The opportunistic organism Pneumocystis carinii (Pc) produces a life-threatening pneumonia (PcP) in patients with low CD4+ T cell counts. Animal models of HIV-AIDS-related PcP indicate that development of severe disease is dependent on the presence of CD8+ T cells and the TNF receptors (TNFR) TNFRsf1a and TNFRsf1b. To distinguish roles of parenchymal and hematopoietic cell TNF signaling in PcP-related lung injury, murine bone marrow transplant chimeras of wild-type, C57BL6/J, and TNFRsf1a/1b double-null origin were generated, CD4+ T cell depleted, and inoculated with Pc. As expected, C57 → C57 chimeras (donor marrow → recipient) developed significant disease as assessed by weight loss, impaired pulmonary function (lung resistance and dynamic lung compliance), and inflammatory cell infiltration. In contrast, TNFRsf1a/1b−/− → TNFRsf1a/1b−/− mice were relatively mildly affected despite carrying the greatest organism burden. Mice solely lacking parenchymal TNFRs (C57 → TNFRsf1a/1b−/−) had milder disease than did C57 → C57 mice. Both groups of mice with TNFR-deficient parenchymal cells had low bronchoalveolar lavage fluid total cell counts and fewer lavageable CD8+ T cells than did C57 → C57 mice, suggesting that parenchymal TNFR signaling contributes to PcP-related immunopathology through the recruitment of damaging immune cells. Interestingly, mice with wild-type parenchymal cells but TNFRsf1a/1b−/− hematopoietic cells (TNFRsf1a/1b−/− → C57) displayed exacerbated disease characterized by increased MCP-1 and KC production in the lung and increased macrophage and lymphocyte numbers in the lavage, indicating a dysregulated immune response. This study supports a key role of parenchymal cell TNFRs in lung injury induced by Pc and a potential protective effect of receptors on radiosensitive, bone marrow-derived cells.

Pneumocystis carinii (Pc)3 is an opportunistic pathogen that is widely disseminated in the general population (1, 2). Under normal circumstances, Pc produces a mild, subclinical pulmonary infection. In the immunocompromised host, however, Pc becomes a pathogen capable of inducing significant morbidity and mortality. Pc pneumonia (PcP) is a predominant presenting complaint in a majority of AIDS patients, in association with a reduced CD4+/CD8+ T lymphocyte ratio, and is a common complication for other immunosuppressed patients. Mortality rates reported for PcP remain as high as 50% and have changed little in recent years (3). Despite the prevalence and severity of PcP, the pathophysiological mechanisms by which Pc produces disease are poorly understood. Although the organism itself can be directly toxic to cells of the lung, eventually resulting in death of lymphocyte-deficient mice or globally immunodeficient patients, PcP occurs most rapidly and severely in those mice or humans in which partial lymphocytic immune function remains or is restored. In the absence of functional CD4+ and CD8+ T cell-mediated immunity, lung injury becomes apparent when the number of Pc organisms becomes overwhelming. However, if either CD4+ or CD8+ T cells are present in the absence of the other, lung injury in response to smaller Pc burdens is profound, typically characterized by necrosis and sloughing of type I alveolar cells, accumulation of alveolar edema, foamy alveolar exudates, and inflammatory cell infiltration. For example, a recent study indicates that in AIDS patients with profound reductions in CD4+ T cell number, the remaining CD8+ T cells mediate PcP-related lung injury (4). Also, preferential reconstitution of CD4+ T cells, as in recovery from chemotherapy or successful highly active antiretroviral therapy for AIDS, can induce an intense pulmonary response to Pc or Pc Ag and lead to PcP-related immune reconstitution inflammatory syndrome. The pathogenic potential of CD4+ and CD8+ T cells has also been demonstrated in animal models of PcP in which treatment with either anti-CD4 or anti-CD8 Abs exacerbated inflammatory injury in an immune reconstitution model of PcP (5, 6). An unbalanced or overzealous host response to Pc is now recognized as largely responsible for PcP-related lung injury.

The cytokine TNF-α is centrally involved in the generation of both innate and adaptive immune responses, and it has been extensively studied in the context of PcP. Clinical reports indicate that long-term anti-TNF therapy can render previously resistant patients susceptible to Pc infection (7). Similar findings have been reported in animal models (8). However, TNF signaling through its receptors is also directly involved in the immunopathogenesis of PcP. During PcP, peak lung concentrations of TNF correlated temporally to rapid progression of lung injury. The importance of TNF signal transduction in PcP was further supported by observations of Pc infection in mice genetically null for both TNF receptors (TNFRs). CD4+ T cell-depleted TNF receptor sf1a/1b−/− mice displayed reduced inflammatory cell recruitment to the lung when compared with CD4-depleted wild-type (WT) mice similarly infected with Pc (9). Quantitatively, the TNFR-deficient mice had significantly reduced pulmonary RANTES, MCP-1, MIP-2, and cytokine-induced neutrophil chemoattractant responses, as well as reduced histological evidence of PcP-related alveolitis as compared with infected WT mice. Less severe pulmonary inflammation correlated with improved surfactant activity and improved pulmonary function in the TNFR-deficient mice.

While robust TNF responses have been documented in both bronchoalveolar lavage fluid (BALF) from Pc-infected mice and culture supernatants of Pc-stimulated alveolar macrophages, the relevant cellular target of TNF action during PcP has not been determined. TNFRsf1a and TNFRsf1b are ubiquitously expressed at varying, cell type-dependent ratios. Ligand binding of the TNFRs has pleiotropic effects depending on cell type and cell priming, ranging from induction of apoptosis or proliferation to induction of secondary cytokines. TNFRs are present on most cells, and lung parenchymal (including epithelial) cells, as well as classical immune cells, respond to both Pc and TNF. The present study was performed to determine whether PcP-related lung injury is dependent on TNFR function on lung parenchymal cells, which are largely radio-resistant at low dosage, or in radiosensitive, bone marrow-derived immune cells. Sublethal total body irradiation followed by bone marrow transplantation was used to create chimeric mice in which either hematogenous cells (radio-sensitive, donor morrow), parenchymal cells (radio-resistant, recipient), or both are null for TNFRs. The chimeric mice were CD4+ T cell depleted before Pc infection, and the progression of disease was assessed.

The original studies of PcP in TNFR-null mice were performed in commercially available TNFRsf1atm1Imx/sf1btm1Imx double-null mice generated as intercrosses of TNFRsf1atm1Imx−/−, created by homologous recombination in C57BL/6-derived ES cells, and TNFRsf1btm1Imx−/− generated in AB1 (129S5/SvEvBrd) ES cells (The Jackson Laboratory). C57BL/6.129SF2 hybrids were the recommended controls (10). To reduce potentially confounding genetic differences in the current experiments, TNFRsf1a/1b double-null mice were regenerated in a C57BL/6J background. TNFRsf1atm1mak/j−/− and TNFRsf1btm1Mwm/J−/− single-knockout (KO) mice were each backcrossed 12 generations onto C57BL/6J and then interbred to produce double-deficient mice. Wild-type C57BL/6J mice were used as controls (The Jackson Laboratory). Mice used in the current protocol were bred and maintained in microisolator cages in specific pathogen-free rooms in the animal care facility at the University of Rochester Medical Center (Rochester, NY). Sentinel animals maintained in the same rooms, on bedding mixed with bedding taken from other cages within the room, routinely tested negative for common murine pathogens, including murine hepatitis, pinworm, and Sendai virus. All animal care and experimental protocols were approved by the University of Rochester Committee on Animal Research and follow the guidelines of IUCAC.

To distinguish the role of parenchymal cell TNFR signaling in Pc-induced lung injury, BMT chimeras (donor → recipient) C57 → C57, C57 → TNFRsf1a/1b−/−, TNFRsf1a/1b−/− → C57, and TNFRsf1a/1b−/− → TNFRsf1a/1b−/− were generated by radioablation of female recipients (total body irradiation, 6 Gy × 2 doses, Shepherd irradiator, 6000 Ci 137Cs source) followed by reconstitution with male donor bone marrow. The radiation dose was previously demonstrated to induce agranulocytosis without evidence of radiation-induced injury at >3–4 wk. Bone marrow was extracted from a minimum of three donors of the appropriate mouse strain by flushing femurs and tibias into HBSS with 1% FCS, dispersing through a 21-gauge needle, and pooling. Erythrocytes were removed by hypotonic lysis. The cells were counted, resuspended in media at 5 × 107/ml, and delivered to recipient mice by tail vein injection (1 × 107 per mouse). Following BMT, animals were allowed to reconstitute for 8 wk under microisolator conditions, supplied with high efficiency particulate-filtered air, sterilized food, acid water, and bedding. To confirm chimerism, fluorescent in situ hybridization with a probe directed against a region of Sry, the sex-determining gene of the Y chromosome, was used on cells obtained by bronchoalveolar lavage (BAL) in some mice to confirm the male origin of cells transplanted to female recipients (11). Additionally, analysis of soluble TNFR present in peripheral blood and FACS analysis of cells isolated from spleens and BAL were consistent with the chimeric design (Table I).

