blood cells were lysed with 0.85 M ammonium chloride, as described previously (35). The splenocytes were washed and suspended in RPMI, and MSG was added (final cell concentration, 106/ml; final MSG concentration, 50.0 μg/ml). The cell cultures were prepared in 225-cm2 culture flasks (total volume, 250.0 ml) for adoptive transfer or in six-well plates (total volume, 8.0 ml) for in vitro tumor necrosis factor alpha (TNF-α) analysis and were incubated at 37°C and 5% CO2. In vitro analysis also included splenocyte incubation with either concanavalin A (ConA) (final concentration, 5 μg/ml) or RPMI. For in vitro TNF-α protein analysis, the supernatant was collected at selected time points and stored at −70°C. For adoptive transfer, the cells were harvested after 4 days, washed, and resuspended in RPMI at various concentrations depending on the experiment (the cell concentration was based on the total number of reconstituted cells per rat in 0.4 ml of RPMI). The cells were delivered to recipient Pneumocystis pneumonia rats (after 7 weeks of immunosuppression) by 0.4-ml tail vein injections. Control Pneumocystis pneumonia rats received 0.4 ml of RPMI. Methyl salicylate (Sigma) was used to dilate the tail vein to improve injection success, as described on the University of California, Irvine, Guidelines for Collection of Blood from Laboratory Animals website (http://www.rgs.uci.edu/as/blood.htm ).
Proliferation assay.Donor spleen cells were suspended in 96-well plates at 1.0 × 105 splenocytes/well and cultured for 4 days in the presence of MSG (6.0 μg/well), ConA (1.0 μg/well; Sigma) as a positive control for proliferation, or RPMI, as previously described (35). After incubation at 37°C and 5% CO2 for 4 days, the cultures were pulsed with 1.0 μCi of [3H]thymidine/well (20 Ci/mmol; New England Nuclear, Boston, Mass.) for 6 h prior to being harvested onto Filtermat filter disks with a semiautomatic 12-well cell harvester (Skatron Instruments, Lier, Norway). The samples were then counted in a Beckman LS 6500 liquid scintillation counter. The data were expressed as the stimulation index, determined as follows: (mean counts per minute of triplicate cultures with mitogen or antigen − mean counts per minute of triplicate cultures in medium alone)/mean counts per minute of triplicate cultures in medium alone.
Lung processing and Pneumocystis enumeration.Following adoptive transfer, the animals were monitored for a total of 3 weeks and sacrificed at various times. The rats were anesthetized intraperitoneally with 16 mg of Telazol (1:1 tiletamine HCl and zolazepam HCl; Lederle Parenterals, Inc., Puerto Rico)/kg of body weight and 4.8 mg of xylazine HCl (2-[2,6-dimethylphenylamino]-4H-5,6-dihydrothiazine; Sigma)/kg, and the total body weight (in grams) was determined. The chest of each rat was opened to expose the lungs and trachea, a stainless steel catheter was inserted through an excision made in the trachea, and the lung was lavaged via the trachea with 10.0-ml aliquots up to a total of 30.0 ml of PBS. The BALF was spun at 350 × g on a Sorvall RT6000B rotor (model no. H1000B; Dupont, Wilmington, Del.) to remove cells and was stored at −70°C. Then, the lung was excised, weighed, and sectioned for selective analysis. A portion of the left lung lobe was excised and immersed in 10% buffered formalin. Lung histological sections were stained with Grocott-Gomori methenamine-silver nitrate and hematoxylin and eosin, as described elsewhere (37). The remainder of the left lung was placed in 2.0 ml of TRizol (Invitrogen, Carlsbad, Calif.) and homogenized. mRNA was isolated according to the TRizol protocol, resuspended in diethylpyrocarbonate-treated water, and stored at −70°C. The remaining four lobes of the lung were minced in 10.0 ml of PBS, homogenized in a Stomacher 80, and filtered through gauze, as previously described (30). The preparation was spun at 2,000 × g (Sorvall RT600B), and the supernatant was collected for TNF-α protein analysis and stored at −70°C. The pellet was then treated with 0.85 M ammonium chloride and washed, and slides were prepared for organism enumeration. The stains used to identify Pneumocystis were cresyl echt violet, which selectively stains the cell wall of the cyst, and Diff-Quick, which stains the nuclei of all developmental forms (cysts, precysts, and trophic forms). Three 10-μl drops were placed on a glass slide, the slides were stained, and 10 randomly scanned fields were read from each drop; the lower limit of detection by this evaluation method is log10 5.24 (5.26 × 105) organisms per lung. The organism burden was reported as the mean log10 organisms (± standard error of the mean [SEM])/lung for the rats in a group.
