Comparative atomic force and scanning electronmicroscopy: an investigation on fenestrated endothelial cellsin vitro

F. BRAET,*1 W. H. J. KALLE,† R. B. DE ZANGER,* B. G. DE GROOTH,‡ A. K. RAAP,†H. J. TANKE† & E. WISSE**Laboratory for Cell Biology and Histology, Free University of Brussels (VUB),Laarbeeklaan 103, 1090 Brussels-Jette, Belgium†Department of Cell Biology, Laboratory for Cytochemistry and Cytometry,State University of Leiden (RUL), Wassenaarseweg 72, 2333 AL Leiden, The Netherlands‡Department of Applied Physics, Technical University of Twente, Twente PO Box 217,7500 AE Enschede, The Netherlands

Key words. Liver, endothelial cells, fenestrae, sieve plates, serotonin, ethanol,cytochalasin B, scanning electron microscopy, atomic force microscopy, imageanalysis.

Summary

Rat liver sinusoidal endothelial cells (LEC) contain fenestrae,which are clustered in sieve plates. Fenestrae control theexchange of fluids, solutes and particles between thesinusoidal blood and the space of Disse, which at its backside is flanked by the microvillous surface of the parenchy-mal cells. The surface of LEC can optimally be imaged byscanning electron microscopy (SEM), and SEM images canbe used to study dynamic changes in fenestrae by comparingfixed specimens subjected to different experimental con-ditions. Unfortunately, the SEM allows only investigationof fixed, dried and coated specimens. Recently, the use ofatomic force microscopy (AFM) was introduced for analysingthe cell surface, independent of complicated preparationtechniques. We used the AFM for the investigation ofcultured LEC surfaces and the study of morphologicalchanges of fenestrae. SEM served as a conventionalreference.

AFM images of LEC show structures that correlate wellwith SEM images. Dried-coated, dried-uncoated and wet-fixed LEC show a central bulging nucleus and flat fenestratedcellular processes. It was also possible to obtain heightinformation which is not available in SEM. After treatmentwith ethanol or serotonin the diameters of fenestraeincreased (�6%) and decreased (ÿ15%), respectively. The

same alterations of fenestrae could be distinguished bymeasuring AFM images of dried-coated, dried-uncoated andwet-fixed LEC. Comparison of dried-coated (SEM) and wet-fixed (AFM) fenestrae indicated a mean shrinkage of 20% inSEM preparations. In conclusion, high-resolution imagingwith AFM of the cell surface of cultured LEC can beperformed on dried-coated, dried-uncoated and wet-fixedLEC, which was hitherto only possible with fixed, dried andcoated preparations in SEM and transmission electronmicroscopy (TEM).

Introduction

Sinusoidal liver endothelial cells (LEC) line the hepaticsinusoids and are characterized by the presence of multiplefenestrae arranged in sieve plates. Fenestrae lack adiaphragm and there is no basal lamina underneath theendothelium. The fenestrae lumina constitute an openconnection between the sinusoidal lumen and the space ofDisse (Wisse, 1970; Wisse et al., 1985). One role of thefenestrae is filtration or sieving of the plasma (Fraser et al.,1978; Naito & Wisse, 1978; De Zanger & Wisse, 1982),allowing only fluid and particles smaller than the fenestraeto reach the parenchymal cells or to leave the space of Disse.By using scanning electron microscopy (SEM) it has beendemonstrated that several agents modulate fenestrae dia-meter and number, in vivo and in vitro (for reviews see Arias,1990; Smedsrød et al., 1994; Fraser et al., 1995). Recently, itwas demonstrated that fenestrae are surrounded by a fila-mentous, fenestrae-associated cytoskeleton ring. Serotonin

Journal of Microscopy, Vol. 181, Pt 1, January 1996, pp. 10–17.Received 15 May 1995; accepted 8 July 1995

10 # 1996 The Royal Microscopical Society

1Filip Braet is an Aspirant of the National Fund for Scientific Research—Belgium.

Correspondence: Filip Braet, Laboratory for Cell Biology and Histology, Free

University of Brussels (VUB), Laarbeeklaan 103, 1090 Brussels-Jette, Belgium.

Email: filipbra@cyto.vub.ac.be.

reduced the diameter for these fenestrae-associated cyto-skeleton rings by 20%, whereas ethanol treatment enlargedthe diameter by 5% (Braet et al., 1995).

