INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bleomycin-induced fibrosis has been extensively used as a model to study the pathophysiology of lung fibrosis and the role of the collagen-elastin-proteoglycan matrix in this disease process (1). The lung injury in this model is progressive, characterized by an initial alveolitis with edema and recruitment of inflammatory cells into the air spaces, followed by fibroblast proliferation during the second week after exposure. After 20 d, lesions are focal, consisting of scars of connective tissue matrix with few inflammatory cells (4). In the early phase of bleomycin-induced lung injury, there is accumulation of hyaluronan in the alveolar space and interstitial compartment (5). Collagen synthesis reaches its peak during the second week and then gradually subsides (3). Other components of the extracellular matrix that are increased include fibronectin, laminin, biglycan, and elastin (2, 4, 6).

The mechanical changes that accompany these structural changes have been incompletely characterized. Although changes in the quasi-static mechanical properties have been measured (1, 7, 8), relatively little is known about changes in the dynamic mechanical properties. The dynamic properties refer not only to the flow-resistive behavior of the airways, but also to the viscoelastic or resistive behavior of the lung parenchymal tissues (9). Indeed, in many species, tissue resistance (Rti) accounts for a major proportion of overall lung resistance (9). A major structural element responsible for tissue resistance is the extracellular matrix, i.e., the stress-bearing collagen and elastic fibers and the proteoglycans, glycosaminoglycans, and glycoproteins that constitute the surrounding "ground substance" (12, 13). The mechanical properties of the air-liquid interface and peripheral contractile elements likely also contribute (12, 14). Insofar as the matrix is so profoundly altered in lung fibrosis, one would predict important changes in the dynamic mechanical behavior. Moreover, Horiuchi and colleagues (8) have recently shown changes in the physical properties of pulmonary surfactant in the bleomycin model. Several years ago, Bachofen and Scherrer (15) measured tissue resistance indirectly in human subjects with pulmonary fibrosis and showed that Rti was increased as compared with normal control subjects. More recently, Verbeken and colleagues (16) have studied autopsy lungs from patients with pulmonary fibrosis, measuring Rti with alveolar capsules to directly measure alveolar pressure. They also showed that specific tissue resistance was increased. Moreover, they found a positive correlation between Rti and their index of alveolar wall thickness, and a negative correlation between Rti and the size of air spaces. This structure-function correlation is in contrast to the results of Goldstein and colleagues (7) who were unable to document a relationship between changes in the static mechanical properties of bleomycin-induced fibrotic hamster lungs and changes in the composition of the lung connective tissues.

We were interested in defining changes in the viscoelastic properties of the lung parenchyma in the bleomycin model, examining changes both in vivo and in vitro. The advantage of making measurements in an in vitro model is that contributions to the mechanical behavior related to surface film, airway closure, and ventilation heterogeneities can be excluded (17). Therefore we made measurements of tissue resistance in intact rats previously exposed to bleomycin. After in vivo measurements were completed, we excised subpleural parenchymal strips and studied oscillatory mechanics in the organ bath, measuring resistance and elastance during sinusoidal oscillation of the tissues. After physiologic measurements we fixed the parenchymal strips and, using histochemical staining, made measurements of collagen and elastin in order to determine whether changes in lung function were correlated with changes in lung structure.

We were also interested in the contractile response of the bleomycin-treated lungs. An increase in myofibroblast content has been demonstrated in this model (18). Myofibroblasts have been suggested as the key cell driving the increased secretion of extracellular matrix proteins (21). Moreover, Evans and colleagues (18, 22) have shown that increases in isometric tension after contractile stimulation are enhanced in parenchymal strips from bleomycin-treated animals. We questioned whether there would also be increases in the dynamic indices after induced contraction, and whether this augmented response would be correlated with the amount of smooth-muscle-specific actin present in the tissue. Therefore, we stimulated parenchymal strips with acetylcholine and quantitated -smooth muscle actin-positive interstitial cells (AIC) in fixed tissue using conventional morphometry.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In Vivo Experiment