RPA.Cytokine mRNA levels were quantified with a multicytokine RNase protection assay (RPA) as previously described (40, 41), with modifications. The rCK-1 template kit (BD Pharmingen, San Diego, Calif.) was used to transcribe 32P-labeled antisense riboprobes for rat IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, transforming growth factor β, TNF-α, the rat ribosomal protein L32, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase). The protected mRNA was purified by phenol-chloroform extraction and ethanol precipitation, resuspended in the kit loading buffer, and electrophoresed on 6% Tris-borate-EDTA-urea gels (Invitrogen) at 180 V for 85 min. The gels were allowed to adhere to filter paper, covered in plastic wrap, and placed against PhosphorImager screens (Molecular Dynamics, Sunnyvale, Calif.). The intensity of each band was translated into a density value using a computer-linked PhosphorImager and ImageQuant software (Molecular Dynamics). To correct for unequal RNA loading, each cytokine band density was normalized to the L32 band density. The relative cytokine mRNA density values were reported as the mean ± SEM.
TNF-α chemiluminescent ELISA.TNF-α protein was measured in lung homogenate supernatants, BALF, and in vitro eMSG-stimulated splenocytes using the Rat TNF BD OptEIA enzyme-linked immunosorbent assay (ELISA) set (BD Biosciences Pharmingen) according to the manufacturer's instructions. Briefly, 100.0 μl of the capture Ab in coating buffer (0.1 M carbonate, pH 9.5) was transferred to wells of Greiner (Longwood, Fla.) Bio-One 96-well Lumitrac plates and left overnight at 4°C. The plates were washed, and nonspecific binding sites were blocked with assay diluent (PBS with 10% fetal bovine serum) for 1 h. After the plates were washed, 100.0 μl of standards, supernatant, or BALF was added to the wells for 2 h. The plates were washed, and 100.0 μl of detection Ab was added to the wells for 1 h. Following another wash, 100.0 μl of streptavidin-horseradish peroxidase was added to the wells for 30 min, and then the plate was washed one final time. One hundred microliters of SuperSignal ELISA Pico Chemiluminescent Substrate (Pierce, Rockford, Ill.) was added to the wells, and the plates were read with a FLOUstar OPTIMA microplate reader (BMG Labtechnologies, Durham, N.C.). A standard TNF-α curve was used to determine the TNF-α concentration, and the samples were reported as the mean ± SEM.
Statistical analysis.Statistical analysis was performed using Instat and Prism (GraphPad Software Inc., San Diego, Calif.) software. Data were compared by analysis of variance, followed by a t test with the Neuman-Keuls test for multiple comparisons (30). A P value of <0.05 was considered significant.
RESULTS
Adoptive transfer of 107 splenocytes results in lowered organism burden.Various doses (1 × 106 to 6 × 107 cells) of splenocytes were administered to immunosuppressed Pneumocystis pneumonia rats in order to find the minimum number of cells required to reduce the Pneumocystis burden. The number of donor splenocytes administered to Pneumocystis pneumonia recipient rats was crucial to achieving a reduction in the Pneumocystis cyst burden. We found that ≥107 splenocytes yielded the most successful results in our experiments (Tables 1 and 2). Overall, the cyst burden of Pneumocystis pneumonia rats was reduced by 0.5 log units following the adoptive transfer of 1 × 107 to 6 × 107 splenocytes (combined data from nine experiments) sensitized to MSG (P < 0.005). The adoptive transfer of 1 × 106 to 9.9 × 106 splenocytes (combined data from four experiments) sensitized to MSG did not appear to have an effect on Pneumocystis cysts in the lungs of Pneumocystis pneumonia rats. There was no effect on the Pneumocystis nucleus burden (all developmental forms) of recipient Pneumocystis pneumonia rats following the adoptive transfer of 1 × 106 to 9.9 × 106 splenocytes, but 1 × 107 to 6 × 107 splenocytes caused a nonsignificant 0.3-log-unit nucleus reduction. There was variation in the cyst reduction per experiment, yet the nucleus burden results were very consistent among experiments.
