of each airway was standardized for size by dividing by the Ideal airway lumen (LuBldeal), which ..... E 0.4 o. '-" 0.3. Morphologic and Morphometric Studies. Sections from both the left and right lungs ... (n = 8,0.242 ± 0.024 em H20/mlls).
Morphometric Changes during the Early Airway Response to Allergen Challenge in the Rat 1- 3
T. DU, L. J.
xu,
M. LEI, N. S. WANG, D. H. EIDELMAN, H. GHEZZO, and J. G. MARTIN
Introduction
The early response (ER) to inhalation of allergen by sensitized subj ects appears within minutes of the exposure and resolves spontaneously within 1 to 3 h (1). The ER is not only a feature of human asthma, but also follows allergen inhalational challenge in sheep (2, 3), guinea pig (4), rabbit (5), monkey (6-8), dogs (9-11), and rat models (12). It is associated with IgE-mediated mast cell activation (1, 13) and bronchospasm (14). The reversibility of the ER by /3-agonists has been frequently interpreted as indicating that it is caused principally by airway smooth muscle contraction (15, 16). However, airway narrowing, hyperemia, and edema have been observed by bronchoscopy after local challenge of allergic asthmatic patients with allergen (17). By 30min after specificinhalational challenge of sensitized rabbits, edema and vasodilation can be observed histologically in large airways (5). Even though published data (5, 17) suggest that edema as well as smooth muscle contraction may be involved in the pathogenesis of airway narrowing during the early response, quantitation of their relative contributions does not appear to have been attempted. One of the difficulties inherent in such a study is that the ER is short, making it hard to demonstrate the pathologic changes in vivo using conventional methods of fixation. To overcome this difficulty, a quick-freeze technique was used in conjunction with morphometric techniques to quantitate the airway changes in vivo during the ER in sensitized Brown Norway rats. Methods Animals and Sensitization Nineteen male Brown Norway rats, 6 to 7 wk old and ranging in weight from 210 to 290 g, were obtained from Charles River Inc. (St. Constant, Quebec, Canada). The animals were actively sensitized with a single subcutaneous injection of a sterile suspension of 1 mg ovalbumin (OA) and 200 mg of albu-
SUMMARY The purpose of this stUdy was to determine the relative contributions of airway wall edema and smooth muscle contraction to the early response (ER) of allergic bronchoconstriction. Brown Norway rats, 6 to 7 wk old, were sensitized with ovalbumin (OA). Anesthetized rats were challenged with either OA or saline 2 wk later. Pulmonary resistance (RL)was measured every minute until either It Increased to 150% of the baseline, defined as a significant ER, or until 15 min elapsed. Eight OA-challenged test rats with a significant ER and eight saline-challenged control rats were used for morphometric studies. The lungs werequick-frozen with liquid nitrogen, processed with freeze substitution. and sagittal sections (5 IJ,m) were stained with hematoxylin and eosin. The airway lumen subtended by the epithelial basement membrane (LuB) and cross sectional airway wall area (AW) of all airways were measured by camera luclda and dlgltlzatlon. The LuB and AW of each airway was standardized for size by dividing by the Ideal airway lumen (LuBldeal), which was calculated from the length of basement membrane, assuming a perfect circle in the unconstricted state. The cumulative frequency distribution of the LuB/LuBldeal for the airways from test rats was shifted to the left compared with the control rats (p < 0.01), Indicating airway narrowing after challenge. Airway narrowing increased as a function of airway size. Cumulative frequency distributions of AW/LuBldeal showed that there was a significant Increase In the wall thickness of only the small airways of test animals. We conclude that in the Brown Norway rat, the ER to Inhalatlonal challenge with OA is mainly caused by smooth muscle contraction rather than airway wall edema. AM REV RESPIR DIS 1992; 146:1037-1041
min hydroxide in 1 ml of 0.9070 saline. Simultaneously, 1ml of Bordetellapertussis vaccine containing 6 x 109 heat-killed organisms was administered intraperitoneally as an adjuvant.
