Hayes1975.pdf

The

Journal of Pathology Vol. 117 No. 1 T H E PATHOLOGY OF ELASTASE-INDUCED PANACINAR EMPHYSEMA IN HAMSTERS J. A. HAYES,AGNESKORTHY AND GORDON L. SNIDER Mallory Institute of Pathology, Boston City Hospital, Boston, Mass. 02118 Department of Pathology and Medicine, Boston University School of Medicine, Boston, Mass. 02118 Veterans Administration Hospital, Boston, Mass. 02130

PLATES I-VI HOMOZYGOUS alpha I antitrypsin deficiency is known to be associated with the premature onset of severe panacinar-type emphysema which is widespread throughout the lungs (Eriksson, 1965; Orell and Mazodier, 1972), although probably more severe in the lower lobes (Greenberg et al., 1973). Gross et al. (1964, 1965) developed an experimental model of lung damage simulating emphysema by the intratracheal instillation of papain, a finding subsequently confirmed by others (Goldring, Greenberg and Ratner, 1968; Johanson, Pierce and Reynolds, 1971). Using different commercial types of papain we have found that there is a marked variation in the severity of emphysema induced by the different products, the severity being proportional to the ability of each type of papain to induce dissolution of elastin from bovine ligamentum nuchae (Snider et al., 1974). This finding suggested that emphysema might be produced experimentally by exposing the lung to porcine pancreatic elastase, an enzyme of well-defined chemical composition (Shotten and Hartley, 1973), which acts predominantly, although not exclusively, on elastin (Mandl, 1961). Emphysematous change has recently been demonstrated in hamsters by intratracheal instillation of pancreatic elastase (Kaplan, Kuhn and Pierce, 1973). These workers found that the emphysema-producing effect of elastase was abolished by administering elastase mixed with normal serum (PiMM), but not when mixed with serum deficient in alpha1 antitrypsin (PiZZ). The present paper confirms that the intratracheal injection of elastase produces a lesion with a striking resemblance to panacinar type emphysema Received 17 Sept. 1974; Accepted 11 Oct. 1974 Correspondence to J. A. Hayes, Mallory Institute of Pathology, Boston City Hospital, Boston, Massachusetts 021 18. J. PATH.-VOL.

117 (1975)

1

A

2

J. A. HAYES, AGNES KORTHY AND GORDON L. SNIDER

(Ciba Symposium, 1959) and describes the pathology of the lesion, its relation to the amount of elastase administered, its reproducibility and evolution to a stable “ repaired ” state. MATERIALSAND METHODS All experiments were carried out on male golden hamsters (Cricetufusgriseus) weighing between 90 and 130 g. Transoral intratracheal cannulation was performed using a fine metal cannula under light anaesthesia produced by intraperitoneal injection with sodium methohexital solution (Brevital SodiumR, Eli Lillie, Indianapolis), dosage 0.1 mg per 100 g body weight. Experiment I . Relation of elastase dosage to severity of emphysema and mortality Hamsters were given intratracheal injections of crystalline, porcine pancreatic elastase (pancreatico-peptidase (E.C.3.4.4.7) Whatman) with a specific activity of 26.8 units per mg.

TABLE I Deaths occurring after injection of different concentrations of elastase Elastase dosage-mg/lOO

g

body weight Total

0.05

0.1

0.2

0.3

0.4

0.5

-----___-

Number of animals injected

*

10

16

120

5

7

Deaths

0

0

2

1

4

Percentage of deaths

0

0

2

20

57

27 56 48

I j

34 214*

16

The numbers include animals injected in other studies and not reported in this paper.

The dose of elastase varied between 0.05 mg to 0.5 mg per 100 g body weight and was administered as a solution in physiological saline at a volume of 0.5 ml per 100 g body weight. Two control groups were used. The &st was given a comparable volume of intratracheal saline; the second group was anaesthetised and the trachea cannulated, but no intratracheal injection was made. All the animals were killed 16 days after elastase injection by an overdose of pentobarbitol administered intraperitoneally. The chest was opened and the aorta and venae cavae ligated, divided and the lungs removed. The trachea was intubated and the lungs distended with 10 per cent. neutral, phosphate-buffered formalin or paraformaldehyde-gluteraldehyde (Karnovsky, 1967) at a pressure of 25 cm for 24 h. Sagittal sections from parahilar and lateral regions of both lungs were embedded in paraffin. Histological sections were prepared at 5 pm, 10 pm and 40 p m thickness and stained with hematoxylin and eosin, Foot’s reticulin, Verhoeff‘s elastic with van Gieson’s stain, and the Prussian-blue reaction for hemosiderin. All sections were examined for parenchymal abnormalities, particularly for emphysema, which was considered to be present when there was a significant increase in alveolar diameter, usually accompanied by alveolar wall thinning, diminished prominence of alveolar septa and abnormality of the elastic fibers on Verhoeff’s stain. These subjective judgments were substantiated by measurement of the mean linear intercept (Dunnill, 1962), internal surface area (Weibel, 1962), and mean alveolar density (Weibel), assessed in 20 randomly selected fields on each section. Measurements were made on a representative number of animals from each group (table I). In each animal measurements were made on two slides, one from the lateral, the other from the medial (parahilar)