Since its invention by Binnig et al. (1986), atomic forcemicroscopy (AFM), also called scanning force microscopy(SFM), has been used to image a wide variety of samples (forreviews see Hansma & Hoh, 1994; Morris, 1994). Dried,wet-fixed and living cells have formed three major subjects ofinvestigation over recent years (Haberle et al., 1991;Henderson et al., 1992; Kasas et al., 1993; Barbee et al.,1994). One of the advantages of AFM over SEM is the factthat samples require no coating, no vacuum is needed andelectrons are avoided (Morris, 1994). In addition, the AFMcan be successfully operated in an aqueous environment(Oberleithner et al., 1993; Beckmann et al., 1994; LeGrimellec et al., 1994; Schoenenberger & Hoh, 1994). Itwas demonstrated that AFM scanning under wet conditionsprovides high-resolution imaging of biological samples(Henderson et al., 1992; Radmacher et al., 1992). Evidencefor the successful application of AFM for biological imagingcomes from studies in which biological samples werepreviously imaged by SEM or transmission electron micro-scopy (TEM) (Horber et al., 1992; Kordylewski et al., 1994).

The above mentioned advantages of AFM prompted us tostudy LEC under different experimental circumstances, invitro, by AFM. The aim of this study was to investigatesurface and morphological changes of LEC and fenestrae byAFM. Fenestrae have a critical dimension of the order of150–200 nm, making it necessary to use high-resolutionmicroscopes, other than light microscopes. SEM served as areference method in the acquisition of AFM data, becauseSEM examinations of cultured LEC have already extensivelybeen described (Shaw et al., 1984; De Leeuw et al., 1990;Braet et al., 1994).

Materials and methods

Isolation, purification and culture of LEC

The isolation of LEC has been described earlier (Braet et al.,1994) as a modification of the method by Smedsrød & Pertoft(1985). Briefly, rat liver cells were suspended by collagenaseperfusion of the rat liver. Low-speed centrifugation of the cellsuspension results in the removal of the parenchymal cells.Purification of the supernatant, containing a mixture ofsinusoidal liver cells, was carried out on a two-step Percollgradient (25/50%). After centrifugation for 20 min at 900 g,the intermediate layer between the two density layers wasenriched in LEC. LEC purity was further enhanced duringselective adherence on plastic dishes, removing adherentcells, such as Kupffer cells. The LEC were cultivated in 24-multiwell plates on thermanox coverslips (N.V. Technologies,; 12 mm, Belgium) previously coated with collagen as asubstrate for the culture of LEC (Collagen-S, Boehringer

Mannheim, Belgium). Endothelial cell culture mediumconsisted of RPMI-1640 with 2 mM L-glutamine, 100 U/mLpenicillin and 100�g/mL streptomycin. The purity of theLEC cultures was examined by SEM (Braet et al., 1994), andwas estimated to be >95%. Less than 5% of the cells weredevoid of fenestrae.

Alterations of sinusoidal LEC fenestrations

LEC were cultured for 8 h and treated with 2 g/L ethanol99.8% (v/v) for 90 min (Braet et al., 1995). A second groupof LEC was treated with 1�g/mL serotonin (Sigma Chemi-cals, H5755, U.S.A.) for 7 min (Braet et al., 1995). In a thirdgroup LEC were treated with 10�g/mL cytochalasin B(Sigma Chemicals, C6762, U.S.A.) for 120 min (Steffan et al.,1987). Control LEC were incubated in culture mediumwithout ethanol or serotonin. In the case of the cytochalasinB experiment, control culture medium contained 0.4%dimethylsulphoxide but no cytochalasin B. After incubationwith ethanol, serotonin or cytochalasin B, LEC were preparedfor examination in SEM or AFM as described below.

Scanning electron microscopy

For the routine examination of LEC in SEM, cells were fixedwith 2% glutaraldehyde in 0.1 M Na-cacodylate buffer with0.1 M sucrose for 12 h. They were treated with filtered 1%tannic acid in 0.15 M Na-cacodylate buffer for 1 h and post-fixed with 1% osmium tetroxide in 0.1 M Na-cacodylate for1 h. SEM samples were dehydrated in graded ethanolsolutions, critical point dried and sputter coated with10 nm gold. The samples were examined with a PhilipsSEM 505 at an accelerating voltage of 30 kV. From eachexperimental variable 20 images were taken in randomlyselected fields. For automatic analysis of fenestrae, digitalimages were acquired at 20 000� magnification and a largespotsize (20 nm) to obtain images with low noise content.Digital images were processed on a Masscomp 5520Scomputer as previously described (Braet et al., 1994). Themagnification of the SEM was regularly calibrated with aPolaron 30 000 L/inch replica, with the specimen ineucentric position.

