proton mri as a noninvasive tool to assess elastase-induced lung damage in spontaneously breathing...

Proton MRI as a Noninvasive Tool to Assess Elastase-Induced Lung Damage in Spontaneously Breathing Rats

Harry Karmouty Quintana,1,3 Catherine Cannet,1 Stefan Zurbruegg,1

Francois-Xavier Ble,1,2,4 John R. Fozard,2 Clive P. Page,3 and Nicolau Beckmann1*

Elastase-induced changes in lung morphology and functionwere detected in spontaneously breathing rats using conven-tional proton MRI at 4.7 T. A single dose of porcine pancreaticelastase (75 U/100 g body weight) or vehicle (saline) was ad-ministered intratracheally (i.t.) to male Brown Norway (BN) rats.MRI fluid signals were detected in the lungs 24 hr after admin-istration of elastase and resolved within 2 weeks. These resultscorrelated with perivascular edema and cellular infiltration ob-served histologically. Reductions in MRI signal intensity of thelung parenchyma, and increases in lung volume were detectedas early as 2 weeks following elastase administration and re-mained uniform throughout the study, which lasted 8 weeks.Observations were consistent with air trapping resulting fromemphysema detected histologically. In a separate experiment,animals were treated daily intraperitoneally (i.p.) with all-trans-retinoic acid (ATRA; 500 �g/kg body weight) or its vehicle (tri-glyceride oil) starting on day 21 after elastase administrationand continuing for 12 days. Under these conditions, ATRA didnot elicit a reversal of elastase-induced lung damage as mea-sured by MRI and histology. The present approach comple-ments other validated applications of proton MRI in experimen-tal lung research as a method for assessing drugs in rat modelsof respiratory diseases. Magn Reson Med 56:1242–1250, 2006.© 2006 Wiley-Liss, Inc.

Key words: air trapping; elastase; emphysema; lung MRI;chronic obstructive pulmonary disease (COPD)

A critical component of chronic obstructive pulmonarydisease (COPD) is emphysema, which arises from the de-struction of lung parenchymal tissue by a variety of pro-teases, such as neutrophil elastase and matrix metallopro-teinases (1). Destruction of the parenchyma leads to en-largement of the air spaces and loss of lung elasticity,which ultimately impairs gas exchange (2). Additionalconsequences are decreased maximum expiratory airflow,hyperinflation, and air trapping (3).

Until recently, assessment of these disease processes re-quired the examination of resected tissue. However, with theuse of computed tomography (CT) one can measure the struc-ture of the lung parenchyma and airway wall without having

to remove tissue (see Ref. 4 for a recent review). This isrelevant for studying the pathogenesis of COPD and differ-entiating patients with predominantly emphysema fromthose with airway wall remodeling, as well as for assessingthe effects of therapeutic interventions.

Diffusion-weighted 3He MRI provides an important al-ternative to assess the presence and severity of the pulmo-nary airspace enlargement that occurs in pulmonary em-physema. In humans, normal apparent diffusion coeffi-cients (ADCs) of 3He are less than 0.25 cm2/s, andincreased ADC values are observed in emphysema (0.3–0.9 cm2/s) (5).

Features of human panacinar emphysema can be in-duced in rats by porcine pancreatic elastase (PPE) (6,7)administered intratracheally (i.t.). Chen et al. (8) showedthat the ADC of hyperpolarized 3He in the lungs of suchanimals was larger than that in control, untreated rats.This observation was in line with elastase-induced alveo-lar expansion detected by histology.

In the present work we applied proton MRI techniqueswith the aim of detecting morphological and/or functionalchanges of the lungs related to experimental emphysemainduced in Brown Norway (BN) rats by i.t. instillation ofPPE. The readouts were 1) fluid signals reflecting the acuteedematous response to elastase, 2) lung volume, and 3)parenchymal signal intensity changes due to functionalimpairment of the lung (9,10). Measurements were per-formed on spontaneously breathing rats, and no cardiac orrespiratory gating was used.

Evidence from animal studies suggests that emphysema-tous changes can be reversed by treatment with all-trans-retinoic acid (ATRA) (11,12). Using proton MRI we soughtto verify whether ATRA was able to reverse the emphyse-matous changes induced by PPE in the BN rat, by adoptingthe same treatment protocol described by Massaro andMassaro (11,12).

MATERIALS AND METHODS

Experiments were carried out with the approval of the Vet-erinary Authority of the City of Basel (license number 567).

Animals

Male BN rats (N � 57), weighing 270–300 g, were suppliedby IFFA CREDO (L’Arbresle, France). Upon arrival theanimals were kept at an ambient temperature of 22°C �2°C under a 12-hr, normal-phase light-dark cycle and fedNAFAG® pellets (Nahr und Futtermittel AG, Gossau,Switzerland) for at least 1 week before the experimentswere started. Drinking water was freely available.

1Discovery Technologies, Novartis Institutes for BioMedical Research, Basel,Switzerland.2Respiratory Diseases Area, Novartis Institutes for BioMedical Research,Basel, Switzerland.3Sackler Institute of Pulmonary Pharmacology, King’s College, London, UK.4Faculty of Pharmacy, University Louis Pasteur, Illkirch, France.Grant sponsor: 3R Research Foundation; Grant number: 82/02.*Correspondence to: Nicolau Beckmann, Discovery Technologies, NovartisInstitutes for BioMedical Research, Lichtstr. 35; WSJ-386.2.09, CH-4002Basel, Switzerland. E-mail: [email protected] 13 January 2006; revised 14 June 2006; accepted 12 July 2006.DOI 10.1002/mrm.21051Published online 6 October 2006 in Wiley InterScience (www.interscience.wiley.com).