ELASTASE-IND UCED EMPHYSEMA

3

aspect of the left lung, and were made without knowledge of the prior subjective microscopical assessment. The mean value and standard error was calculated for each group, and the values compared with those from the controls. Experiment 2. Early stages of elastase damage and repair Hamsters were injected with 0.2 mg elastase per 100 g body weight, which was enough to produce a lesion of moderate severity with minimal mortality, as indicated by the results from Experiment 1. The exposed hamsters were killed at 4 and 8 hr, 1, 2,4, 8, and 16 days after injection. The methods of anaesthesia, injection, preparation and morphometry were identical with those described above. Experiment 3. Late evolution of elastase damage Hamsters were injected intratracheally with 0.2 mg elastase per 100 g body weight and were killed 16, 45 and 90 days later. Preparation and morphometry were carried out as described previously. In all experiments unexposed and saline-exposed hamsters were used as controls.

Elastase dose, mg/lOOg body weight

FIG.1.-Mortality

related to dose of elastase administered.

RESULTS Experiment I . Relation of elastase dosage to mortality and lesion severity There was an overall mortality of 15.4 per cent. irrespective of elastase dosage. Mortality was clearly related to dosage because below 0.3 mg elastase per 100 g body weight there were only two deaths, whereas at the 0.5 mg dose

4

J. A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

mortality approached 50 per cent. (fig. 1). What this does not show is that as our experience with intratracheal injection increased, the mortality fell, even with higher doses. In particular, the points plotted for the 0.3 mg and 0.4 mg doses represent only small numbers of animals in an early experiment (table I) in which the mortality for the 0.5 mg level was approximately 80 per cent. With repeated use of the 0.5 mg dose, however, the mortality dropped to about 50 per cent., so the LD,, is probably a dose greater than 0.5 mg per 100 g body TABLE I1 Mean linear intercept values for control and elastase-exposed hamsters Mean linear intercept Dose of elastase/ 100 g body wt.

Number of animals

“t” test

Group mean (cm)

~

Significance(P)

Standard error

}

Unexposed control

6

0.00594

04002

Saline control

6

0.00591

O~OoO1

Combined controls

12

040593

0.0003

Elastase 0.05 mg

10

040730

04003

3.231t

0.10 mg

6

0.00996

O.Oo08

4.729

0.20 mg

6

0.01386

0.0006

11.80

P<0.001

0.30 mg

4

0.01273

00094

13.61

P<0.001

0.40 mg

3

0.01340

0.0014

5.23

P < 0.001

0.50 mg

15

0.01447

0.0009

9.01

P
* t

NS

0.1439*

...

...

0.01>P>OOol

P<

om1

Unexposed controls compared with saline-injected controls. Elastase-treatedcompared with combined control group.

weight, using Whatman purfied elastase. No further attempt was made to define the mortality for the 0.3 mg and 0.4mg levels because the results show that a dosage of 0.2 mg produces negligible mortality, yet yields a consistent widespread and easily discernible lesion. No subjective nor objective difference could be detected between the unexposed and saline-exposed animals (table I]), so that the measurements from these two groups were combined as a single control group for statistical analysis. Strikingly, on opening the pleural cavity, the majority of the lungs in elastasetreated animals showed increased volume and failed to collapse as does the normal hamster lung. This was mostly seen with the high dose of elastase. Subpleural bullae were often visible to the naked eye. Microscopically there was alveolar enlargement with thinning of the wall and increased shallowness of individual alevoli. An intriguing feature was the widespread uniformity of the

5

ELASTASE-IND UCED EMPHYSEMA

alveolar change which was invariably present at the 0.5 mg level, but was also usually seen with the 0.2 mg/100 g dose level (fig. 2).