Atomic force microscopy

The AFM used in this study is the ExplorerTM (TopometrixTMX 2000, Darmstadt, Germany). The cantilever of the AFMwas positioned in the optical axis of a Leica invertedmicroscope with a home-made adaptation. The designallows movement of the AFM independently of the invertedmicroscope. Standard silicon nitride tips (Topometrix SFM-Probes, Ref. 1520-00, U.S.A.) with a spring constant of0.032 N and a 4-�m on 4-�m pyramidal base were used. LECcultured for 10 h on cover slips were scanned in contactmode, either dry in air or wet in 0.1 M Na-cacodylate buffer

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

ATOMIC FORCE MICROSCOPY AND ENDOTHELIAL CELLS 11

supplemented with 0.1 M sucrose. The first set of prepara-tions for AFM were prepared in identical manner to thesamples for SEM (dried-coated). A second set was preparedidentically, except that the cells were not sputter coated withgold (dried-uncoated). A third set of cells were only fixed withglutaraldehyde as described in the ‘scanning electronmicroscopy section’, and immediately examined with theAFM (wet-fixed).

In each experiment, 20 arbitrarily chosen areas offenestrae were scanned with a scan range of 10�m; fiveimages were recorded with a scan range of 20�m and 5�m.

Images, taken in sensor current mode or topographicmode, were analysed by the Topometrix Image AnalysisSoftware (Version V3.04). Images of preparations, scannedin sensor current mode, were displayed on the computermonitor. Photographs were taken directly from the computermonitor. The x–y–z calibration was regularly checked with acalibration grid.

Statistical analysis

All experiments were repeated three times. Statisticalanalysis was performed using the Mann–Whitney U-test.

Results

Scanning electron microscopy

After 10 h of culture, LEC are flattened and attached to thesubstrate and show a central, bulging nucleus, surroundedby clustered fenestrae in sieve plates (Fig. 1A,B). Singlefenestrae occur in peripheral zones of the cytoplasm (Fig.1A). At all magnifications, LEC show good preservation oftheir ultrastructural characteristics, i.e. fenestrae grouped insieve plates (Fig. 1A–C). Fenestrae have a diameter of213� 57 nm (n � 1000). LEC treated with ethanol showenlargement of the fenestrae diameter by approximately 6%,whereas serotonin treatment reduced the diameter by 15%(Table 1).

Atomic force microscopy

In a first group of experiments, LEC prepared for SEM wereexamined by AFM (Fig. 1D–F). The results obtained arecomparable with those obtained by SEM. Figure 1(D) showsan AFM picture at a high scan range (20�m� 20�m) of anLEC after 10 h culture. The LEC shows a round to ovalnucleus with a black shadowed side and smooth processes ofthe cytoplasm with well-preserved fenestrae. At highermagnification (Fig. 1E) the cytoplasm shows fenestrae,some arranged in sieve plates, others lying single at theborder of the cytoplasm. The AFM images also showedtip artefacts, i.e. pyramidal tip images on the cell surface(Fig. 1D,E). The cell surface at high magnification displayed

well-depicted fenestrae (Fig. 1F). In addition, the surface ofthe cell showed a certain roughness, probably reflecting thegold-sputter layer on top of the LEC. Measurements offenestrae diameter in these preparations, with the help ofAFM software, gave values of 211� 40 nm (n � 100).Ethanol treatment enlarges the fenestrae diameter by 6%,whereas serotonin reduces the diameter by 20% (Table 1).Height measurements between the lowest point of the sieveplates and the surrounding cytoplasm reveal that sieve plateslie approximately 200 nm lower than the surface of thenearby cytoplasm (Table 2).