Animal preparation and protocol. Eighteen Fischer rats were anaesthetized intraperitoneally with pentobarbital sodium (50 mg/kg) and intubated with a tracheal cannula (ID = 2 mm, length = 5 cm). Twelve rats received 1.5 U of bleomycin sulfate intratracheally. Control rats (n = 6) received saline. Four weeks later animals were anaesthetized intraperitoneally with pentobarbital sodium (50 mg/kg) and tracheostomized. A metal cannula (ID = 2 mm) was inserted into the trachea. The animals were mechanically ventilated (Model 683; Harvard Apparatus, South Natick, MA) at a tidal volume (VT) of 8 ml/kg, a frequency of 1 Hz, and a PEEP of 4 cm H₂O. The thorax was opened by means of a midline sternotomy. A heating pad was used to maintain body temperature.

Two alveolar capsules were glued to the pleural surface with cyanocrylate. The pleura was punctured with an electrocautery needle through the central port of the capsule to bring the underlying alveoli into communication with the capsule chamber. A piezoresistive microtransducer (Endveco 8510; Endveco, San Juan Capistrano, CA) was placed in the port of the capsule to measure alveolar pressure (PA). Tracheal pressure (Ptr) was measured by a piezoresistive microtransducer (Endevco 8510B-2) placed in the lateral port of the tracheal cannula and tracheal flow () was measured with a Fleisch pneumotachograph (No. 00; Instrumentation Associates, New York, NY). Volume (V) was calculated by digital integration of the flow signal. The signals were amplified, filtered at a cutoff frequency of 100 Hz, converted by a 12-bit analog-digital converter (DT2801-A; Data Translation Inc., Marlborough, MA), sampled at a rate of 200 Hz, and stored on an AT compatible computer.

After two deep inflations (peak pressure of 30 cm H₂O Ptp), the lungs were allowed to stabilize for 10 min, and measurements of 20-s duration were sampled twice.

Calculation of lung mechanics. The tracheal pressure was corrected for both the tube resistance and the Bernoulli effect. Total lung resistance (RL) and lung elastance (EL) were calculated by fitting the equation of motion to changes in Ptr:

(1)

where K is a constant term reflecting PEEP and the error linked to the residuals of least squares adjustment method.

Rti was calculated by adjusting the equation of motion to PA:

(2)

Airway resistance (Raw) was calculated by subtraction:

(3)

where Rti was the average of the values obtained from two capsules.

In Vitro Experiment

Tissue preparation. Immediately after the physiologic measurements were completed, the animals were killed by exsanguination and the lungs and heart were dissected from the thorax. The left lung was filled via the main bronchus with 10% buffered formalin. The right lung was filled via the main bronchus with Krebs solution (mM: NaCl, 118; KCl, 4.5; NaHCO₃, 25.5; CaCl₂, 2.5; MgSO₄, 1.2; KH₂PO₄, 1.2; glucose, 10) (Sigma, St. Louis, MO). Two or three subpleural parenchymal strips (10 × 2 × 2 mm) from each rat were placed in iced Krebs solution, continuously bubbled with 95% O₂/5% CO₂ at pH 7.40. Resting length (L_r) and wet weight (W_o) of each strip were recorded.

Experimental apparatus. Metal clips were glued to either end of the tissue strip with cyanoacrylate. Steel music wires (diameter, 0.5 mm) were attached to the clips and the strip suspended vertically in an organ bath. A mercury bead was placed in the bottom of the organ bath, allowing the music wire to pass through the bath but preventing the Krebs solution from leaking out. The bath was filled with 15 ml Krebs solution, maintained at 37° C, and continuously bubbled with 95% O₂/5% CO₂. One end of the strip was attached to a force transducer (Model 400A; Cambridge Technologies, Watertown, MA) that had an operating range of ± 10 g, resolution of ± 200 mg, and compliance of 1 mm g, whereas the other end was connected to a servo-controlled lever arm (Model 300B; Cambridge Technologies). The lever arm was capable of peak to peak length excursions of 8 mm, and length resolution of 1 mm and was in turn connected to a function generator (Model 3030; BK Precision, Chicago, IL) that controlled the frequency, amplitude, and waveform of the oscillation. The resting tension (T) was set by movement of a screw thumb wheel system that effected slow vertical displacements of the force transducer. Length and force signals were low-pass filtered (8-pole Bessel 902LPF; Frequency Devices, Haverhill, MA) with a corner frequency of 30 Hz, converted from analog to digital with an analog to digital converter (DT2801-A; Data Translation Inc., Marlborough, MA) and recorded on an A/T compatible computer.

