www.elsevier.com/locate/humpath

Human Pathology (2008) 39, 437–443

Original contribution

Collagen is reduced and disrupted in human aneurysms anddissections of ascending aortaB

Luciano de Figueiredo Borges PhDa,b, Rodrigo Gibin Jaldin MDc,Ricardo Ribeiro Dias MD, PhDd, Noedir Antonio Groppo Stolf MD, PhDd,Jean-Baptiste Michel MD, PhDa, Paulo Sampaio Gutierrez MD, PhDb,*

aInserm Unit 698, Cardiovascular Hemostasis, Bio-Engineering and Remodeling, Hôpital Bichat, Paris, FrancebLaboratory of Pathology, Instituto do Coração, University of São Paulo School of Medicine, São Paulo, BrazilcSão Paulo State University Medical School, Botucatu, BrazildDivision of Surgery, Instituto do Coração, University of São Paulo School of Medicine, São Paulo, Brazil

Received 24 May 2007; revised 2 July 2007; accepted 8 August 2007

0CA

S

0d

Keywords:Aorta;Aneurysms;Aneurysm dissection;Collagen;Metalloproteinase;Collagenolysis

Summary In ascending aorta aneurysms, there is an enlargement of the whole vessel, whereas aorticdissections (ADs) are characterized by the cleavage of the wall into 2 sheets at the external half. We searchedif alterations in collagen could be related to these diseases. Sections of aortas from 14 case patients with acutedissections, 10 case patientswith aneurysms, and 9 control subjects were stainedwith picrosirius. Slideswereanalyzed under polarized microscopy to evaluate the structure of collagen fibers. The proportion of collagenwas calculated in each half of the medial layer by color detection in a computerized image analysis system.Collagen appearance under polarized light was consistent with collagenolysis. The mean collagenproportions at the inner and outer halves, respectively, were 0.50 ± 0.13 and 0.40 ± 0.08 in the control group,0.20 ± 0.10 and 0.18 ± 0.12 in theADgroup, and 0.33 ± 0.12 and 0.19 ± 0.12 in the aneurysm group. TheAD(P b .01) and control (P = .04) groups had less collagen at the external half; no difference was found in theaneurysm group (P = .71). In both halves, there was less collagen in the case patients than in the controlsubjects (all P b .01), but at the internal half, the decrease was significantly greater in the case patients withaneurysms than in those with dissections (P = .03; at the external half, P = .99). Aortic dissections andaneurysms show a decrease in collagen content that could be related to a weakness of the wall underlying thediseases, but the locations of the decrease differ: in dissections, it is situated mostly at the external portion ofthemedia (site of cleavage), whereas in aneurysms, it is more diffuse, consistent with the global enlargement.© 2008 Elsevier Inc. All rights reserved.

☆ This study was supported by the State of São Paulo Research Foundation (Fundação de Amparo à Pesquisa do Estado de São Paulo, Brazil), through grant4/13177-7 and by the French National Research Agency (France). Luciano de Figueiredo Borges and Paulo Sampaio Gutierrez received fellowships from theonselho Nacional de Desenvolvimento Científico e Tecnológico (Brazil). Luciano de Figueiredo Borges received a fellowship from Coordenação deperfeiçoamento de Pessoal de Nível Superior (Brazil) as well.⁎ Corresponding author. Laboratory of Pathology, Instituto do Coração, University of São Paulo School of Medicine, Av Dr Eneas C. Aguiar 44, 05412-020

ão Paulo, Brazil.E-mail address: anppaulo@incor.usp.br (P. S. Gutierrez).

046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved.oi:10.1016/j.humpath.2007.08.003

438 L. de Figueiredo Borges et al.

1. Introduction

Aortic dissection (AD) is an acute life-threateningcondition characterized by intimal-medial tear and long-itudinal cleavage of the aortic wall along the medial layer,thus creating a false lumen [1]. Aneurysm is an insidiouschronic illness that corresponds to a segmental enlargementof the vessel wall, including its 3 layers. Aortic dissectionsand ascending aorta aneurysms (AscAA) share the samehistopathologic features, situated at the medial layer—namely, increase in mucoid substance, fragmentation ofelastic fibers, and apparent diminution in the number ofsmooth muscle cells; inflammatory infiltrates are usuallyscanty, in contrast to abdominal aortic aneurysms, whichare more commonly atherosclerotic/inflammatory. Further-more, there are patients with aneurysms who later sufferfrom dissection, and AD and AscAA have a commonassociation with other diseases, mostly systemic arterialhypertension, and genetic disturbances, such as the Marfansyndrome, the Ehlers-Danlos syndrome, and bicuspidaortic valve.