The second group, consisting of dried-uncoated LEC,provided images with lower quality. Images frequentlyshowed streaks (Fig. 1G–I), an indication that material ofthe soft cell surface is being removed by the tip. The numberof streaks increased when the scanned area was hetero-genous in height (Fig. 1G,H). The typical bulging nucleus isstill visible. However, the image of this region is rich inartefacts, i.e. linear streaks along the nucleus (Fig. 1G). Onthe other hand, scanning of a smooth area of cytoplasmprovides images with a small number of streaks (Fig. 1I).Additionally, in highly corrugated areas, white bumpsappeared on the images (Fig. 1G,H). However, fenestraeand sieve plates could be recognized (Fig. 1G,H). Fenestrae incontrol LEC revealed an average diameter of 197� 40 nm(n � 100). Ethanol-treated LEC showed an average diameterof 222� 47 nm (n � 100) and serotonin-treated LECrevealed a diameter of 176� 35 nm (n � 100) (Table 1).Height measurements on sieve plates indicated that sieveplates were lying approximately 200 nm lower as comparedto the surrounding cytoplasm (Table 2).

The third group of cells examined with the AFM consists ofwet-fixed LEC (Fig. 2). Results obtained are essentiallycomparable with those obtained after critical point dryingand gold sputtering (Fig. 1A–C). Scanning under bufferreduces the number of artefacts, i.e. the number of pyramidaltip images decreased significantly (Fig. 2). In addition, areasaround the bulging nucleus and other high regions weredevoid of streaks (Fig. 2A). However, white bumps were morefrequently observed and were related to high regions in thesample (Fig. 2B,C). LEC could be clearly recognized at lowmagnification based on their fenestrated areas (Fig. 2A,C).The diameter of control, wet-fixed fenestrae was significantlylarger (269� 44 nm [n � 100]) as compared to thoseobtained with dried LEC (see Table 1). Differences infenestrae diameter between ethanol- and serotonin-treatedLEC could be discerned (Table 1). Measurements on sieveplates of control and ethanol- and serotonin-treated LECreveal that there is no significant difference in height (Table2). LEC treated with cytochalasin B show an enormouslyincreased number of fenestrae (Fig. 2C). At high magnifica-tion, the fenestrated cytoplasm reveals the presence of typicalsmall cytoplasmic unfenestrated dots which occur aftertreatment with the microfilament inhibiting drug (Fig. 2C,D).

12 F. BRAET ET AL.

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

Fig. 1. (A–C) Scanning electron images of rat liver sinusoidal endothelial cells at increasing magnification. (A) Micrograph of a well-spread liverendothelial cell after 10 h of culture: nucleus (N), sieve plates (arrowhead). White precipitates caused by sample preparation are observed ontop of the cell (arrow); scale bar�2.5�m. (B) High magnification of grouped fenestrae or sieve plates (arrowhead); scale bar�1�m. (C) Fenes-trae (arrowhead) are present in the cytoplasm; scale bar�500 nm. (D–F) Atomic force images of dried-coated liver endothelial cells at increas-ing magnification. (D) A liver endothelial cell after 10 h culture, with fenestrae clustered in sieve plates (arrowhead); nucleus (N), notice theblack shadow at one side of the nucleus (asterisk). Pyramidal tip images are depicted (arrow); scale bar�2.5�m. (E) Higher magnificationshows clearly grouped fenestrae (arrowhead) and pyramidal tip images (arrow); scale bar�1�m. (F) High-resolution image of a sieveplate, showing fenestrae (arrowhead) and the ‘snowy’ structure of sputter-coated gold layer; scale bar�500 nm. (G–I) Atomic force imagesof dried-uncoated liver endothelial cells at increasing magnification. (G) Overview of a spread liver endothelial cell after 10 h culture: nucleus(N), sieve plates (arrowhead). This low-quality image is rich in artefacts, i.e. linear streaks (arrow) in the scan direction (from left to right),white bumps (asterisk); scale bar�2.5�m. (H) Higher magnification showing grouped fenestrae (arrowhead). Notice also the presence of arte-facts; streaks (arrow) and white bumps (asterisk) were regularly observed; scale bar�1�m. (I) High magnification of a sieve plate with fenes-trae (arrowhead). Streaks (arrow) were observed and black shadowing areas (asterisk) without structural information were present; scalebar�500 nm.

ATOMIC FORCE MICROSCOPY AND ENDOTHELIAL CELLS 13

Discussion

During imaging with the AFM, the sample is scanned with asharp tip mounted on a soft cantilever with a low forceconstant. The force between the tip and the sample causes adisplacement of the cantilever. This displacement is mea-sured by a laser beam which is reflected from the cantileverand visualized as image information on a monitor. Weobserved details of LEC by AFM with a minimum ofpreparation steps (Fig. 2). Our recently described isolationprocedure of LEC (Braet et al., 1994) allows us to visualizecultured cells at high resolution within 1 day by AFM. In thepast, isolation of LEC by using elutriation and visualizingcultured LEC by SEM took days of work and great technicalexpertise was needed (De Leeuw et al., 1982; Steffan et al.,1987).