Magnetic Resonance in Medicine 56:1242–1250 (2006)

© 2006 Wiley-Liss, Inc. 1242

Intratracheal Administration of PPE or Saline

The animals were anesthetized (4% isoflurane; Abbott,Cham, Switzerland) in a chamber and then treated withPPE (Calbiochem, Darmstadt, Germany; 75 U/100 g bodyweight dissolved in 0.2 ml saline; N � 16) (6) or vehicle(0.2 ml saline; N � 12). The anesthetized rat was sus-pended at an approximately 45° angle by the two frontupper teeth by a rubber band attached to a metal support.A laryngoscope was used to lift the lower jaw of the rat andkeep the mouth open and the tongue displaced, to allow aclear view of the tracheal opening. During this direct vi-sualization, a curved cannula (diameter � 1.6 mm) at-tached to a 1-ml syringe was inserted into the trachea atthe level above the carina (the ridge separating the open-ings of the left and right main bronchi at their junctionwith the trachea). Immediately after the cannula was in-serted, 0.2 ml of PPE or saline was carefully instilled.

Treatment With ATRA Following PPE or SalineAdministration

In separate experiments we aimed to verify in BN ratswhether ATRA could reverse elastase-induced lung dam-age, an effect that has been described previously inSprague-Dawley rats (11,12). We based our experiments onthe original protocol developed by Massaro and Massaro(11). Thus, starting on day 21 after elastase instillation, therats (N � 15) received intraperitoneally (i.p.), once dailyand continuously for 12 days, either ATRA (500 �g/kgbody weight; N � 8) or its vehicle (triglyceride oil; N � 7).The ATRA (Sigma Aldrich, Steinheim, Germany) was dis-solved in triglyceride oil. The same protocol was repeatedin 14 saline-treated BN rats. Eight of these animals re-ceived ATRA, and six received triglyceride oil.

MRI Acquisition Protocols

The rats were examined by MRI prior to PPE (or saline)administration (baseline), and at 24 hr as well as at 2, 4, 6and 8 weeks after treatment. Imaging was performed lon-gitudinally, and three to six animals were killed at thedifferent time points for histological assessment. In sepa-rate studies to determine the effect of ATRA, the animalswere scanned by MRI before PPE (or saline) administration(baseline) and at 14, 21, 28 and 32 days after enzyme (orsaline) instillation. In this part of the study the animalswere imaged longitudinally and killed at the end of theexperiment (32 days following PPE or saline) for histolog-ical analysis.

During the MRI acquisitions the rats were placed in asupine position in a cradle made of Plexiglas. Body tem-perature was maintained at 37°C � 1°C using warm air,regulated by a rectal temperature probe (DM 852; Ellab,Copenhagen). The rats were anesthetized with 2% isoflu-rane in a mixture of O2/N2O (2:1), administered via a nosecone. All measurements were performed on spontaneouslybreathing animals, and no cardiac or respiratory triggeringwas applied. As demonstrated in previous studies (13,14),averaging over several respiratory cycles suppressed arti-facts caused by movements of the chest and the heartwithout the necessity of triggering the data acquisition.Measurements were carried out with a Biospec 47/40 spec-

trometer (Bruker Medical Systems, Ettlingen, Germany)operating at 4.7 T and equipped with an actively shieldedgradient system capable of generating a gradient of200 mT/m. The operational software of the scanner wasParavision (Bruker).

In line with previous experiments (10,13–15), during theMRI acquisitions the rats breathed spontaneously, andthus no intubation or mechanical ventilation was neces-sary. This technique easily enables repetitive measure-ments and an increased throughput, and is less invasivethan intubation and/or mechanical ventilation. Moreover,we wanted to circumvent possible interference resultingfrom lung injury complications that might potentially becaused by mechanical ventilation. Indeed, it has been re-ported that mechanical ventilation of healthy rats cancause an increase of neutrophils in bronchoalveolar lavage(BAL) fluid, pulmonary edema, and even hypoxemia thatmay lead to progressive circulatory failure and death (16).

Detection of Fluid Signals (Performed Only in the PPE8-Week Study)

For detection of fluid signals (which can arise from edemaor mucus hypersecretion in the lungs (13,15)), we applieda gradient-echo sequence with the following parameters(13,15): repetition time (TR) � 5.6 ms; echo time (TE) �2.7 ms; flip angle of the excitation pulse � �15°; field ofview (FOV) � 6 � 6 cm2; matrix size � 256 � 128; andslice thickness � 1.5 mm. We obtained a single-slice imageby computing the 2D Fourier transform (FT) of the aver-aged signal from 45 individual image acquisitions andinterpolating the data set to 256 � 256 pixels. There wasan interval of 530 ms between individual image acquisi-tions, resulting in a total acquisition time of 59 s for asingle slice. The entire lung was covered by 18 consecutivetransverse slices.