Table I.

Surface markers of cells in BAL and spleen confirm TNFR chimera constitution and CD4 ablationa

MarkerChimera (% Recovered Cells)
WTKOWT to WTKO to KOKO to WTWT to KO
BALF       
 TNFRsf1b+ 30.1 ± 8.6 0.8 ± 0.3b 41.2 ± 11.6 0.4 ± 0.2b 1.0 ± 0.4b 33.3 ± 8.6 
 CD4+ 0.05 ± 0.0 0.6 ± 0.0 0.02 ± 0.02 0.1 ± 0.04 0.04 ± 0.04 0.03 ± 0.02 
 CD8+ 21.8 ± 5.6 10.0 ± 6.4 22.2 ± 15.7 19.5 ± 6.6 13.5 ± 4.8 23.6 ± 9.7 
Spleen       
 TNFRsf1b+ 9.8 ± 1.2 0.5 ± 0.0b 8.3 ± 2.2 0.4 ± 0.05b 0.6 ± 0.14b 11.4 ± 4.1 
 CD4+ 0.1 ± 0.0 0.01 ± 0.0 0.03 ± 0.06 0.01 ± 0.01 0.0 ± 0.0 0.01 ± 0.01 
 CD8+ 7.1 ± 2.8 7.8 ± 2.3 6.6 ± 1.7 8.0 ± 2.2 6.5 ± 2.7 6.1 ± 3.5 
MarkerChimera (% Recovered Cells)
WTKOWT to WTKO to KOKO to WTWT to KO
BALF       
 TNFRsf1b+ 30.1 ± 8.6 0.8 ± 0.3b 41.2 ± 11.6 0.4 ± 0.2b 1.0 ± 0.4b 33.3 ± 8.6 
 CD4+ 0.05 ± 0.0 0.6 ± 0.0 0.02 ± 0.02 0.1 ± 0.04 0.04 ± 0.04 0.03 ± 0.02 
 CD8+ 21.8 ± 5.6 10.0 ± 6.4 22.2 ± 15.7 19.5 ± 6.6 13.5 ± 4.8 23.6 ± 9.7 
Spleen       
 TNFRsf1b+ 9.8 ± 1.2 0.5 ± 0.0b 8.3 ± 2.2 0.4 ± 0.05b 0.6 ± 0.14b 11.4 ± 4.1 
 CD4+ 0.1 ± 0.0 0.01 ± 0.0 0.03 ± 0.06 0.01 ± 0.01 0.0 ± 0.0 0.01 ± 0.01 
 CD8+ 7.1 ± 2.8 7.8 ± 2.3 6.6 ± 1.7 8.0 ± 2.2 6.5 ± 2.7 6.1 ± 3.5 
a

n = 3–8, mean ± SD

b

, p < 0.05 vs WT donor chimeras.

Eight weeks after reconstitution, the BMT chimeras were made susceptible to Pc infection by twice-weekly i.p. injections of anti-CD4+ mAb (0.25 mg, clone TIB 207; American Type Culture Collection) as previously described (9). Injections were begun at least 4 days before Pc inoculation and were continued for the duration of the experiment. After initiation of CD4 depletion, each experimental mouse was anesthetized with a mixture of ketamine (68 mg/kg body weight) and xylazine (6 mg/kg body weight) and then intranasally inoculated with Pc (5 × 105 as determined by cyst count) in 100 μl of sterile saline. Mice were then housed under standard, microisolator technology until sacrifice at ∼28 days after Pc inoculation. The timing of end-point analysis was determined by appearance of illness and elevated respiratory rates. Despite close observation, 10 mice died before harvest, as detailed in the Results. The results of four independent experiments, each with all chimera groups included, were combined and are presented in aggregate.

Dynamic lung compliance and resistance were measured in live mice, anesthetized by sodium pentobarbital injection (100 mg/kg body weight, i.p.), using a previously described method with modifications (5). A tracheotomy was performed and a 20-gauge cannula was inserted 3 mm into an anterior nick in the exposed trachea. To ensure that the mice tolerated the procedure, they were examined for spontaneous respirations before proceeding further. Mice were immediately placed into a plethysmograph designed for anesthetized mice (Buxco Electronics) and connected to a Harvard rodent ventilator (Harvard Apparatus). Mice were ventilated with a tidal volume of 0.01 ml/g body weight at a rate of 150 breaths per minute. Data were collected and analyzed using the Biosystems XA software package (Buxco Electronics). Dynamic lung compliance (ml/cm H2O), normalized for peak body weight, and lung resistance (ml/cm H2O/sec) were calculated by the method of Amdur and Mead (12) from air flow and pressure signals transduced from the chamber and passed through an analog-to-digital converter.

Pretreatment serum was collected by mandibular bleed. At peak PcP in the most severely affected mice, all mice were euthanized by injection with pentobarbital (130 mg/kg, i.p.). Serum was collected by cardiac puncture and allowed to clot in serum separator microtainers. Single-cell suspensions from spleen were obtained by homogenization in DMEM medium and successive filtration through 100- and 40-μm cell strainers and through 25-μm nylon gauze. For BAL, the trachea was intubated, the anterior chest wall removed, and the lungs were lavaged with normal saline (room temperature, 2 ml). BALF was centrifuged at 250 × g for 10 min at 4°C. The supernatant was placed on ice before assay for lactate dehydrogenase and then frozen at −80°C until assayed for total protein by bicinchoninic acid assay (Pierce), as well as for MIP-2, MCP-1, TNF, and soluble TNFRs by ELISA (DuoSet, R&D Systems), using commercial reagents and protocols. The pelleted BALF cells were resuspended in 1 ml 0.15 M NH4Cl, 0.01 M NaHCO3 solution for 10 min to lyse RBC, then washed twice with HBSS and resuspended in DMEM. A total cell count of resuspended BALF cells was performed by hemocytometer. BALF cellular differential was determined on 50 μl cytospins stained with Diff-Quik (Dade Behring) as previously described. MIP-2, MCP-1, KC, and TNF were also assessed by ELISA in lung homogenate prepared in antiprotease buffer and normalized to total protein concentration of the samples.

CD4, CD8, and TNFRsf1b-positive leukocyte populations in BAL and spleen were identified by multiparameter flow cytometry as previously described (13). Cells obtained from lung lavage were washed, resuspended in PBS containing 1% BSA-0.1% sodium azide, and stained for 30 min at 4°C with anti-CD4-FITC (clone RM4-4), anti-CD8a-peridinin chlorophyll-a protein (clone 53-6.7, PerCP), and anti-TNFRsf1b-PE (BD Pharmingen). The anti-CD4 clone RM4-4 was used to confirm that CD4+ T cells were depleted in experimental mice because it is not blocked by the anti-CD4-depleting Ab used (TIB207). Surface marker phenotypes were detected on a FACSCalibur cytofluorometer and analyzed using CellQuest (BD Biosciences) and FlowJo (Tree Star) software. At least 10,000 events were routinely analyzed from the BAL of each Pc-infected experimental mouse. At least 5000 events were analyzed from uninfected control mice. Where indicated, analysis was confined to the lymphocyte population gate based on forward and side scatter characteristics.

Lung tissue was harvested, homogenized in guanidine isothiocyanate buffer, immediately frozen on liquid nitrogen and stored at −80°C. RNA was isolated by extraction with acid/phenol, resuspended in RNase-free water, and quantified by absorbance at 260 nm. Changes in the abundance of specific RNAs were assessed by RNase protection analyses (data not shown) using riboprobes radiolabeled to high specific activity with [γ-32P]UTP as previously described (14) and confirmed by quantitative real-time PCR performed by commercial protocol. Total mouse lung mRNA (1 μg) was used to synthesize first strand cDNA following the GeneAMP protocol (Applied Biosystems). Quantitative real-time PCR was performed with Assays-on-Demand Primer/MGB Probes (Applied Biosystems) for MCP-1, MIP-2, KC, and ribosomal protein L32 (rpL32). Standard curves were made with serial dilutions of a mixed pool of all cDNA assayed. All chemokine values were normalized to the rpL32 mRNA content of each sample.