Dried-coated LEC observed by AFM (Fig. 1D–F) showdetails comparable to SEM (Fig. 1A–C). The glutaraldehyde–tannin–osmium fixation method induced precipitates on thecell surface, which were visualized in SEM as small whiteglobules on the cell surface (Fig. 1A). In AFM these structuresgive rise to artefactual, pyramidal tip images (Fig. 1D), due totip/specimen interactions on structures which have sharpercontours than the AFM tip. In other words, in this caseprecipitates visualize the AFM tip, rather than vice versa (Hoh

& Schoenenberger, 1994; Manne et al., 1994). In addition, theshadow of structural details was observed on AFM images; forinstance elevated regions, such as the nucleus, produce ablack shadow (Fig. 1D). Horber et al. (1992) postulated thatthis scanning artefact occurs mainly on strongly corrugatedsurfaces, causing a phase shift in the imaging system whenthe tip has to move up or down. The ‘snowy’ structure of thegold-sputter layer on top of the LEC was also observed (Fig.1F). This structure is seen in SEM only at its highestmagnification and resolution (40 000� and 6 nm, respec-tively).

The scanning of dried-uncoated LEC provides AFM imagesof different quality (Fig. 1G,H). At high scan range(20�m� 20�m) it was impossible to obtain images withoutstreaks. The streaks have the direction of the scanningmovement, from left to right (Fig. 1H). This type of artefact isprobably caused by an interaction of the tip and the soft cellsurface (Putman et al., 1993). However, by using a lowerscan rate on a small, smooth area of cytoplasm, the numberof streaks decreased significantly (Fig. 1H,I). A secondartefact in dried-uncoated LEC was related to high,corrugated regions, resulting in white bumps (Fig. 1G,H).This can be explained as smearing or lateral deformationwhich is most probably caused by the cantilever indentingthe cell membrane (Beckmann et al., 1994). This artefact

Table 2. Differential height values (nm) between the cytoplasm and sieve plates, with atomic force microscopy (AFM).

Treatment Dried-coated LEC Dried-uncoated LEC Wet-fixed LEC

Control 206:1� 41:2 (n � 25) 202:8� 37:9 (n � 25) 183:2� 47:9 (n � 25)Ethanol n.d. n.d. 165:6� 45:2 (n � 25)Serotonin n.d. n.d. 192:1� 57:5 (n � 25)

Results are expressed as mean� SD. No significant differences were found in the wet-fixed LEC between the control, ethanol- and serotonin-treated LEC at the 0.05 significance level (Mann–Whitney U-test [two-sided]). These data were obtained by examining parallel sectionsthrough the lowest point of the sieve plates and the highest point of the surrounding cytoplasm. n � number of sieve plates measured,n:d: � not determined.

Table 1. Comparison of morphometric values on fenestrae diameter (nm) obtained by scanning electron microscopy (SEM) and atomic forcemicroscopy (AFM).

SEM; AFM; AFM; AFM;Treatment dried-coated LEC dried-coated LEC dried-uncoated LEC wet-fixed LEC

Control 213� 57 (n � 1000) 211� 40 (n � 100) 197� 40 (n � 100) 269� 44 (n � 100)Ethanol 226� 70*** (n � 1000) 231� 41** (n � 100) 222� 47** (n � 100) 291� 43* (n � 100)Serotonin 182� 54*** (n � 1000) 175� 36** (n � 100) 176� 35** (n � 100) 202� 45** (n � 100)

Results are expressed as mean �SD. Notice the significant difference, indicated by asterisks, in diameter of fenestrae between the control,ethanol- and serotonin-treated LEC; *P < 0:01, **P40.002, ***P < 0:001 (Mann–Whitney U-test [two-sided]). n � number of fenestraecounted.

14 F. BRAET ET AL.

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

could not be resolved by decreasing the scan rate or forcebetween the tip and cell surface. Shadowing of structuraldetails was also observed, as described for dried-coated LEC(Fig. 1I). We suppose that the main problem for imagingdried-uncoated LEC is the softness of the samples. From theseresults, it was clear that a thin layer of gold improved theimage quality of dried LEC.