The pathogenesis of AD/AscAA remains unclear.Weakness of the wall is presumed to be present, but theactual factors underlying it have not been disclosed [2]. It isremarkable that in AD, the blood, tearing from the lumen,passes through a great proportion of the wall beforecleaving it: the separation into sheets occurs almost alwaysat the external half (in fact, at the external third) of themedial layer, near the adventitia [1]. This is the reasonbehind the worse prognosis of AD—rupture is usual,causing cardiac tamponade or hemorrhage to other sites.Although the main feature of AD is the intima-medial tear,this external localization of the cleavage probably has anoutstanding influence in the pathogenesis of the disease,although it is rarely taken into account in the literatureconcerning it.

Considering that the extracellular matrix constitutes agreat part of the arterial wall, it is probable that lesions ofthis compartment are involved in AD/AscAA. As men-tioned, alterations in mucoid content (probably proteogly-cans) and in the elastic system have been described [3], butthe possible changes in collagen fibers in the aortic wallhave been poorly explored despite their pivotal role as astructural protein. The amount, quality, and organization ofcollagen fibers vary significantly depending on theinteraction of collagen production and degradation, thusaffecting the tensile strength of the arterial wall. Sariola etal [4], in a histologic and nonquantitative study, found nodifference concerning collagen I and collagen III betweenaortas in patients with ADs and those in control subjects; incontrast, a biochemical analysis detected an increase incollagen content but a decrease in its concentration [5].Increase in collagen was also found by Wang et al [6]. Inthese last 2 series, either most patients had chronicdissection [6] or information concerning the intervalbetween onset of symptoms and taking of specimens was

not given [5]. Considering that in dissection repair, as inany scar in the body, fibrosis—a massive deposition ofcollagen—is the main characteristic, these results must beinterpreted with caution.

In the present study, we compared the amount andmorphological characteristics of collagen between patientswith aneurysms and those with acute dissections of theascending aorta, comparing the inner and outer regions of theaortic media.

2. Materials and methods

This study was approved by the Instituto do CoraçãoScientific and Ethical Committee (University of São PauloSchool of Medicine, São Paulo, Brazil).

Samples from the ascending aorta were collected duringsurgery for the correction of aneurysms (10 patients, 5 menand 5 women, whose ages ranged from 46 to 79 years [mean =62 years, median = 62.5 years]) or dissection (14 patients,including 1 necropsy case patient, 12 men and 2 women[dissection type A according to the Stanford classification [7]or type I or II according to Debakey et al's [8] classification],whose ages ranged from 23 to 82 years [mean = 56.29 years,median = 55.5 years]). Concerning this last disease, weincluded cases with no more than 3 days of evolution(interval between onset of symptoms and surgery) onlybecause if the time of evolution was greater, collagen couldbe deposited as part of the scarring process. Considering thatthe inflammatory infiltrate is usually absent or scanty inAscAA/AD, cases with prominent inflammation wereexcluded, as were cases with established syndromes, suchas bicuspid aortic valve and the Marfan syndrome. Therationale behind this last exclusion is, on one hand, to focuson the most common cases in which such syndromes areabsent; on the other hand, in some of those syndromes, incontrast to patients without them, a known connective tissuealteration may explain the aortic illness. In the Marfansyndrome, the genetic defect concerns fibrillin [9].

From the control subjects, 9 fragments of the ascendingaorta were taken during coronary artery bypass surgery. Allsubjects fulfilled our criteria, according to which aortas withmild intimal thickening or fat streaks were accepted butspecimens exhibiting either macroscopically or microscopi-cally any other lesion, such as moderate or severeatherosclerosis, were excluded. The ages of the controlsubjects (6 men and 3 women) ranged from 43 to 72 years(mean = 65.22 years, median = 69 years).