The intensity of infection in mouse lungs was determined by real-time PCR and by Pc cyst count in mice at the time of experimental harvest. For real-time PCR, crude lung homogenates were boiled for 15 min and centrifuged at 13,000 × g for 15 min. Portions (2.5 μl) of the supernatants, used fresh or stored at −70°C, were diluted 1:3 to minimize PCR inhibition, then assayed by quantitative PCR using TaqMan primer-fluorogenic probe chemistry (Applied Biosystems). The primer-probe set used was obtained from Integrated DNA Technologies and is specific for a 96-bp region of the Pc kex1 gene (GenBank accession no. AF_093132) (15). The PCR reactions (total volume, 25 μl) consisted of universal PCR master mixture (Applied Biosystems), 900 nM forward primer (5′-GCACGCATTTATACTACGGATGTT-3′ sequence positions 1192–1215), 900 nM reverse primer (5′-GAGCTATAACGCCTGCTGCAA-3′ sequence positions 1268–1288), 150 nM TaqMan kex1 probe (5′-/56-FAM/CAGCACTGTACATTCTGGATCTTCTGCTTCC/36-TAMSp/-3′ sequence positions 1230–1260), and 2.5 μl diluted lung sample. To generate a standard curve for the assay, a section of the mouse Pc kex1 gene was subcloned into the pRSET B plasmid (Invitrogen). The copy number of the plasmid vector was calculated from the DNA concentration determined by A260 spectrophotometric measurement. The thermocycler profile used was 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Quantization of the organism burden was performed with the ABI Prism 7000 sequence detection system and its associated SDS software (version 1.0; Applied Biosystems), extrapolating the amplification curve threshold cycle against the threshold cycles of a standard curve constructed with serial 10-fold dilutions of predetermined copy numbers of the pRSETB:kex1 vector. The quantitative real-time PCR method was found in our laboratory to give results that closely approximated enumerations of Pc cysts determined by conventional staining techniques. The Pc counts reported are those obtained by quantitative PCR, which were normalized for each experiment to the mean Pc count measured in the WT to WT chimera lungs to account for interexperimental variability in organism burden. The mean and SE of Pc counts measured in each chimera group are also reported.

Where appropriate, statistical significance was determined by Student’s t test or by ANOVA with Fisher’s protected least significant difference post hoc test (StatView, SAS Institute).

At peak injury, ∼28 days after Pc instillation, the left lung was inflation fixed with 10% buffered formalin overnight, dehydrated to 70% ethanol, and paraffin embedded. Paraffin sections (4–5 μm thick) were deparaffinized and rehydrated through graded ethanol and stained with H&E. Photomicrographs were taken by a SPOT RT digital camera. Representative low- and high-power fields are presented from three mice for each chimera group.

Previous study of TNFR-null mice demonstrated a significant reduction in inflammatory cell recruitment, chemokine production, and histological evidence of PcP in the global absence of TNFRs (9). To begin to define the cells that, via the TNFRs, mediate the inflammatory response leading to lung injury, chimeric WT/TNFRsf1a/sf1b double-null mice, expressing receptors either on radio-resistant parenchymal cells or on radiosensitive bone marrow donor-derived cells, were generated. Following reconstitution, the bone marrow chimeras were treated with anti-CD4 Ab to induce susceptibility to Pc. Flow cytometry of the BALF cells demonstrated that <1% of the recruited lymphocytes were CD4-positive, confirming the uniform CD4 depleted state of the chimeras (Table I). The absence of CD4+ T cells was confirmed in FACS analysis of splenocytes. FACS analysis of splenocytes and BALF cells for TNFRsf1b receptor also confirmed the design and success of the BMT-induced chimerism. Less than or equal to 1% of lymphocytes in BALF or spleen were TNFRsf1b-positive in KO to WT and KO to KO chimeras where these cells should have originated from donor receptor-null bone marrow. In contrast, there was no significant difference between percentage of spleen and BALF lymphocytes that were TNFRsf1b-positive in WT to WT and WT to KO chimeras, nor between these chimeras and nontransplanted WT animals, demonstrating full reconstitution (Table I). The percentage of splenocytes and BALF cells that were CD8-positive did not differ significantly between the chimera groups and did not correlate with severity of illness.

Pc-induced pulmonary inflammation has been shown to correlate well with surfactant dysfunction and physiologic impairment of lung function as measured by lung resistance and dynamic compliance in mice (9). Lung resistance and compliance were slightly altered by the BMT procedure, with both resistance and compliance tending to be lower in transplanted WT to WT and KO to KO mice as compared with nontransplanted WT and KO mice, respectively. These subtle (p > 0.05) BMT effects were not TNFR dependent. In the present study, resistance and compliance measures in all Pc-infected chimeric mice were significantly impaired as compared with noninfected, nontransplanted WT and KO controls, as well as with noninfected, transplanted WT to WT and KO to KO controls (Fig. 1, gray bars in C and D vs noninfected represented in A and B), consistent with a TNFR-independent component of PcP. However, pulmonary function of WT to WT mice was significantly more compromised by Pc infection compared with KO to KO mice, reinforcing the importance of TNFR signal transduction in the development of PcP. Consistent with the hypothesis that parenchymal cell TNFR function has an important role in PcP, the greatest decrement in lung function, measured as reduced compliance and increased resistance, was noted in the KO to WT chimeras, trending greater than that measured in WT to WT chimeras and significantly greater than all other chimera groups (Fig. 1, C and D). The obtained measurements on WT donor to KO recipient chimeras showed slight improvement in lung function over WT to WT animals that did not reach statistical significance (p > 0.05). However, the mortality rate in the WT to WT group was significantly greater than in either KO recipient group, consistent with more severe disease. While overall mortality rate was fairly low in these experiments, 10 mice died, all within days of the 4-wk experimental end-point. Five were WT to WT chimeras out of 20 tested (25%), 4 were KO to WT chimeras out of 19 tested (21%), only 1 was a WT to KO chimera out of 14 (7%) tested, and none was from a KO to KO chimera, reinforcing the increased severity of disease in mice whose parenchymal cells expressed TNFRs as compared with recipient mice who were receptor null. When average lung function measurements of all surviving mice with WT parenchymal cells (WT to WT and KO to WT) were compared with the average measurements of all mice with KO parenchymal cells (WT to KO and KO to KO), the role of parenchymal TNFRs was unmistakable. The mean lung compliance for mice with WT parenchymal cells was 0.68 ± 0.04 as compared with 0.93 ± 0.05 for the mice with KO parenchymal cells (mean ± SEM, p < 0.0001). Similarly, mean lung resistance values for the two groups were 2.36 ± 0.9 and 1.84 ± 0.04, respectively (p < 0.0001). These data are also supported by the overall mortality rates of ∼23% for mice with WT parenchymal cells and 3% for mice with KO parenchymal cells.

FIGURE 1.

Pulmonary function in TNFR WT, KO, and BMT chimeric mice. Lung resistance (A and C) and specific dynamic lung compliance (B and D) measurements were made in live, intubated mice 4 wk after (A and B) saline instillation or (C and D) Pc infection. Black bars, nontransplanted, n = 7–12; gray bars, transplanted (donor → recipient), n = 4 per noninfected group, n = 16–26 per Pc-infected group. Values are means ± SEM; p values are in brackets. NS = p > 0.05. There were no significant differences among saline-treated controls. Resistance and compliance of Pc-infected chimeras were all significantly worsened as compared with noninfected mice either transplanted or not.

FIGURE 1.

Pulmonary function in TNFR WT, KO, and BMT chimeric mice. Lung resistance (A and C) and specific dynamic lung compliance (B and D) measurements were made in live, intubated mice 4 wk after (A and B) saline instillation or (C and D) Pc infection. Black bars, nontransplanted, n = 7–12; gray bars, transplanted (donor → recipient), n = 4 per noninfected group, n = 16–26 per Pc-infected group. Values are means ± SEM; p values are in brackets. NS = p > 0.05. There were no significant differences among saline-treated controls. Resistance and compliance of Pc-infected chimeras were all significantly worsened as compared with noninfected mice either transplanted or not.

