Inhibitory Effect of TIMP Influences the Morphology of Varicose Veins

12 Pages·2010·2.35 MB·English

by  B. Aravind

EurJVascEndovascSurg(2010)40,754e765
Inhibitory Effect of TIMP Influences the Morphology
of Varicose Veins
B. Aravinda,b, B. Saunders a,b, T. Navina,b, A. Sandison c, C. Monaco a,b,
E.M. Paleolog a,b, A.H. Davies a,*
aDepartment of Surgery andCancer,Facultyof Medicine,Imperial College, London, UK
bKennedyInstituteof Rheumatology,Facultyof Medicine,Imperial College, London, UK
cDepartment of Histopathology,Charing CrossHospital, London, UK
Submitted24October 2009; accepted25 April 2010
Available online 3July2010
KEYWORDS Abstract Objectives:Imbalanceofmatrixmetalloproteinaseenzymes(MMP)andtheirinhib-
Varicoseveins; itors (TIMPs) may contribute to the development of varicose veins. We hypothesised that,
TIMP; histologicalchangesinvaricoseveinwallcorrelatewithalterationsinexpressionofMMP/TIMP.
Matrix turnover; Methods:Varicoseveins(nZ26)werecomparedwithgreatsaphenousvein(GSV)segments
Morphology (nZ11)fromarterialbypass,andwitharmandneckveinsfromfistulaandcarotidoperations
(nZ13).Varicoseveinwallthicknesswasmeasured,enablingcategorisationasatrophicand
hypertrophic.MMP-2,MT1-MMP,TIMP-2,andTIMP-3expressionwerequantitativelyanalysed
byimmunohistochemistry.
Results:TherewassignificantlyhigherexpressionofTIMP-2(immunopositivearea4.34%versus
0.26%),linkedwith connective tissueaccumulationin thetunicamediaofvaricoseveinsas
comparedwitharmandneckveincontrols.TIMP-2andTIMP-3expressionwashigherinhyper-
trophicthanatrophicsegments(3.2%versus0.99%forTIMP-2,1.7%versus0.08%forTIMP-3).
Similarly, TIMP-2 and TIMP-3 had elevated expression in the thicker proximal varicose vein
segmentscomparedtodistal(4.3%versus1.3%forTIMP-2and0.94%versus0.41%forTIMP-3).
Conclusions:This study linkedmorphological changes in varicoseveinwalls with MMP/TIMP
balance.AhigherTIMPexpressionfavoursdepositionofconnectivetissueandthusthickervein
wall,reducingmatrixturnoverbysuppressionofproteaseactivity.
ª2010EuropeanSocietyforVascularSurgery.PublishedbyElsevierLtd.Allrightsreserved.
* Correspondingauthorat:DepartmentofVascularSurgery,CharingCrossHospital,FulhamPalaceRoad,LondonW68RF,UK.Tel.:þ44
02087461234;fax:þ4402087467362.
E-mailaddress:[email protected](A.H.Davies).
1078-5884/$36ª2010EuropeanSocietyforVascularSurgery.PublishedbyElsevierLtd.Allrightsreserved.
doi:10.1016/j.ejvs.2010.04.028
TIMPRegulateMatrix Turnoverin VaricoseVeins 755
Introduction Materials and Methods
Primary dilatation of the vein wall,1 causing weakness and Tissue acquisition and processing
resultant separation of valve cusps leading to reflux, has
gained wide acceptance as a cause of varicose vein. The AllpatientswereassessedandinvestigatedinCharingCross
primary basis for this theory, are the abnormalities found Hospital, London. Ethical approval was obtained from the
onthestructuralmatrixcomponentsoftheveinwall,such Riverside Research Ethics Committee (RREC3092), London.
ascollagenandelastin.2e4Remodellingoftissueisanormal The data collected from recruited subjects were anony-
dynamic process and abnormalities of the extracellular mised and protected according to the Human Tissue Bill
matrix, as above, suggest a defect in this process. It is (2004).
already known that remodelling is a fine balance between A total of 26 patients with varicose veins underwent
proteases and its inhibitors, the most prominent of which duplex ultrasound assessment to confirm the diagnosis of
are matrix metalloproteinase enzymes (MMP) and tissue varicose veins, prior to “High tie and stripping of the GSV
inhibitors of MMP (TIMP).5 An increase in the protease with multiple avulsions”. Informed consent and clinical
activitywillpromotematrixdegradation,whileanincrease details were obtained, including history of deep vein
in its inhibitors would have the reverse effect. The resul- thrombosis, diabetes mellitus, rheumatoid arthritis,
tantchangeinmatrixcomponentswillalsobereflectedin connective tissue disorders and cancers. Drug history was
morphology as it changes the basic structural elements. obtained, to include statins, doxycycline and steroids.