'

0

4

-2

*3

*4

.5

Elastase dose, mg/lOOg body weight

FIG.3.-Alveolar density in relation to dose of injected elastase. Each dot represents observations on a single hamster. Mean values for different dose levels are joined by the continuous line. o Mean value with standard error for control animals. f A

U

n I

-0100

*

.-

L

c e

a0050

1 0.05 FIG.4.-Mean

1

0.2 0.3 04 Elastase dose, mg/l Wg body weight

01

05

linear intercept at differing elastase doses represented by dark line drawn through points. Interrupted line represents mean control value.

These changes correlated closely with the measurements of mean linear intercept (M.L.I.) which showed larger values with increasing dosage (table 11) and all elastase dose levels showed a highly significant increase in M.L.I. when compared with the control values. Coincident with increased alveolar size, there was a prominent diminution of the mean alveolar count per unit area (fig. 3). Plotting the values for M.L.I. at different dose levels gave a curve

J. A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

6

showing that a dose of 0.2 mg produced major damage, larger doses producing only marginally more severe damage (fig. 4). The internal surface area was calculated using the mean linear intercept (Tomkieff, 1945) and the volume of air in the lungs at a transpulmonary hydrostatic pressure of 25 cm as a measure of total lung capacity. The mean internal surface area with standard error in 15 untreated hamsters was 0.34 (k0.017 m2) compared with a value of 0.25 (&0.008 m2) in eight hamsters treated with 0.2 mg elastase per 100 g body weight ( t = 4.39, P
Mode of death

4 hr

'

8 hr

1 day 25

Spontaneous

1

2 days

4 days

6 days

3

...

Barbiturate overdosage Total

8

8 days

16 days

_ _ _ ~ _ _ _ _ _ _ _ _ _

1

2

...

...

...

5

5

5

5

1

Total

30

63

being shown in table 111. In hamsters dying spontaneously, the majority of deaths occurred within 3 days of injection, most being within 24 hr. The lungs of these animals were invariably heavy and red due to diffuse haemorrhage. Microscopically there was patchy intra-alveolar haemorrhage which was confluent in many instances. Often haemorrhage was seen in the adventitial tissues of pulmonary vessels and bronchi. Perivascular oedema was invariably present even in areas where haemorrhage was not conspicuous (fig. 5), but intra-alveolar oedema was only seen focally. This haemorrhagic phase with increased lung weight was transitory and disappeared by the 4th day after injection, as shown in fig. 6. Lungs from hamsters dying spontaneously showed more severe changes than those killed, but the process was essentially the same in both groups. By 4 hr the alveolar walls showed a mild infiltration of polymorphonuclear leucocytes (PML) which increased to a maximum by 1 day. In hamsters dying spontaneously there were focal areas showing alveolar wall necrosis and dense PML infiltrate (fig. 7). The PML infiltrate decreased rapidly in the lungs of the surviving animals and had disappeared by the 2nd day. However, it must be emphasised that a PML infiltrate was a focal lesion and was not a major feature

7

ELASTASE-IND UCED EMPHYSEMA

of the early changes. At 8 hr the lungs showed increasing numbers of macrophages, both about the bronchi and arteries and in alveolar walls, and these reached a maximum by 24 hr. The interstitial infiltrate had virtually cleared by the 4th day. However, intra-alveolar collections of macrophages persisted up to the 16th day. The cytoplasm of alveolar macrophages gave a strongly positive reaction for haemosiderin. Macrophages aggregated in groups in bronchioles and some alveoli, although most alveoli were cell free (fig. 8) and resembled the focal aggregations seen when blood is injected into normal animal lungs (Margarey, 1951). Saline-injected animals did not show this phenonemon. In two hamsters given the 0.5 mg dose of elastase and dying within 24 hr of injection some alveoli showed structures resembling hyaline membranes.

4 8

24 4 hours days

8

16

Time after elastase exposure

FIG.6.-Mean lung weight at intervals following injection (elastase injection 0-0, saline injection 0-0). Transitory weight increase affects only elastase-treatedhamsters.