Close modal

Weight loss following Pc infection is another clinically significant marker of pneumonia severity. As seen in Fig. 2, the mice expressing TNFRs only on radio-resistant, parenchymal cells demonstrated significantly greater weight loss than any of the other Pc-exposed chimeric mice (p < 0.01). Importantly, both groups of mice with TNFR expressing parenchymal cells (WT to WT and KO to WT) exhibited significantly greater weight loss than did mice lacking parenchymal TNFRs (KO to KO and WT to KO), again highlighting the role of parenchymal TNFRs during PcP-related injury. As with the pulmonary function measurements, accounting for those mice that died before harvest at 4 wk of infection, average weight loss would be expected to be even greater in the groups with WT parenchymal cells. Thus, considering morbidity and mortality, the presence of TNFRs on only radio-resistant cells was sufficient to induce clinically significant PcP.

FIGURE 2.

Percentage of pretreatment weight loss or gain in TNFR BMT chimeric mice. The change in body weight was determined from measurements made at the time of harvest, 4 wk after intratracheal Pc infection (gray bars) or control saline instillation (black bars), as compared with the preinfection weight assessed 8 wk after BMT. Values are means ± SEM; n = 17–26 per group. p values in brackets compare Pc-infected groups. There were no significant differences between saline-treated controls.

FIGURE 2.

Percentage of pretreatment weight loss or gain in TNFR BMT chimeric mice. The change in body weight was determined from measurements made at the time of harvest, 4 wk after intratracheal Pc infection (gray bars) or control saline instillation (black bars), as compared with the preinfection weight assessed 8 wk after BMT. Values are means ± SEM; n = 17–26 per group. p values in brackets compare Pc-infected groups. There were no significant differences between saline-treated controls.

Close modal

No significant difference was measured in resistance, compliance, or weight loss in WT to KO as compared with WT to WT chimeras. The values for these measures fell between those of the WT to WT and KO to KO chimeras, demonstrating that TNFRs restricted to marrow-derived cells are capable of generating PcP similar to but milder than globally TNF responsive mice (when considering the differences in mortality).

Total cell count and differential of cells available by BAL were determined to quantify differences in the alveolar inflammatory cell recruitment in the presence or absence of parenchymal TNFRs (Fig. 3). There was a marked recruitment of lavageable cells in the WT to WT chimeras, as well as a significantly greater increase in the lavaged cell count of the KO to WT mice. In comparison, the remaining two chimera groups with receptor-null parenchymal cells had significantly reduced total BALF cell counts, not differing from saline-treated chimeric controls. Differential cell counts on lavaged cells demonstrated a small, TNFR-independent increase in lymphocytes in saline control chimeras as compared with nontransplanted, noninfected mice (p < 0.05). The increase in total cell number in Pc-infected mice with WT parenchymal cells was accounted for mainly by increases in lymphocytes and neutrophils (Fig. 3,B). Furthermore, KO to WT mice showed a dramatic increase in the number of lavageable AMs compared with all other Pc-infected groups. That the numbers of BALF alveolar macrophages in all other chimeras were reduced from baseline (saline treated) suggests margination of resident and, potentially, recruited macrophages. Although the percentage of BALF cells that were CD8+ did not differ between the chimeras (Table I), the absolute numbers of CD8+ T cells recruited by the WT to WT and KO to WT mice were significantly greater than those recruited by the WT to KO or the KO to KO mice, suggesting that the presence of TNFRs on radio-resistant, parenchymal cells is sufficient for CD8+ T cell recruitment and is necessary for peak recruitment.

FIGURE 3.

Analysis of lavageable white blood cells in BALF of TNFR BMT chimeric mice. BALs were performed 4 wk after intratracheal saline (black bars) or Pc (gray bars) instillation. Total cell count per milliliter of BALF (A) and WBC differential cell count (B) were determined. Values are means ± SEM; n = 12–17 each Pc-treated group, n = 4–11 each saline control group. In brackets, p values reflect comparisons of Pc-treated groups. There were no significant differences between saline-treated chimeras.

FIGURE 3.

Analysis of lavageable white blood cells in BALF of TNFR BMT chimeric mice. BALs were performed 4 wk after intratracheal saline (black bars) or Pc (gray bars) instillation. Total cell count per milliliter of BALF (A) and WBC differential cell count (B) were determined. Values are means ± SEM; n = 12–17 each Pc-treated group, n = 4–11 each saline control group. In brackets, p values reflect comparisons of Pc-treated groups. There were no significant differences between saline-treated chimeras.

Close modal

As with the pulmonary function tests, there was a small but in this case significant increase in BALF protein content, as well as a trend toward increased LDH, in noninfected control chimeras compared with noninfected, nontransplanted controls, which was TNFR independent (Fig. 4, A and B). Consistent with enhanced capillary leak and cell death, the BAL total protein was elevated >10-fold in the WT (see Ref. 9) and chimeras after Pc infection; no distinction could be made between chimeras based on protein concentration in the alveolar space, although the trend was lower in KO recipient lungs as compared with WT recipients (Fig. 4,C). LDH concentrations in BALF were also elevated in all Pc-treated mice (Fig. 4,D); in this case, LDH was significant higher in the WT to WT chimeras and trended higher in the KO to WT mice, suggesting enhanced cellular necrosis and release of LDH dependent on parenchymal cell response to TNF (Fig. 4 D). The average BALF protein and LDH concentrations were greater in the TNFR WT recipients than in the receptor-null recipients: 1.45 ± 0.55 vs 1.08 ± 0.58 (p = 0.01) and 304.5 ± 46.5 vs 132.1 ± 20.8 (p < 0.005), respectively.

FIGURE 4.

Analysis of BALF for (A and C) total protein and (B and D) LDH indicating capillary leakage and cellular death in (A and B) saline-treated chimeras (gray bars) vs nontransplanted controls (black bars) and in (C and D) Pc-infected TNFR chimeric mice. Values are means ± SEM; n = 7–8, p values are in brackets.

FIGURE 4.

Analysis of BALF for (A and C) total protein and (B and D) LDH indicating capillary leakage and cellular death in (A and B) saline-treated chimeras (gray bars) vs nontransplanted controls (black bars) and in (C and D) Pc-infected TNFR chimeric mice. Values are means ± SEM; n = 7–8, p values are in brackets.

Close modal

Analysis of lung sections from the Pc-infected TNFR chimeras by light microscopy demonstrated differences in degree of lung injury that were consistent with those suggested by the physiologic and cellular indices (Fig. 5). Both the WT to WT and KO to WT lungs demonstrated dense, mononuclear cell infiltration that was maximal around distal bronchioles and pulmonary vessels; the pattern was somewhat more intense, interstitial, and diffuse in the absence of TNFRs in the marrow-derived cells of the KO to WT mice. Inflammation was also present in the absence of parenchymal cell TNFRs, but cellularity was markedly reduced when compared with parenchymal WT lungs. The architecture of the WT to KO chimera lungs, and especially the alveolar spaces, were relatively spared in comparison to all other chimeras. Interestingly, the TNFR-null mice (KO to KO) were not free of injury but demonstrated a relatively disorganized cellular infiltration and large amounts of bland alveolar debris that may be related to direct injury from the high Pc burden.

FIGURE 5.

PcP in TNFR chimera mice demonstrated by H & E stain in representative sections (4 μm) of inflation fixed left lung of three mice per chimera group. Gray bar = 100 μm.

FIGURE 5.

PcP in TNFR chimera mice demonstrated by H & E stain in representative sections (4 μm) of inflation fixed left lung of three mice per chimera group. Gray bar = 100 μm.

