Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The clinical success rate of allogeneic bone marrow transplantation (BMT) has increased steadily. Yet graft-versus-host disease (GVHD) and pulmonary complications remain serious threats to survival after transplant. Pulmonary complications account for 40-60% of the morbidity and mortality in BMT patients (1), and up to 85% of mortalities post-BMT have been reported to result from interstitial pneumonitis (both infectious and idiopathic) (2). Noninfectious idiopathic pneumonia syndrome (IPS) accounts for as much as 50% of the total cases of interstitial pneumonitis after BMT (3). The incidence of IPS usually occurs from 2 wk to 6 mo post-BMT (1). The main pathologic features of IPS are diffuse interstitial pneumonitis and diffuse alveolitis in the absence of an identifiable infectious agent; however, other reported manifestations include interstitial edema, interstitial fibrosis, lymphocytic bronchiolitis, and alveolar hemorrhage (7). In contrast, bronchiolitis obliterans (BO), which is characterized as conductive airway occclusion caused by intense mononuclear cell accumulation, can occur anytime after 3 mo post-BMT (1). Whether IPS and BO are temporally distinct syndromes caused by the same disease process is currently unknown. A hypothesis that an immunopathologic mechanism underlies IPS is based on the observation that the severity of IPS was reduced in patients receiving immunosuppressive therapy for GVHD (1). GVHD is clinically distinguished as either an acute or chronic disease on the basis of the time of occurrence of symptoms post-BMT. Acute GVHD generally occurs in the first 100 d post-BMT, whereas chronic GVHD, which can occur months to years after transplant, is usually less severe but more progressive (1). Whether GVHD in the lungs contributes to the morbidity of IPS has been difficult to pinpoint because IPS can occur in the setting of several distinct lung abnormalities. The immunologic defects produced by an underlying malignancy, the pretransplant chemotherapy, irradiation, and the development of GVHD, may all contribute to the pulmonary abnormalities encountered in BMT patients. Unfortunately, it has been difficult to evaluate the cellular and molecular mechanisms of IPS post-BMT in the absence of a suitable animal model that mimics the disease progression seen in humans.

We have used a parental  F1 model of GVHD to investigate the development of lung disease following allogeneic bone marrow transplantation. Lethally irradiated [DBA/2 × C57BL/6] F1 mice (B6D2F1) were transplanted with C57BL/6 (B6) bone marrow containing a small number of B6 spleen cells as a constant source of alloreactive T-cells. In this article, we demonstrate induction of a mild form of GVHD, which is associated with the development of pulmonary inflammation that displays many of the signs of IPS in humans.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Mice

Female C57BL/6 ("parental" strain, H-2^b) and B6D2F1 (F1 strain, H-2^b,d) mice were purchased from the National Cancer Institute (Frederick, MD) and maintained in sterile microisolator cages (Lab Products, Inc., Maywood, NJ) with sterile rodent chow and acidified water ad libitum. Mice were maintained by the Division of Laboratory Animal Resources at the University of Kentucky according to the guidelines in the Animal Welfare Act. Sentinel mice were held in the same room and periodically screened for serologic evidence of infection for a panel of common mouse pathogens. Donor and recipient mice were 5-6 wk of age at the beginning of each study.

Induction of GVHD

B6D2F1 (F1) mice were lethally irradiated (9 Gy) in a Mark I ¹³⁷Cs irradiator (JL Shepard and Associates, Glendale, CA) and were transfused within 4-6 h with a single injection of 1 × 10⁷ T-cell-depleted C57BL/6 (B6) bone marrow cells (BMCs) or B6 BMCs plus 5 × 10⁶ spleen cells for the BMT control and GVHD groups, respectively. Preparation of BMCs and spleen cells was done as previously described (8). Cells were suspended in sterile phosphate-buffered saline (PBS, pH 7.2), and 0.1 ml was injected into the tail vein of each recipient mouse. Following transplantation, mice were housed without additional treatment until killed for tissue analysis. Groups of three mice were killed at weeks 3, 5, 7, 9, and 12 as indicated in each experiment, at which time body weights and spleen weights were recorded.