The volume of fluid signals was quantified using a semi-automatic segmentation procedure implemented in theIDL (Interactive Data Language Research Systems, Boul-der, CO, USA) environment (version 5.1) on a Linux sys-tem. The procedure was extensively described by Beck-mann et al. (13). The segmentation parameters were thesame for all analyzed images, and were chosen to segmentregions corresponding to high-intensity signals. Becausethe fluid signals and those from vessels were of compara-ble intensities, the volume corresponding to the vesselswas assessed on baseline images and then subtracted fromthe volumes determined on post-treatment images.

Determination of Lung Volume (Performed Only in theATRA Study)

To determine the lung volume, a gradient-echo sequencewith the following parameters was applied: TR � 5.6 ms;TE � 2.7 ms; flip angle of the excitation pulse � �15°;FOV � 10 � 6 cm2; matrix size � 256 � 128; and slicethickness � 1.5 mm. We obtained a single coronal imageby computing the 2DFT of the averaged signal from 30individual image acquisitions and interpolating the dataset to 256 � 256 pixels. There was an interval of 530 msbetween individual image acquisitions, resulting in a totalacquisition time of 37.5 s for a single slice. The entire lungwas covered by nine to 10 consecutive coronal slices.

Proton MRI for Elastase-Induced Lung Damage 1243

We determined the lung volume for a coronal slice bymanually delineating a region-of-interest (ROI) coveringthe entire lung, using the software from the spectrometer,and multiplying the extracted ROI area by the slice thick-ness. The total lung volume was obtained by adding thevolumes determined for the individual and consecutivenine to 10 coronal slices.

Detection of Modulations in the Intensity of Signals Fromthe Lung Parenchyma

The parameters of a gradient-echo sequence were chosenwith the aim of detecting signals from lung parenchyma(designated here as “parenchymal signals”). T2*-weightedimages were acquired using the following parameters(10,13): rise time of the gradients � 60 �s; matrix size �36 � 128; FOV � 6 � 6 cm2 for transverse slices or 6 �10 cm2 for coronal ones; slice thickness � 2 mm; TR �1.95 ms; and TE � 555 �s; 80 averages. The acquisition ofan individual image took 250 ms. A single slice image wasobtained by computing the 2DFT of the averaged signalfrom 80 individual image acquisitions and interpolatingthe data set to 256 � 256 pixels. There was an interval of500 ms between individual image acquisitions. The totalacquisition time for a single slice was 60 s. Five transverse

slices and five coronal slices covering the upper airwayswhile avoiding signals from the heart were acquired.

The mean signal intensities were determined for ROIslocated in the left and right sides of the lung in transverseimages, and in the upper (right and left) and lower (rightand left) lung areas of coronal images. Additionally, foreach slice the mean signal intensity over a ROI delineatinga large area of external skeletal muscle served as a refer-ence. Thus, for every slice we computed the ratio betweenthe mean signal intensity of an ROI positioned in theparenchyma and the mean signal intensity of the ROIlocated in the muscle. For a given lung region and animal,the mean ratio calculated for the five images acquiredconsecutively constituted what we designated as the pa-renchymal signal intensity in that area. We either com-pared this parenchymal signal intensity with histologicalreadouts obtained for the same rat in the same region, orwe obtained mean values of the parenchymal signal foranimals that belonged to the same group.

Histology

Preparation of the Tissue and Staining

The animals were killed with an overdose of pentothal(250 mg/kg i.p.) immediately after the MRI acquisitions

FIG. 1. a: Transverse MR images of the same animal acquired prior to (baseline) and at 24 hr and 2 weeks after i.t. administration of PPE(75U/100 g b.w.). Fluid signals (white arrow) were seen at 24 hr. Areas of weak signal intensity within the fluid signals (bulb end arrow) arerelated to the presence of blood following hemorrhage, 24 hr after PPE. At week 2 the fluid signals had subsided. b: Fluid signal volumes(means � SEM, N � 6–12) assessed by MRI in BN rats treated with PPE. c: Hematoxylin/eosin-stained lung slice of a rat previously treatedwith PPE (75U/100 g b.w., i.t.) and killed 24 hr after enzyme instillation. Increased alveolar areas are outlined by the black lines. The blackarrows highlight areas where gross hemorrhage was observed, and the blue arrow demonstrates the presence of perivascular edema.

1244 Quintana et al.

(i.e., at 24 hr and 2, 4, 6, and 8 weeks after PPE/salineadministration, and in the ATRA study 32 days after PPE/saline instillation). The lungs were then perfused via thetrachea with 10% neutral phosphate-buffered formalin(pH 7.2) at a constant rate of 5 ml/min using a perfusionpump (Perfusor IV; Bender & Holbein, Zurich, Switzer-land) until full inspiration volume was reached. The tra-chea was clamped and the lungs immersed in formalin fora maximum of 72 hr. After fixation, one transverse sectionof approximately 4 mm thickness from the left, right api-cal, median, and caudal lobes was trimmed and embeddedin paraffin wax. Paraffin blocks were sectioned at 3 �m.Reticulin staining according to the Gordon-Sweet method(17) was used to demonstrate collagen reticular fibers, ashistochemical staining of reticulin and elastin have beendemonstrated to closely reflect alveolar architecture (17).Using the Gordon-Sweet method, the fibers were first sen-sitized with a silver nitrate solution. Upon treatment witha reducing agent, silver taken up by the tissue in unre-duced form was converted to metallic silver, which wasdeposited at the sensitized sites. Any remaining unreactedsilver was removed by treatment with sodium thiosulfate.For a permanent preparation, the silver was partially im-pregnated with gold by treatment with gold chloride. Sec-tions were counterstained with eosin (17).