The linearity and hysteresis of the system were tested by measuring the moduli of a steel spring of stiffness comparable with that of the tissue strip. The spring was suspended in the bath by music wire in the same manner as the strip. The frequency and amplitude-dependence of the system were assessed over a range of frequencies (0.1 to 10 Hz). The spring stiffness did not show any dependence upon oscillatory frequency below 5 Hz. The hysteresivity of the system was independent of frequency and had a value < 0.003.

Protocol. Strips were preconditioned by slowly cycling tension from 0 to 2 g three times. On the third cycle the strip was unloaded to a tension of 1 g, and sinusoidal length oscillations of 2.5% L_r at a frequency of 1 Hz were applied. After 60 min of stress relaxation, the final resting tension was approximately 0.7 g. Baseline recordings were obtained and then, acetylcholine (ACh, 10³ M) (BDH Inc., Poole, UK) was added to the organ bath. Measurements of length and T were collected continuously for an additional 15 min.

Measurement of strip mechanics. Elastance (E) and resistance (R) were estimated by applying the recursive least squares algorithm to the equation of motion (23).

(4)

where l = length, l/t is the length change per unit time, and K is a constant reflecting resting tension. Results were standardized for strip size. The unstressed cross-sectional area (A_o) of the strip was obtained from the formula:

(5)

where r is the mass density of the tissue taken as 1.06 g/cm³, W_o is the wet weight in grams, and L_r is the unloaded length in cm. Values of E and R were multiplied by L_r/A_o. Hysteresivity, , a dimensionless variable coupling the dissipative and elastic behavior, was calculated by the equation:

(6)

where f is frequency (12).

Morphometric study. Sagital slices from the left lung were embedded in paraffin and 5-µm-thick sections were stained with H&E and Sirius Red for routine morphologic observation.

After physiologic study, strips were fixed with 4% paraformaldehyde and embedded in paraffin for morphometric and immunohistochemical study. Five micrometer-thick sections were stained with Sirius Red and Weigert's Resorcin-fuchsin (24, 25) for collagen and elastic fibers, respectively. Collagen fibers were identified in Sirius-Red-stained tissue using polarized light. Immunohistochemical staining using the APAAP method was performed using a monoclonal antibody to -smooth muscle actin (DAKO, Mississauga, ON, Canada). Sections were deparaffinized, hydrated, and incubated in 2% normal rat serum (NRS) for 1 h at room temperature. Sections were then rinsed with TBS (0.5 M TRIS; pH, 7.6; 1.5 M NaCl) and incubated with antiactin (1:400 in TBS) overnight at 4° C. After washing with TBS, the tissue was incubated with an unconjugated rabbit antibody against mouse immunoglobulin (DAKO) (1:30 in 20% NRS) for 30 min, washed again, and incubated with APAAP (soluable complexes of calf alkaline phosphatase and murine monoclonal antibody to alkaline phosphatase; DAKO) (1:30 in 20% NRS) for 30 min. After further washing, sections were developed with Fast Red salt (Sigma) (1 mg/ml in alkaline phosphatase substrate) for 10 min at room temperature. Sections were counterstained with Harris Haematoxylin for 1 min. Negative controls were made by exclusion of the primary antibody.

A semiquantitative analysis was performed by applying point-counting in one strip per animal. Using a 121-point grid, we calculated the volume proportion of collagen, elastic fibers, and actin in airways, vessels, and parenchyma as the relation between the number of points falling on stained and nonstained tissue. Measurements were performed in 20 fields (actin) or 10 fields (collagen and elastic fibers) per slide, using a magnification ×200. Using the same method, we also measured the fractional area of tissue constituents. Fractional areas were measured for bronchial wall (BW), blood vessel wall (BVW), and alveolar wall (AW). BW and BVW were counted when the point fell on the smooth muscle, the epithelial layer, the endothelial layer, or its associated connective tissue. Points falling on airway lumen and blood vessel lumen were excluded.