Close modal

TNFRs, active as cell surface transmembrane receptors, are solubilized by enzyme cleavage in response to inflammatory stimuli (for review, see Refs. (16 , 17). Elevated serum sTNFRs in patients with acute respiratory distress syndrome correlated with mortality (18). In the current study, sTNFRsf1a and sTNFRsf1b were analyzed in serum and BALF to test the hypothesis that TNFRs are also shed and accumulate in response to Pc infection. As controls, sTNFR levels were measured in serum harvested 8 wk after bone marrow reconstitution, just before beginning CD4+ T cell depletion, and were consistent with the desired chimeric composition (Fig. 6, A and B, black bars). In these pretreatment samples, sTNFRsf1a and sTNFRsf1b concentrations in WT to WT transplants were comparable to WT, nonirradiated, nontransplanted mice, demonstrating that solubilization was not an artifact of bone marrow transplantation. TNFRsf1a/1b−/− to TNFRsf1a/1b−/− chimeras had no detectable sTNFRs in serum. Serum sTNFR concentrations of the mixed chimeras were, as expected, intermediary between WT and KO animal levels. The relative levels of the two receptors were dependent on the genotype of the donor and recipient mice. Comparable serum sTNFRsf1a levels were detected in noninfected WT, WT to WT, and KO to WT chimeras. The receptor was very low in the WT to KO chimeras, suggesting that the primary origin of this soluble receptor, constitutively in circulation, is the parenchyma. Serum TNFRsf1b was comparable to WT control in the WT to KO chimeras but reduced in the KO to WT chimeras, suggesting that bone marrow-derived cells are a significant basal source of this circulating receptor. Four weeks after Pc infection, sTNFRsf1a and sTNFRsf1b concentrations were increased in both serum and BALF of WT mice when compared with saline-treated controls (Fig. 6), suggesting that sTNFRs could be a marker of PcP. There was minimal elevation of serum sTNFRsf1a in KO to WT chimeras but a significant rise in WT to KO chimeras in response to Pc, consistent with most stimulated serum sTNFRsf1a originating from bone marrow-derived cells. An increase in BAL sTNFRsf1a after Pc in both mixed chimeras suggests stimulated release of the receptor from both marrow-derived and parenchymal cells in the airways. All chimeras except KO to KO (undetectable) had an increase in serum and BALF sTNFRsf1b in response to Pc, suggesting that although much of basal serum sTNFRII originates from circulating, bone marrow-derived cells, Pc-induced shedding originates from both donor marrow and recipient parenchymal sources in the BMTs.

FIGURE 6.

Soluble TNFRs in serum and BALF of chimeric TNFR mice before and 4 wk after Pc infection. Serum sTNFRsf1a (A) and sTNFRsf1b (B) concentrations were assessed preinfection (black bars) and at 4 wk after Pc instillation (gray bars). BALF levels of sTNFRsf1a (C) and sTNFRsf1b (D) were also measured in noninfected WT controls (black bars) and at 4 wk post-Pc infection (gray bars). ND indicates none detected; p values are as indicated in brackets; ∗, p < 0.01 vs all pre-Pc infection. Values are means ± SEM; n ≥ 11 per serum group, n = 6–8 per BAL group.

FIGURE 6.

Soluble TNFRs in serum and BALF of chimeric TNFR mice before and 4 wk after Pc infection. Serum sTNFRsf1a (A) and sTNFRsf1b (B) concentrations were assessed preinfection (black bars) and at 4 wk after Pc instillation (gray bars). BALF levels of sTNFRsf1a (C) and sTNFRsf1b (D) were also measured in noninfected WT controls (black bars) and at 4 wk post-Pc infection (gray bars). ND indicates none detected; p values are as indicated in brackets; ∗, p < 0.01 vs all pre-Pc infection. Values are means ± SEM; n ≥ 11 per serum group, n = 6–8 per BAL group.

Close modal

Although KO to KO mice were the healthiest of the TNFR chimera mice, determination of Pc burden in the lungs at 4 wk after infection demonstrated the highest number of organisms in this group (10.1 ± 2.3 × 106 per lung vs 6.6 ± 1.8 × 106 per lung in WT to WT mice averaged across four experiments), a 2-fold difference when normalized to WT to WT burden in each experiment (Fig. 7). These data are consistent with previous reports that TNF exerts some control over Pc growth in SCID mice (19) and in CD4-depleted mice (8), and they also demonstrate that the host’s immune response, and not Pc burden, is the critical factor in determining the severity of PcP. Interestingly, while KO to WT mice had the most intense pulmonary inflammatory response and most severe PcP, they did not have reduced Pc burden (6.4 ± 2.0 × 106 per lung). This indicates that an inappropriate immune response can exacerbate disease without the beneficial effect of killing Pc. The lowest Pc burden was detected in the mice with WT bone marrow but TNFR-deficient parenchymal cells, the WT to KO chimeras (1.6 ± 0.5 × 106 per lung). These mice consistently maintained a significantly lower Pc burden than did the remaining chimeras.

FIGURE 7.

Relative Pc burden determined by quantitative real-time PCR for kex1 gene copies in Pc-infected TNFR BMT chimeric mice at the time of harvest, 4 wk after Pc infection. kex1 gene copies are presented normalized to the average kex1 gene copy number in WT to WT chimeras in each experiment. Values are means ± SEM; n = 18–20 for each group, averaged over four independent experiments, p values are in brackets.

FIGURE 7.

Relative Pc burden determined by quantitative real-time PCR for kex1 gene copies in Pc-infected TNFR BMT chimeric mice at the time of harvest, 4 wk after Pc infection. kex1 gene copies are presented normalized to the average kex1 gene copy number in WT to WT chimeras in each experiment. Values are means ± SEM; n = 18–20 for each group, averaged over four independent experiments, p values are in brackets.

Close modal

Recruitment of inflammatory cells to the lung during PcP has been correlated to chemokine production, in particular to MIP-2, MCP-1, KC, and cytokine-induced neutrophil chemoattractant, in addition to TNF (9). The chemokine elevation was largely dependent on TNFR expression, as it was blunted in TNFR-null mice (9). To determine whether the variation in PcP severity observed in the presence or absence of parenchymal cell TNFRs correlated with altered chemokine production, both lung tissue homogenates and BALF from TNFR chimeric mice were analyzed for chemokine protein. Previous studies demonstrated that MCP-1, MIP-2, KC, and TNF were below the limit of detection in lung and BALF of uninfected mice. No significant difference in homogenized lung tissue TNF was detected between chimeras, while MCP-1 was modestly, and KC was markedly, increased in KO to WT mice in comparison to TNFR-null recipient chimeras (Fig. 8,B). MIP-2 protein concentration was significantly increased in all mice expressing TNFRs on either (or both) marrow-derived or parenchymal cells, with relatively reduced levels in fully null chimeras. Measurement of the cytokines in BALF detected chemokine concentration differences dependent on TNFR distribution (Fig. 8,A). MCP-1 protein was significantly elevated in BALF of KO to WT chimeras, with those animals also having been documented to have more severe PcP (Figs. 1 and 6) and a dramatic increase in lavageable macrophages (Fig. 3). BALF TNF levels were highest in the KO to KO mice and lowest in the WT to WT mice, with TNF levels in the mixed chimeras falling between. This result likely reflects alteration of TNF-negative feedback loops that require the expression of TNFRs on both compartments for normal control of TNF production. Relatively low concentrations of MIP-2 were detected in the BALF, with the greatest levels found in the WT to TNFR-null mice, suggesting dysregulation of the chemokine with stimulation of marrow-derived cells in the absence of parenchymal TNF response.

FIGURE 8.

Cytokines in BALF and whole lung isolated from Pc-infected TNFR chimeric mice 4 wk after infection. TNF, MIP-2, MCP-1, and KC were assessed by ELISA in (A) BAL fluid and (B) homogenized lung tissue. C, MIP-2, MCP-1, KC, RANTES, and IL-10 were measured by quantitative real-time PCR in isolated lung RNA. Values are means ± SEM; n = 7–8, p values in brackets.

FIGURE 8.

Cytokines in BALF and whole lung isolated from Pc-infected TNFR chimeric mice 4 wk after infection. TNF, MIP-2, MCP-1, and KC were assessed by ELISA in (A) BAL fluid and (B) homogenized lung tissue. C, MIP-2, MCP-1, KC, RANTES, and IL-10 were measured by quantitative real-time PCR in isolated lung RNA. Values are means ± SEM; n = 7–8, p values in brackets.

Close modal

Analysis of mRNA for the cytokines MIP-2, MCP-1, and KC in whole lung preparations by quantitative real-time PCR demonstrated patterns of expression similar to the protein concentrations measured in BAL and lung homogenates. MCP-1 and KC mRNA concentrations were significantly elevated in KO to WT chimeras when compared with all Pc-treated chimeras and controls, while MIP-2 mRNA was significantly less concentrated in the TNFR-null chimeras, intermediate in the mixed chimeras, and several-fold greater in the WT to WT chimeras (Fig. 8 C). RANTES, also previously shown to be stimulated in response to Pc (9), was increased in all cases from noninfected controls (0.2 ± 0.1 ratio to rpL32) but with no significant difference between chimeras. The expression of IL-10, an antiinflammatory cytokine, was also tested as a potential explanation for differences in PcP severity. There were significant differences in IL-10 mRNA levels between the chimeras, in part correlated with severity of disease, as highest concentrations were measured in fully TNFR-null chimeras. However, those most injured, the KO to WT mice, had IL-10 mRNA levels comparable to WT animals, while they were significantly lower in moderately injured WT to KO mice.

