INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The inflammatory response is a complex multicell process that evolved to protect the host from invading organisms, in the majority of cases resulting in the complete resolution of effects and no permanent injury. However, there are circumstances in which inflammatory processes have deleterious results. These processes underlie a number of pulmonary diseases that are today increasing in frequency. Chronic inflammation often leads to injury and permanent scarring of the involved tissue, with consequent catastrophic loss of function. Chronic inflammatory disease of the lung is an important cause of premature death and loss of time from work, but the mechanisms that underlie persistent inflammation remain obscure. This lack of understanding has hampered the design of treatments for such disease, which is further complicated by the lack of means by which inflammatory processes may be monitored in the patient. Current methods of assessment, such as lung biopsy, sample only a small fraction of the organ in what is often patchy disease, and are too hazardous to repeat unless absolutely necessary. Chest X-rays show changes in lung density but can give no indication of disease activity. Bronchoalveolar lavage (BAL) accesses only the air spaces, when the disease process may be largely in the interstitium.

External imaging of radiolabeled markers of inflammatory processes has great potential for providing noninvasive and repeatable methods of monitoring inflammatory-cell behavior. Neutrophils radiolabeled ex vivo with ¹¹¹In and monitored by gamma-scintigraphy have given valuable information about neutrophil migration and emigration in response to inflammatory stimuli (1), but this trafficking of cells gives no indication of their postmigratory metabolic activity.

Neutrophils become highly activated in response to inflammatory stimuli, and undergo a respiratory burst. The consequent increase in their energy requirements results in an increased uptake and metabolism of glucose. The use of radiolabeled glucose analogues enables the increase in glucose uptake to be monitored by external imaging. In previous studies, we have shown that in an experimental animal model of acute lung inflammation induced by localized instillation of Streptococcus pneumoniae into the right upper lobe of rabbit lung, imaging with positron emission tomography (PET) following injection of (¹⁸F)-2-fluoro-2-deoxy-D-glucose (¹⁸FDG) provides a remarkably specific index of neutrophil metabolic activity in vivo (6). This increased signal occurred at a time when neutrophil migration had effectively ceased, and most likely reflected the respiratory-burst activity of the cells. Autoradiography with tritiated deoxyglucose [(³H)-deoxyglucose] confirmed that the signals obtained reflected activity almost exclusively of the neutrophil in both this model and also in a model of chronic inflammation following bleomycin instillation. The involvement of the neutrophil in lung scarring is unclear, but the persistent presence of these cells in the lung tissue of patients with pulmonary fibrosis suggests that they may play a key role in modulation of the fibrotic process (7). The metabolic activity of the neutrophil measured in situ with PET may give an indication of disease activity and provide a technique for monitoring the efficacy of intervention with appropriately directed therapies. We have now extended our studies to the relationship between the persistence of neutrophil metabolic activity and the development of pulmonary fibrosis in animal models following pulmonary challenge with fibrogenic and nonfibrogenic inflammatory stimuli.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Models of Localized Pulmonary Inflammation

New Zealand white rabbits (2.5 to 3.5 kg) were anesthetized with intramuscular ketamine and xylazine. Fibrogenic or nonfibrogenic substances in 0.5 to 0.8 ml of normal saline (NS) were instilled into the right upper lobe of the lungs under direct vision, via the biopsy channel of a neonatal pediatric bronchoscope (Olympus BF 3C 10; Olympus, Tokyo, Japan).

The nonfibrogenic stimuli consisted of 1,000 nM C5a (recombinant human; Sigma Chemical Company Ltd, Poole, UK), a standard dose of S. pneumoniae, or 50 mg of 5 µm amorphous silica (Hypersil; Shandon Southern Products Ltd., Runcorn, UK). The fibrogenic stimuli consisted of 0.75 mg bleomycin (Lundbeck Ltd., Milton Keynes, UK) or 50 mg of 5 µm microcrystalline silica (Sigma).