Pneumocystis organism burden (cysts) following splenocyte adoptive transfer
Pneumocystis organism burden (nuclei) following splenocyte adoptive transfer
A common finding in our experiments was that some immune-reconstituted rats exhibited signs of clinical illness (e.g., labored breathing or neglected grooming) and died early or were sacrificed. These animals also demonstrated evidence of HIR (elevated LW/BW ratio and multiple cytokines). However, since not all tests were performed on all animals, it was difficult to draw definitive conclusions about the clinical and laboratory findings. Therefore, we designed experiments to analyze these parameters at different time points after adoptive transfer. Two experiments were performed, and the results were pooled to provide sufficient numbers of animals for analysis.
Association of clinical illness and increased LW/BW ratio following splenocyte adoptive transfer in Pneumocystis pneumonia rats.Donor splenocytes were first sensitized to MSG. Figure 1 shows the proliferative responses of the donor cells to Pneumocystis antigen MSG and the positive control mitogen ConA. Approximately 107 of these MSG-sensitized splenocytes were administered to Pneumocystis pneumonia rats by tail vein injection; control Pneumocystis pneumonia rats (group C; n = 50) received RPMI only (Tables 3, 4, and 5). Splenocyte recipient rats that appeared healthy were categorized as group A (n = 49), while rats that showed signs of clinical illness were categorized as group B (n = 11; 18% of splenocyte recipient rats) (Tables 3 to 5). Animals were sacrificed at fixed time points (days 1, 2, 3, 7, 10, 14, and 21 following adoptive transfer) and at time points created to accommodate the sacrifice of rats that appeared clinically ill. This satisfied the anticipation that some rats would perish throughout the 3 weeks following adoptive transfer. When group B rats exhibited signs of clinical illness and required sacrifice, an appropriate number of rats from group A and group C (control) were also sacrificed. None of the group C rats that received RPMI exhibited the signs of clinical illness observed in group B rats.
Proliferative responses of splenocytes stimulated with Pneumocystis antigen MSG. Shown are the stimulation indices of 1.0 × 105 treated splenocytes exposed to 1.2 μg of ConA (C) or RPMI (R) or 6.0 μg of eMSG (M)/well. Each column represents the mean stimulation index + SEM of eight wells. ***, P < 0.001 compared to RPMI negative control.
LW/BW ratio following adoptive transfer
Pneumocystis cyst burden following adoptive transfer
Pneumocystis nucleus burden following adoptive transfer
Table 3 shows the increase in lung weight following splenocyte transfer in rats that showed the clinical-illness phenotype. Recipient rats that did not appear ill (group A) had LW/BW ratios similar to those of control rats (group C) that received RPMI for all 3 weeks. In contrast, the rats that were observed to be clinically ill (group B) demonstrated progressively higher LW/BW ratios than group C controls over time, and they reached statistical significance at weeks 2 (25 ± 0.40 versus 17 ± 1.8 [mean ± SEM]; P < 0.05) and 3 (28 ± 2.6 versus 16 ± 1.9; P < 0.01).