Thus,theobservedmorphologyofvaricoseveinwallatany Patients with a history of DVT were excluded from the
timeofitsdevelopmentwillbetheresultofsuchamatrix study, considering the risk of cellular infiltration on the
imbalance,andmaybereflectedinthe expressionof MMP vein wall from inflammation.13 CEAP (Clinical, Etiologic,
andTIMPandviceversa.Thelatterassertioncanbeproved Anatomic, and Pathophysiology) classification system was
only by comparing the changes in proteases with that of used to standardize the assessment of severity of
morphology.Inthecurrentstudy,thealterationofMMPand disease.14
TIMP are correlated to morphological feature of thickness Vein wall changes along the length of the varicose vein
of veinwall. were compared by harvesting 2e3 cm long vein segments,
In this study, it was hypothesised that the observed inpairs,onefromsapheno-femoraljunction andthe other
alteration in varicose vein morphology was the result of from the thigh end of the stripped varicose vein segment,
imbalanceinMMP/TIMPexpression.OftheseveralMMPand designated proximal and distal, respectively. To compare
TIMP, we analysed MMP-2, an important gelatinase with varicose veins, GSVs from primary arterial bypass opera-
unique ability to degrade collagen and elastin6 and MT1- tions were collected from the distal end of the harvested
MMP, which interacts with TIMP-2 in activation of MMP-2. length of vein (n Z 11). Neck veins from carotid endar-
TIMP-2 was analysed for the same reason. TIMP-3 was terectomy(commonfacialveinligatedandexcisedforsafe
analysed considering its ability to inhibit all MMP and its accesstobifurcationofcarotidartery)andarmveinsfrom
uniqueanti-apoptotic activity. segments left after fistula formation for haemodialysis
SeveralMMPsandTIMPshavebeenanalysedinpublished (nZ13), served asadditional controls.
literature on varicose veins, with most considering GSV
from bypass patients as control. However, these vein
segments showed intimal and smooth muscle cell (SMC) Special staining of vein segments
hypertrophy, increase in intervening connective tissue,
fragmentation of elastic tissue and changes of phlebo- Vein segments preserved in 4% formalin were used for
sclerosis from old age.7e12 Hence, in this study, arm and histological analysis and comparison with special stains
neck veins were considered as an alternative control for namely,haematoxylinandeosin(H&E),Masson’sTrichrome
varicose veins,inaddition to GSV frombypassgrafts. andElastic VanGieson(EVG).
Table 1 Demographic features of venous tissue: The average age of the arm and neck vein group was closer to those of
varicoseveinsthanGSVfrombypass.Morepatientsofthelattergroupshadco-morbiditiesthanthatofvaricoseveins.
Parametersmeasured Varicoseveins GSVfrombypass Armandneckveins
Numberofpatients 26 11 13
Age(range)* 51(29e81) 76(62e88) 63(22e82)
Smokers 9 8 4
Coronaryarterydisease 0 3 3
Diabetes 0 3 2
Peripheralvascular
disease 0 5 8
Hypertension 0 5 3
Statin 0 2 3
Aspirin 0 8 4
*P<0.01.
VaricoseveinversusGSVfrombypass,by1-wayANOVAwithBonferroni’sMultipleComparisonTest.