Measurement of Early Responses At 14 to 18days after sensitization, rats were anesthetized with urethane (l g/kg, intraperitoneally). A tracheostomy was performed using an 11-cm length of polyethylene tubing (PE 240). The animals were mechanically ventilated (Model 680; Harvard Apparatus Co., South Natick, MA) with a tidal volume of 1.2 ml and a frequency of 90 breaths/min. A positive end-expiratory pressure of 2 cm H 2 0 was applied. To facilitate removal of the lungs at the end of the experiment, the abdominal cavity was opened by a midline incision and the thorax was opened through the diaphragm. A heating lamp was used to maintain a constant body temperature. The rat was placed in a constant volume body plethysmograph for the physiologic measurements. Volume was obtained by measuring changes in pressureinside the box using a differential pressure transducer (Validyne MP-45 ± 5 ern H 2 0 ; Validyne Corp., Northridge, CA). Airflow was obtained by differentiation of volume. The pressure inside the box was restored to atmospheric pressure periodically. Tracheal pressure was measured through a sideport in the tracheostomy tube
using a differential pressure transducer (Validyne MP-45 ± 100 em H 2 0 ). Aerosols were delivered by ultrasonic nebulizer (Model 100HC; DeVilbiss, Somerset, PA) into the intake port of the ventilator. An airflow of 11 mIls was used with a nebulizer output of 0.18 mIlmin. Each animal was challenged with aerosolized saline for 5 min and a measurement of RL was taken, which was considered as baseline. Test animals were then challenged with aerosolized OA for 5 min, and RLwas then measured every minute for a maximum of a further 10 min. Each (Received in original form August 23, 1991 and in revised form March 11, 1992) 1 From the Meakins-Christie Laboratories and the Department of Pathology, Royal Victoria Hospital, the Respiratory Health Network of Centres of Excellence, McGill University, Montreal, Quebec, Canada. 2 Supported by the J. T. Costello Memorial Research Fund and Grant No. 10381 from the Medical Research Council of Canada. Dr. Eidelman is an investigator of the Respiratory Health Network of Centers of Excellence. Dr. Martin is the recipient of an MRC Scientist Award. 3 Correspondence and requests for reprints should be addressed to Dr. J. G. Martin, MeakinsChristie Laboratories, McGill University, 3626 St. Urbain Street, Montreal, Quebec H2X 2P2, Canada.
1037
1038
DU, XU, LEI, WANG, EIDELMAN, GHEZZO, AND MARTIN
control rat was challenged with aerosolized saline and killed at the same time point as its matched test rat. Analog signals were conditioned by antialiasing filters and sampled at 200 Hz with an analog-to-digital board (DT280l-A; Data Translation, Marlboro, MA) installed in a microcomputer (Deskpro 286; Compaq, Houston, TX). Data were collected for two lO-s periods. RLwas calculated using multiple linear regression by obtaining the best fit for the equation: P tr
=
VRL X VEL X K,
where Ptr is tracheal pressure which is equal to transpulmonary pressure, V is flow, V is volume, ELis elastance, and K is a constant (18).
Tissue Preparation After the final measurement of RL, the endotracheal tube was clamped at end-expiration and the animal was removed from the box. The endotracheal tube was reconnected through a T-piece to a constant flow of air at a pressure of 4 em H 2 0 . This pressure was chosen so as to achieve a lung volume during fixation within the range of tidal breathing. Liquid nitrogen was poured into the thoracic cavity. The lungs were removed 10 min later and fixed in Carnoy's solution at - 80° C overnight. Carnoy's solution was prepared by mixing 60070 ethanol (E), 30% chloroform (C), and 10% acetic acid (A). Three modified Carnoy's solutions (MC) were made as follows: MC-l (E: C: A = 70: 22.5: 7.5); MC-2 (E: C: A = 80: 15: 5); MC-3 (E: C: A = 90: 7.5: 2.5). After fixation, the lung tissue was immersed for consecutive l-h periods in MC-l, MC-2, MC-3, and finally 100% ethanol at - 20° C. Ethanol at - 20° C wasreplacedwith ethanol at 4° C for 15 h. After the above procedure, the lungs were maintained in ethanol at room temperature for 2 h until sagittal blocks were cut about 1 mm from the hila of the lungs. The blocks were embedded in paraffin and 5-llm thick sections were stained with hematoxylin and eosin. Morphologic and Morphometric Studies Sections from both the left and right lungs were examined by light microscopy. For each airway,the airway structure was traced on paper using a drawing tube attachment (Leitz, WetzIar,Germany) and images wasmeasured using a commercial digitizing program (Sigma Scan; Jandel Scientific, Corte Madera, CA). Airwayswhose ratio of maximal to minimal internal diameter was equal to or larger than 2 were considered to be cut obliquely and were not measured. All other airways present on the slides from both lungs (mean, 18 airways/animal; range, 12 to 29 airways) were used for the following measurements: (1) airway lumen defined by the basement membrane of epithelium (LuB); (2)the length of the airway basement membrane (BM); and (3) the airway defined by the outer border of the airway wall (Ae). The area of the airway
wall (AW)was calculated as the difference between Ae and LuB.