Reduced Pneumocystis cyst burden following splenocyte adoptive transfer in Pneumocystis pneumonia rats.A significant reduction in the Pneumocystis cyst burden was observed at week 3 following donor splenocyte adoptive transfer in rats that developed clinical illness (group B; 6.7 ± 0.10; P < 0.05) compared to control Pneumocystis pneumonia rats (group C; 7.5 ± 0.10) (Table 4). Splenocyte recipient rats that did not become ill (group A) also show a lower cyst burden at week 3 (6.9 ± 0.20), but this difference did not reach statistical significance. While there was variation in Pneumocystis cyst levels for group A and group B, there was a time-dependent increase in the group C cyst burden throughout the 3 weeks of the experiment. This suggests that there were developing Pneumocystis infections in group C rats that were disturbed by adoptive transfer in group A and B rats. Despite a reduction in the cyst form of Pneumocystis, adoptive transfer did not appear to have the same effect on the other Pneumocystis developmental stages. Nucleus quantifications were similar among all groups of rats and were higher in number than cyst quantifications alone (Table 5).
Association of clinical illness and increased proinflammatory cytokines following splenocyte adoptive transfer in Pneumocystis pneumonia rats.Representative lanes comparing the relative cytokine mRNA densities in the lung homogenates of rats are shown in Fig. 2. The level of IL-1α mRNA in group B rats was increased over control levels at weeks 1 (P < 0.01) and 2 (P < 0.05) (Fig. 3). The IL-1β and IL-6 mRNA levels were increased at week 2 in group B rats compared to control rats as well (P < 0.05). These cytokines were also amplified at other time points, but the increases were not significant. Rats that received splenocytes without developing HIR (group A) did not show increased IL-1α, IL-1β, and IL-6 mRNA levels.
Representative lanes from an RPA showing pulmonary cytokine mRNA levels in rats following sensitized-splenocyte adoptive transfer. Two rats each from groups A (lanes A1 and A2), B (lanes B1 and B2), and C (lanes C1 and C2) were sacrificed during week 1 following adoptive transfer. The migration of cytokine-protected fragments is shown on the right.
Relative proinflammatory cytokine (IL-1α, IL-1β, and IL-6) mRNA density values following sensitized-splenocyte adoptive transfer. Group A rats received splenocytes but did not develop clinical illness; group B rats received splenocytes and did become clinically ill; group C rats received RPMI. The rats were sacrificed at selected time points following adoptive transfer. Data at separate time points were combined on a weekly basis: week 1 (days 1 to 7), week 2 (days 8 to 14), and week 3 (days 15 to 21). The numbers of group A rats represented at weeks 1, 2, and 3 were 34, 6, and 9, respectively; the numbers of group B rats represented at weeks 1, 2, and 3 were 4, 4, and 3, respectively; the numbers of group C rats represented at weeks 1, 2, and 3 were 34, 9, and 7, respectively. *, P < 0.05, and **, P < 0.01 (compared to control rats [C]). The error bars indicate the SEM.
Another proinflammatory cytokine, TNF-α, was increased in group B recipient rats with HIR compared to group C control rats. This was reflected in lung homogenate mRNA levels and also in lung homogenate and BALF protein levels measured by ELISA (Fig. 4). Both lung homogenate mRNA and protein TNF-α levels were increased at weeks 1 (P < 0.05 for both) and 2 (P < 0.01 for both) in group B rats. Nonsignificant increases in TNF-α were also observed in lung homogenate mRNA and protein at week 3. Group B BALF TNF-α protein level increases were observed in all weeks, with a significant increase at week 3 (P < 0.05). There were no significant increases in TNF-α levels in group A recipient rats without HIR.
Relative TNF-α mRNA density, lung homogenate TNF-α protein levels, and BALF TNF-α protein levels following sensitized-splenocyte adoptive transfer. Group A rats received splenocytes but did not develop clinical illness; group B rats received splenocytes and did become clinically ill; group C rats received RPMI. The rats were sacrificed at selected time points following adoptive transfer. See the legend to Fig. 2 for the numbers of rats per group per week for relative mRNA densities. The numbers of rats per group per week are the same for the lung homogenate TNF-α levels. For BALF TNF-α levels, the numbers of group A rats represented at weeks 1, 2, and 3 were 25, 5, and 8, respectively; the numbers of group B rats represented at weeks 1, 2, and 3 were 1, 3, and 3, respectively; and the numbers of group C rats represented at weeks 1, 2, and 3 were 28, 8, and 5, respectively. *, P < 0.05, and **, P < 0.01 (compared to control rats [C]). The error bars indicate the SEM.