756 B. Aravind etal.
Figure1 Varicoseveinhistology:comparisonwithGSVandbasilicveins.(aed)RepresentativeEVG-stained(a,c)andTrichrome
stained(b,d)varicoseveinsections,showingconnective tissueinfiltration (I)intunicamedia(a,b).Smoothmusclecellhyper-
trophyisseenaswell-definedinnercircular(SMC)andouterlongitudinalbundles(d),withinternalelasticlamina(IEL)inappo-
sition to endothelium lining the lumen(c, d). (e,f) Bypass GSV, showinga well-defined circular smooth muscle cell (SMC) layer
stainedredwithTrichrome(f)withscatteredstainingintunicaintima.EVGstaining(e)demonstratesdisruptedexternalelastic
lamina (EEL) stained black. The well-defined internal elastic lamina (IEL) highlights the intimal hyperplasia (distance between
luminal endothelium and IEL). (g,h) Basilic (arm) vein, withTrichrome stained slide (h), it was possible to clearly differentiate
between connective tissue and smooth musclecell (SMC) layer.EVG-stained section (g) shows well-demarcated external elastic
lamina(EEL)andinternalelasticlamina(IEL).(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderis
referredtothewebversionofthisarticle).
TIMPRegulateMatrix Turnoverin VaricoseVeins 757
serum (Serotec, Oxford, UK). Tissue sections were incu-
Table 2 Vein wall thicknesses: Measured as distance
bated with mouse monoclonal anti-human antibody recog-
between intimal lining and external elastic lamina, this
nising MMP-2, TIMP-2, MT1-MMP and TIMP-3 (Chemicon,
compared paired segments of proximal and distal varicose
Chandlers Ford, Hampshire, UK). Biotinylated horse anti-
veins.
mouse antibody (Chemicon, Chandlers Ford, Hampshire,
Veintype Rangeof Median(mm) UK) followed by streptavidin biotin with HRP (horseradish
thickness(mm) peroxidase) was used, and sections were developed with
Varicoseveins 271e1841 755 DAB (3,30-di aminobenzidine) and counterstained with
Proximalvaricosevein 406e1841 767* haematoxylin.
Distalvaricosevein 271e1588 574 Colourimageswerecaptured(at4(cid:2)magnificationusing
GSVfrombypass 274e1114 848** a JVC KY-F55BE) using a video camera (Victor Company of
Armandneckveins 115e758 308 Japan, Tokyo, Japan) connected to the microscope and
projected to a computer screen (GATEWAY 2000, North
*PZ0.07,versusdistalvaricoseveinsegments.
**P<0.001versusvaricoseveinsandP<0.01versusarmand Sioux City, USA). The immunopositive areas were sampled
for red, blue, green, hue, saturation and intensity using
neckveins.
AnalySIS software (Soft Imaging Software GmbH, Munster;
Germany). Minimum and maximum values were further
assignedby the softwarefor each ofthe characteristics of
H&E stained slides were used for screening of sections,
the identified positive pixels. For uniformity, these
and for cross-referencing with immunostained slides. EVG
measurementswererepeatedin10projectionseachonfive
specialstainshowedthesmoothmusclecelllayerasredand
different tissue immunostained sections. Bland Altman
elastinasblack,helpingdelineationofelasticlaminaonthe
analysis was performed for inter-observer variability (bias
vein wall aiding measurement of vein wall thickness. Mas-
(cid:3)13.3SD 8.9, r2 Z0.02, PZNS).
son’sTrichromestainedconnectivetissuegreen,whilethe
For quantitative analysis of immunostained sections,
smooth muscle cell layer was stained red/purple, and was
the software binarises the immunopositive areas to grey
usedinthevisualassessmentofchangesinarchitecture.
scale, accounts for erosion and dilatation of pixels and
Staining techniques used in this study were used and
provides the area of immunopositivity in the defined area
standardised in Charing Cross Hospital, Pathology Depart-
of interest as a percentage. The mean of all such reading
ment, London. Tissue collected in formalin were set in
paraffin blocks and cut into 4 mm sections. Routine H&E on a given section of vein was considered as %
immunopositivity.
staining was performed and Masson’s Trichrome with
ChromazoneRed(SurgipathEuropeLimited)andTrichrome
mixture(2partsof1%ChromazoneRedin1%aceticacid,4 Statistical analysis
parts 1% phosphomolybdic acid and 6 parts 1% Light Green
in1%aceticacid).EVGstainingwasperformedwithMillers Data were analysed using the Graph Pad Prism software
elasticstainfollowedbyVanGiesonstain[saturatedpicric package (Graph Pad Software, CA, USA). For 2 groups of
acidand1% acidfuchsin (SurgipathEurope Limited)]. data, Wilcoxon rank test was used to compare non-para-
metrically distributed paired data, and paired-test for
Vein wall thickness normally distributed paired data. ManneWhitney test was
usedforcomparisonofnormallydistributedunpaireddata.