The bronchial and bronchiolar walls showed similar infiltrates of PML in which few or no eosinophils could be identified. This infiltrate cleared in the same time-course as did the alveolar wall infiltrate. In the hamsters killed at 4 hr, and in the spontaneous deaths within 24 hr, the bronchial and bronchiolar epithelium generally remained intact, although focal ulceration did occur and was quite extensive in some early spontaneous deaths. Animals dying spontaneously within the 1st day after injection showed clear evidence of intra-arterial thrombosis in the main pulmonary artery and its larger branches, none being seen in the smaller arteries or arterioles. The thrombi showed a lamellar pattern of fibrin deposition and a smooth luminal surface, without obvious endothelialisation. The deep portion of the thrombus had a broad mural attachment (fig. 9). Elastic stains showed segmental necrosis of the arterial wall underlying the thrombus with destruction of both inner and outer elastic membranes and the muscle coat (fig. 10). In some instances only the inner elastic lamina was destroyed. There was a variable amount of PML infiltration at these sites, the infiltrate at times being minimal, at others marked. Mural thrombosis was generally restricted to the area of underlying necrosis, although several areas of segmental necrosis occurred without overlying

8

J. A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

thrombi. Careful search revealed no evidence of pulmonary venous lesions of a similar nature. Half of the 28 hamsters dying in the first 24 hr after injection showed mural thrombosis. In hamsters surviving a 0.5 mg elastase dose, 60 per cent. showed thrombi in pulmonary arteries in the first 24 hr, but none was seen after 4 days. A single focal arterial wall lesion was the only evidence which suggested a late, healed lesion. This consisted of focal reduplication and thickening of the inner elastic membrane in a hamster exposed to 0-5 mg elastase and killed 16 days after injection (fig. 11). Experiment 3. Late evolution of elastase damage All animals exposed to a dose of 0 3 mg elastase showed widespread, uniform emphysematous change of the panacinar type with little subjective

0

20

40

60

80

100

Days after exposure

FIG.12.-Mean linear intercept (fstandard error) in control hamsters and elastase-treated hamsters at differing time intervals after injection ( $ control, $ elastase 0.5 mg/100 g body weight).

difference between lungs examined 16 and 90 days after injection of elastase. The values for mean linear intercept in each group confirmed this interpretation (fig. 12), but as the numbers are small these should be looked on as preliminary results. The measurements showed a highly significant increase over the control values (table IV), but the difference between the mean linear intercept at 16 and 90 days was not significant ( t = 1.338, P not significant). Elastic stains on the lungs of elastase-exposed animals showed striking alterations when compared with those of controls. In control animals alveolar walls show moderate numbers of fine, crossed elastic fibrils while the pleura shows a cascade-like pattern of fine, interwoven fibrils (fig. 13). In contrast, elastic fibres in the alveoli of elastase-exposed hamsters are coarse, nodular and few in number, while some appear to be ruptured, although this interpretation is difficult to substantiate. The changes are more readily appreciated in the pleura where the cascade of fine fibrils is replaced by scanty coarse fibres with a clumped and fragmented appearance (fig. 14).

PLATEI

HAYES,KORTHYAND SNIDER ELASTASE-INDUCED EMPHYSEMA

Fig. 20.-Control

hamster injected with saline alone. Haematoxylin and eosin (HE).

FIG.2b.-Hamster

injected with 0.2 mg elastase per 100 g body weight. HE.

x 12.

x 12.

PLATEI1

HAYES,KORTHYA N D SNIDER ELASTASE-INDUCED EMPHYSEMA

FIG.2c.-Control

at higher magnification. HE. x 35.

FIG.2d.-Elastase-damaged magnified. HE.

x 35.

PLATE111

HAYES,KORTHY AN D SNIDER ELASTASt-INDUCED tMPHYSEh4A

FIG.5.-Lung

of hamster dying spontaneously 24 h after 0.5 mg/100 g elastase injection showing prominent perivascular oedema and alveolar haemorrhage. Only small numbers of PML are present. HE. x100.

FIG.7.-Lung of hamsters dying 24 h after elastase injection showing alveolar haemorrhage and focal necrosis of alveolar wall associated with P ML aggregation. HE. x 240.

PLATE1V

HAYES,KORTHYAND SNIDER ELASTASE-INDUCED EMPHYSEMA

FIG.8.-Lung

16 days after 0.05 ing/lOO g elastase injection showing focal aggregations of haemosiderin containing macrophages. Prussian-blue reaction. x 180.

FIG.9.-Spontaneous

death at 24 h. Thrombus in pulmonary artery. Note adjacent PML infiltrate in wall. HE. x110.

HAYES,

PLATEV

KORTHYAND SNIDER ELASTASE-INDUCED EMPHYSEMA

FIG.10.--Comparable lesion showing thrombus overlying zone i n arterial wall with focal absence of inner and outer elastic laminae. Elastic and van Cieson. x 200.