Association of clinical illness and increased IL-4, IL-5, and IFN-γ following splenocyte adoptive transfer in Pneumocystis pneumonia rats.Splenocyte adoptive transfer increased the mRNA levels of Th2 cytokines in the lung homogenates of group B rats (Fig. 5). Group B rats showed a nonsignificant IL-4 mRNA level increase throughout, while IL-5 mRNA was increased at week 2 (P < 0.05) and IL-10 mRNA levels were increased at weeks 1 and 2 (P < 0.05). Group A rats showed Th2 cytokine mRNA levels similar to those of group C control rats throughout the experiment.
Relative Th2 cytokine (IL-4, IL-5, and IL-10) mRNA density values following sensitized-splenocyte adoptive transfer. Group A rats received splenocytes but did not develop clinical illness; group B rats received splenocytes and did become clinically ill; group C rats received RPMI. The rats were sacrificed at selected time points following adoptive transfer. See the legend to Fig. 2 for the numbers of rats per group per week for relative mRNA densities. *, P < 0.05 compared to control rats (C). The error bars indicate the SEM.
The multilineage colony-stimulating cytokine IL-3 and the Th1 cytokines IL-2 and IFN-γ were also analyzed by RPA following donor splenocyte adoptive transfer (Fig. 6). Group A recipient rats without HIR showed IL-3, IL-2, and IFN-γ mRNA levels similar to those of control rats. Group B recipient HIR rats had an increase in IFN-γ mRNA at week 3 (P < 0.05) compared to control rats but showed no significant increases in IL-3 and IL-2 mRNA levels.
Relative IL-2, IL-3, and IFN-γ mRNA density values following sensitized-splenocyte adoptive transfer. Group A rats received splenocytes but did not develop clinical illness; group B rats received splenocytes and did become clinically ill; group C rats received RPMI. The rats were sacrificed at selected time points following adoptive transfer. See the legend to Fig. 2 for the numbers of rats per group per week for relative mRNA densities. *, P < 0.05 compared to control rats (C). The error bars indicate the SEM.
In vitro TNF-α increase in splenocytes following Pneumocystis eMSG sensitization.Specific stimulation with the Pneumocystis antigen MSG resulted in increased splenocyte production of TNF-α on days 1 (P < 0.001), 3 (P < 0.01), and 6 (P < 0.01) in vitro (Fig. 7). Donor splenocyte stimulation with the positive mitogen ConA resulted in TNF-α increases on days 1 (P < 0.05) and 3 (P < 0.01) of culture. The day 2 time point did not show increased TNF-α production with either ConA or MSG splenocyte stimulation. It is unclear why this phenomenon on day 2 occurred in our experiment. The increased TNF-α production of MSG-stimulated splenocytes in vitro suggests that donor splenocytes are at least partially responsible for the TNF-α production observed in the lungs of group B recipient rats with HIR (Fig. 4).
TNF-α increases in splenocytes following MSG sensitization. Cells were incubated with ConA, RPMI, or MSG in six-well plates and incubated at 37°C and 5% CO2. The cell supernatant was collected, and a sandwich TNF-α ELISA was performed. C, ConA stimulation; Rm, RPMI stimulation; M, MSG stimulation. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 (compared to the RPMI control). The error bars indicate the SEM.
Histopathology.Lung sections stained with Grocott-Gomori methenamine-silver nitrate revealed clusters of Pneumocystis cysts within alveoli. Hematoxylin-and-eosin-stained lung sections showed alveoli filled with the typical eosinophilic exudates and macrophages that contained Pneumocystis. There was also a mild to moderate nonspecific mononuclear interstitial infiltrate that accumulated in the lung vessels, bronchi, and bronchioles. These changes were slightly more pronounced in the clinically ill group B rats, but there was such overlap among the groups that it is difficult to draw definite conclusions.