For3ormoregroupsofdata,one-wayanalysisofvariance
Thickness of the vein wall was considered as the index of
(ANOVA)withBonferronipost-testformultiplecomparisons
morphology as it was easily measured, and reproducible
(seeFig.2aeb).Itwasmeasuredasthedistancebetween was used to compare different groups of normally distrib-
uted data. For non-parametrically distributed data, Krus-
endothelium and external elastic lamina using AnalySIS
kaleWallis ANOVA with Dunn’s post-test for multiple
software (Soft Imaging Software GmbH, Munster;
comparisons was used. Correlations were assessed using
Germany).Theadventitiawasexcludedtoreducepotential
Spearman’s co-efficient for non-parametrically distributed
inconsistency resulting from dissection close to the vein
data. Bland Altman analysis was performed to analyse
wall at surgery. Intimal thickness was measured using the
inter-observer variation.
same principle, as the distance between endothelium and
internal elastic lamina. To further correlate the extremes
ofthicknessesofvaricoseveinwall,theyweresub-divided Results
into atrophic, which was inside the 25th quartiles and
hypertrophic, which was beyond the 75th quartile. Bland Patient demographics
Altman analysiswasperformed for inter-observer variance
(bias19.9SD 15.2 mm, r2 Z0.14, PZNS).
Patientswithvaricoseveinsweresignificantlyyoungerthan
thosewhocontributedGSVfrombypassoperations(Table1),
ImmunohistochemistryforMMPandTIMPandimage
butsimilartothatofarmandneckveingroup.Therewasno
analysis significantgenderdifference(byc2test)betweenthegroups
(varicosevein11male,15female;GSVfrombypass7male4
Immunohistochemical staining was performed for MMP-2, female;armandneckvein10male,3female).
MT1-MMP, TIMP-2, and TIMP-3 on paraffin sections. Non- Venoustissueharvestedfromnon-varicosecontrolgroups
specific antibody sites were blocked with normal horse wasvariablyexposedtodiabetes,statins,andaspirinunlike
758 B. Aravind etal.
Figure2 Measurementofveinwallthickness.Definitionsofveinwallthicknessandmeasurements.(a)Theouterboundaryfor
measurement of vein wall thickness was defined as external elastic lamina, which limits the tunica media. (b) The distance
betweentheintimalendotheliumandtheexternalelasticlamina,whichspansthetunicaintimaandtunicamedia,wasmeasured
astheveinwallthickness.Anaverageof6measurementspersectionwereobtained,andameanwascalculated.(c)Thickness
(mm)ofproximalanddistalvaricoseveins(nZ26pairedsegments),comparedtobypassGSV(nZ11),andtoarmandneckveins
(nZ13).Veinsweresub-dividedfurtherintoatrophic(thickness<507mmbasedon25thpercentileofvaricoseveinwallthickness)
andhypertrophic(>1022mmbasedon75thpercentileofvaricoseveinwallthickness).Shadedareashowsregionbetween25thand
75th percentiles. (dee) The variation in relative thickness in a representative pair of proximal (d) and distal (e) segments of
varicosevein,showninsectionsatsamemagnification,stainedwithhaematoxylin/eosin.
TIMPRegulateMatrix Turnoverin VaricoseVeins 759
Figure3 MMP-2expressioninvaricoseveins.Leftpanel:Proximalvaricoseveinsegments(a)werestainedforMMP-2,andcompared
withGSVsegmentsfromarterialbypassoperations(b)andwitharmveins(c).Rightpanel:CorrespondingTrichromestainedslidesare
shownfordefiningsmoothmuscle(red/purple)andconnectivetissue(green).(d)PercentageimmunopositivityofMMP-2expressionin
proximalanddistalvaricoseveinsegments(nZ19),comparedwithGSVfrombypass(nZ10),andneckandarmveins(nZ13).Data
wereanalysedby1-wayANOVA(KruskaleWallis)withDunn’spost-hoctestformultiplecomparisons.Barsindicatemedianvalues.(For
interpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle).