The animals remained essentially healthy throughout the time of study, as the challenge was localized to a small region of lung. Care of the animals was in accordance with United Kingdom Home Office and institutional guidelines.

Monitoring Cell Metabolism In Situ by PET Scanning

At various timed intervals following the instillation of the test substances, the rabbits were reanesthetized and the ear veins were cannulated. Each rabbit was positioned with its thorax within the field of view of the PET scanner (Siemens 931 positron scanner; Brackwell, UK).

A scan of thoracic density distribution was performed by measuring the transmission of radioactivity through the field of view of the PET scanner from a ring source containing ⁶⁸Ge. The ring source was then retracted and a bolus of 0.5 to 2 mCi ¹⁸FDG (< 100 µg) in 2 ml of NS was injected into the ear vein of each rabbit. Emitted radioactive counts were accumulated by the PET scanner for six successive 15-min time-frames. These data were corrected for attenuation with the transmission scan data, and were reconstructed to give 15 transaxial tomographic images of the distribution of radioactivity for each time-frame.

To define the time-course of response to the inflammatory challenges, PET scanning was done at intervals up to a maximum of 8 wk or until any signal had returned to baseline. For each model, at least eight animals were studied, and individual animals were scanned up to six times.

Analysis of PET Data

PET data were analyzed as described in a previous communication (6). Regions of interest (ROIs) were drawn around the right and left lung areas on the transmission (density) images, using an image-analyzing software system (10). The mean ¹⁸FDG radioactivity for these ROIs was calculated for each emission time-frame. Selection of the ROIs on the basis solely of tissue density from the transmission scan eliminated any bias in the quantification of the ¹⁸FDG emission data. All tomographic slices containing lung fields were analyzed for each study of each rabbit. The accumulated radioactive counts from each of the ROIs in the challenged right upper lobe were divided by those in the same tomographic slices of the unchallenged left upper lobe. The rate of accumulation of ¹⁸FDG at each time point after challenge was calculated by linear regression, as the slope of the right/left radioactive counts against the times at which the six time-frames of each individual PET scan were taken. These rates of accumulation were then plotted against time after challenge for each model. Curves were fitted to these data points, using a computerized, iterative, nonlinear regression technique, to yield a biexponential function (Siphar pharmacokinetics package; Simed, Creteil, France). The decay constants were compared through Student's t test.

Histology

At intervals following challenge, animals were killed and the lungs rapidly excised and inflated to a pressure of 20 cm with formol saline. After 24 h of fixation, specimens were taken from challenged and control areas and the specimen material was processed into paraffin- embedded sections by conventional methods. Serial sections (5 µm) were cut for light microscopy and autoradiography. Sections were stained with hematoxylin and eosin (H&E), reticulin stain, and elastic Van Gieson stain. All microscope slides were coded, and cell counting was done in a blind manner by an experienced histopathologist. Cells were counted in 10 consecutive high-power fields (hpf) of 0.2 mm², and cells were classified by type and position (intraalveolar or interstitial). Neutrophils were further classified as normal or pyknotic.

³H Autoradiography for Identification of Metabolically Active Cells

Deoxyglucose and fluorodeoxyglucose are both taken up by the same mechanism as is glucose, are converted to their 6-phosphate forms, and can then undergo no further metabolism, hence accumulating within metabolically active cells in proportion to their glucose uptake (11). ³H-labeled deoxyglucose was used for microautoradiography to identify the cell type responsible for the increased ¹⁸FDG uptake demonstrated by PET scanning.

Selected animals from each model received 5 mCi of tritiated 2-deoxy-D-glucose [(³H)-DG] intravenously immediately after exhibiting a positive ¹⁸FDG PET scan. Forty-five minutes later the animals were killed and the lungs excised and fixed as described earlier. Representative sections of lung were coated in autoradiographic emulsion (Ilford K2 nuclear emulsion size A; Ilford Ltd, Knutsford, UK). The sections were exposed for 6 wk, developed (Kodak D19; Hemel Hempstead, UK), fixed (AMFIX), stained with H&E, and examined light microscopically.