Isolation of Lung Lymphoid Cells

Lung lymphocytes were isolated by a modified method of Abraham and coworkers (9) as we have previously described (10). Briefly, after perfusion of the lungs with sterile PBS to remove blood and circulating lymphoid cells, one lobe was snap frozen in liquid nitrogen and then stored at 80°C for subsequent reverse transcription-polymerase chain reaction (RT-PCR) analysis. The remaining lung tissue was used for isolation of lymphoid cells for analysis by flow cytometry. Lungs were dissected into small fragments and incubated for 60 min at 37°C in RPMI 1640 medium (supplemented with 5% fetal calf serum [FCS], 2 mM L-glutamine, 100 U/ml penicillin, 0.1 mg/ml streptomycin, and 0.05 mM 2-mercaptoethanol) containing 20 U/ml collagenase (Sigma Chemical Co., St. Louis, MO) and 40 µg/ml DNase I (Sigma). Cells were released from the tissue by a 4-min treatment in a Stomacher tissue disrupter (Seward Medical, London, UK), centrifuged on a discontinuous Percoll gradient (Sigma), and lymphoid cells were then collected at the 40-80% interface. Cells were adjusted to equal concentrations and stained for flow cytometric analysis.

Assessment of Immunodeficiency

Immunodeficiency was assessed by testing the proliferative capacity of splenocytes in response to concanavalin A (ConA) and lipopolysaccharide (LPS). Single-cell suspensions of spleen cells were obtained by processing each spleen separately in supplemented RPMI medium using a Stomacher tissue disrupter (Seward Medical). Cells (5 × 10⁵) were cultured in triplicate in each well of a 96-well culture plate (Becton Dickinson, Franklin Lakes, NJ) with either ConA (5 µg/ml; Sigma) or LPS (10 µg/ml; Sigma) for 48 h at 37°C in the presence of 5% CO₂. Proliferation was determined by incorporation of [³H]thymidine (ICN Pharmaceuticals, Inc., Costa Mesa, CA) during the last 4 h of culture.

Flow Cytometry

Lung lymphoid cells and spleen cells from individual mice were assessed by single- or two-color flow cytometry using the following monoclonal antibodies: anti-CD3-fluorescein isothiocyanate (FITC) (Sigma), anti-CD4-FITC (Sigma), anti-CD8-phycoerythrin (PE) (Sigma), anti-polyvalent immunoglobulin (IgA, IgG, IgM)-FITC (Sigma), anti-H-2D^d-PE (Pharmingen, Inc., San Diego, CA), and anti-H-2K^b-FITC (Pharmingen). Data are expressed as the percentage of gated lymphocytes positive for a given cell surface marker.

Reverse Transcriptase-Polymerase Chain Reaction

The RT-PCR was used to detect cytokine mRNA in lung tissue. Total RNA was isolated from frozen lung tissue as previously described by Cohen and colleagues (10). One microgram of RNA was reverse transcribed into cDNA with the Promega reverse transcription system (Promega Corp., Madison, WI). One-tenth volume of the reaction product was subsequently amplified with Taq polymerase (Promega) for 30 cycles for all cytokines, whereas the housekeeping gene -actin was amplified for 25 cycles. Note that the number of cycles was chosen on the basis of preliminary studies to identify conditions that were subsaturating so that quantitative comparisons between samples could be established. A 50-µl PCR contained 2 µl of the cDNA mix, 0.2 mM dNTP, 1.5 mM MgCl₂, 0.75 µM each of sense and antisense primers, and 1.5 U of Taq polymerase in 1× reaction buffer (Promega). Amplification was performed by a 1-min denaturation step at 94°C, primer annealing for 1 min at 55°C, and elongation at 72°C for 2 min in a Perkin-Elmer thermal cycler. PCR products were separated by electrophoresis on 2% agarose gels and were visualized by ethidium bromide staining. PCR band densities were determined by the OneDScan analytical program (Scanalytics, Billerica, MA) on unaltered computer-scanned images. Band densities of cytokines from each mouse were divided by the density of -actin for the same mouse and results are expressed as normalized integrated optical densities. The sequences of primers used to amplify cytokine messages in this study were as follows:

IL-1   Sense: 5'-CAGGATGAGGACATGAGCACC-3'

  Antisense: 5'-CTCTGCAGACTCAAACTCCAC-3'

TNF-   Sense: 5'-ATGAGCACAGAAAGCATGATC-3'

  Antisense: 5'-TACAGGCTTGTCACTCGAATT-3'

IL-6   Sense: 5'-GACAAAGCCAGAGTCCTTCAGAGAG-3'

  Antisense: 5'-CTAGGTTTGCCGAGTAGATCTC-3'

IFN-   Sense: 5'-TACTGCCACGGCACAGTCATTGAA-3'

  Antisense: 5'-GCAGCGACTCCTTTTCCGCTTCCT-3'

IL-12 (p40)   Sense: 5'-GGAGACCCTGCCCATTGAACT-3'

  Antisense: 5'-CAACGTTGCATCCTAGGATCG-3'

Histology

For histologic examination, lungs were obtained from groups of three mice. Lungs were perfused via tracheal instillation with neutral buffered formalin (10% formalin, 46 mM Na₂HPO₄, and 29 mM NaH₂PO₄) for 5 min. Lungs and heart were then removed en bloc and stored in neutral buffered formalin until further use. Histologic data shown were derived from sections of the right cardiac lobe of the lungs, which was removed from fixative and embedded in paraffin for sectioning. Multiple 4-µm sections were stained with hematoxylin and eosin by the Histology Services at the University of Kentucky Medical Center.

Statistical Analysis

To determine significant differences in the expression of cytokines from groups of three mice, the band densities obtained from densitometry of RT-PCR products were compared by Student's independent (unpaired) t test using SigmaStat software (Jandel Scientific Co., San Rafael, CA). A value of P  0.05 was considered to be significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Induction of Murine Graft-versus-Host Disease

GVHD was induced in B6D2F1 mice by administration of a single intravenous injection of 1 × 10⁷ T-cell-depleted B6 bone marrow cells and 5 × 10⁶ B6 spleen cells (as a source of T cells) to lethally irradiated B6D2F1 (F1) recipients. Controls consisted of F1 mice that were either untreated or received only T-cell-depleted bone marrow (BMT controls). Following BMT, mice were maintained without additional treatment for up to 12 wk. Groups of three mice were killed at weeks 3, 5, 7, 9, and 12 post-BMT, and tissues were analyzed as indicated below. To determine engraftment, at each time point, recipient spleen cells were analyzed by flow cytometry for expression of the parental and F1 haplotypes using anti-H-2K^b and anti-H-2D^d antibodies. Approximately 90-95% of cells in the recipients were routinely found to be of the donor haplotype (not shown).

Characteristics of acute GVHD (11) such as lymphoid atrophy and moderate to severe mononuclear cell infiltrates were seen in tissue sections of skin in animals transplanted with BM plus 1 × 10⁷ spleen cells generally by week 3 (data not shown). However, in GVHD animals receiving BM plus a lower dose of 5 × 10⁶ spleen cells, symptoms of acute GVHD were still apparent but less severe. Overall acute GVHD at week 3 post-BMT was assessed as mild using a previously established histology grading procedure (12), which indicated grade I/II histopathology in the liver and skin (data not shown). Importantly, GVHD remained mild at week 12, because the histopathologic grade in liver and skin did not exceed grade II for any GVHD mouse at this time point. Loss of body weight, which was greater in the GVHD group than in the BMT controls, was observed at week 3 post-BMT and recovered to normal by week 12 (Figure 1). In addition, spleen weights of GVHD mice were approximately half that of BMT controls at week 3 and failed to return to normal levels throughout the entire period of the study (Table 1), suggesting the onset of a persistent form of GVHD. Immunosuppression, a hallmark of acute GVHD, typically develops early in the course of disease (13). Figure 2 shows that spleen cells from GVHD mice at 3 wk posttransplantation failed to respond to the T-cell and B-cell mitogens, ConA and LPS, respectively. The low proliferative response of the BMT controls could be accounted for by the lower numbers of mature donor lymphocytes at this time point (data not shown). However, by week 12, proliferative responses in GVHD mice had returned toward normal and were equivalent to that of BMT controls.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Loss of body weight during GVHD. Body weights of GVHD and BMT control mice were recorded at weeks 3, 5, 7, 9, and 12 posttransplantation. Data shown represent the mean ± SEM (n = 3) from a single experiment.