2.1. Tissue preparation and light microscopy

All samples were fixed in 10% buffered formalin,submitted to routine histologic processing, and embeddedin paraffin. Five-micrometer-thick sections were stainedwith orcein to the visualization of elastic fibers or using the

Fig. 1 Histologic sections of human aorta medial layer stained by orcein (A and D) and picrosirius (B and E) at light microscopy or atfluorescence microscopy (C and F). Control group sections are shown in panels A–C, whereas AscAA/AD group sections are shown in panelsD–F. A, Normal arrangement and distribution of elastic lamellae. B, Normal arrangement and distribution of collagen bundles.C, Simultaneous demonstration of elastic and lamellae fibers in green and collagen fibers in yellow/orange colors. D, Disrupted elastic lamellaein the AscAA/AD group. E, Disorganized and weakly stained collagen fibers. F, Presence of disruption of elastic lamellae (green) andcorrosion of collagen bundles (yellow/orange). Bar = 50 μm; objective magnification ×20.

439Collagen in ascending aorta diseases

440 L. de Figueiredo Borges et al.

Table 1 Percentages of area occupied by collagen in bothhalves (near the intima and near the adventitia) of the aorticmedial layer

Control group Aneurysm group AD group

Neartheintima a

Neartheadventitia b

Neartheintima c

Neartheadventitia d

Neartheintima e

Neartheadventitia f

0.59 0.39 0.20 0.27 0.31 0.240.31 0.35 0.19 0.20 0.46 0.120.59 0.39 0.14 0.07 0.32 0.140.56 0.50 0.31 0.16 0.47 0.340.45 0.27 0.38 0.24 0.38 0.250.66 0.44 0.23 0.10 0.50 0.420.53 0.51 0.08 0.06 0.19 0.090.49 0.33 0.23 0.34 0.46 0.370.30 0.42 0.20 0.01 0.14 0.07

0.04 0.36 0.20 0.070.19 0.080.30 0.110.41 0.250.28 0.09

a Mean ± SD = 0.50 ± 0.13.b Mean ± SD = 0.40 ± 0.08.c Mean ± SD = 0.20 ± 0.10.d Mean ± SD = 0.18 ± 0.12.e Mean ± SD = 0.33 ± 0.12.f Mean ± SD = 0.19 ± 0.12.

441Collagen in ascending aorta diseases

picrosirius method (1-hour bath in a 0.1% Sirius redsolution [Direct Red 80, Aldrich, Milwaukee, Wis] insaturated picric acid), followed by rapid washing in runningtap water without any counterstaining for the demonstrationof collagen.

2.2. Picrosirius polarization method

Picrosirius-stained sections were evaluated by polariza-tion microscopy. Picrosirius polarization is a method thatallows study of the arrangement and aggregation of collagenfibers caused by a normal birefringence exhibited by theirmolecules disposed orderly in a parallel orientation in thetissues. The picrosirius polarization method has been appliedfor histopathologic diagnosis of collagenolysis because thismethod detects morphologically not only the presence ofintegral collagen bundles but also the fragmented collagenbundles [10].

2.3. Picrosirius fluorescence method

Picrosirius-stained sections were also evaluated byfluorescence microscopy in the channel of fluorescein. Thismethod efficiently detects collagen and elastic fibers whenthese 2 structures are present simultaneously in the tissue andhas valuable applications in the processes that involve bothfibers [11].

2.4. Morphometry for collagen

Picrosirius-stained sections were examined in a Leicamicroscope coupled to a Quantimet 500 (Leica, Bannockburn,IL) image analysis system using a ×20 lens magnification. Thearea occupied by collagen was quantified by the color-detecting mode of the computer program in the inner (near theintima) and outer (near the adventitia) halves of the aorticmedial layer. Each half usually fit the screen field; those inwhich there were zones without the medial layer were notconsidered. The same number of fields (at least 20) wasmeasured in each half; thus, the amount of collagen reflectsdirectly its proportion. In the case patients with ADs, the2 sheets were measured, and regions containing blood or othermaterials were discharged if present.