Inflammatory cells were counted in 10 consecutive hpf of 0.2 mm² and were classified by type, position (interstitial or intra alveolar), and grain density. Grains were also counted over areas of tissue with no neutrophils, to obtain background levels. Those cells with five or more associated grains were counted as positively labeled.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Time Course of Metabolic Response to Inflammatory Challenge

Figure 1 shows the overall time courses calculated from the PET scans done at different times after instillation of the different inflammatory challenges. The time course and intensity of the ¹⁸FDG signal varied markedly. Following instillation of C5a there was a rapid and marked increase in the uptake of ¹⁸FDG in the challenged area. The time from instillation to peak signal (T_max) was 10 h, and the rate constant for its decline was 0.045 ± 0.030 h¹. The response to Streptococcus pneumoniae was not as sharp, the peak signal not being reached until 15 h after instillation. The rate of decline of the signal was also slightly slower (k = 0.068 ± 0.012 h¹, p = 0.4 versus C5a), although return to baseline values still occurred within a few days. In contrast, it took longer to reach maximum ¹⁸FDG uptake values with the fibrogenic agents, the periods being 28 h with bleomycin and 90 h with microcrystalline silica. The ¹⁸FDG signal in these models remained elevated for some considerable time (as judged from the time at which the lower 95% confidence interval [CI] for the regression lines reached 0), the rate constants for the decline being 0.002 ± 0.001 h¹ for bleomycin (p < 0.05 versus C5a; p < 0.001 versus S. pneumoniae) and 0.0012 ± 0.0007 h¹ for microcrystalline silica (p < 0.05 versus C5a; p < 0.001 versus S. pneumoniae), following which the signal was still increased at 8 wk, when the study was terminated. No increase in ¹⁸FDG uptake was detected following instillation of the nonfibrogenic amorphous silica spheres. Figure 2 shows the fitted curves for all the challenges.

View larger version (16K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   View larger version (10K): [in this window] [in a new window]   Figure 1.   Time course of 18FDG uptake following localized instillation of C5a (A), Streptococcus pneumoniae (B), bleomycin (C ), microcrystalline silica (D), and amorphous silica (E ). Curves were fitted to the individual data points (closed circles) for each model using a computerized, iterative, nonlinear regression technique to yield a biexponential function (Siphar). The 95% confidence limits are shown as dotted lines, except for the amorphous silica challenge, in which there was no significant increase in 18FDG uptake at any time. Insets in A and D show expanded time scale. At least eight animals were studied for each model, each being scanned up to six occasions.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Composite of time courses of 18FDG uptake following localized instillation of all challenges. C5a (solid line), Streptococcus pneumoniae (long dashes), bleomycin (short dashes) microcrystalline silica (loosely dotted line), and amorphous silica (tightly dotted line).

Histology

In all the models, the microautoradiographic threshold for determining that a cell was positively labeled was five grains. This was well above the average background of 0.1 grains for cells other than neutrophils. The number of grains associated with any cell other than a neutrophil was never greater than two.