View larger version (36K): [in this window] [in a new window]   Figure 2.   Deficient mitogenesis during GVHD. Spleen cells from GVHD and BMT control mice were stimulated in vitro with 5 µg/ ml ConA or 10 µg/ml LPS and proliferative responses were measured 48 h later by [3H]thymidine uptake. Data shown represent the mean ± SEM (n = 3) from a single experiment.

Histologic Evaluation of Lungs during GVHD

Standard histologic analysis was performed to examine lungs for the presence of IPS. Hematoxylin and eosin-stained sections of the right lobe of the lungs from GVHD mice at 3 wk posttransplantation demonstrated only minor histopathologic changes, consisting of slight edema in interstitial tissue with little or no infiltration of inflammatory cells (data not shown). At 12 wk post-BMT the lungs of untreated normal F1 mice or those of syngeneic BMT controls (F1  F1 controls) showed no signs of pathology (Figures 3A and 3B). Lungs of allogeneic BMT control mice showed minimal histopathologic changes at week 12 (Figure 3C), whereas those of GVHD animals displayed prominent perivascular and peribronchiolar inflammation and diffuse infiltrates of mononuclear cells throughout the alveolar interstitium (Figures 3D and 4A). GVHD lungs also exhibited focal extension of similar mononuclear infiltrates beyond the limiting plate into alveolar interstitial tissues (Figure 4A), which were not evident in allogeneic BMT control lungs (data not shown). Regions of organizing inflammatory lesions were evident by the presence of focal regions of prominent alveolar congestion owing to intraalveolar hyperplastic macrophage-like cells, found only in the GVHD mice at week 12 (Figure 4B). Whether these intraalveolar cells are indeed macrophages or metaplastic alveolar epithelial cells remains to be determined.

View larger version (140K): [in this window] [in a new window]   Figure 3.   Histopathology of interstitial pneumonitis is evident in GVHD mice. Untreated normal F1 mice (A), F1  F1 syngeneic BMT control mice (B), allogeneic BMT control mice (C), and GVHD mice (D) were killed at the end of the study (12 wk after transplant) and paraffin-embedded lung sections were stained with hematoxylin and eosin. (Original magnification: ×400.)

View larger version (127K): [in this window] [in a new window]   Figure 4.   Organizing inflammatory lesions are evident in GVHD mice. Peribronchiolar inflammation (A) and alveolar congestion (B) are evident in lungs of GVHD mice. Paraffin-embedded lung sections were stained with hematoxylin and eosin and observed by light microscopy. [Original magnification: (A) ×500; (B) ×400.]

Analysis of tissue sections with Masson's trichrome stain demonstrated that lungs of GVHD mice 12 wk post-BMT had increased deposition of collagen, which was found mainly in perivascular and peribronchiolar regions (data not shown). Such fibrosis was observed to a lesser degree in lungs of BMT control mice and was completely absent in untreated mice and F1  F1 control mice (data not shown). Furthermore, lungs of GVHD mice displayed localized interstitial fibrosis that was not observed in any of the control groups of mice. These results indicate that a single injection of allogeneic bone marrow cells containing a small number of mature T cells leads to the development of focal interstitial pneumonitis with some accompanying fibrosis during the later stages of GVHD.