Assessment of Elastase-Induced Lung Damage by PointGrid Morphometry

Five semiconsecutive pictures from each section of theleft, right caudal, and right median lobes were captured at

�5 magnification (videocamera Prog/Res/3008; JenopticLOS, Eching, Germany) on Gordon-Sweet-stained slides,only in the alveolar parenchymal areas, avoiding largeblood vessels and bronchi. Morphometric analyses wereperformed with Adobe Photoshop software (Adobe Sys-tems Inc., San Jose, CA, USA). A point-counting method ofplanimetry (18) was used to estimate the relative amountof alveolar septa in each lung section. A 192-point grid(Image processing tool kit PlugIns, version 5.0; ReindeerGraphic Inc., Asheville, NC, USA) was superimposed overeach binarized picture, and the number of points that hitalveolar ducts or alveolus septa, lumen of alveolar ducts oralveolus, and vessel walls was recorded respectively. Pic-tures that presented more than 10% of points hitting thevessel walls were excluded from the statistics. The amountof the alveolar septum was calculated for each picturefrom the number of points hitting the alveolar ducts oralveolus septa, divided by the total number of test points �100, and referred to as the percent alveolar wall. Tenpictures per animal were evaluated.

Statistical Analysis

ANOVA comparisons (with Bonferroni corrections) withrespect to control groups were performed for MRI paren-chymal signal intensities and for changes in lung volumeusing the SigmaStat 3.1 (Systat Software, Inc., Point Rich-mond, CA, USA) software. Histological data using pointmorphometry were also analyzed by ANOVA.

FIG. 2. Coronal MR images of the same animalacquired before (baseline) and at 2, 4, 6, and 8weeks after a single dose of PPE (75U/100 g bodyweight i.t.). The black arrows show the decline insignal intensity of the lung parenchyma in both theleft and right sides of the lung following elastasetreatment.


RESULTS

Prominent MRI fluid signals were observed in the BN ratlung 24 hr following the administration of elastase, and

these signals resolved completely 2 weeks after instillationof the enzyme (Fig. 1a and b). The appearance and inten-sity of these signals were comparable to MRI fluid signalsin a previous study that strongly correlated with the for-mation of edema detected in the lungs of actively sensi-tized BN rats challenged with allergen (13). However, incontrast to the allergen-elicited signals, the PPE-instilledanimals presented areas of very low signal intensities inareas rich in fluid signals. These attenuated areas wereprobably due to hemorrhage in the lung parenchyma, asdescribed previously (19). Histological analysis revealedthe presence of gross cellular infiltration and perivascularedema 24 hr after elastase instillation (Fig. 1c), whichresolved 2 weeks later. Also hemorrhage was detected atthis time point (Fig. 1c).

As early as 2 weeks after enzyme administration (i.e.,once the acute edematous response caused by PPE hadsubsided), reductions in the MR signal intensity of theparenchyma were detected in most regions of the lung andwere maintained until the end of the study (8 weeks afterenzyme instillation). The decrease in signal intensities isillustrated in Fig. 2, which shows coronal images of thelung of the same animal acquired prior to and at 2, 4, 6,and 8 weeks after PPE administration. The most prominentchanges were observed in the lower regions of the lung(Fig. 3a). In transverse slices of the lung, changes weremore apparent in the left anatomical side of the animalcompared to the right side (data not shown). Saline-treatedanimals did not present any significant changes in signalintensities of the lung parenchyma over the experimentalperiod (Fig. 3b).

Enlargement of the alveolar spaces was clearly apparentin the histological assessment 24 hr after enzyme admin-istration. However, due to the presence of edema andcellular infiltration, quantification by point morphometryat this time point was not feasible. Quantitative histologi-cal assessments at 2 weeks after instillation and thereafterrevealed the development of emphysema in the left andright caudal lobes characterized as a decrease in the per-centile alveolar parenchyma (Fig. 4a and b). A slight re-covery in percentile alveolar parenchyma was observed atweek 8 in the left lobe. A highly significant correlation(Fig. 4c) was found between the MRI parenchymal signaland the percent alveolar parenchyma assessed histologi-cally for the right (r � 0.81; P � 0.008; N � 9) and left (r �0.92; P � 0.001; N � 9) sides of the lung.

Treatment with ATRA (500 �g/kg i.p.) or vehicle (tri-glyceride oil) consecutively for 12 days starting at day 21post-PPE administration, as described by Massaro andMassaro (11), did not result in a recovery of the parenchy-mal signal intensity at 7 or 12 days after initiation oftreatment (Fig. 5a). Also, the significant increase in lungvolume observed 14 days after administration of PPE wasnot altered following therapy with ATRA or vehicle (Fig.5b). No changes regarding either parenchymal signal in-tensities or lung volumes were observed in saline-admin-istered animals treated with ATRA or triglyceride solution(data not shown). The histological analyses were in linewith the MRI data, which showed that no regeneration oramelioration of lung structure occurred following ATRAadministration (Fig. 5c).