2.5. Statistical analysis

Analysis of variancewas used to verify if possible differencesbetween halves and groups were independent and to compare

Fig. 2 Histologic sections of human aorta media stained by picrosiriusin a control subject (A) and in the AscAA/AD group (B) in which the wascattered fibers. C, The region of mucoid “lake” shows an absence of bmagnification the pattern of arrangement and distribution of fibers. Notecolors correspond to collagen fibers. Bar = 50 μm; objective magnificati

the groups; paired t test was used to compare the halves withineach group. Significance was established at P ≤ .05.

3. Results

3.1. Histologic observations

Sections of aortas stained with orcein showed in thecontrol subjects a normal arrangement of the media in whichthe layers of smooth muscle cells are separated by prominentelastic lamellae, which are interconnected by a network ofsmall elastic fibers and collagen fibers (Fig. 1A). Theconnective tissue that surrounds smooth muscle cells iscomposed of thick collagen fibers and bundles that werestained dark red in the picrosirius-stained sections (Fig. 1B).These patterns were confirmed by analyzing the slides influorescence microscopy (Fig. 1C). In this method, thecollagen fibers presented the characteristic red-orangefluorescence, whereas elastic fibers showed strong green

and observed under polarized light (picrosirius polarization method)ll has a disrupted appearance and collagen appears as thin and moreirefringence caused by the lack of collagen. Insets show in higherthat, differently from Fig. 1C and F, the green and yellow/orangeon ×20. In all insets, bar = 20 μm; objective magnification ×40.

Fig. 3 Percentages of area occupied by collagen in each half ofthe ascending aorta medial layer in control subjects and in thepatients with ascending aorta dissections and aneurysms.

442 L. de Figueiredo Borges et al.

intrinsic fluorescence. Under polarized light, the thickcollagen fibers and bundles presented a parallel arrangementat the medial layer, intensely bright, exhibiting irregular andanomalous dispersion of birefringence in red and/or yellowpatterns of the birefringence spectrum. The other few and thincollagen fibers disposed perpendicularly showed a greenbirefringence (Fig. 2A).

Representative cases of AD/AscAA are shown in Fig. 1D,E, and F. In these groups, the orcein-stained samples showeddisrupted and irregular elastic lamellae; in some areas, anelastic fiber framework was absent (Fig. 1D and F).Similarly, when analyzed by picrosirius, the collagennetwork disclosed dramatic morphological changes incollagen bundles (Fig. 1E and F), with the fibers disorga-nized and weakly stained. When analyzed under polarizedlight, collagen appeared as thin, pale (weak birefringence),greenish, and more scattered fibers, whereas the histoarch-itecture of the aortic wall showed a disrupted appearance(Fig. 2B). Fig. 2C shows a region of mucoid deposit in whichcollagen is almost absent, surrounded by disorganized fibers.

3.2. Morphometry for collagen

Table 1 and Fig. 3 show the percentages of area occupiedby collagen in both halves (near the intima and near theadventitia) of the aortic medial layer. The variationsconcerning halves and groups were independent. Patientswith ADs (P b .01) and control subjects (P = .04) had lesscollagen at the external half; in contrast, no difference wasfound in patients with AscAAs (P = .71). Patients with eitherAD or AscAA had less collagen as compared with thecontrol subjects in both halves (all P b .01), but at the internalhalf, the decrease was significantly greater in the patientswith AscAA than in those with AD (P = .03; at external half,P = .99).

4. Discussion

Extracellular matrix alterations are considered as a keyelement in the pathogenesis of aortic diseases, but there aredifferent patterns. In abdominal aortic aneurysms, there are

changes in collagen and massive elastic destruction, possiblycaused by the action of inflammatory cell–derived metallo-proteases (MMPs) [12,13], whereas in AscAA/AD, there isan accumulation of mucoid material (likely proteoglycans)and elastic fiber–delicate and localized fragmentation. Inspite of the role of collagen in giving structural resistance tothe tissues, its actual situation has not been totally describedyet. We focused our study on the amount and organization ofcollagen in the aortic media and on its distribution along withthe halves of this layer, considering the major importancethat this protein has in tissue structure and that dissectionsoccur at the external third of the artery.