Data Analysis

Unpaired two-tailed t test was used to compare the different baseline mechanical parameters in vivo and in vitro in bleomycin versus control rats. Paired two-tailed t test was used to compare the strips before and after ACh challenge. Unpaired two-tailed t test was used to compare the percent increase in T, E, R, and , and the volume proportion of collagen, elastic fibers, and actin in the parenchyma, vessels, and airways in the two groups of strips (bleomycin and control rats). Finally, a matrix of correlation (Pearson's correlation) was performed to identify correlations between the in vitro mechanical parameters and the volume proportion of collagen, elastic fibers, and actin in the parenchymal strips. Results were considered statistically significant at a probability level of 5%. Values are reported as means ± standard error.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mean baseline values of the different mechanical parameters in bleomycin and control rats in vivo and in vitro, respectively, are shown in Table 1 and 2. In intact animals RL, Raw, Rti, and EL were all significantly increased in bleomycin-treated rats as compared with control rats (p < 0.05). In parenchymal strips, R, E, and  were similarly increased in lung tissue from bleomycin-treated rats (p < 0.0001). The percent increase in T, R, E, and  after ACh treatment is shown in Figure 1. There were significant increases in these parameters in bleomycin and control strips (p < 0.001), all of which were significantly greater in the bleomycin strips (p < 0.01).

View larger version (40K): [in this window] [in a new window]   Figure 1.   Percent increase in tension, resistance, elastance, and hysteresivity () after treatment with acetylcholine in parenchymal strips from bleomycin-treated and control rats. *p < 0.05 as compared with baseline; op < 0.05 as compared with control.

Morphologically, bleomycin-treated lungs showed patchy, diffuse distribution of interstitial fibrosis. There was also a discrete amount of inflammatory interstitial infiltrate consisting, primarily, of mononuclear cells. The severity of the lesions varied from one region to another and among rats. In some animals, there was a tendency toward bronchocentric distribution of fibrosis (Figure 2D). Lung sections from control animals showed no evidence of fibrosis or infection; alveolar architecture and bronchial walls were intact (Figures 2A, 2C, 2E, and 2G). With specific staining of parenchymal strips we observed increased amounts of collagen and elastic fibers, which were deposited in a disoriented fashion in areas of fibrosis (Figures 2B and 2H). As well, an increased amount of AIC was observed (Figure 2F).

View larger version (103K): [in this window] [in a new window]   Figure 2.   Photomicrographs of parenchymal strips. (A) Control strip stained with Sirus Red with polarization for collagen. Collagen fibers are seen in bright orange within alveolar ducts and alveolar walls (arrows). (B) Bleomycin-treated strip stained with Sirus Red with polarization. Note large area of interstitial fibrosis (F). (C ) Control strip stained with Sirus Red with polarization showing normal vessel and airway with stained collagen fibers in adventitial layer (arrows). (D) Bleomycin-treated strips stained with Sirus Red with polarization. Note intense peribronchiolar fibrosis (F). Arrow indicates airway. (E) Immunostaining for -smooth muscle actin in control strip. AIC are seen in alveolar ducts (arrows). (F ) Immunostaining for -smooth muscle actin in bleomycin-treated strip. Increased number of AIC are seen within area of fibrosis (arrows). (G) Control strip stained with Weigert's Resourcin-fuchsin. Delicate elastic fibers are seen in black within alveolar walls (arrows). (H ) Bleomycin-treated strip stained with Weigert's Resourcin-fuchsin. Note increased amount of elastic fibers (arrow) within fibrotic area. Scale bars = 50 µm.

The morphometric analysis of the fractional area of tissue constituents within the lung strips showed that alveolar tissue comprised 82 ± 2%, small blood vessel, 9 ± 1%, and small airways, 7 ± 1%, of the tissue. These data indicate that the lung strips represented subpleural parenchyma. The volume proportion of the different ECM elements in the alveolar wall in control and bleomycin-treated lung strips is shown in Figure 3. There was a significant increase in the volume proportion of collagen, elastin, and actin in the parenchyma of bleomycin versus control rat lungs (p < 0.001).

View larger version (25K): [in this window] [in a new window]   Figure 3.   Volume proportion of collagen, elastin, and actin in parenchymal strips from bleomycin-treated and control rats. *p < 0.05 as compared with control.