cartilaginous airways were not always present in all of the lung sections, these airways were excluded from analysis. A total of 136 airways from eight test rats were analyzed, as were 149airways from eight matched controls. These airways were divided into three groups based on the length of the basement membrane: small (BM 0 to 0.99 mrn), medium (BM 1 to 1.99 mm), and large (BM > 2 mm) (19). The LuB of each airwaywas standardized for airway size by expressing LuB as a fraction of the ideal airway lumen (LuBideaI), assuming a circular airway in the unconstricted state. AW was also standardized with LuBideaI. Cumulative frequency distributions for LuB/LuBideal and AW/LuBideal were compared between test and control animals. The data are expressed throughout as the mean ± SEM. Comparison of means between test and control rats was made by unpaired t tests. A Wilcoxon signed rank test was used
Data Analysis A significant early airway response was defined as an increase in RL to a value of at least 150% of the baseline. Animals in the test group that showed such a response were retained for further analysis.Data are reported for eight ofthe 11 test animals that had a significant ER. A sample of 10randomly selected airways was measured by each of two observers independently. The correlation of measurements between and within observers was excellent (r 2 > 0.97) and the results for each observer were highly reproducible. However, small but systematic differences were found between observers so that the morphometric measurements weremade by a singleindividual (T. D.) blinded to the group status. Because
A
O.S
Tl
T2
0.3 0.2 0.1
,
r-;
..•-...-..•.•.....•
•....•......... •.. • ... ............... • -
T4
T3
0.4
1···-·····-···
•
•I
-
/
/'""0
(fJ
0.0
I
E 0 N
:r: E 0 '-:::J
o.s 0.40.3
ct:
0.2
/....•.• ...
1.0
T6
T5
-
/
.-.....-.
0.6
•
•
..._• •-•.•.••-.1.0
0.2
0.0 7
9
11
13 15
B 5
7
9
11 13
15
0.0
1.6 1.3
.-.-
O.
0.1 B 5
T7
0.8
/ •
0.7 0.4-
T8
r
•
0.0 B 5
7
9
11 13 15
8 5
7
9
11 13 15
Time (min)
B O.S
0.4
C2
Cl
C3
C4
0.3 0-0-0'0-0_
0_2 ~
0-0-0-0-0'0'0-0-0-0-0'0
0-0-0'0-0'0
0. 0'0'0.0-0.0.0.0.0'0
0-0-0-0-0'0'0-0-0-0-0-0
0.1
(fJ
I
E
o N :r: E o
'-"
O.S
CS
C6
C7
C8
0.4 0'0_0-0-0-0
0.3
0-0.0-0'0
0-0'0-0-0-0'0'0-0-0-0-0
0.1
MORPHOMETRIC ER CHANGES IN THE RAT
1039
0.02). However, after saline challenge the RL did not change significantly from the baseline value (0.235 ± 0.028; p > 0.05). The data for individual test and control animals are shown in figures lA and lB.
Morphologic Findings On light microscopy, the airway walls generally appeared normal. Basement membrane undulation, indicating narrowing of the airway lumen, was present and very obvious in some airways, especially in the test group (figures 2A and 2B). Although a small amount of mucus was present on the surface of the epithelium, mucous plugs were not observed. Morphometric Findings The distributions of airway size for control and test animals were similar (data not shown). ANOVA showed that the variance between blocks was from four to six times smaller than the withinanimal variance for both LuB/LuBideal and AW/LUBideal, indicating homogeneity in findings between animals and heterogeneity within the animals. Thus, the individual airways were virtually independent of each other because the blocking effect of the animal was considerably diluted. Therefore, wechose to examinethe entire distribution of airways to better see where the differences between the test and control animals lay. For all airways, the degree ofbronchoconstriction is shown in the form of cumulative frequency distributions in figures 3 and 4. The ratio of LuB/LuBideal was significantly smaller for the test versus control airways, as indicated by the leftward shift of the frequency distribution (p < 0.01; figure 3). A signifi-
0- -0
Fig. 2. Two representative airways from an allergen-challenged test rat (A) and a saline-challenged control rat (8). The epithelial infolding in the airway from the test rat indicates airway narrowing. In contrast, the airway from a control rat appears normal (x140).