C5a

At 4 h after instillation of C5a, there was a marked increase in the number of neutrophils in the interstitium and alveolar spaces, at 22 and 70 per hpf, respectively. At the light-microscopic level it was not usually possible to differentiate between neutrophils in the interstitium and those within capillaries. In this model, most neutrophils classified as interstitial were probably within dilated capillaries. None appeared pyknotic. In contrast, about 47% of neutrophils in alveolar spaces were pyknotic. Large numbers of neutrophils were observed marginating and migrating through the arteriolar walls (Figure 3). At 8 h, when the ¹⁸FDG PET signal was near its peak, 44% of all neutrophils (110 per hpf) were shown to be activated by autoradiography of ³H-DG (Figure 4A). These metabolically active cells represented nearly all of the nonpyknotic, postmigrational neutrophils in the air spaces. Only a quarter of neutrophils in the interstitium showed increased deoxyglucose uptake. By 15 h following instillation, the majority (86% of 79 per hpf) of neutrophils in the alveolar spaces were pyknotic, but again, effectively all nonpyknotic cells were labeled. Nonpyknotic neutrophils were still present in the interstitium (22 per hpf), but showed little activation. Macrophage numbers were increased (18 per hpf). By 35 h very few neutrophils were present (a total of 10 per hpf), but macrophage numbers remained increased for more than 2 d, after which the lung tissue returned to normal, with no disruption of lung architecture.

View larger version (116K): [in this window] [in a new window]   Figure 3.   Photomicrograph of challenged lung region 4 h after instillation of C5a. Arteriolar wall shows margination of neutrophils and migration through the wall. Original magnification: ×970, H&E stain.

View larger version (128K): [in this window] [in a new window]   Figure 4.   Autoradiography of (3H)-DG distribution of challenged lung regions 8 h after C5a instillation (A), 8 h after Streptococcus pneumoniae instillation (B), 2 wk after bleomycin instillation (C ), and 3 wk after microcrystalline silica instillation (D) in which the refractile microcrystalline silica particles present in the alveoli and interstitium were mainly within macrophages. Grains developed by autoradiography represent the distribution of deoxyglucose. In all cases radioactivity is localized only to neutrophils. Left panels, original magnification: ×970; right panels, original magnification: ×2,425. Counterstained with H&E.

Streptococcus pneumoniae

There was an initial interstitial accumulation of neutrophils, with marked interstitial edema. Twenty-four percent of neutrophils (87 per hpf) were in alveolar spaces, but only 8% of neutrophils showed increased metabolic activity at 8 h following instillation (Figure 4B). By 15 h the lung tissue was consolidated, with 81% of neutrophils in alveolar spaces (164 per hpf) showing increased activation; only 15% of neutrophils in the interstitium (41 per hpf) were activated. Tissue debris, pyknotic neutrophils, and large numbers of macrophages (48 per hpf) were present. By 30 h after instillation, neutrophil numbers had fallen (25 per hpf) and 64% of these cells were pyknotic. Macrophages containing debris were present (32 per hpf). Very few neutrophils were observed from 50 h after instillation onward, but macrophage numbers remained high for about 4 d. Full resolution of the inflammatory process occurred by 1 wk, with no evidence of scarring.

Bleomycin

Consolidation as a result of interstitial infiltration of large numbers of neutrophils (94 per hpf) was apparent at 6 h, with accumulation of edema fluid. By 24 h the interstitium was expanded by mixed inflammatory cells (principally neutrophils), few of which were activated. At 3 d and 5 d, fibroblasts and macrophages (131 and 86 per hpf, respectively) became evident, with large numbers of neutrophils (99 and 35 per hpf, respectively) throughout the consolidated tissue. At 1 wk, neutrophils, macrophages, and fibroblasts were present, with evidence of reticulin deposition. Neutrophil numbers remained increased for over 2 wk with curiously few pyknotic cells (2%). By 2 wk there was disturbance of the lung architecture, with marked reticulin deposition. There was moderate interstitial infiltration, principally by neutrophils (157 per hpf), but also by some macrophages (70 per hpf). Autoradiography of (³H)-DG showed that 31% of the neutrophils were activated (Figure 4C). Neutrophils, macrophages, and debris were present in the air spaces where the latter were preserved. At 4 wk, macrophage numbers were still increased, but neutrophil numbers had fallen. No fibroblasts could be detected, but reticulin deposition remained in the grossly thickened interstitium. No mature collagen was detected by elastic Van Gieson staining.