A correlation matrix was performed separately on data from control and bleomycin strips. The correlation coefficients between mechanical parameters and morphometric indices in the bleomycin strips are presented in Table 3. The only significant correlations were between baseline R and E and volume proportion of collagen in bleomycin-treated lungs (r = 0.93 and 0.88; p < 0.001 and < 0.01, respectively) (Figure 4).

View larger version (12K): [in this window] [in a new window]   Figure 4.   Correlations between volume proportion of collagen and (A) resistance, (B) elastance in lung parenchymal strips. Circles indicate control lungs; diamonds indicate bleomycin-treated lungs. Linear regression for bleomycin datapoints only is shown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The collagen-elastin-proteoglycan matrix is believed to play a major role in determining the viscoelastic behavior of the parenchymal tissues and is potentially responsible for tissue resistance (12, 13). In this study we used the bleomycin rat model to study parenchymal oscillatory mechanics in vivo and in vitro, and to compare the latter to morphometric changes in the extracellular matrix. After 4 wk of bleomycin treatment, we observed diffuse distribution of patchy interstitial fibrosis, with variable severity of injury among rats. This observation is in accordance with other studies that have demonstrated interstitial fibrosis and increase in collagen deposition in animal lungs after bleomycin (2, 3, 7).

As expected, there were marked changes in the dynamic mechanical properties of the lungs both in vivo and in vitro. In intact animals, RL, Raw, Rti, and EL were all increased as compared with control animals. Making measurements of dynamic mechanical behavior is important as resistance measurements reflect the energy cost of breathing. Moreover, the mechanisms that give rise to changes in dynamic mechanical parameters may be different from those that contribute to changes in quasi-static behavior (12). The increase in Rti in this study is consistent with data previously reported by Bachofen and Scherrer (15) who made measurements of Rti in human subjects with lung fibrosis. In their study, they measured total lung resistance using an esophageal balloon and subtracted airway resistance measured plethysmographically. Such an approach presents some problems in terms of Raw being measured at nonphysiologic breathing frequencies, and the importance of frequency to the contribution of Rti to RL (26). Nonetheless, Bachofen and Scherrer (15) reported significant increases in Rti in these subjects. More recently, Verbeken and colleagues (16) made direct measurements of Rti during forced oscillations, applying alveolar capsules to autopsy lung specimens of patients with pulmonary fibrosis. They also found that Rti was increased. In contrast, indices of airway resistance were not generally changed in these two studies (although Verbeken and colleagues did describe an increase in peripheral airway impedance in a subgroup of patients with fibrosis and air-space enlargement). The discrepancy between the results of studies in humans and the current study in rats may relate to the precise nature of the injury induced. We observed substantial peribronchial fibrosis likely related to the mode of bleomycin delivery, which may have contributed to the increase in Raw (Figure 2D). Indeed, this raises the issue of the pertinence of the bleomycin model to human disease. As stated above, bleomycin instillation results in a rather heterogeneous response; moreover, the progression of the injury is relatively rapid, more closely resembling the time frame of Hamman Rich syndrome or fibrosis after ARDS. Nonetheless, in affected areas, the nature of the injury is quite similar to human disease (1). Moreover, the patchy distribution of disease was a major rationale to repeat physiologic and histologic measurements in isolated strips.

Similar to mechanical changes in vivo, we also observed changes in the dynamic mechanical parameters measured in vitro. R, E, and  were all increased in parenchymal strips from bleomycin-treated animals. Making measurements in isolated parenchymal strips confers certain advantages. In intact lungs, surfactant and airway heterogeneities may contribute in an important way to measurements of parenchymal mechanics (27, 28). Indeed, Horiuchi and colleagues (8) have recently shown in the bleomycin rat model that BAL fluid, isolated surfactant, and organic solvent lipid extracts of surfactant all demonstrated elevated minimal surface tension. Hence, making measurements of dynamic mechanics in a fluid-filled preparation is especially pertinent. One question that arises, however, in a model characterized by heterogeneous distribution of injury, is whether the strips sampled were, in fact, abnormal. The degree of change in the in vitro mechanics were similar to that measured in vivo. Moreover, we performed morphometric evaluation on these strips, in part, to address this precise issue.