FIG. 3. a: Representative coronal image showing the delineation ofthe ROIs that were created to evaluate the images. Signal intensities(means � SEM) from coronal slices corresponding to the upperright, upper left, lower right, and lower left anatomical sides prior to(baseline) and 2, 4, 6, and 8 weeks after treatment with elastase (b)or saline (c). Significance levels *P � 0.01, **0.001 � P � 0.01,***P � 0.001 refer to ANOVA comparisons (Bonferroni test vs.control) between the different time points and baseline values, foreach region.


DISCUSSION

Using conventional proton MRI techniques, we observedcharacteristic signal changes in the lungs of spontaneouslybreathing BN rats following i.t. administration of PPE. Theacute response, 24 hr after elastase, was characterized by

the presence of pronounced fluid signals in MRI imagesthat correlated with extensive cellular infiltration accom-panied by hemorrhage, as evidenced histologically. Theseobservations are consistent with previous results demon-strating that administration of elastase to rats acutely leadsto hemorrhage in the lungs (19). They are also consistentwith the fact that neutrophil elastase is implicated in acutelung injury through increased permeability of the lungmicrovascular barrier, which leads to edema formation(20). The basis for this increased permeability may be thefact that elastase enhances the degradation of protein com-ponents, particularly those of the basement membrane ofvessels (21). Furthermore, Houtz et al. (22) showed thatelastase may produce an increased hydraulic conductivityin the interstitium surrounding large pulmonary vesselsthat can contribute to the formation of pulmonary edema(20), and is an important factor in the development ofpulmonary diseases, such as acute respiratory distress syn-drome.

Two weeks after PPE administration, the MRI-detectedfluid signals had subsided. On the other hand, from thistime point onward, reductions in the parenchymal signalintensity were apparent in several regions of the lung,particularly in inferior areas. Beckmann et al. (10) demon-strated a negative correlation between the parenchymalsignal in a gradient-echo image and the partial pressure ofoxygen in the blood for different amounts of oxygen ad-ministered. Therefore, the reduction in parenchymal sig-nal after PPE observed in the present experiments suggestsan increase of the lung oxygen content. This would beconsistent with air trapping occurring in enlarged airspaces in the lungs of PPE-instilled animals. Indeed, theaccompanying destruction of elastic fibers due to PPEcompromises the elastic recoil of the lung, leading to airentrapment (23). In addition to air entrapment, the loss oftissue and subsequent impairment of blood perfusion inareas of alveolar enlargement may also have contributed toa reduction in the lung parenchymal signal as conse-

FIG. 4. a: Histological slices stained with the Gordon-Sweetmethod (top) and corresponding binarization with the 192-point gridused for analysis superimposed (bottom). The alveolar area from theleft lobe of rats administered with saline (0.2 mL i.t.) and killed 24 hrlater (left images), and treated with PPE (75U/100g b.w) and killed 2weeks after i.t. enzyme instillation (right images) is shown. In theseslides 129 intercepts were counted for the saline-treated rat and 59for the elastase-treated animal. The black arrow shows the in-creased area of alveoli, and the dotted arrow shows the increasedterminal bronchi in the PPE-administered rat. b: Percent alveolarparenchyma (means � SEM) assessed histologically from the leftand right caudal lobes from animals treated with PPE and killed priorto (baseline) and 2, 4, 6, and 8 weeks after i.t. application. Signifi-cance levels *P � 0.01, and ***P � 0.001 refer to ANOVA compar-isons (Bonferroni pairwise test) between the different time pointsand baseline values, for each lobe. Significance levels #P � 0.01refer to differences between the time point after elastase adminis-tration in the left lobe. c: Correlation (r � 0.84, P � 0.001) betweenthe MRI parenchymal signal intensity and percent alveolar paren-chyma assessed histologically from 18 samples corresponding tothe same animals analyzed by MRI. Control values (N � 12) groupin the upper part of the graph, whereas PPE-treated animals (N � 6)cluster in the lower end of the graph.


quence of 1) an increased air volume relative to tissuemass, and 2) impairment of oxygen transfer from the alve-olar space to the blood circulation, leading to consistentlyhigher levels of oxygen in the alveolar space and hypox-emia compared to healthy animals in which normal ex-change rates between CO2 and O2 are present (24). The factthat more-pronounced parenchyma signal reductions wereobserved in lower regions of the lung was as expected,because these areas have a higher alveolar density thanupper regions of the lung. Consistent with this, extensiveenlargement of the alveoli was observed in alveoli-richsections of histological slices. Parenchymal signal changes

detected by proton MRI remained stable between weeks 2and 6 following the insult. A tendency toward recovery ofthe MRI signal intensity was apparent at week 8, whichcorrelated with a reduction of the emphysematous damageassessed histologically by point morphometry. Related tothis reduction in damage could be the fact that followingPPE administration the elastin content initially decreases,but appreciable elastic fiber deposition, and granulation ofthe alveolar airspaces containing fibroblasts, endothelialcells, and a provisional collagen matrix are observedweeks after injury (25). These events are hallmarks of thewound-healing process (26) and may mark the start offibrotic tissue deposition or, in some cases, the restorationof the gas-exchange apparatus (26).

Although instillation of PPE usually results in rapid andsignificant airspace enlargement (25), as exemplified byrabbit studies indicating that changes in lung function andmorphology were apparent 24 hr post-PPE administrationand remained constant for 8 weeks (27), we began to assessparenchymal signal intensity only at week 2 following PPEto ensure that there would be no interference with inflam-matory signals.