Characterization of the Lymphocytic Infiltration of Lungs during GVHD

To determine the relative distribution of CD4⁺ and CD8⁺ T cells in the lungs of mice with GVHD, flow cytometric analysis was performed on isolated lung lymphoid cells from GVHD and BMT control mice at 3, 5, 7, 9, and 12 wk post-BMT. For comparative purposes, T-cell subset distribution in spleens of the same groups was also evaluated by flow cytometry (Figure 5). Lungs of BMT control mice displayed reduced levels of CD4⁺ T cells and CD8⁺ T cells at week 3, which were likely reduced owing to incomplete reconstitution at this time point. The CD8⁺ T-cell level in BMT control mice was elevated above normal at week 5, but at later time points both CD4⁺ and CD8⁺ subsets were at normal levels. A similar response was observed in spleens of BMT control mice in that reduced T cells were found at week 3 for both subsets and elevated levels at week 5, which returned to normal levels at later time points. Lungs of GVHD mice displayed a different response than BMT controls. At week 3, CD8⁺ T cells were significantly elevated above normal in the GVHD mice, but were at or below normal levels at all later time points. In contrast, CD4⁺ T cells were at normal levels at weeks 3 and 5, but were significantly elevated above normal at later time points. A similar pattern of T-cell distribution was seen in the spleens of GVHD mice. These data indicate that GVHD mice develop a progressive interstitial pneumonitis, which is dominated by CD8⁺ T cells early in the course of disease and is replaced at later time points by a prominent infiltration of CD4⁺ T cells. Moreover, the cellular response in the lungs appears identical to the pattern of GVHD that develops in the spleens of transplanted mice.

View larger version (24K): [in this window] [in a new window]   Figure 5.   T-cell subset distribution in lungs and spleen of GVHD and BMT control mice. Lung lymphoid cells and splenocytes were isolated from GVHD and BMT control mice at the indicated time points, stained with fluorochrome-conjugated antibodies specific for cell surface CD4 (squares) and CD8 (circles) molecules, and analyzed by flow cytometry. Data shown represent the mean ± SEM (n = 3) from a single experiment. Dashed and dotted lines indicate the average percentage of CD4+ and CD8+ lymphocytes, respectively, in normal untreated F1 mice. (Asterisk indicates n = 2.)

Cytokine Expression in Lungs during GVHD

Lung lobes of individual mice from GVHD and BMT control groups were analyzed at weeks 3 and 12 posttransplantation by RT-PCR for steady state levels of mRNA for several cytokines to determine if mRNA expression of particular cytokines correlated with the development of interstitial pneumonitis during GVHD. Band densities of cytokines from each mouse were divided by the density of -actin for the same mouse and results are expressed as normalized integrated optical densities. In all experiments shown, the densities of -actin for individual mice did not vary significantly from one another (Figure 6).

View larger version (33K): [in this window] [in a new window]   Figure 6.   Proinflammatory cytokine profile in lungs of GVHD and BMT control mice. Total lung RNA was extracted and subjected to RT-PCR analysis for IL-6, IL-1, and TNF. Signal strength of ethidium bromide-stained bands on the agarose gel was determined by densitometry and the optical densities thus obtained for each cytokine from individual mice were normalized to their corresponding -actin levels. Data are shown as normalized optical densities and represent the mean ± SEM (n = 3) from a single experiment. (ND = not detectable.)

The proinflammatory cytokines interleukin (IL)-1 and tumor necrosis factor  (TNF-) were significantly higher (P  0.05) in lungs of GVHD animals at 3 wk following transplantation compared with untreated controls (Figure 6). BMT control mice also displayed elevated expression of these cytokines, although only IL-6 was significantly higher than untreated controls. At 12 wk posttransplantation, whereas IL-6 and IL-1 mRNA levels subsided to the basal (normal) level of expression in both GVHD and BMT control mice, TNF- mRNA levels remained elevated in GVHD animals. The mRNA level of the helper T cell type 1 (Th1) cytokine, interferon  (IFN-), was increased in the lungs of both GVHD and BMT control mice compared with untreated mice at 3 wk post-BMT (Figure 7). However, the level in GVHD mice was significantly greater than in BMT controls (P = 0.0002). IL-12 mRNA levels were also elevated at week 3 in GVHD and BMT control mice compared with untreated mice, but were not significantly different from each other. At 12 wk post-BMT, IFN- levels returned to basal levels in both GVHD and BMT control mice, whereas IL-12 levels in GVHD lungs remained significantly elevated (P < 0.005) compared with BMT control and untreated mice. These results indicate that proinflammatory cytokine gene expression in lungs is upregulated early after BMT, probably owing to the cumulative effects of radiation conditioning and hematopoietic cell reconstitution, but that elevated levels of IL-12 and TNF- mRNA at week 12 post-BMT occurred only in those mice that developed GVHD. The mRNA for the Th2 cytokines IL-4, IL-5, and IL-10 were undetectable in all groups of mice at all time points studied (data not shown).