It was demonstrated earlier that scanning under wetconditions enables high-magnification imaging of biologicalsamples (Henderson et al., 1992; Radmacher et al., 1992), asconfirmed by our results on glutaraldehyde-fixed LEC (Fig. 2).It was reported that fixation increases the rigidity of cells(Beckmann et al., 1994; Hoh & Schoenenberger, 1994;Schoenenberger & Hoh, 1994), resulting in an improvedimage quality, as opposed to living cells. Artefactualshadowing (Fig. 2A,C) and white bumps (Fig. 2B,C) werepresent, as discussed earlier. However, the images weredevoid of streaks.

Fenestrae visualization has hitherto been restricted to SEM(Wisse et al., 1985) and TEM (Wisse, 1970), due to theirlimited size (150–200 nm in diameter). We consider, there-fore, the possibility of visualizing the fenestrae by AFM as animportant achievement, particularly because preparationsteps beyond fixation are not needed. AFM measurementsreveal an average fenestrae diameter of 269� 44 nm forwet-fixed LEC. In contrast, for dried-coated LEC an averagediameter of 213� 57 nm was found (Table 1). This

indicates, in accordance with preliminary observations(Wisse et al., 1985), that dehydration and critical pointdrying causes a considerable shrinkage at the level offenestrae. In addition, it is well known that 20–40%shrinkage occurs as a result of dehydration and criticalpoint drying (Boyde & Maconnachie, 1984). Recently, wewere able to reduce this shrinkage of fenestrae by 7–10%using the glutaraldehyde–tannin–osmium fixation method(Braet et al., 1994). However, the shrinkage of dried-coatedLEC fenestrae still remains at about 20%.

In addition, dynamic changes of fenestrae after treatmentwith ethanol and serotonin could be observed (Table 1), asshown in our previous studies (Wisse et al., 1980; Braetet al., 1994, 1995). Measurements on AFM images showedchanges in fenestrae diameter, i.e. a decrease after serotoninand an enlargement after ethanol, in accordance withmeasurements on SEM images (Table 1). Our AFM resultsalso confirm the increase in the number of fenestrae aftercytochalasin B treatment (Fig. 2C) and the presence of smallcytoplasmic dots lying in the highly fenestrated cytoplasm(Fig. 2D), previously observed by SEM (Steffan et al., 1987).

Cell height data were obtained by AFM (Table 2),indicating that sieve plates lie approximately 200 nm lowerthan the surrounding cytoplasm. No significant heightdifference was found in sieve plate depressions betweencontrol, ethanol- and serotonin-treated cells. Such datahave not been reported previously. Indeed, topographical

Fig. 2. Atomic force images of wet-fixed endothelialcells. (A) Low-magnification image of a wet-fixedliver endothelial cell, showing the central bulgingnucleus (N) and sieve plates (arrowhead); scalebar � 2:5�m. (B) Higher magnification showinga sieve plate (arrowhead) in surrounding cyto-plasm. Notice also the white bumps (asterisk),scale bar � 1�m. (C) Liver endothelial cells weretreated with 10�g/mL of cytochalasin B for 2 h,inducing a highly fenestrated cytoplasm (asterisk).White bumps (]) and shadowing of structures(arrow) are also present; scale bar � 1�m. (D)Higher magnification of the highly fenestratedcytoplasm after cytochalasin B treatment, whichshow fenestrae (arrow). Notice also the presenceof typical small unfenestrated dots (asterisk) thatoccur after treatment with the microfilament-inhi-biting drug; scale bar � 500 nm.

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

ATOMIC FORCE MICROSCOPY AND ENDOTHELIAL CELLS 15

measurements in SEM are technically complicated and aresubject to artefacts, such as shrinkage, whereas TEM givesheight information only in vertically orientated sections,which are sometimes difficult to obtain (Williams, 1977).

In conclusion: (1) dried-coated LEC are optimal samples forexamination with AFM and the results are comparable indetail and resolution with SEM; (2) images of dried-uncoatedLEC could be obtained, but contained a number of artefacts;(3) glutaraldehyde-fixation of LEC allows high-resolutionimaging of cultured LEC, without the necessity of furtherpreparative steps.