760 B. Aravind etal.
Figure 4 MT1-MMPexpressioninvaricoseveins.Leftpanel:Proximalvaricoseveinsegments(a)werestainedforMT1-MMP,and
comparedwitharterialbypassGSVsegments(b)andwitharmveins(c).Rightpanel:CorrespondingTrichromestainedslidesareshown
fordefiningsmoothmuscle(red/purple)andconnectivetissue(green).(d)PercentageimmunopositivityofMT1-MMPexpressionin
proximalanddistalvaricoseveinsegments(nZ20),comparedwithGSVfrombypass(nZ10),andneckandarmveins(nZ13).Data
wereanalysedby1-wayANOVA(KruskaleWallis)withDunn’spost-hoctestformultiplecomparisons.Barsindicatemedianvalues.(For
interpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle).
TIMPRegulateMatrix Turnoverin VaricoseVeins 761
Figure 5 TIMP-2 expression in varicose veins. Left panel: Proximal varicose vein segments (a) were stained for TIMP-2, and
comparedwitharterialbypassGSVsegments(b)andwitharmveins(c).Rightpanel:CorrespondingTrichromestainedslidesare
shownfordefiningsmoothmuscle(red/purple)andconnectivetissue(green).(d)PercentageimmunopositivityofTIMP-2expression
inproximalanddistalvaricoseveinsegments(nZ19),comparedwithGSVfrombypass(nZ10),andneckandarmveins(nZ13).
Data were analysed by 1-way ANOVA (KruskaleWallis) with Dunn’s post-hoc test for multiple comparisons. Bars indicate median
values.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle).
762 B. Aravind etal.
thoseinvaricosegroup(Table1).Whenvaricoseveinswere TIMP-3 in the smooth muscle of the tunica media of vein
classifiedforclinicalseverity(CofCEAPscoringdevisedby wall. MMP-2, MT1-MMP and TIMP-2 were, however, not
PorterandMoneta),82%belongedtoC2(simplevaricosevein identifiable in the intervening connective tissue of tunica
only), with the remainder classified as C4 (skin pigmenta- mediaortheendotheliumoftheveinwall(Figs.3e5aec).
tion)andC6(venousulcer). Vein segments also showed TIMP-2 immunopositivity on
endothelium lining the blood vessels in tunica adventitia,
Histology of varicose versus non-varicose veins butnotthoseliningthelumenofthevein.TIMP-3,inaddi-
tiontoitspresenceinsmoothmuscle,wasidentifiedonthe
Themostconspicuousandconsistentchangeswerenotedin elastictissueoftheveinwall.Thiswasconfirmedbycross-
tunicamedia of varicose veins.Histology sections of all 26 referencingwithEVG-stainedcorrespondingsectionswhere
patients showed loss of architecture of this layer due to theelastictissuewasstainedasblack(Fig.6aec).Therewas
infiltration of connective tissue (green in Trichrome). nodifferenceintheimmunohistochemicalstainingpattern
Tunica intima showed hypertrophy due to similar infiltra- between the proximal or distal varicose veins or between
tionofconnectivetissueinnearlyhalfofthesections(12of differenttypesofnon-varicoseveinsections.
26). In all the sections, there was extensive disruption of Quantitative analysis showed a significantly higher
external elastic lamina, the outer limiting layer between expressionofMT1-MMPandTIMP-2invaricoseveinsthanarm
tunica media and adventitia. The internal elastic lamina, andneckveincontrols(Figs.4e6dand;Table3).Asimilar
however, wasrelatively well-preserved (Fig. 1aed). trend was observed for MMP-2, but was not statistically
Theabovechangesinthevaricoseveinscontrastedwith significant. TIMP-2 expression was significantly higher in
those observed in GSV from bypass surgery, which showed varicoseveinscomparedtoneckandarmveins.Expression
predominant smooth muscle hypertrophy of tunica media ofTIMP-3wascomparablebetweenallthreegroups.