View larger version (34K): [in this window] [in a new window]   Figure 7.   Th1 cytokine profile in lungs of GVHD and BMT control mice. Total lung RNA was extracted and subjected to RT-PCR analysis for IFN- and IL-12. Signal strength of ethidium bromide-stained bands on the agarose gel was determined by densitometry and the optical densities thus obtained for each cytokine from individual mice were normalized to their corresponding -actin levels. Data are shown as normalized optical densities and represent the mean ± SEM (n = 3) from a single experiment.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The development of interstitial pneumonitis remains a significant threat to the long-term survival of bone marrow transplant recipients. The incidence of IPS following BMT has been directly correlated with the dose of radiation administered to the lung and with the incidence of GVHD (14). Why the lungs become a target for this post-BMT inflammatory response and whether this response represents a unique disease process or another tissue site for GVHD remain unknown. This study demonstrates that B6D2F1 mice that are conditioned with whole-body irradiation and transplanted with semiallogeneic bone marrow containing mature T cells develop a progressive pulmonary inflammation similar to IPS in human BMT recipients. Transplantation of a small number of semiallogeneic C57BL/6 spleen cells (5 × 10⁶) as a source of T cells along with C57BL/6 bone marrow cells into lethally irradiated B6D2F1 recipients was shown to induce a mild, sublethal form of GVHD that remained mild throughout the 12 wk of study. Interstitial pneumonitis developed late in mice undergoing GVHD, whereas mice that were transplanted only with T cell-depleted bone marrow, and thereby did not develop GVHD, showed no signs of interstitial pneumonitis. The progression of IPS was associated with an acute influx of CD8⁺ T cells during the first 3 wk after transplant, a chronic accumulation of CD4⁺ T cells that began after week 5 post-BMT, and increased expression of mRNA for proinflammatory and Th1 cytokines in the lungs of GVHD mice. It should be noted that appearance of histopathologic lesions, such as collagen deposition and alveolar macrophage accumulation, was observed only at the 12-wk time point when CD4⁺ T cells predominated within the lungs. Whether CD4⁺ T cells and/or the cytokines produced by these cells are responsible for the development of IPS in this model remains to be determined.

The GVHD model that was developed for these studies was modified from a standard protocol to induce acute GVHD (15). Because most of the pathology associated with graft-versus-host reactions is initiated by mature donor-derived T cells that react against allotypic determinants on host cells (16), reduction of the number of donor spleen cells to 5 × 10⁶ allowed for a mild, less lethal form of GVHD. Previous studies employing P  F1 transplant strategies to induce acute or chronic GVHD have involved the injection of 10⁷-10⁸ parental spleen cells as a source of T cells (17). Depending on the strain of the parental donor spleen cells, either acute or chronic GVHD was induced in recipient mice (21, 22). Despite the histologic similarities of these two murine models to acute and chronic GVHD in humans, neither model accurately reflected the temporal progression of GVHD in humans (23). Acute GVHD has been described as the most important risk factor for chronic GVHD, and an increased probability of chronic GVHD has been reported to correlate with the incidence and severity of acute GVHD (26). Indeed, whether acute and chronic GVHD are two separate phenomena or represent two stages of a single process has been a matter of controversy (25, 27). The murine model described in this article not only displays a mild pattern of GVHD, but also an interstitial pneumonia-like syndrome, which develops in the chronic phase as is seen in humans.

Signs of acute GVHD (13, 17) occurred at week 3 post-BMT as indicated by weight loss, splenic atrophy, and lack of proliferative responses of lymphocytes to ConA and LPS. By week 12 post-BMT, body weights had recovered to normal while splenic atrophy persisted. Proliferative responses of splenic B and T cells were, however, normal at this time point, a finding that has also been reported in other mouse models of acute GVHD (13, 17, 19). It may be that the proliferative capacity of T cells is critical for the development of IPS, because it is likely that activation of alloreactive cells in the lung is necessary for sustained cytokine release or other functions by these cells.