Because the bulk of interstitial collagen is disposedorderly in a parallel orientation, a normal birefringence is aclassic characteristic of collagen fibers. Sirius red is anelongated dye molecule that reacts with collagen andpromotes an enhancement of its birefringence, allowing adetailed description of the molecular arrangement of thisprotein by analyzing the section under polarized light [10].Using this method, we detected collagenolysis—the cor-roded collagenous framework viewed in the aortic wall inAscAA/AD.

Sirius red stain also provides good evaluation of theamount of collagen present in a tissue: similar results werefound by quantifying collagen by hydroxyproline contentand by a procedure using this dye [14], a reason for its useand collagen quantification in other organs [15,16]. Con-sidering that it is not reliable to separate inner and outerhalves of the aorta to perform biochemical determinations,we thus used a morphometric approach in samples stainedwith the picrosirius method to measure collagen.

We found less collagen at the external lamellae of theaortas. According to theoretical calculations [17], thecapacity to sustain stress is greater at the internal two thirdsof the media, leaving an abrupt interface at the external third.The decrease in collagen, added to the increase in wall stressrepresented by systemic arterial hypertension (present in70%-90% of patients), could play a role in the pathogenesisof ADs. It is interesting to note that the difference betweenthe halves is also present in control cases, meaning that theouter part possibly has less tensile strength; worsening of thisnormal situation by additional decrease in collagen couldfacilitate the dissection.

In patients with AscAAs, the collagen content is alsolower than that in control subjects. However, the diminutionspans the whole wall, without a difference between itshalves. This fact could be related to the global enlargement ofthe artery as a result of diffuse weakening of the media.

Della Corte et al [18] had an alternative approach: theydid not compare inner and outer halves of the samesegment of the aortic wall and instead compared regions ofthe artery more or less affected by dilations linked to thebicuspid aortic valve. They also found diminished collagencontent in areas related to the disease, supporting the ideathat a decrease in collagen plays a role in the pathogenesisof AD/AscAA.

443Collagen in ascending aorta diseases

Despite the inflammatory infiltrate being usually scanty,MMPs, possibly produced by smooth muscle cells, havebeen demonstrated by others in the diseases we focusedon [19-24]. Among them were found MMP-1 [24], whichdegrades interstitial collagen (collagen I and collagen III), aswell as MMP-2 and MMP-9, which digest fibrillar collagenalready denatured (gelatin) and elastin. However, it is worthyto note that MMPs may be present in a tissue without anyaction. Although an important proof of MMP activity byzymography is lacking, the alterations we found in collagenamount and structure are probably consequences of theactions of these enzymes.

In conclusion, derangements in the metabolism of medialsmooth muscle cells should be involved in the pathogenesisof AscAAs and dissections, causing an increase in MMPexpression as well as activity and, consequently, collagendegradation. Some of these disturbances would be commonin both diseases, whereas others would not, leading to eitherdiffuse or more external alterations and, ultimately, determi-nation of the disease type.

Acknowledgments

We thank Adriana Psota and Solange Aparecida Consortifor their technical support in the preparation of the slides.

References

[1] Roberts WC. Aortic dissection: anatomy, consequences, and causes.Am Heart J 1981;101:195-214.

[2] Wilson SK, Hutchins GM. Aortic dissecting aneurysms: causativefactors in 204 subjects. Arch Pathol Lab Med 1982;106:175-80.

[3] Gutierrez PS, Reis MM, Aiello VD, Higuchi ML, Stolf NAG, LopesEA. Distribution of hyaluronan and dermatan/chondroitin sulfateproteoglycans in human aortic dissection. Connect Tissue Res 1998;37(3-4):151-61.

[4] Sariola H, Viljanen T, Luosto R. Histological pattern and changes inextracellular matrix in aortic dissections. J Clin Pathol 1986;39:1074-81.

[5] Whittle MA, Hasleton PS, Anderson JC, Gibbs AC. Collagen indissecting aneurysms of the human thoracic aorta. Increased collagencontent and decreased collagen concentration may be predisposingfactors in dissecting aneurysms. Am J Cardiovasc Pathol 1990;3:311-9.