FIG.11.-Sixteen days after elastase injection. Pulmonary artery shows segmental reduplication of elastic lamina resulting in intimal thickening (small arrows). Underlying there is breaching of both outer and inner elastic laminae (large arrow). EVC. Y 200.

PLATEVI

HAYES,KORTHYAND SNIDER

EMPHYSEMA ELASTASE-INDUCED

F I G.13. --Control hamster lung showing fine elastic fibres in pleura and alveolar walls. EVG. x 400.

FIG.14.-Elastase-cxposed

hamster lung showing coarse granular pattern of elastic fibres in pleura and adjacent alveolar wall. EVG. x 400.

ELASTASE-IND UCED EMPHYSEMA

9

DISCUSSION These experiments demonstrate a striking difference between the lungs of elastase-exposed animals and control groups, a conclusion supported both by the illustrations and by the measured indices. The widespread structural change is accompanied by physiological abnormalities in the anaesthetised living animal, similar to those seen in human emphysema (Koo et al., 1973, 1974). Our studies showed that the quasi-static deflation volume-pressure curve is shifted upward and to the left with an increase in compliance measured just above the total lung capacity. The functional residual capacity and lung volume at 25 cm water transpulmonary pressure (analogous to total lung capacity in humans) are markedly increased (fig. 15). These physiological and structural changes fulfil criteria which have been suggested for the identification of emphysema in experimental animals (Wright, 1968). TABLE IV Mean linear intercept at later stages of elastase damage M.L.I. (cmx 105) Time of killing

Number of animals

6Gt,,

Mean

SE

test

16 days

5

0.01335

04008

45 days

4

0.01458

04006

12.91

90 days

4

0.01575

04016

6.082

Controls

12

0.00593

04003

...

8.983

Comparison with controls (P)

<0.001 <0~001 < 0.001 ...

The evolution of the lesion is relatively rapid, the stable, " healed " state being established within a few days, thus making identification of the initial site and nature of the injury difficult. There is an initial, transitory increase in lung weight, pulmonary hemorrhage and mild polymorphonuclear infiltration in alveolar walls which contrasts markedly with the prominent PML infiltrate seen in papain-induced emphysema (Goldring et al., 1972). The finding suggests that there is some underlying tissue destruction. However, focal destruction was only seen in animals dying spontaneously from large doses of elastase. It is not clear whether the initial haemorrhagic changes represent tissue injury associated with cell death, or whether they simply result from a transitory phase of increased vascular permeability not associated with cell death. So far, our data do not help resolve this question. An important aspect of the mechanism of elastase damage lies in the possibility that human emphysema may result from the liberation of enzymes from damaged cells within alveoli, as in patients with homozygous alpha I antitrypsin deficiency. Proteolytic enzymes are present within PML (Janoff and Scherer, 1968) and alveolar macrophages (Janoff, 1972), and it has been shown

10

J. A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

that lung destruction resembling human emphysema can be produced in dogs by aerosols of enzymes extracted from dog and human PML and from dog alveolar macrophages (Mass et al., 1972). Recently, the same group of workers has demonstrated that washings from normal lungs have antiprotease activity which can be neutralised by excess enzyme (Weinbaum et aE., 1974a). Consequently the minor PML exudate seen both in the observations reported here and in reports by others (Weinbaum et al., 1974b) does not seem to adequately account for the severe, widespread changes produced. These considerations suggest that the lesion is directly due to the introduction of the elastase solution into the lung rather than from proteases liberated from cells evoked as part of an inflammatory reaction to the injection. Lung volumes and volume-pressure diagrams in 15 untreated and 8 elastase treated hamsters

PL (cm H@)

FIG.15.-Quasi-static deflation volume-pressure curves from control and elastase-exposed hamsters. Elastase damage shifts the curve upward and to the left. This is associated with a marked increase in functional residual capacity and lung volume at 25 cm water transpulmonary pressure.