1.0
Control (n=147)
. - . Test (n=136)
0.8 0.6 ~
o
to compare the baseline RLand the peak value of RLafter challenges (Rtrnax) within the test and control groups because of nonnormal distribution of data. Similarly, comparisons of the baseline RL and Rtmax between the two groups were made using the Mann-Whitney U test. The significance of differencesbetween cumulative frequency distributions was tested by the KolmogorovSmirnov test. We performed a restricted maximum likelihood analysis of variance (ANOVA) (19) in order to take into account the random blocking effect of the animals and the different number of airways contributed
by each animal. Valuesof p < 0.05 were considered significant.
c
o
0.4
:;:J
u o
0.2
G:: 0.2
Results Early Responses The baseline values of RL were similar
in both the control group (n = 8, 0.239 ± 0.022em H 20/mlls) and the test group (n = 8,0.242 ± 0.024 em H 20/mlls). In test animals, RL increased after OA challengeto a peak median value of 0.323 (range, 0.22 to 1.192 em H 20/mlls; p
,
DU, XU, LEI, WANG, EIDELMAN, GHEZZO, AND MARTIN
10
0
~
« 0
E
a - a Control (n=70) . - . Test (n=63)
D.!!
0.6
VJ
'0
0.4
c 0
:;:; (J
0.2
t'
u,
0.0 O.D
D.2
0.4
D.6
D.8
10
lu Bflu Bideal
{JJ
>,
0 3:
~
10
a - a Control (n=52) . - . Test (n=52)
D.8
E
OJ
'6
0.6
V
:::!'
'0
0.4
c 0
0.2
u,
0.0
~ t'
DO
D.2
D.4
D.6
lD
0.8
luB/luB,deal
(JJ
>,
10
0
~
«
a-o 0-0
Control (n=25) Test (n=21)
0.8
V
t"
0
0.6
--'
'0
04
c
'"
U
0.2
0
~
0.0 O.D
II /l
~ 0.2
D.4
0.6
Substantial heterogeneity in the degree of airway narrowing was also evident in the frequency distributions. The dispersion of airway lumen sizes increased as a function of airway size. The cumulative frequency distribution of airway wall area corrected for size (AW/LuBideal) for the small airways was shifted to the right in the test animals, indicating that the airway wall was significantly thicker (p < 0.05). There was a similar shift in the frequency distribution for the large airways, but the magnitude of the shift was not statistically significant. The AWfor the medium sized airways weresimilar for the test and control animals (figures 5A, 5B, and 5C). The results obtained using the Kolmogorov-Smirnov test to examine differences between frequency distributions of LuB/LuBideal and AW/LuBideal were in agreement with results obtained using the restricted maximum likelihood ANOVA, which in general showed a slightly larger significance level for the comparisons. We did not find any significant correlation between the increase in RL in the test animals and the mean or median of LuB/LuBideal of either all airways or the airways considered according to different categories of size.
1.0
0.8
Discussion
lu Bflu B,deal Fig. 4. The degree of airway narrowing for small- (A), medium- (B) and large- (C) sized airways. For each airway, the area of airway lumen (luB) was standardized by the ideal area (lUBideal). Airway narrowing was present in each size category (p < 0.01). However, the difference between the distributions in test and control rats increased progressively from small to large airways.The degree of airway narrowing was greatest in the large airways and least in the small airways.
cant reduction in airway lumen (LuB/ LuBideal) was present for airways of all sizes in test animals (figure 4). Furthermore, the maximal difference in frequency distributions between test and control animals increased progressively from the small to the large airways (table 1).
We found a progressive airway narrowing from small to large airways in the actively sensitized Brown Norway rat during the ER after allergen challenge. In the medium- and large-sized airways, the airway narrowing appeared to be mainly caused by airway smooth muscle contraction because, in the test group, airway walls of OA-challenged rats were not significantly thicker than those of control animals. Increased airway wall thickening was demonstrated in the small airways of allergen-challenged animals and may have contributed to luminal narrowing in these airways. The most striking morphologic difference between the airways of test and con-
TABLE 1 MEDIANS AND MAXIMAL DIFFERENCES OF CUMULATIVE FREQUENCY DISTRIBUTIONS OF LuB/luBideal FOR TEST AND CONTROL RATS
Small-sized airways Medium-sized airways large-sized airways All airways
Control Rats (median)
Test Rats (median)
0.84 0.80 0.76 0.71
0.77
• p < 0.01 compared with control group.
0.68 0.60 0.81
Maximal Difference between Control and Test Rats 0.24" 0.35* 0.45* 0.32*
co
10
.Z':{.
0.8
0
D.6
.>, 0
E
(J)
'0
0.4
c () ~
0 - 0 Control (n=70) . - . Test (n=53)
0.2
(J
t'
"-
0.0
06
0.0
0.8
AW/luB ideal
'" c-, 0 !"
«
10
r
0.8
E
:J
0.6
TI