(Fig.1eef).Thearmandneckveins,however,showedno The expression of the analysed MMP and TIMP were
such changes of tunica intima or tunica media with well- similar in proximal and distal varicose vein segments
preservedelasticlaminaandgeneralarchitectureinallthe although there was a trend of higher expression of TIMP-2
sections(Fig. 1gehfor armandneck veins). (4.3% versus 1.3%) and TIMP-3 (0.94% versus 0.41%) in
proximal segments.
Measurement of vein wall thickness
Comparison of expression between hypertrophic
and atrophic varicose vein segments
GSV from bypass was thicker than both varicose and arm
and neck vein segments (Table 2, Fig. 2c and e). Compar-
Expression of the protease MT1-MMP was higher in the
isonmadebetweenpairedproximalanddistalvaricosevein
thinner, atrophic segments compared to the hypertrophic
segments showed that the distal segments were thinner
than proximal segments (Fig. 2 cee and Table 2). Further, varicoseveinsegments.However,inhypertrophicsegments,
theexpressionofproteaseinhibitorsTIMP-2andTIMP-3was
there were more atrophic segments (42%, thickness
<507 mm based on 25th percentile of mean varicose vein higherthanatrophicvaricoseveinsegments(Table4).
wall thickness) among these distal varicose vein segments
thanintheirproximalcounterparts(Fig.2c).Theproximal Discussion
segments, on the other hand, showed predominantly
hypertrophic changes (27%, thickness >1022 mm based on
The key histologicalfindinginthisstudy was infiltrationof
75thpercentile).
connective tissue in the tunica media of varicose veins.
There was no histological difference in the constitution
Most of the change was noted in tunica media, especially
of tunica media between distal and proximal or atrophic
thesmoothmuscles,whichalsolocalisedtheanalysedMMP
and hypertrophic varicose vein segments. There was infil-
and TIMP, emphasising their importance in the matrix
tration of connective tissue in all these sections. The
turnover. Moreover, the split measurement suggested that
relative thickness of the tunica media is proportionately
the thickness of tunica media determine the overall
lowerindistal andatrophic.
morphology of the veinwall.
Intimalthicknesswasmeasuredtofindthecontribution
Quantitative analysis in previous studies have showed
of tunica intima and media separately to the observed
elevatedamountofcollagenwhileelastinwasdegradedand
thickness of the vein wall, and was expressed as reduced on varicose vein wall.15e17 The accumulation of
a percentage of total vein wall thickness. Tunica intima
connective tissue on varicose vein wall in our study could
accounted for 5.6% of the total thickness of the proximal
account for this elevated amount of collagen. TIMP-2
veinwalls,whileitwas7.6%fordistalsegments.Similarly,
expression of varicose veins was found to be significantly
4.3% of the wall thickness of GSV from bypass was its
higher than in arm and neck veins, with both TIMP-2 and
intima, while arm and neck vein, had no significant
TIMP-3expressionbeinghigherthaninGSVharvestedfrom
measurablecontribution from theintima.
bypass.Theelevatedinhibitorwillfavourtheaccumulation
of matrix components, like collagen, by suppressing the
MMP and TIMP expression in varicose and non-
proteases. On the other hand, MT1-MMP, a protease, was
varicose veins significantlyelevatedinthevaricoseveinsegmentsaswell.
However,itispossiblethatMT1-MMP,beingaweakcollage-
Immunohistochemical staining of the vein wall showed nase, with known ability to degrade collagen, fibronectin,
predominant localisation of MMP-2, MT1-MMP, TIMP-2, and laminin,fibrinandaggrecans,doesnothaveadominantrole
Figure 6 TIMP-3 expression in varicose veins. Left panel: Proximal varicose vein segments (a) were stained for TIMP-3, and
comparedwitharterialbypassGSVsegments(b)andwitharmveins(c).Rightpanel:CorrespondingEVG-stainedslidesareshown,
to demonstrate expression of TIMP-3 predominantly in elastic tissue of internal and external elastic lamina. (d)Percentage
immunopositivity of TIMP-3 expression in proximal and distal varicose vein segments (n Z 20), compared with GSV from bypass
(n Z 9), and neck and arm veins (n Z 12). Data were analysed by 1-way ANOVA (KruskaleWallis) with Dunn’s post-hoc test for
multiplecomparisons.Barsindicatemedian.