Many lines of evidence support the importance of TNF in the development of PcP. Pc stimulates TNF production and release from alveolar macrophages in both immunocompetent and immunosuppressed mice and from monocytes and macrophages in culture (20, 21). Release of TNF from alveolar epithelial cells in response to the organism has also been demonstrated (22). In the reconstituted SCID model of PcP, the onset of reduced compliance and hypoxia is temporally related to peak TNF mRNA in lung tissue and TNF protein in BALF, in association with the influx of neutrophils, macrophages, and lymphocytes. TNF protein is likewise increased in CD4+ T cell-depleted, Pc-infected mice in a CD8+ T cell-dependent manner (9). Our previous studies using TNFR-deficient mice demonstrated that maximal Pc-induced chemokine production, lung injury, and pulmonary dysfunction required intact TNFR signal transduction (9). The current study demonstrates that TNFRs on cells resistant to split dose irradiation are sufficient to mount an inflammatory response to Pc, comparable to or in excess of that generated when all cells express the receptors. Limitation of TNFR distribution to marrow-derived cells improved control of Pc burden and reduced Pc-induced injury.

Pulmonary function and inflammatory markers were analyzed in the TNFR chimeric mice when the sickest treatment group reached clinically peak disease, ∼4 wk after Pc treatment in the chimeric mice; at and beyond this point mortality increased markedly in the KO to WT and WT to WT chimeras. Each of the transplanted mice that expressed TNFRs either on radiosensitive marrow-derived cells or radio-resistant parenchymal cells or both (WT to WT, WT to KO, or KO to WT) demonstrated some evidence of pneumonia at this time point. However, the greatest mortality, decrement in pulmonary function, weight loss, and inflammatory cell recruitment occurred in those mice expressing TNFRs on radio-resistant but not on marrow-derived cells. TNFR-null chimeras (KO to KO) were least injured despite having a relatively greater Pc burden than the remaining chimeras. Interestingly, in vitro Pc-dependent induction of chemokines in alveolar epithelial cells occurs independently of TNFRs (23), which may explain the inflammation and physiologic impairment observed even in the TNFR-null mice. In vivo, TNFR-deficient models indicate that maximal inflammation that correlates with clinically significant loss of weight and pulmonary function and that results in death requires parenchymal TNFR responses. TNFR expression limited to marrow-derived cells was sufficient to generate Pc-stimulated lung injury and to augment loss of pulmonary function and weight loss but was much less apt to cause mortality, and it was associated with reduced MCP-1, KC, and TNF production, intraalveolar cell death (by LDH), and recruitment of lavageable cells. As previously observed, the severity of lung injury was not directly related to Pneumocystis burden (24, 25). Reduction in lung compliance weakly correlated with Pc burden but only in the presence of parenchymal TNFR function, which may reflect inhibition of surfactant production or function (9). The global TNFR-null mice carried the greatest Pc burden but the least evidence of injury, consistent with a role for TNF in clearance of Pc as previously demonstrated (9). In the present study, the parenchymal cell-null, marrow WT animals maintained the lowest burden of organism, suggesting that TNF stimulates immune cells to remove the organism. It has also been suggested that adherence of Pc to lung epithelial cells enhances proliferative growth of the organism (26). The absence of parenchymal TNFRs may reduce Pc-epithelial cell adherence, thereby reducing the organism’s growth potential and, potentially, leaving it more accessible to immune cell clearance. Further study is necessary to clarify this observation.

Accumulation of inflammatory cells and LDH in the bronchoalveolar space was parenchymal cell TNFR dependent. Although each of the chimeras developed some degree of lung injury, there was a marked increase in numbers of inflammatory cells recruited to the lavageable air space in WT recipient mice, particularly when the bone marrow-derived, recruited inflammatory cells did not express the TNFRs. No increase in total lavageable cell counts was detected in the absence of parenchymal TNF signal transduction. The differential of the BALF cells was altered in response to Pc; infiltration by neutrophils and lymphocytes occurred in all chimeras but was exacerbated by the presence of parenchymal TNFRs. Flow cytometry suggested marked increases in the percentage of lymphocytes that are CD8+ in all Pc-exposed groups, with no differences found between chimera groups. However, when considered as absolute numbers of lavageable cells, the maximal increase in recruitment of CD8+ T cells was seen in the TNFR WT recipient chimeras. This result is consistent with a significant role for parenchymal TNFR signal transduction in the recruitment of these cells, which have been shown to mediate Pc-induced lung injury in this CD4+ T cell-depleted model. Note that despite markedly increased inflammatory cell infiltration mediated by parenchymal TNFRs, the KO to WT chimeras were unable to control the Pc burden any better than did the fully WT mice. This is consistent with the failure of sensitized CD8+ T cells to control organism burden; however, the failure of enhanced macrophage numbers in this chimera to control the Pc again supports the role of TNF stimulation of the immune cells in Pc clearance.

A significant contribution of parenchymal cell TNF responses to lung injury induced by pathogens has been previously demonstrated. For example, CD8+ T cell recognition of alveolar cells expressing a specific viral Ag triggered MCP-1 and MIP-2 expression by the lung epithelial cells in large part due to T cell transmembrane TNF (tmTNF) and the presence of TNFRsf1a on the epithelium (27, 28). Additionally, a study of alveolar macrophages in patients with acute respiratory distress syndrome demonstrated enhanced tmTNF correlated with severity of disease, although soluble TNF concentrations in the BAL did not (29), suggestive of the importance of inflammatory cell-bound TNF interacting with parenchymal receptors. Because tmTNF is an active signaling molecule, the expression of TNFRs on parenchymal cells constitutes a mechanism for intercellular communication with tmTNF-expressing immune cells. In this way, TNF expressed by inflammatory cells may have enhanced capacity to either induce cytotoxicity or to stimulate proinflammatory protein production by the structural cells of the lung. In the case of Pc, the epithelial cells are anchors for the organism and so are best positioned to stimulate or amplify a local host response. Prior studies also suggest that CD8+ T cell mediated Pc-induced lung damage is dependent on MHC class I expression by radio-resistant cells (6). Intracellular Pc Ag processing or presentation by alveolar epithelial cells has not yet been demonstrated but is feasible. Intercellular TNF-TNFR interactions may enhance such lymphocyte-epithelial cell interactions.

As well as TNF-α, the TNFRs also bind and transduce signal of the homotrimeric ligand, lymphotoxin-α (LT-α), previously known as TNF-β. LT-α also signals by forming heterotrimers with LT-β that bind the LT-β receptors. Removal of the TNFRs therefore prevents signal transduction by homotrimeric LT-α, as well as by TNF-α. The phenotypes observed in the present study are then the result of manipulating both TNF-α and LT-α activity. Due to the strong relationship previously established between TNF-α and progression of PcP, it is thought likely that it is primarily the effect of TNF-α signal transduction via TNFR activity that has been altered in this study. No regulation or role of LT-α in PcP has been clearly demonstrated to date. Whether PcP is altered by LT-α binding of TNFRs on parenchymal or bone marrow cells, however, remains to be studied potentially by anti-ligand Ab or ligand-specific knockout transgenic models.

The TNF sf1a receptor has been demonstrated to be necessary for normal development and maintenance of splenic B cell follicles and germinal centers (30, 31, 32). In mice deficient in either LT-α or the type I TNFR, but not the type II TNFR, germinal centers failed to develop in peripheral lymphoid organs. While the chimeras in the present study have not been tested for germinal centers, it is highly unlikely that a reduction or failure of Ab production in response to Pc accounts for the differences in PcP between the CD4+ T cell-depleted chimeras studied given the known role of these lymphocytes in producing anti-Pc Abs. Even when Pc-infected SCID mice are given Pc-sensitized lymphocytes, CD4+ T cells are required for an Ab response to be generated (25). Additionally, nonimmunized, CD4-depleted mice, directly analogous to the animals tested in the present study, do not make an Ab response to Pc (33). Therefore, none of the chimeras tested in this study would have had an Ab response to Pc, regardless of the presence or absence of TNFRs.