Acknowledgments

Mr T. Balink (Intersurface, The Netherlands) and Mr H. VanHumbeeck (Techmation, Belgium) are gratefully acknowl-edged for their technical support. We are also indebted toMrs Chris Derom for her photographic assistance. Thisinvestigation was partially supported by the ‘BelgianNational Fund for Scientific Research (NFWO)’, grant no.3.0053.92 and no. 3.0050.95; and partially by the ‘DutchOrganization of Scientific Research (NWO)’, grant no. 900-534-107 (WK, AKR) and grant no. 900-538-041 (HJT,BdG).

References

Arias, I.M. (1990) The biology of hepatic endothelial cell fenestrae.Progress in Liver Disease (ed. by F. Schaffer and H. Popper), pp. 11–26. Saunders, Philadelphia.

Barbee, K.A., Davies, P.F. & Lal, R. (1994) Shear stress-inducedreorganization of the surface topography of living endothelial cellsimaged by atomic force microscopy. Circ. Res. 74, 163–171.

Beckmann, M., Kolb, H.A. & Lang, F. (1994) Atomic forcemicroscopy of peritoneal macrophages after particlephagocytosis. J. Membrane Biol. 140, 197–204.

Binnig, G., Quate, C.F. & Gerber, C.H. (1986) Atomic forcemicroscope. Phys. Rev. Lett. 56, 930–933.

Boyde, A. & Maconnachie, E. (1984) Not quite critical point drying.Scanning Electron Microsc. 1, 71–75.

Braet, F., De Zanger, R., Baekeland, M., Crabbe, E., Van Der Smissen,P. & Wisse, E. (1995) Structure and dynamics of the fenestrae-associated cytoskeleton of rat liver sinusoidal endothelial cells.Hepatology, 21, 180–189.

Braet, F., De Zanger, R., Sasaoki, T., Baekeland, M., Janssens, P.,Smedsrød, B. & Wisse, E. (1994) Assessment of a method ofisolation, purification and cultivation of rat liver sinusoidalendothelial cells. Lab. Invest. 70, 944–952.

De Leeuw, A.M., Barelds, R.J., De Zanger, R. & Knook, D. L. (1982)Primary cultures of endothelial cells of rat liver. Cell Tissue Res.223, 201–215.

De Leeuw, A.M., Brouwer, A. & Knook, D.L. (1990) Sinusoidalendothelial cells of the liver: fine structure and function inrelation to age. J. Electron Microsc. Techn. 14, 218–236.

De Zanger, R. & Wisse, E. (1982) The filtration effect of rat liverfenestrated sinusoidal endothelium on the passage of (remnant)

chylomicrons to the space of Disse. Sinusoidal Liver Cells (ed. by D.L. Knook and E. Wisse), pp. 69–76. Elsevier, Amsterdam.

Fraser, R., Bosanquet, A.G. & Day, W.A. (1978) Filtration ofchylomicrons by the liver may influence cholesterol metabolismand atherosclerosis. Atherosclerosis, 29, 113–123.

Fraser, R., Dobbs, B.R. & Rogers, G.W.T. (1995) Lipoproteins and theliver sieve: the role of the fenestrated sinusoidal endothelium inlipoprotein metabolism, atherosclerosis, and cirrhosis. Hepatology,21, 863–874.

Haberle, W., Horber, J.K.H. & Binnig, G. (1991) Force microscopy onliving cells. J. Vac. Sci. Technol. 9, 1210–1213.

Hansma, H.G. & Hoh, J.H. (1994) Biomolecular imaging with theatomic force microscope. Ann. Rev. Biophys. Biomol. Struct. 23,115–139.

Henderson, E., Haydon, P.G. & Sakaguchi, D.S. (1992) Actinfilament dynamics in living glial cells imaged by atomic forcemicroscopy. Science, 257, 944–946.

Hoh, J.H. & Schoenenberger, C.A. (1994) Surface morphology andmechanical properties of MDCK monolayers by atomic forcemicroscopy. J. Cell Sci. 107, 1105–1114.

Horber, J.K.H., Haberle, W., Ohnesorge, F., Binnig, G., Liebich, H.G.,Czerny, C.P., Mahnel, H. & Mayr, A. (1992) Investigation of livingcells in the nanometer regime with the scanning forcemicroscope. Scan. Microsc. 6, 919–930.

Kasas, S., Gotzos, V. & Celio, M.R. (1993) Observation of living cellsusing atomic force microscope. Biophys. J. 64, 539–544.