The magnitude of the parenchymal signal changes ob-served here presumably cannot be explained by the para-magnetic effects of molecular oxygen only. Diamagneticeffects in the foam-like structure of the inflated lung causeinhomogeneous line-broadening (equivalent to a short-ened free-induction decay (FID)) that is not found in to-tally collapsed lungs or in other tissues (28). The anoma-lously short FID is caused by local perturbations in themagnetic field near air–tissue interfaces (28), which areproduced by the differences in the magnetic susceptibilityof tissue and air, and therefore T2* values reflect morpho-metric features of the microarchitecture of the lung at thealveolar level (29). Since lung inflation and oxygenationare intrinsically related, it is conceivable that the paramag-netic effects of molecular oxygen and the line-broadening

FIG. 5. a: Signal intensities (means � SEM) from coronal slicescorresponding to the upper right, upper left, lower right, and lowerleft anatomical sides prior to (baseline) and 14, 21, 28, and 32 daysafter instillation of elastase. ATRA (500 �g/kg i.p.) therapy began 21days after PPE and was administered daily for 12 days. Significancelevels *P � 0.01, **0.001 � P � 0.01, ***P � 0.001 refer to ANOVAcomparisons (Bonferroni test vs. control) between the different timepoints and baseline values, for each region. b: Lung volume(means � SEM) of animals administered with elastase-retinoic acid,and animals that received elastase-triglyceride oil (vehicle) 14, 21,28, and 32 days after enzyme instillation. Significance levels ***P �0.001, refer to ANOVA comparisons (Bonferroni test) between thedifferent time points and baseline values. c: Percent alveolar paren-chyma (means � SEM) assessed histologically from the left lobefrom rats treated with saline (PPE vehicle) and ATRA (N � 8), PPEand ATRA (N � 8), PPE and triglyceride oil (ATRA vehicle, N � 7),and saline and triglyceride oil (N � 7). All of the animals were killed32 days after PPE or saline administration, and 12 days after ATRAor triglyceride oil. Significance levels ***P � 0.001 and # # #0.001 �P � 0.0 refer to ANOVA comparisons (Bonferroni pairwise test)between the different time points and baseline values, for each lobe.Consistent with our previous experiments, a significant reduction inthe MRI signal intensity from the parenchyma was evident 14 and 21days following elastase (but not saline) administration.


caused by an inflated lung would add up. Thus, the effectsof changes in tissue oxygen content on image contrastwould be more evident in the lung than in other tissues.Consequently, the outcome effects on the 1/T2* relaxationrates may have been amplified, resulting in larger contrastchanges in the gradient-echo images than would be ex-pected to be caused solely by the paramagnetic propertiesof oxygen.

An increase in lung volume assessed by MRI was appar-ent in elastase-treated animals, reflecting a feature of em-physema (3). The increase in pulmonary volume detectedin the present work is consistent with the destruction andenlargement of alveoli induced by elastase, and was re-ported previously in studies using whole-body pressureplethysmography (26) and MRI (8).

MRI and the histological parameters of emphysema de-velopment showed that there were no significant differ-ences between ATRA and vehicle (triglyceride)-treated an-imals after 12 consecutive days of therapy, following theprotocol of Massaro and Massaro (11). This is in contrast toexperiments carried out with Sprague-Dawley rats, whichshowed that ATRA may cause a marked and rapid reversalof elastase-induced alveolar damage (11,26). In those stud-ies, following PPE instillation, a regeneration of alveoliwas observed in the lungs of rats treated with ATRA com-pared to vehicle-treated animals (11,26). Similar to ourfindings, studies carried out on Fisher 344 rats (30), mice(31), and rabbits (27), as well as a pilot clinical trial in-volving 20 subjects with severe emphysema (32), did notdemonstrate beneficial effects of ATRA treatment in re-versing physiological and structural changes in the lungscaused by emphysema.

It is possible that the effects of ATRA in lung paren-chyma regeneration are species- or strain-dependent.However, it cannot be ruled out that in the present studies,lung regeneration might have been observed if the treat-ment time had been increased beyond 12 days, e.g., up today 56 after elastase treatment. A further aspect to con-sider is the fact that the mechanism by which ATRA pro-motes the formation of new alveoli is poorly understood(26,33). Evidence suggests that retinoids may play a piv-otal role in alveolar septation. Previous findings indicatethat 1) fibroblasts rich in vitamin A (retinol) accumulate inthe alveolar wall before septation, but decrease after com-pletion of the division process (33); 2) during alveolarseptation the concentrations of cellular retinoic acid bind-ing protein I and cellular retinol binding protein I(CRBP-1, which is involved in the synthesis of retinoicacid) are elevated (33); 3) treatment with dexamethasoneabrogates septation in premature rats and mice, and dimin-ishes the concentration of CRBP-1 mRNA (33); and 4)treatment with ATRA up-regulates the concentration ofCRBP-1 mRNA and partially rescues septation in rats andmice previously treated with dexamethasone (33). Thus, itis plausible that the mechanisms involved in repair andregeneration of alveolar tissue vary among species andaccording to the severity of the alveolar damage. Takentogether, these results suggest that ATRA therapy mayrequire higher doses and/or more prolonged treatment indifferent strains and species.