Microcrystalline Silica

By 6 h after instillation there was a pronounced neutrophil infiltration into the alveoli, and some interstitial neutrophil accumulation. Increased numbers of macrophages (25 per hpf) were observed. By 60 h, the lung was consolidated, with silica particles in the thickened interstitium. There was a large amount of nuclear debris and a few fibroblasts evident in the interstitium, with the beginning of reticulin deposition; neutrophil numbers remained high (172 per hpf, of which 66% were pyknotic). Macrophages containing silica particles and debris were present. From 5 d after instillation, collagen deposition was demonstrated by elastic Van Gieson staining. At 8 d the lung tissue was consolidated, with marked fibroblastic activity and necrosis. Some lymphocytes and plasma cells were also present. From 15 d, aggregates of silica particles and necrotic debris were walled off by highly organized fibroblasts and collagen fibers (Figure 5). At 3 wk after instillation, autoradiography showed that 24% of the neutrophils, present in persistently elevated numbers (59 per hpf), were activated. Labeled neutrophils were mainly within alveolar spaces at the periphery of granulomatous regions (Figure 4D). At 8 wk (the final time point in this study), there was continuing fibroblast presence, with large amounts of collagen in the extracellular matrix. Neutrophil and macrophage numbers remained elevated (37 per hpf and 104 per hpf, respectively). Nineteen percent of neutrophils (52 per hpf) were activated.

View larger version (147K): [in this window] [in a new window]   Figure 5.   Photomicrograph of challenged area of lung 15 d after instillation of 5-µ particles of microcrystalline silica. The silica was found in large aggregates surrounded by a dense fibrous reaction containing neutrophils at the periphery. Upper panel, original magnification: ×600, H&E. Lower panel shows high-power view of periphery of nodule; original magnification: ×970, H&E.

Amorphous Silica

Although neutrophil numbers rose to 88 per hpf at 15 h after-instillation, very few neutrophils could be detected at later time points. Macrophages were present at 15 h (24 per hpf), and their numbers remained increased for the whole of the 3-wk time course. Macrophages, most of which contained phagocytosed silica spheres, were initially present in the alveolar spaces only. By 4 d, sphere-bearing macrophages were present in the interstitium as well as in the air spaces. A few neutrophils (10 per hpf) were also observed in the interstitium but not in air spaces. At 13 d after instillation, occasional neutrophils were observed, but the numbers of macrophages remained elevated (53 per hpf), with most of these cells carrying spheres. These sphere-bearing cells were distributed throughout the alveolar spaces and interstitium, with a marked perivascular accumulation, probably in the lymphatics. Throughout the remainder of the 3 wk of the study, the amorphous silica spheres were slowly removed from the lungs. There was no disruption of lung architecture nor any evidence of fibrosis.

Time Course of Inflammatory-cell Accumulation

¹⁸FDG uptake correlated well with total neutrophil numbers in models treated with C5a and S. pneumoniae (Figure 6, upper panels), but not in the fibrogenic models or those treated with amorphous silica (Figure 6, lower panels). There was no correlation between the number of macrophages present and the uptake of ¹⁸FDG in any of the models. Microautoradiography of (³H)-DG in the models with an increased ¹⁸FDG signal showed radioactivity localized almost entirely to neutrophils, indicating that increased deoxyglucose uptake is a remarkably specific measure of neutrophil metabolic activity.

View larger version (19K): [in this window] [in a new window]   Figure 6.   Relationship between neutrophil numbers (mean of 10 hpf of 0.2 mm2) and 18FDG uptake for each challenge. The lines on the upper two panels, C5a and S. pneumoniae, represent the lines of best fit by linear regression analysis. There is no apparent correlation between neutrophil numbers and 18FDG uptake for the other challenges.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have developed a range of well-defined models of pulmonary inflammation in the rabbit, each with characteristic profiles in terms of the cellular response and subsequent resolution or progression to fibrosis and scarring. These models are likely to be useful for the development of techniques for noninvasive imaging of the different components of the inflammatory process in the lungs, and the therapeutic modification of these components. We have previously shown, by PET imaging of intravenously injected ¹⁸FDG in vivo combined with microautoradiography of lung tissue obtained following intravenous injection of (³H)-DG, that postmigrational neutrophils are responsible for the deoxyglucose uptake signal (6). Neutrophils exhibit several types of behavior in response to different stimuli, and it was of interest to determine which phase of response we were detecting by PET scanning of ¹⁸FDG.