Histologic analysis of lungs from GVHD mice at 12 wk post-BMT provided evidence of severe IPS, particularly inflammation around blood vessels and airways, with some fibrosis and focal alveolar congestion. It should be noted that bronchiolitis obliterans (BO) was not readily apparent in lung sections up to 12 wk after BMT. BO is a common finding in individuals with severe BMT-related IPS (28). Although frank BO was not observed in this study, areas of minor bronchiolar inflammation were seen (data not shown). Thus, it will be critical to determine if BO is an end point of the progressive lung disease in this GVHD model. Future studies are planned to examine transplanted mice up to 20 wk post-BMT, which is necessary to determine if BO will eventually develop in this model.

A substantial increase in the percentage of CD8⁺ T cells in both GVHD spleen and lungs occurred at 3 wk post-BMT, which is normally associated with cytotoxicity and tissue destruction. However, it has been previously reported that cells of the CD8⁺ phenotype occur in acute GVHD and mediate immunosuppression (25, 29). Analyses in various GVHD models have revealed that CD8⁺ cells are chiefly associated with acute GVHD, whereas CD4⁺ cells are associated with chronic GVHD (16, 25). The influx of CD8⁺ T cells into spleen and lungs was rapid and transient in that the major T-cell subset at later time points was CD4⁺, a feature commonly associated with chronic GVHD.

The pathogenesis of many chronic lung diseases is thought to involve local secretion of cytokines, such as IFN-, which are known to have modulatory effects on the function and adhesiveness of endothelial cells and epithelial cells. The activation and proliferation of T cells, as occurs in acute GVHD, is normally associated with Th1 cytokine expression, whereas T cells from chronic GVHD lesions usually display a Th2 cytokine pattern (22, 30, 31). Messenger RNA for the Th1 cytokine IFN- was elevated at week 3 post-BMT in lungs of GVHD animals compared with BMT controls, suggesting that early immunologic changes in the lungs may be due to an acute GVHD response. Interestingly, IL-12 mRNA levels were elevated in the lungs of GVHD mice at 12 wk following transplantation in the present study, but IFN- expression had returned to normal levels. Whether this difference reflects an inability of IL-12-responder cells such as T cells and/or NK cells to produce IFN- or a decrease in the number of IFN- producer cells remains to be determined. The inability to detect IL-4, IL-5, and IL-10 by PCR prevents us from proposing a role of Th2 cytokines in the lung pathology associated with chronic GVHD at this stage. Additional studies to increase the level of detection of these cytokines by southern blot or ELISA (35) may help resolve this issue.

The lungs of mice with GVHD also expressed significantly elevated levels of mRNA for the proinflammatory cytokine TNF- at 12 wk post-BMT. The involvement of TNF- in GVHD has been demonstrated in a study in which administration of anti-TNF antibodies to mice with GVHD led to significant inhibition of GVHD mortality (32). Increased TNF- mRNA expression was also found to be associated with alveolitis in GVHD lungs (33). Interestingly, the expression of a TNF- transgene in murine lungs can lead to the development of pulmonary fibrosis (34). Elevated expression of TNF- and IL-12 in GVHD lungs at 12 wk post-BMT in the present study may reflect the persistence of activated macrophages in lungs after transplantation of allogeneic bone marrow plus T cells. The cellular source of cytokines in this model is currently unknown. A variety of cell types can synthesize IL-12, TNF-, and IFN-, and cell types other than CD4⁺ T cells have been implicated in the pathology of GVHD, including NK cells. Additional studies to identify the cytokine-producing cell type are required to answer this question.

In conclusion, these studies demonstrate that allogeneic bone marrow transplantation in this model results in a progressive interstitial pneumonitis that is characterized by an acute influx of CD8⁺ T cells and is eventually replaced by a chronic accumulation of CD4⁺ T cells. Furthermore, these cells and/or other lung cells produce cytokines known to activate macrophages, which may contribute to the pathogenesis of this lung disease.