DISCUSSION
While the nude (9), SCID (23, 41), and lymphocyte-depleted (3, 39, 42) mouse models are useful for investigating specific immune functions in host defenses against Pneumocystis, the CS-treated-rat model provides insights into the global effects of ongoing disease processes, such as HIV infection, or immunosuppressive therapy. The present study has shown that ≥107 splenocytes sensitized to MSG and administered to immunosuppressed rats with Pneumocystis pneumonia result in a 0.5-log-unit fall in the organism burden, whereas <107 splenocytes do not. These results differ from previous reports, which showed that 106 sensitized spleen cells resulted in a 1- to 2-log-unit decline in the Pneumocystis burden (30, 34). Several differences in the experimental designs of the present and former studies may have contributed to these results: housing (microisolator cages in a protected environment versus open cages in a conventional colony), differences in associated microbial flora, and sensitization with enriched versus purified MSG. We found that sensitizing splenocytes with the mitogen ConA did not have an effect on lung inflammation or on the organism burden in rats with Pneumocystis pneumonia (data not shown), implying that the results obtained in this study are a consequence of specific Pneumocystis antigen.
The results presented here demonstrate a link between clinical illness, increased LW/BW ratio, and HIR. Interestingly, only 11 (18%) of 60 rats developed this HIR following adoptive transfer, while other rats appeared healthy during Pneumocystis clearance. HIR was also noted in a portion of rats in early experiments in this study and in previous studies (30, 34), although its previous frequency varied. The reasons why HIR occurs in only some rats are unclear but may be related to the organism burden or level of immunosuppression. The slightly higher degree of perivascular inflammation and mononuclear cell infiltrate we observed in rats with HIR compared to other rats in the study was not consistent enough to draw conclusions. The small piece of lung tissue isolated from each rat for histological analysis may not represent the condition of the whole lung sufficiently, but this allowed the remaining lung to be used for other analyses in the study.
Our rats with HIR demonstrated significantly elevated proinflammatory cytokine levelss (IL-1α, IL-1β, IL-6, and TNF-α) in their lungs by RPA. Significant levels of TNF-α protein were also found in the lungs and BALF by ELISA. By contrast, no significant elevation of these cytokines was found in rats without HIR or in the controls. Perenboom et al. found increased IL-1β and IL-6 in the lung homogenates and/or BALF, but not in the blood, of rats with CS- induced Pneumocystis pneumonia without HIR (19). Mice with Pneumocystis pneumonia that developed HIR following adoptive transfer exhibited increased proinflammatory cytokines and chemokines and impaired respiratory function (23, 40, 41).
In addition to IL-8, HIV and non-HIV patients with Pneumocystis pneumonia have increased levels of proinflammatory cytokines, eicosanoids, and other inflammatory mediators in their BALF; however, these patients also have increased levels of anti-inflammatory cytokines in their peripheral blood (4, 5, 20, 21). The beneficial effects of CS in HIV+ patients with Pneumocystis pneumonia who are receiving anti-Pneumocystis drugs were thought to be due either to their anti-inflammatory properties or to their effects on the surfactant system (36). Steroids reduce cytokine production in peripheral blood, but their effects on inflammatory mediators and surfactant components in BALF have not been consistent (8, 12, 15, 18).
In addition to increased proinflammatory cytokine levels, our studies demonstrated enhanced Th1, Th2, and colony-stimulating factor cytokine levels. This suggests that an uncontrolled overproduction of these cytokines at least partially contributed to the HIR in these animals. Both Th1 and Th2 cytokine classes are involved in host defense in the mouse (10). Studies with rodent models show that the Th1 cytokine IFN-γ is important in host defense and for controlling the tissue damage caused by the immune attack of the host on Pneumocystis (3, 14, 24, 28). Except for enhanced TNF-α protein levels observed in the BALF, all of the other cytokine increases observed in this experiment occurred in the first 2 weeks following adoptive transfer, as opposed to a week 3 increase in IFN-γ mRNA.