Macrophage predominance persisted and was amplified in the BALF of KO to WT mice, in contrast to the other chimeras. This occurred in association with exaggerated production of the CC chemokine MCP-1, documented by increased tissue mRNA as well as BAL protein concentrations. Because MCP-1 is a chemoattractant for lymphocytes as well as for macrophages, it is possible that parenchymal response to TNF results in enhanced induction of MCP-1 from epithelial cells, for example, that in turn would increase recruitment of these cells. This model would be consistent with the viral model in which CD8+ T cell tmTNF stimulates alveolar epithelial cells to produce MCP-1 (34). It is also possible that activation of parenchymal cells by Pc-induced TNF causes release of other chemoattractants that enhance the recruit of inflammatory cells that become the source of MCP-1. Further analysis of the present model will determine the source of the CC chemokine. Previous studies suggest that CD8+ T cells, even stimulated by specific recognition of alveolar epithelial cells, do not produce MCP-1, whereas the target cells do (28). In contrast, in situ hybridization in the SCID mouse model of PcP demonstrated the primary location of MCP-1 mRNA to be the type II epithelial cells (23). Maximal MCP-1 production occurs in the absence of hematogenous cell TNFRs, and thus direct TNF stimulation of macrophages is not the source.

The alveolar macrophage numbers and MCP-1 induction was not as marked in the WT to WT chimera as in the KO to WT chimera, which is suggestive not only of a role for TNFR stimulation of parenchyma in production of this chemokine in response to Pc, but also of a suppressive effect of TNFRs in the marrow-derived cells. One potential explanation for immune cell TNFR-mediated suppression of Pc-induced lung inflammation is that in the absence of TNFRs, inflammatory cells recruited to the lung fail to undergo TNF-induced apoptosis, an important system of regulation of inflammation. Further studies of cellular turnover in the present model are indicated.

MCP-1 and MIP-2 have both been implicated in the pathogenesis of PcP, having been shown to be induced by Pc and by TNF, as well as correlating with severity of disease in WT and TNFR KO mice. In this Pc chimera model, however, independent regulation of the two chemokines was observed. In vitro studies with Pc-stimulated primary alveolar type II cells suggested that MCP-1 was induced by Pc directly and that this induction was dependent on both NF-κB and JNK activity (23). Preliminary data suggest a synergistic induction of MCP-1 from lung epithelial cells exposed to both Pc and TNF (data not shown). In contrast, MIP-2, similarly induced by direct Pc interaction with epithelial cells, was unaffected by JNK inhibitors, consistent with differential regulation of the two chemokines and perhaps greater dependence of MIP-2 gene expression on NF-κB, of which TNF is a most potent stimulant. Considering the RNA measurements, while maximal MCP-1 induction was dependent on parenchymal cell signaling in the absence of immune cell receptors, MIP-2 induction was maximal when both cell compartments could respond to TNF, inducing a 3-fold increase over that in the mixed chimeras.

One or both of the two distinct receptors, TNFRsf1a and TNFRsf1b, have been identified on most lung cells tested, including type II pneumocytes, although their relative ratio varies by cell type and can be altered by stimulation (35, 36, 37, 38). In most studies, TNFRsf1a is constitutively expressed while TNFRsf1b expression is inducible. The predominant receptor, by mRNA, in mouse and human lung is TNFRsf1a, although TNFRsf1b is induced by many stimuli including TNF delivery and Pc infection (9). A differential role of the two receptors in PcP has not yet been clarified. As for TNF, the TNFRs are also targets for matrix metalloproteinases. sTNFR may act as a reservoir of soluble TNF or as circulating inhibitors of both soluble and tmTNF. In this study, we demonstrated that Pc infection induces shedding of these receptors and accumulation of sTNFRsf1a and sTNFRsf1b both in serum and BAL of mice. Whether soluble TNFR levels may be useful as biomarkers reflecting the severity of PcP or may have a physiological role in disease is not yet determined. Interestingly, BMT chimera experiments suggest that in the basal, healthy state, most circulating sTNFRsf1a originates from parenchyma, whereas bone marrow-derived cells appear to be the source of >50% of sTNFRsf1b. In PcP, both sTNFRsf1a and sTNFRsf1b originated from both parenchymal and marrow-derived cells. Although the quantities of sTNFR present in serum or BALF have been shown to mirror the severity of other diseases, the regulation and function of TNFR solubilization remain unclear. sTNFRs may be involved in controlling the TNF response during the generation of an immune response. The KO to WT chimeras had reduced sTNFR and enhanced inflammatory response relative to WT to WT chimeras. It is possible that the lack of soluble receptors acting as competitive inhibitors contributes to the enhanced injury documented in the mixed chimeras. Alternatively, expression of membrane-bound TNFRs may create a feedback loop that regulates TNF transcription. Increased TNF production was documented in this study but only in the KO to KO mice, suggesting that TNF feedback on either marrow-derived or parenchymal cells is sufficient to regulate the ligand.

The present study demonstrates that TNFRs on parenchymal cells, those resistant to split dose irradiation, are sufficient to mount an inflammatory response to Pc, even if marrow-derived cells are receptor null, comparable to or in excess of that generated when all cells express the receptors. The relative importance of epithelial, endothelial, or mesenchymal cell TNFR responses, as well as the source of the stimulating TNF, remains to be determined. Limitation of TNFR distribution to marrow-derived cells improved control of Pc burden and reduced the injurious host response, best demonstrated in this study by improved survival and cellular recruitment as compared with normal receptor expression. Limitation of TNFR distribution to parenchymal cells markedly worsened the inflammatory response and resulting injury. The results of this study support the notion that therapeutic interventions that inhibit parenchymal cell TNF signal transduction, while maintaining or enhancing immune cell TNFR responses (analogous to the WT to KO chimeras), could be effective both in limiting Pc burden and in limiting the injurious host inflammatory response. Alternatively, global anti-TNF treatment, if given as an adjunctive therapy to the currently used and effective anti-Pc drugs, could also have benefit for patients with severe PcP by reducing the immune aspects of PcP-related lung injury.

The technical assistance of Jane Malone and Min Yee is greatly appreciated.

The authors have no financial conflicts of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by National Institutes of Health Grants P01HL71659, R01HL077415, R01HL064559, and P30ES01247 and the Strong Children’s Research Center.

3

Abbreviations used in this paper: Pc, Pneumocystis carinii; BAL, bronchoalveolar lavage; BALF, bronchoalveolar lavage fluid; BMT, bone marrow transplantation; KO, knockout; LDH, lactate dehydrogenase; LT, lymphotoxin; PcP, Pneumocystis carinii pneumonia; rpL32, ribosomal protein L32; sTNFR, soluble TNF receptor; tmTNF, transmembrane TNF; TNFR, TNF receptor; WT, wild type.