Measurements of the diffusion of hyperpolarized 3Hegas enabled the detection of changes resulting from en-

largement of the distal air spaces induced by PPE in rats(8,34) and mice (35). Significantly increased diffusion co-efficients, indicating alveolar expansion, were detected 4weeks or later in rats, and 2 weeks later in mice, afterelastase administration. The main advantage of hyperpo-larized techniques is their high sensitivity, which is animportant feature for lung studies in small rodents (8,34–37). However, they require specialized, costly equipmentand are subject to patent issues. Measurements of thediffusion of the SF6 gas in the lungs with 19F MRI may bean interesting alternative (38); however, application of thetechnique to an elastase model has not yet been demon-strated.

In summary, we have shown that proton MRI can detect,in spontaneously breathing rats, responses that reflect thecourse of inflammatory processes and microstructuralchanges at the alveolar level induced in the lungs by PPE.Since the measurements are easily repeatable, the ap-proach is potentially useful for the noninvasive assess-ment of compounds that may be beneficial for COPD (2). Inthe model adopted here, severe emphysema was inducedby a single i.t. dose of PPE. It remains to be demonstratedwhether proton MRI can assess the progression of diseasein more subtle models of emphysema, such as those in-volving smoke exposure (17) or inhibition of vascular en-dothelial growth factor leading to apoptosis and emphy-sema (39). The present approach complements some of thepreviously validated applications of proton MRI to exper-imental lung research, in view of routine drug testing (40),covering noninvasive assessments of inflammation(13,14), mucus hypersecretion (15), and of airway remod-eling (10) in several injury models of the rat. These distinctparameters of lung disease make proton MRI a particularlyinteresting tool for assessing animal models of airway dis-ease in which complex inflammatory mechanisms arepresent.

ACKNOWLEDGMENT

N.B. received an award from the 3R Research Foundation,Muensingen, Switzerland (grant 82/02).

REFERENCES

1. Barnes PJ. Mechanisms in COPD: differences from asthma. Chest 2000;117:10S–14S.

2. Barnes PJ. New treatments for COPD. Nat Rev Drug Disc 2002;1:437–446.

3. Sharp JT, Hammond MD. Pressure volume relationships. In: CrystalRG, West JB, Barnes PJ, Cherniack NS, Weibel ER, editors. The lung:scientific foundations. New York: Raven Press Ltd.; 1991. p 839–854.

4. Chooi WK, Morcos SK. High resolution volume imaging of airways andlung parenchyma with multislice CT. Br J Radiol 2004;77(spec no1):S98–S105.

5. Kauczor HU. Hyperpolarized helium-3 gas magnetic resonance imagingof the lung. Top Magn Reson Imaging 2003;14:223–230.

6. Busch RH, Lauhala KE, Loscutoff SM, McDonald KE. Experimentalpulmonary emphysema induced in the rat by intratracheally adminis-tered elastase: morphogenesis. Environ Res 1984;33:497–513.

7. Eidelman DH, Bellofiore S, Chiche D, Cosio MG, Martin JG. Behavior ofmorphometric indices in pancreatic elastase-induced emphysema inrats. Lung 1990;168:159–169.

8. Chen XJ, Hedlund LW, Moeller HE, Chawla MS, Maronpot RR, JohnsonGA. Detection of emphysema in rat lungs by using magnetic resonancemeasurements of 3He diffusion. Proc Natl Acad Sci USA 2000;97:11478–11481.


9. Edelman RR, Hatabu H, Tadamura E, Li W, Prasad PV. Noninvasiveassessment of regional ventilation in the human lung using oxygen-enhanced magnetic resonance imaging. Nat Med 1996;2:1236–1239.

10. Beckmann N, Cannet C, Zurbruegg S, Rudin M, Tigani B. Proton MRI oflung parenchyma reflects allergen-induced airway remodelling andendotoxin-aroused hyperresponsiveness: a step toward ventilationstudies in spontaneously breathing rats. Magn Reson Med 2004;52:258–268.

11. Massaro GD, Massaro D. Retinoic acid treatment abrogates elastase-induced pulmonary emphysema in rats. Nat Med 1997;3:675–677.

12. Massaro D, Massaro GD. Retinoids, alveolus formation, and alveolardeficiency: clinical implications. Am J Respir Cell Mol Biol 2003;28:271–274.

13. Beckmann N, Tigani B, Ekatodramis D, Borer R, Mazzoni L, Fozard JR.Pulmonary edema induced by allergen challenge in the rat: noninva-sive assessment by magnetic resonance imaging. Magn Reson Med2001;45:88–95.

14. Beckmann N, Tigani B, Mazzoni L, Fozard JR. MRI of lung parenchymain rats and mice using a gradient-echo sequence. NMR Biomed 2001;14:297–306.

15. Beckmann N, Tigani B, Sugar R, Jackson AD, Jones G, Mazzoni L,Fozard JR. Noninvasive detection of endotoxin-induced mucus hyper-secretion in rat lung by MRI. Am J Physiol Lung Cell Mol Physiol2002;283:L22–L30.

16. Walder B, Fontao E, Totsch M, Morel DR. Time and tidal volume-dependent ventilator-induced lung injury in healthy rats. Eur J Anaes-thesiol 2005;10:785–794.

17. Rangasamy T, Cho CY, Thimmulappa RK, Zhen L, Srisuma SS, KenslerTW, Yamamoto M, Petrache I, Tuder RM, Biswal S. Genetic ablation ofNrf2 enhances susceptibility to cigarette smoke-induced emphysema inmice. J Clin Invest 2004;114:1248–1259.