The increase in the ¹⁸FDG signal measured by PET does not correlate with the time-course of migration of intravenously injected ¹¹¹In-labeled white cells, the peak signal being obtained after migration has effectively ceased after challenge with either C5a (12) or S. pneumoniae (13). Although glucose uptake by migrating cells is probably increased, this increase is likely to be small relative to that due to the postmigratory metabolic burst. At present there is no model of neutrophil migration without subsequent activation, and it is therefore difficult to determine whether chemotaxis itself contributes to the PET signal. Microautoradiography shows that the majority of labeled neutrophils are in the air spaces rather than capillaries. A proportion of the (³H)-DG uptake by the few labeled cells present in the capillaries (C5a 26%, S. pneumoniae 10%) might be due to emigrating cells. The major contributor to the increased uptake of ¹⁸FDG in pulmonary inflammation is likely to be associated with respiratory-burst activity of neutrophils. Neutrophils depend on glucose as their major or sole energy source, and there is a dramatic increase in the uptake and phosphorylation of deoxyglucose when neutrophils are stimulated in vitro (14). In vivo, cells other than neutrophils presumably take up ¹⁸FDG, albeit at a rate undetectable by microautoradiography of (³H)-DG. In the amorphous silica model, in which there are no neutrophils but large numbers of macrophages present, there is no concomitant increase in ¹⁸FDG signal. Monocytes cannot increase their glucose uptake to the same extent as can neutrophils (15). Once monocytes are differentiated into macrophages, this ability is further reduced (16). Resident lung macrophages have weak respiratory-burst activity, although stimulation has been shown to induce a slight increase in superoxide anion production (17). However, in our models there is clear evidence both of a continuing presence of neutrophils some weeks after the instillation of chronic inflammatory stimuli and the exhibition by a significant proportion of these cells of increased deoxyglucose uptake, which is indicative of persistent metabolic activation.

The intensity of the ¹⁸FDG signal following challenge with C5a or S. pneumoniae in our study was proportional to the number of neutrophils present in lung-tissue sections (Figure 6, upper panel), a relationship not seen in the models of chronic inflammation leading to scarring (Figure 6, lower panel). This was most probably due to a sampling problem. The individual cells counted histologically represented only a small region of the lung, whereas the ¹⁸FDG signal was an integral of the whole region, and the neutrophils tended to be grouped around the edges of fibrotic regions (Figure 5). This resulted in a large variability between the numbers of cells counted in the 10 hpf, the SD typically being greater than 50% of the mean. Microautoradiography of (³H)-DG showed localization of silver grains only over neutrophils and never over macrophages, even when neutrophils showed more than 10 grains per cell. This confirms the observations made in our previous study, which suggested that deoxyglucose is a remarkably specific indicator of neutrophil activity in these models (6).