Cytokine mediators of Th2 responses, IL-4, IL-5, and IL-10, may play a role in Pneumocystis host defense, as indicated by rodent models (10, 22, 33), and in humans with Pneumocystis pneumonia (31). It is possible that a controlled response to Pneumocystis in host immune defense is constructed from a balance of Th1 and Th2 cytokines that can counter each other while the immune system clears the infection. The Th2 immune component seen in our studies may also be required to ameliorate harmful proinflammatory responses. It is also interesting that the CS that we use in our rat model may themselves be responsible for the Th2 response (7), but Th2 responses found in nonsteroid Pneumocystis pneumonia mouse models argue against this idea (10, 22).
The colony-stimulating factor IL-3 was also found to be slightly up regulated in our studies. Wright et al. found that this cytokine was also increased in splenocyte-reconstituted SCID mice with Pneumocystis pneumonia (40). This cytokine is involved in leukocyte proliferation and differentiation and therefore may be the host's attempt to build immune defenses against Pneumocystis.
Our in vitro studies suggest that donor splenocytes are at least partially responsible for increased TNF-α levels. The donor splenocytes were administered to recipient rats after 4 days of culture, and TNF-α production in vitro was shown to continue by day 6. Significant increases in lung homogenate TNF-α mRNA and protein at week 1 following adoptive transfer were observed in recipient rats with HIR. Donor cell in vitro stimulation with MSG was not extended beyond 6 days, so assumptions about the responsibility of splenocyte TNF-α production for increased lung homogenate TNF-α mRNA and protein at week 2 following adoptive transfer cannot be made.
One of the surprising findings in the present report was the differential effects of immune reconstitution on the morphological forms of Pneumocystis. The MSG-sensitized donor splenocytes used in this model had a statistically significant reducing effect on the Pneumocystis cyst but not on the trophic form. We typically observe a 1- to 2-log-unit-higher nucleus count than cyst count. Whether the differential effect of the splenocytes reflects differences in organism numbers or in susceptibility to immune factors among cyst and trophic-form stages (16, 32) will require further investigation. It must be noted here that Pneumocystis nucleus enumeration is more subjective than cyst counting, so interpretation of the data should be done with caution. However, the discrepancy of differential effects on Pneumocystis cyst and nucleus levels has been observed by other investigators (25, 26, 29, 39).
In conclusion, the present study has shown that the adoptive transfer of MSG-sensitized splenocytes to rats with Pneumocystis pneumonia results in a decrease in the cyst burden and in harmful effects characterized by clinical illness, HIR, and cytokine cascade in a portion of the rats. These complex effects most likely involve contributions from MSG and other Pneumocystis cell wall constituents, the cells and their cytokines that are adoptively transferred, and the effects of Pneumocystis and the transferred cells on recipient lung cells and cytokines. Future studies involving the adoptive transfer of CS-treated non-Pneumocystis-infected rats may shed light on the relative contributions of donor splenocytes and the organism to increased pulmonary cytokine production. The important information from our model is that these effects occur despite the presence of CS. Despite many studies of animal models and humans, the fact remains that the host-parasite relationship in Pneumocystis infection is complex, and our knowledge of how CS affect this relationship is still at a rudimentary level. Immunologic approaches are potentially important adjuncts to the treatment and prophylaxis of Pneumocystis infection. However, before studies of humans are contemplated, much more information is needed about the interaction of Pneumocystis with the host in different animal models, which aspects of interaction in these models best reflect the situation in humans, and which aspects of the interaction are unique to humans.
ACKNOWLEDGMENTS
We gratefully acknowledge the technical assistance of Sandy Rebholtz, Maggie Collins, and Reiko Tanaka.
This work was supported by the Medical Research Service, Department of Veterans Affairs, and by the public service contracts AI 75319 and AI 25467 and grant RO1 HL64570 from the National Institutes of Health.
FOOTNOTES
- Received 25 August 2003.
- Returned for modification 24 September 2003.
- Accepted 10 November 2003.
- Copyright © 2004 American Society for Microbiology