18. Elias H, Hyde DM. An elementary introduction to stereology (quanti-tative microscopy). Am J Anat 1980;159:411–446.

19. O’Byrne PM, Postma DS. The many faces of airway inflammation. Am JRespir Crit Care Med 1999;159:S41–S63.

20. Kawabata K, Hagio T, Matsuoka S. The role of neutrophil elastase inacute lung injury. Eur J Pharmacol 2002;451:1–10.

21. Negrini D, Passi A, De Luca G, Miserocchi G. Proteoglycan involvementduring development of lesional pulmonary edema. Am J Physiol LungCell Mol Physiol 1998;274:L203–L211.

22. Houtz PK, Jones PD, Aronson NE, Richardson LM, Lai-Fook SJ. Effect ofpancreatic and leukocyte elastase on hydraulic conductivity in lunginterstitial segments. J Appl Physiol 2004;97:2139–2147.

23. Yuan H, Kononov S, Cavalcante FS, Lutchen KR, Ingenito EP, Suki B.Effects of collagenase and elastase on the mechanical properties of lungtissue strips. J Appl Physiol 2000;89:3–14.

24. Lucey EC, Stone PJ, Snider GL. Consequences of proteolytic injury. In:Crystal RG, West JB, Barnes PJ, Cherniack NS, Weibel ER, editors. Thelung: scientific foundations. New York: Raven Press Ltd.; 1991. p1789–1801.

25. Shapiro SD. Animal models for COPD. Chest 2000;117:223S–227S.26. Belloni PN, Garvin L, Mao CP, Bailey-Healy I, Leaffer D. Effects of

all-trans-retinoic acid in promoting alveolar repair. Chest 2000;117:235S–241S.

27. Nishi Y, Boswell V, Ansari T, Piprawala F, Satchi S, Page CP. Elastase-induced changes in lung function: relationship to morphometry andeffect of drugs. Pulm Pharmacol Ther 2003;16:221–229.

28. Morris AH, Blatter DD, Case TA, Cutillo AG, Ailion DC, Durney CH,Johnson SA. A new nuclear magnetic resonance property of lung.J Appl Physiol 1985;58:759–762.

29. Case TA, Durney CH, Ailion DC, Cutillo AG, Morris AH. A mathemat-ical model of diamagnetic line broadening in lung tissue and similarheterogeneous systems: calculations and measurements. J Magn Reson1997;73:304–314.

30. March TH, Cossey PY, Esparza DC, Dix KJ, McDonald JD, Bowen LE.Inhalation administration of all-trans-retinoic acid for treatment ofelastase-induced pulmonary emphysema in Fischer 344 rats. Exp LungRes 2004;30:383–404.

31. Fujita M, Ye Q, Ouchi H, Nakashima N, Hamada N, Hagimoto N,Kuwano K, Mason RJ, Nakanishi Y. Retinoic acid fails to reverse em-physema in adult mouse models. Thorax 2004;59:224–230.

32. Mao JT, Goldin JG, Dermand J, Ibrahim G, Brown MS, Emerick A,McNitt-Gray MF, Gjertson DW, Estrada F, Tashkin DP, Roth MD. Apilot study of all-trans-retinoic acid for the treatment of human emphy-sema. Am J Respir Crit Care Med 2002;165:718–723.

33. Massaro GD, Massaro D. Related Retinoic acid treatment partially res-cues failed septation in rats and in mice. Am J Physiol Lung Cell MolPhysiol 2000;278:L955–L960.

34. Peces-Barba G, Ruiz-Cabello J, Cremillieux Y, Rodriguez I, Dupuich D,Callot V, Ortega M, Rubio Arbo ML, Cortijo M, Gonzalez-Mangado N.Helium-3 MRI diffusion coefficient: correlation to morphometry in amodel of mild emphysema. Eur Respir J 2003;22:14–19.

35. Dugas JP, Garbow JR, Kobayashi DK, Conradi MS. Hyperpolarized(3)He MRI of mouse lung. Magn Reson Med 2004;52:1307–1310.

36. Chen XJ, Chawla MS, Hedlund LW, Moller HE, MacFall JR, JohnsonGA. MR microscopy of lung airways with hyperpolarized 3He. MagnReson Med 1998;39:79–84.

37. Heverhagen JT, Hahn HK, Wegmann M, Herz U, Shaffer Whitaker CD,Matschl V, Alfke H. Volumetric analysis of mice lungs in a clinicalmagnetic resonance imaging scanner. MAGMA 2004;17:80–85.

38. Perez-Sanchez JM, Perez de Alejo R, Rodriguez I, Cortijo M, Peces-Barba G, Ruiz-Cabello J. In vivo diffusion weighted 19F MRI using SF6.Magn Reson Med 2005;54:460–463.

39. Petrache I, Natarajan V, Zhen L, Medler TR, Richter AT, Cho C, Hub-bard WC, Berdyshev EV, Tuder RM. Ceramide upregulation causespulmonary cell apoptosis and emphysema-like disease in mice. NatMed 2005;5:491–498.

40. Beckmann N, Tigani B, Mazzoni L, Fozard JR. Magnetic resonanceimaging of the lung provides potential for noninvasive preclinicalevaluation of drugs. Trends Pharmacol Sci 2003;24:550–554.


proton mri as a noninvasive tool to assess elastase-induced lung damage in spontaneously breathing...

Documents