The damage produced in the rat lung by elastase in vitro appears to affect mainly the elastic framework (Johanson and Pierce, 1972). Morphologically, it is difficult to determine how and where elastic fibres are damaged because alveolar shape and size are so markedly altered. However, damage to elastic fibres in the pleura probably occurs because the interweaving cascades of fine fibrils are replaced by thick, lumpy fibres often with knobs at their ends. It is reasonable to infer that elastic fibres are similarly damaged throughout the lungs, a conclusion supported by the physiological abnormalities mentioned above (fig. 15). The changes in elastic fibres have not yet been studied by electron microscopy, but in papain-induced emphysema, which has some

ELASTASE-IND UCED EMPHYSEMA

11

features of the elastase lesion, damage was thought to be restricted to elastic fibres (Johanson et aZ., 1973). Johanson considered that the elastic fibres were disrupted but not removed. This interpretation agrees with the observations reported here and also with the minimal alterations detected in the total elastin content, and elastin-collagen ratio of aging and emphysematous human lungs (Pierce et aZ., 1961; Pierce and Ebert, 1965). In other words, the elastic fibres could have undergone marked physiological impairment without a significant loss in total lung content of chemically identifiable elastin. It is extremely difficult to produce lung damage by intravenous administration of elastase (Weinbaum et al., 1974). These workers describe no pulmonary arterial lesions similar to those noted in this paper, even after pulmonary arterial infusion of elastase. The only record of vascular damage in enzymeinduced emphysema is a fleeting reference to disrupted and beaded " elastic fibres in the walls of arterioles, venules and bronchioles " within a few hours of exposure to papain aerosol (Johanson et al., 1973). Presumably this is due to the relatively large amounts of antiprotease activity present in the circulating blood. The mechanism of the arterial damage describe here is puzzling. If elastase is directly absorbed across the alveolar walls into the capillaries, lesions might be expected in pulmonary veins, but these were not seen. Moreover, no microscopic evidence of acute capillary destruction was seen in the first 24 hr, the lesions being essentially arterial in location. Absorption of elastase followed by systemic circulation seems unlikely in view of the findings by Weinbaum and his colleagues mentioned above. Transmural absorption with direct damage to the arterial wall seems most unlikely because the arteries affected have thick media covered by an appreciable layer of adventitia and interstitial connective tissue which would present a distinct anatomic barrier. Furthermore, in some lesions the artery showed breaching only of the internal elastic lamina which suggests that injury is originating from within the arterial lumen. This, of course, assumes that the internal elastic lamina and external elastic lamina are equally susceptible to elastase digestion. Lymphatic transportation of elastase to produce injury is unlikely because absorption would be predominantly perivascular and so damage might be expected to affect the external rather than the internal elastic lamina. Theoretically it might be explained by distribution via vasa vasorum except these do not appear to be present in the medium-sized pulmonary arteries of the hamster. Arterial thrombosis including pulmonary arterial involvement has been reported in dogs and rabbits after intravenous injection of dilute trypsin solutions (Tagnon, 1945; Taylor and Wright, 1954). It is thought that the trypsin promotes the conversion of prothrombin to thrombin with subsequent thrombus formation. Neither paper mentioned damage to the arterial wall. A recent publication has noted that in man pulmonary thromboemboli (Salyer et al., 1974) produce damage of the underlying arterial wall which includes damage to and even rupture of the elastic lamina in almost 20 per cent. of the cases studied. In theory, systemic circulation of elastase could produce systemic venous thrombosis and subsequent embolism leading to arterial damage. We think this is also an unlikely

12

J . A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

mechanism. The presence of the single " healed " arterial lesion at 16 days after injection is intriguing. Morphologically it resembles changes seen in elderly human lungs, particularly in persons suffering from chronic obstructive lung disease (Mclean, 1958; Hayes, 1968). Acute pulmonary arterial lesions with thrombosis have not been described in human emphysema, although they may occur in septicemic processes (Rabin et al., 1961) or polyarteritis nodosa (Symmers, 1952). The relationship of these arterial lesions to human emphysema is obscure. Two additional points deserve mention. Firstly, the lesion with the 0.5 mg dose in our hands is invariably uniform throughout the lung, whereas papain damage tends to involve entire segments served by a separate bronchi, the intervening lung being uninvolved. This may in part be due to the dose administered because the 0.2 mg dose level shows many examples of a segmental lesion. However, this is not the entire explanation because papain rarely produces a diffuse lesion even when the LD50 is approached (Snider et al., 1974). Secondly, our data does not clearly establish whether or not the panacinar lesion becomes more severe as the interval following the initial injury increases. Although there is no statistically significant difference in the M.L.l. between 16 days and 90 days after damage (table IV), the points in the diagram (fig. 14) do suggest that the alveoli of the damaged lung become bigger with increasing age. However, as the numbers are small, further study is necessary to clarify this point. It is very difficult to assess the presence of parenchymal destruction histologically, particularly in lungs showing increased alveolar size. It has been suggested that the loss of " spurs " at the mouth of alveoli, and increase in alvoelar diameter indicate tissue destruction (Gross et al., 1964). We have found it difficult to interpret such changes. However, the present experiment shows that there is a decrease in internal surface area which is consistent with alveolar loss. The study reported here shows that the intratracheal injection of elastase provides an animal model with some of the characteristic anatomical and physiological features of the uncomplicatedpanacinar type of emphysema seen in man (Ciba Symposium, 1959). The model is useful for studying the possible structural changes and pathogenesis of human emphysema and for assessing the impact of such factors as infection, smoking amd inhalation of toxic matter of various kinds. SUMMARY A single dose of crystalline, porcine pancreatic elastase injected intratracheally into hamsters induces widespread alveolar enlargement with subpleural bullae. A uniformly severe lesion is consistently induced by 0.2 mg elastase per 100 g body weight and with negligible mortality. Compared with controls, which showed no lesion, elastase-damaged lungs show a highly significant (P ~0.001)increase in alveolar size and a decrease in internal surface area. Taken with the associated physiological abnormalities, these findings closely simulate human emphysema of the panlobular (panacinar) type. Histologically