Kordylewski, L., Saner, D. & Lal, R. (1994) Atomic force microscopyof freeze-fracture replicas of rat atrial tissue. J. Microsc. 173, 173–181.

Le Grimellec, C., Lesniewska, E., Cachia, C., Schreiber, J.P., de Fornel, F.& Goudonnet, J.P. (1994) Imaging of the membrane surface ofMDCK cells by atomic force microscopy. Biophys. J. 67, 36–41.

Manne, S., Zaremba, C.M., Giles, R., Huggins, L., Walters, D.A.,Belcher, A., Morse, D.E., Stucky, G.D., Didymus, J.M., Mann, S. &Hansma, P.K. (1994) Atomic force microscopy of the nacreouslayer in mollusc shells. Proc. R. Soc. Lond. 256, 17–23.

Morris, V.J. (1994) Biological applications of scanning probemicroscopies. Prog. Biophys. Molec. Biol. 61, 131–185.

Naito, M. & Wisse, E. (1978) Filtration effect of endothelialfenestrations on chylomicron transport in neonatal rat liversinusoids. Cell Tissue Res. 190, 371–382.

Oberleithner, H., Giebisch, G. & Geibel, J. (1993) Imaging thelamellipodium of migrating epithelial cells in vivo by atomic forcemicroscopy. Eur. J. Physiol. 425, 506–510.

Putman, C.A.J., van Leeuwen, A.M., de Grooth, B.G., Radosevic, K.,van der Werf, K.O., Van Hulst, N.F. & Greve, J. (1993) Atomicforce microscopy combined with confocal laser scanning micro-scopy: a new look at cells. Bioimaging, 1, 63–70.

Radmacher, M., Tillman, R.W., Fritz, M. & Gauv, H.E. (1992) Frommolecules to cells: imaging soft samples with the atomic forcemicroscope. Science, 257, 1900–1905.

Schoenenberger, C.A. & Hoh, J.H. (1994) Slow cellular dynamics inMDCK and R5 cells monitored by time-lapse atomic forcemicroscopy. Biophys. J. 67, 929–936.

Shaw, R.G., Johnson, A.R., Schultz, W.W., Zahlten, R.N. & Combes,B. (1984) Sinusoidal endothelial cells from normal guinea pigliver: isolation, culture and characterization. Hepatology, 4, 591–602.

16 F. BRAET ET AL.

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

Smedsrød, B. & Pertoft, H. (1985) Preparation of pure hepatocytesand reticuloendothelial cells in high yield from a single rat liver bymeans of Percoll centrifugation and selective adherence. J. Leukoc.Biol. 38, 213–230.

Smedsrød, B., De Bleser, P.J., Braet, F., Lovisetti, P., Vanderkerken, K.,Wisse, E. & Geerts, A. (1994) Cell biology of liver endothelial andKupffer cells. Gut, 35, 1509–1516.

Steffan, A.M., Gendrault, J.L. & Kirn, A. (1987) Increase in thenumber of fenestrae in mouse endothelial liver cells by alteringthe cytoskeleton with cytochalasin B. Hepatology, 7, 1230–1238.

Williams, M.A. (1977) Stereological techniques. Quantitative Meth-ods in Biology (ed. by A. M. Glauert), pp. 5–84. Elsevier/North-Holland Biomedical Press, Amsterdam.

Wisse, E. (1970) An electron microscope study of the fenestrated

endothelial lining of rat liver sinusoids. J. Ultrastruct. Res. 31,125–150.

Wisse, E., De Zanger, R.B., Charels, K., Van Der Smissen, P. &McCuskey, R.S. (1985) The liver sieve: considerations con-cerning the structure and function of endothelial fenestrae,the sinusoidal wall the and space of Disse. Hepatology, 5, 683–692.

Wisse, E., Van Dierendonck, J.H., De Zanger, R.B., Fraser, R. &McCuskey, R.S. (1980) On the role of the liver endothelial filter inthe transport of particulate fat (chylomicrons and their remnants)to parenchymal cells and the influence of certain hormones onthe endothelial fenestrae. Communications of Liver Cells (ed. by H.Popper, F. Gudat, L. Bianchi and W. Reutter), pp. 195–200. MTPPress, Basel.

# 1996 The Royal Microscopical Society, Journal of Microscopy, 181, 10–17

ATOMIC FORCE MICROSCOPY AND ENDOTHELIAL CELLS 17