1
Dei-Cas, E..
2000
. Pneumocystis infections: the iceberg?.
Med. Mycol.
38
: (Suppl. 1):
23
-32.
2
Mansharamani, N. G., R. Garland, D. Delaney, H. Koziel.
2000
. Management and outcome patterns for adult Pneumocystis carinii pneumonia, 1985 to 1995: comparison of HIV-associated cases to other immunocompromised states.
Chest
118
:
704
-711.
3
Curtis, J. Randall, P. R. Yarnold, D. N. Schwartz, R. A. Weinstein, C. L. Bennett.
2000
. Improvements in outcomes of acute respiratory failure for patients with human immunodeficiency virus-related Pneumocystis carinii pneumonia.
Am. J. Respir. Crit. Care Med.
162
:
393
-398.
4
Vahid, B., M. Bibbo, P. E. Marik.
2007
. Role of CD8 lymphocytes and neutrophilic alveolitis in Pneumocystis jiroveci pneumonia.
Scand. J. Infect. Dis.
39
:
612
-614.
5
Wright, T. W., F. Gigliotti, J. N. Finkelstein, J. T. McBride, C. L. An, A. G. Harmsen.
1999
. Immune-mediated inflammation directly impairs pulmonary function, contributing to the pathogenesis of Pneumocystis carinii pneumonia.
J. Clin. Invest.
104
:
1307
-1317.
6
Meissner, N. N., F. E. Lund, S. Han, A. Harmsen.
2005
. CD8 T cell-mediated lung damage in response to the extracellular pathogen Pneumocystis is dependent on MHC class I expression by radiation-resistant lung cells.
J. Immunol.
175
:
8271
-8279.
7
Kaur, N., T. C. Mahl.
2007
. Pneumocystis jiroveci (carinii) pneumonia after infliximab therapy: a review of 84 cases.
Dig. Dis. Sci.
52
:
1481
-1484.
8
Kolls, J. K., D. Lei, C. Vazquez, G. Odom, W. R. Summer, S. Nelson, J. Shellito.
1997
. Exacerbation of murine Pneumocystis carinii infection by adenoviral-mediated gene transfer of a TNF inhibitor.
Am. J. Respir. Cell Mol. Biol.
16
:
112
-118.
9
Wright, T. W., G. S. Pryhuber, P. R. Chess, Z. Wang, R. H. Notter, F. Gigliotti.
2004
. TNF receptor signaling contributes to chemokine secretion, inflammation, and respiratory deficits during Pneumocystis pneumonia.
J. Immunol.
172
:
2511
-2521.
10
Peschon, J. J., D. S. Torrance, K. L. Stocking, M. B. Glaccum, C. Otten, C. R. Willis, K. Charrier, P. J. Morrissey, C. B. Ware, K. M. Mohler.
1998
. TNF receptor-deficient mice reveal divergent roles for p55 and p75 in several models of inflammation.
J. Immunol.
160
:
943
-952.
11
Theise, N. D., S. Badve, R. Saxena, O. Henegariu, S. Sell, J. M. Crawford, D. S. Krause.
2000
. Derivation of hepatocytes from bone marrow cells in mice after radiation-induced myeloablation.
Hepatology
31
:
235
-240.
12
Amdur, M. O., J. Mead.
1958
. Mechanics of respiration in unanesthetized guinea pigs.
Am. J. Physiol.
192
:
364
-368.
13
Meissner, N. N., S. Swain, M. Tighe, A. Harmsen, A. Harmsen.
2005
. Role of type I IFNs in pulmonary complications of Pneumocystis murina infection.
J. Immunol.
174
:
5462
-5471.
14
Pryhuber, G. S., D. P. O'Brien, R. Baggs, R. Phipps, H. Huyck, I. Sanz, M. H. Nahm.
2000
. Ablation of tumor necrosis factor receptor type I (p55) alters oxygen-induced lung injury.
Am. J. Physiol.
278
:
L1082
-L1090.
15
Lee, L. H., F. Gigliotti, T. W. Wright, P. J. Simpson-Haidaris, G. A. Weinberg, C. G. Haidaris.
2000
. Molecular characterization of KEX1, a kexin-like protease in mouse Pneumocystis carinii.
Gene
242
:
141
-150.
16
Wallach, D., H. Engelmann, Y. Nophar, D. Aderka, O. Kemper, V. Hornik, H. Holtmann, C. Brakebusch.
1991
. Soluble and cell surface receptors for tumor necrosis factor.
Agents Actions
35
:
51
-57.
17
Wajant, H., K. Pfizenmaier, P. Scheurich.
2003
. Tumor necrosis factor signaling.
Cell Death Differ.
10
:
45
-65.
18
Parsons, P. E., M. A. Matthay, L. B. Ware, M. D. Eisner.
2005
. Elevated plasma levels of soluble TNF receptors are associated with morbidity and mortality in patients with acute lung injury.
Am. J. Physiol.
288
:
L426
-L431.
19
Chen, W., E. A. Havell, A. G. Harmsen.
1992
. Importance of endogenous tumor necrosis factor alpha and gamma interferon in host resistance against Pneumocystis carinii infection.
Infect. Immun.
60
:
1279
-1284.
20
Kolls, J. K., J. M. Beck, S. Nelson, W. R. Summer, J. Shellito.
1993
. Alveolar macrophage release of tumor necrosis factor during murine Pneumocystis carinii pneumonia.
Am. J. Respir. Cell Mol. Biol.
8
:
370
-376.
21
Hoffman, O. A., J. E. Standing, A. H. Limper.
1993
. Pneumocystis carinii stimulates tumor necrosis factor-alpha release from alveolar macrophages through a beta-glucan-mediated mechanism.
J. Immunol.
150
:
3932
-3940.
22
Evans, S. E., P. Y. Hahn, F. McCann, T. J. Kottom, Z. V. Pavlovic, A. H. Limper.
2005
. Pneumocystis cell wall β-glucans stimulate alveolar epithelial cell chemokine generation through nuclear factor-κB-dependent mechanisms.
Am. J. Respir. Cell Mol. Biol.
32
:
490
-497.
23
Wang, J., F. Gigliotti, S. P. Bhagwat, S. B. Maggirwar, T. W. Wright.
2007
. Pneumocystis stimulates MCP-1 production by alveolar epithelial cells through a JNK-dependent mechanism.
Am. J. Respir. Cell Mol. Biol.
292
:
L1495
-L1505.
24
Gigliotti, F., T. W. Wright.
2005
. Immunopathogenesis of Pneumocystis carinii pneumonia.
Expert Rev. Mol. Med.
7
:
1
-16.
25
Gigliotti, F., E. L. Crow, S. P. Bhagwat, T. W. Wright.
2006
. Sensitized CD8+ T cells fail to control organism burden but accelerate the onset of lung injury during Pneumocystis carinii pneumonia.
Infect. Immun.
74
:
6310
-6316.
26
Kottom, T. J., J. R. Kohler, C. F. Thomas, Jr, G. R. Fink, A. H. Limper.
2003
. Lung epithelial cells and extracellular matrix components induce expression of Pneumocystis carinii STE20, a gene complementing the mating and pseudohyphal growth defects of STE20 mutant yeast.
Infect. Immun.
71
:
6463
-6471.
27
Zhao, M. Q., M. K. Amir, W. R. Rice, R. I. Enelow.
2001
. Type II pneumocyte-CD8+ T-cell interactions: relationship between target cell cytotoxicity and activation.
Am. J. Respir. Cell Mol. Biol.
25
:
362
-369.
28
Zhao, M. Q., M. H. Stoler, A. N. Liu, B. Wei, C. Soguero, Y. S. Hahn, R. I. Enelow.
2000
. Alveolar epithelial cell chemokine expression triggered by antigen-specific cytolytic CD8+ T cell recognition.
J. Clin. Invest.
106
:
R49
-R58.
29
Armstrong, L., D. R. Thickett, S. J. Christie, H. Kendall, A. B. Millar.
2000
. Increased expression of functionally active membrane-associated tumor necrosis factor in acute respiratory distress syndrome.
Am. J. Respir. Cell Mol. Biol.
22
:
68
-74.
30
Alexopoulou, L., M. Pasparakis, G. Kollias.
1998
. Complementation of lymphotoxin α knockout mice with tumor necrosis factor-expressing transgenes rectifies defective splenic structure and function.
J. Exp. Med.
188
:
745
-754.
31
Matsumoto, M., S. Mariathasan, M. H. Nahm, F. Baranyay, J. J. Peschon, D. D. Chaplin.
1996
. Role of lymphotoxin and the type I TNF receptor in the formation of germinal centers.
Science
271
:
1289
-1291.
32
Pasparakis, M., S. Kousteni, J. Peschon, G. Kollias.
2000
. Tumor necrosis factor and the p55TNF receptor are required for optimal development of the marginal sinus and for migration of follicular dendritic cell precursors into splenic follicles.
Cell. Immunol.
201
:
33
-41.
33
Harmsen, A. G., W. Chen, F. Gigliotti.
1995
. Active immunity to Pneumocystis carinii reinfection in T-cell-depleted mice.
Infect. Immun.
63
:
2391
-2395.
34
Zhao, M. Q., M. P. Foley, M. H. Stoler, R. I. Enelow.
2001
. Alveolar epithelial cell chemokine expression induced by specific antiviral CD8+ T-cell recognition plays a critical role in the perpetuation of experimental interstitial pneumonia.
Chest
120
:
11S
-13S.
35
Abdolrasulnia, R., V. L. Shepherd.
1992
. Purification of type I and type II tumor necrosis factor receptors from human lung tissue.
Am. J. Respir. Cell Mol. Biol.
7
:
42
-48.
36
Schimomoto, H., Y. Hasegawa, Y. Nozaki, N. Takagi, T. Shibagaki, A. Nakao, K. Shimokata.
1995
. Expression of tumor necrosis factor receptors in human lung cancer cells and normal lung tissues.
Am. J. Respir. Cell Mol. Biol.
15
:
271
-278.
37
Brockhaus, M., H. J. Schoenfeld, E. J. Schlaeger, W. Hunziker, W. Lesslauer, H. Loetscher.
1990
. Identification of two types of tumor necrosis factor receptors on human cell lines by monoclonal antibodies.
Proc. Natl. Acad. Sci. USA
87
:
3127
-3131.
38
Pryhuber, G. S., H. L. Huyck, R. J. Staversky, J. N. Finkelstein, M. A. O'Reilly.
2000
. Tumor necrosis factor-α-induced lung cell expression of antiapoptotic genes TRAF1 and cIAP2.
Am. J. Respir. Cell Mol. Biol.
22
:
150
-156.