ELASTASE-IND UCED EMPHYSEMA

13

it appears that elastase converts the fine elastic fibres in alveolar walls and pleura into thickened, nodular fibres which may also be broken along their length. With higher doses of elastase, i.e., 0.5 mg/100 g body weight, many pulmonary arteries showed segmental loss of inner and outer elastic laminae, usually with thrombosis on the overlying endothelium. The mechanism of this thrombosis is unclear. These experiments suggest that damage to elastic fibres may be an important element in the development of human panacinar emphysema, and that the damage could be one pathogenetic mechanism which produces damage of elastic fibres. We wish to express our thanks to Olva Delmano, Bradford Milne, Eveline Gillette and Frank Harris for their technical assistance in preparing this paper, and Claire Costello for secreterial assistance. The study was supported by Grants HL15563 (NIH) and MRIS 8019 (VA). REFERENCES 1959. Terminology, definitions and classification of chronic pulCIBAGUESTSYMPOSWM monary emphysema and related conditions. Thorax, 14,286. DUNNILL,M. S. 1962. Quantitative methods in the study of pulmonary pathology. Thorax, 17,320. ERIKSSON, S. 1965. Studies in alpha1 antitrypsin deficiency. Acta Med. Scand., 177,Suppl. 432, 1. GOLDRING, I. P., GREENBERG, L., AND RATNER, I. M. 1968. On the production of emphysema in Syrian hamsters by aerosol inhalation of papain. Arch. Environ. Health, 16, 59. GOLDRING, I. P., PARK,S. S., GREENBERG, L., AND RATNER, I. M. 1972. Sequential anatomic changes in lungs to papain and other proteolytic enzymes. In Pulmonary emphysema and proteolysis, ed. by C. Mittman, New York, p. 389. GROSS,P., BABYAK, M. A., TOLKER, E., AND KASCHAK, M. 1964. Enzymatically produced pulmonary emphysema. A preliminary report. J. Occup. Med., 6,481. GROSS,P., PFITZER,E. A., TOLKER, E., BABYAK, M. A., AND KASCHAK, M. 1965. Experimental emphysema: Its production with papain in normal and silicotic rats. Arch. Envirn. Health, 11, 50. GREENBERG, S. D., JENKINS, D. E., STEVENS, P. M., AND SCHWEPPE, H. I. 1973. The lungs in homozygous alpha I antitrypsin deficiency. Amer. J. Clin. Path., 60,581. HAYES,J. A. 1968. M.D. Thesis, University of Bristol. Cardiopulmonary Disease in Jamaica. A. 1972. Elastase-like proteases of human granulocytes and alveolar macrophages. JANOFF, In Pulmonary emphysema and proteolysis, ed. by C. Mittman, New York, p. 205. JANOFF, A., AND SCHERER, J. 1968. Mediators of inflammation in leucocytic lysosomes. IX. Elastolytic activity in granules of human polymorphonuclear leucocytes. J. Exper. Med., 128, 1137. JOHANSON, W.G., JR., AND FIERCE, A. K. 1972. Effects of elastase, collagenase and papain on structure and function of rat lungs in vitro. J. Clin. Invest., 51,288. JOHANSON, W.G., JR., PIERCE, A. K., AND REYNOLDS, R. C. 1971. The evolution of papain emphysema in the rat. J. Lab. Clin. Med., 78,599. JOHANSON, W. G., JR., REYNOLDS, R. C., SCOTT, T. C., AND PIERCE, A. K. 1973. Connective tissue damage in emphysema. An electron microscopic study of papain-induced emphysema in rats. Amer. Rev. Resp. Dis., 107,589. KAPLAN, R., KUHN,C., AND PIERCE, A. K. 1973. The induction of emphysema with elastase. I. The evolution of the lesion and the influence of serum. J. Lab. Clin. Med., 82, 349. KARNOVSKY, M. J. 1967. The ultrastructural basis of capillary permeability studied with peroxidase as a tracer. J. Cell Biol., 35, 213.