[6] Wang X, LeMaire SA, Chen L, et al. Increased collagen deposition andelevated expression of connective tissue growth factor in humanthoracic aortic dissection. Circulation 2006;114:I200-5.

[7] Daily PO, Trueblood HW, Stinson EB, et al. Management of acuteaortic dissections. Ann Thorac Surg 1970;10:237-47.

[8] Debakey ME, Henly WS, Cooley DA, et al. Surgical management ofdissecting aneurysms of the aorta. J Thorac Cardiovasc Surg 1965;49:130-49.

[9] Dietz HC, Cutting GR, Pyeritz RE, et al. Marfan syndrome caused by arecurrent de novo missense mutation in the fibrillin gene. Nature 1991;352:337-9.

[10] Borges LF, Gutierrez PS, Marana HR, Taboga SR. Picrosirius-polarization staining method as an efficient histopathological tool forcollagenolysis detection in vesical prolapse lesions. Micron 2007;38:580-3.

[11] Borges LF, Taboga SR, Gutierrez PS. Simultaneous observation ofcollagen and elastin in normal and pathological tissues: analysis ofSirius-red–stained sections by fluorescence microscopy. Cell TissueRes 2005;320:551-2.

[12] Jacob MP, Badier-Commander C, Fontaine V, et al. Extracellularmatrix remodeling in the vascular wall. Pathol Biol (Paris) 2001;49:326-32.

[13] Shah PK. Inflammation, metalloproteinases, and increased proteolysis:an emerging pathophysiological paradigm in aortic aneurysm.Circulation 1997;96:2115-7.

[14] Lopez-De Leon A, Rojkind M. A simple micromethod for collagenand total protein determination in formalin-fixed paraffin-embeddedsections. J Histochem Cytochem 1985;33:737-43.

[15] Stumpf M, Cao W, Klinge U, et al. Reduced expression of collagentype I and increased expression of matrix metalloproteinases 1 inpatients with Crohn's disease. J Invest Surg 2005;18:33-8.

[16] De Heer E, Sijpkens YW, Verkade M, et al. Morphometry of interstitialfibrosis. Nephrol Dial Transplant 2000;15(Suppl 6):72-3.

[17] Berry CL, Sosa-Melgarejo JA, Greenwald SE. The relationshipbetween wall tension, lamellar thickness, and intercellular junctionsin the fetal and adult aorta: its relevance to the pathology of dissectinganeurysm. J Pathol 1993;169:15-20.

[18] Della Corte A, De Santo LS, Montagnani S, et al. Spatial patterns ofmatrix protein expression in dilated ascending aorta with aorticregurgitation: congenital bicuspid valve versus Marfan's syndrome.J Heart Valve Dis 2006;15:20-7.

[19] Ishii T, Asuwa N. Collagen and elastin degradation by matrixmetalloproteinases and tissue inhibitors of matrix metalloproteinasein aortic dissection. HUM PATHOL 2000;31:640-6.

[20] Lesauskaite V, Tanganelli P, Sassi C, et al. Smooth muscle cells of themedia in the dilatative pathology of ascending thoracic aorta:morphology, immunoreactivity for osteopontin, matrix metalloprotei-nases, and their inhibitors. HUM PATHOL 2001;32:1003-11.

[21] Koullias GJ, Ravichandran P, Korkolis DP, et al. Increased tissuemicroarray matrix metalloproteinase expression favors proteolysis inthoracic aortic aneurysms and dissections. Ann Thorac Surg2004;78:2106-10.

[22] LeMaire SA, Wang X, Wilks JA, et al. Matrix metalloproteinases inascending aortic aneurysms: bicuspid versus trileaflet aortic valves.J Surg Res 2005;123:40-8.

[23] Schmoker JD, McPartland KJ, Fellinger EK, et al. Matrix metallo-proteinase and tissue inhibitor expression in atherosclerotic andnonatherosclerotic thoracic aortic aneurysms. J Thorac CardiovascSurg 2007;133:155-61.

[24] Segura AM, Luna RE, Horiba K, et al. Immunohistochemistry ofmatrix metalloproteinases and their inhibitors in thoracic aorticaneurysms and aortic valves of patients with Marfan's syndrome.Circulation 1998;98:II331-7.