The brisk but transient response to C5a was probably a function of the short biologic half-life of this cytokine. C5a is directly chemotactic to neutrophils, and there is no delay in the migration of these cells into the challenged region. Activation, although lagging behind migration, peaked at 10 h. Increased labeling of neutrophils with (³H)-DG while still in blood vessels at 8 h may indicate increased hexose uptake due to migration into the air spaces, or direct activation due to C5a. Because the bulk of neutrophil influx following C5a challenge occurs within 2 h (12), the peak ¹⁸FDG uptake at 10 h must represent mainly postmigratory activity. In the absence of a continuing stimulus, it is likely that the ¹⁸FDG signal falls because neutrophils become inactive, undergo apoptosis, and are removed from the challenged region by macrophages. The slightly slower response (15 h to peak) observed after challenge with S. pneumoniae suggests that circulating neutrophils are not responding to primary bacterial signals but probably to mediators produced by the resident cells in the challenged area (18, 19). The short duration of this response, which was followed by a return to baseline within 2 d, implies that the chemotactic factors produced are short-lived, and that after their initial release no more are produced, or that factors are released which inhibit further neutrophil migration or activity. Bleomycin evoked a response that persisted for more than 2 wk. The rate of increase in ¹⁸FDG uptake (28 h to peak) implies that the recruitment and activation of neutrophils was secondary to the initial insult. It is probable that direct oxidative damage by free radicals produced by iron-bleomycin-complex formation (20) initiates a response by resident lung cells. Epithelial necrosis correlates with neutrophil accumulation at all levels of the tracheobronchial tree (21). Intracellular adhesion molecule-1 (ICAM-1), not normally present on alveolar type II cells, is expressed in response to high oxygen levels, and its distribution on type I cells is altered (22). Alveolar type II cells can produce neutrophil chemotactic factors (23), and macrophages release factors that are chemotactic to neutrophils in response to bleomycin (24). Microcrystalline silica in our model caused marked neutrophil infiltration by 15 h, followed by fibrotic changes that were obvious within a few days. However, peak ¹⁸FDG uptake did not occur until 4 d after insult, and still persisted at 8 wk, when the study was terminated. By light microscopy, the macrophages showed numerous granules. These granules contain substances which, when released, perpetuate the accumulation and activation of neutrophils (25), which would be detectable by PET.

Activation of neutrophils is associated with release of agents that cause tissue injury (26) and potentiate the migration of further neutrophils by the production of chemotactic factors through enzymatic cleavage of matrix proteins (27). In our study, following challenges that resolved completely, the ¹⁸FDG signal was either transient (C5a or S. pneumoniae) or absent (amorphous silica). The chronic inflammatory responses that resulted in scarring were consistently marked by a persistently elevated ¹⁸FDG uptake. Increased ¹⁸FDG uptake has also been shown in patients with cryptogenic fibrosing alveolitis (CFA) (28) and sarcoidosis (29). This association suggests that activated neutrophils may be involved in some component of the scarring process. Increased accumulation of neutrophils and macrophages induced by instillation of formyl methionylleucyl peptide (FMLP) into the silica-challenged lungs of mice increased the removal of silica and reduced subsequent fibrosis (30). Neutrophils can synthesize and release several enzymes capable of cleaving matrix proteins (31, 32), and hence are likely to play a major role in the remodeling of extracellular-matrix tissue, either directly or through macrophages which, while incapable themselves of synthesizing neutrophil elastase, are capable of its internalization and secretion, in addition to secreting their own battery of enzymes (33). Appropriate regulation of these enzymes facilitates cell migration and tissue remodeling and repair. Inadequate control has been implicated in the development of fibrosis and emphysema (31). Secretory leukoprotease inhibitor (SLPI) is a major endogenous inhibitor of neutrophil elastase activity, and following bleomycin challenge in hamsters, the administration of a synthetic, truncated SLPI reduces subsequent fibrosis (34). This may lead to its therapeutic use in the presence of increased neutrophil burden in patients with chronic pulmonary inflammation (34).

It is apparent from the studies described here that a persistence of the ¹⁸FDG signal measured by PET is associated with scarring of lung tissue. Only neutrophils in tissue sections were found by autoradiography to take up significant (³H)-DG. This continuing neutrophil activation may be a reflection of tissue damage, and is not necessarily its cause. In any event, the ability to repeatedly and noninvasively monitor this indicator of neutrophil activation will enable the evaluation of different therapeutic strategies in patients with lung disease.