J. A . HAYES, AGNES KORTHY AND GORDON L. SNIDER

14

Koo, K. W., HAYES,J. A., KAGEN,H. M., LEITH,D. E., FRANZBLAU, C., AND SNIDER,G. L. 1973. Lung volumes and mechanics in elastase-induced emphysema. Clin. Res., 21, 986 (Abstract). Koo, K. W., HAYES, J. A., KAGEN,H. M., LEITH,D. E., FRANZBLAU, C., AND SNIDER, G. L. 1974. Lung volumes and mechanics following elastase and collagenase in hamsters. Clin. Res., 22, 508 (Abstract). MCLEAN,K. H. 1958. The significance of pulmonary vascular changes in emphysema. Australas. Ann. Med., 7 , 69. MANDL,I. 1961. Collagenases and elastases. Advan. Enzymol., 23, 163. MARGAREY, F. R. 1951. Experimental pulmonary haemosiderosis. J. Path. Bact., 63, 729. MASS,B., IKEDA, T., MERANZE, D. R., WEINBAUM, G., AND KIMBEL,P. 1972. Induction of experimental emphysema. Cellular and species specificity. Amer. Rev. Resp. Dis., 106, 384.

ORELL,S. R., AND MAZODIER, P. 1972. Pathological changes in alpha I antitrypsin deficiency, In Pulmonary emphysema and proteolysis, ed. by C. Mittman, New York, p. 69. PIERCE, J. A., AND EBERT,R. V. 1965. Fibrous network of the lung and its change with age. Thorax, 20,469. FQRCE, J. A., HOCOTT,J. B., AND EBERT, R. V. 1961. The collagen and elastin content of the lung in emphysema. Ann. Intern. Med., 55, 210. RABIN,E. R., GRABER, C. D., VOGEL,E. H., FINKELSTEIN, R. A., AND TUMBUSCH, W. A. 1961. Fatal pseudomonas infection in burned patients. A clinical, bacteriologic and anatomic study. New England J. Med., 265, 1225. SALYER, W. R., SALYER, D. C., AND HUTCHINS,G. M. 1974. Local arterial wall injury caused by thromboemboli. Amer. J. Path., 75, 285. SHOTTON, D. M., AND HARTLEY, B. S. 1973. Evidence for the amino-acid sequence of porcine pancreatic elastase. Biochem. J., 131, 643. SNIDER, G. L., HAYES,J. A., KORTHY, AGNES,STONE,P., FRANZBLAU, c.,AND KAGAN,H. 1974. Relationship between elastolytic activity and experimental emphysema-inducing properties of papain preparation. Amer. Rev. Resp. Dis., 110, 254. SYMMERS, W. ST. C. 1952 Necrotising pulmonary arteriopathy associated with pulmonary hypertension. J. Clin. Path., 5, 36. TAGNON, H. J. 1945. The nature and mechanism of the shock produced by the injection of trypsin and thrombin. J . Clin. Invest., 24, 1. TAYLOR, A,, AND WRIGHT,1. S. 1954. Intravenous trypsin. Circulation, 10, 331. TOMKIEFF, S. I. 1945. Linear intercepts, areas and volumes. Nature (London), 155, 24. WEIBEL, E. R. 1962. Morphometry of the human lung. New York. WEINBAUM, G., MARCO,V., IKEDA, T., MASS,B., MERANZE, D. R., AND KIMBEL, P. 1974(a). Enzymatic production of experimental emphysema in the dog. Route of exposure. Amer. Rev. Resp. Dis., 109, 351. WEINBAUM, G., TAKAMOTO, M., SLOAN,B., MERANZE, D. R., AND KIMBEL,P. 1974(b). Lung antiprotease: A possible primary defense against emphysema development. Amer. Rev. Resp. Dis., 109, 741 (Abstract). WRIGHT,R. R. 1968. Suggested criteria for confirming pulmonary emphysema in the experimental animal. Yale J. Biol. Med., 40, 576.