Morphometric study of the same strips evaluated physiologically showed that the volume proportion of collagen and elastin was increased in bleomycin-treated lungs. Although increase in collagen has been well documented in this model (2, 3, 7), few studies have focused on changes in elastic fibers. Starcher and colleagues (2) demonstrated a 2.5-fold increase in elastin content in hamster lungs 30 d after bleomycin injection. In our study we observed a 1.7-fold increase in the volume proportion of both elastic and collagen fibers in bleomycin-treated rats.

In addition to measuring changes in viscoelastic behavior under baseline conditions, we also made measurements after induced constriction. Previous investigators have shown that isometric responses to histamine, acetylcholine, and epinephrine are enhanced in parenchymal strips from bleomycin- injured lungs (18, 22). Our study demonstrated increases in the dynamic mechanical response to agonist-induced constriction. The cell responsible for this enhanced response may be the myofibroblast. In regions of fibrosis, cells that share many morphologic characteristics of both fibroblasts and smooth muscle cells are prominent (19). These cells are very similar to the myofibroblast described as the key cell in the general processes of wound healing and tissue remodeling (29, 30). An increase in myofibroblast content has been demonstrated in bleomycin-induced fibrotic lungs (19); this cell has also been postulated to play a key role in collagen secretion (21). In the current study, in strips from bleomycin-treated rats, there was a 2.2-fold increase in the volume proportion of AIC. This positivity was observed in alveolar ducts, alveolar septa, and within areas of fibrosis.

We examined potential correlations between the mechanical and morphometric parameters to investigate the role of the ECM components in determining parenchymal mechanics. In parenchymal strips from control lungs, there were no correlations between the mechanical parameters R, E, and  and collagen or elastin. These findings agree with preliminary data reported from this laboratory in rats of different ages (31). However, in parenchymal strips from bleomycin-treated rats, baseline values of R and E were very significantly correlated with volume proportion of collagen in the parenchyma. These results suggest that, in this model, the increase in collagen fibers is important in determining parenchymal mechanics. These data are in contradistinction to those described by Goldstein and colleagues (7) who could find no correlation between changes in quasi-static mechanics and total protein, collagen, and elastin content in bleomycin-exposed hamster lungs. Verbeken and colleagues (16), in their study that correlated oscillatory mechanics with lung morphometry in autopsy lungs, were able to show a positive correlation between tissue resistance and alveolar wall thickness. These results suggest that changes in structure are more closely correlated with changes in the dynamic mechanical indices.

We were unable to demonstrate correlations between agonist-induced increases in R, E, and  and the amount of AIC. Previous investigators have documented enhanced isometric responses in parenchymal strips from bleomycin-exposed animals as well as increases in -smooth muscle actin-positive cells (18, 22). Although these investigators hypothesized that these cells were responsible for the augmented contractile response, no direct correlation has previously been shown. In a previous study in human parenchymal tissue obtained at the time of surgical cancer resection, we were also unable to document a correlation between contractile responses and the amount of AIC positive cells (32). We postulated that the lack of correlation was due to the inability of this specific monoclonal antibody to sample all "contractile interstitial cells" resident in the alveolar wall (33). However, it has been shown in bleomycin-induced fibrosis, that staining with this antibody within the alveolar wall and in areas of fibrosis is greatly enhanced (19, 20). Therefore, it is more difficult to reconcile the lack of correlation between AIC and response. Perhaps the contractile response depends more on the effects of constriction-induced changes in alveolar geometry on parenchymal mechanics (34) than on the absolute amount of "contractile cells" present.

In summary, we have shown that structural modification of the extracellular matrix induced by bleomycin caused changes in the oscillatory mechanics of the lung tissue both in vivo and in vitro. We have also shown that the response of rat subpleural parenchymal strips to contractile stimulus is increased after bleomycin treatment. Finally, the increase in collagen content induced by bleomycin exposure is correlated with changes in the viscoelastic behavior of the pulmonary parenchyma. Other components of the extracellular matrix such as proteoglycans and glycosaminoglycans are known to be altered in this model (5, 6). Their contribution to changes in dynamic parenchymal mechanics also warrants investigation.