Ultrastructural Pathology, 34:226–231, 2010Copyright & Informa Healthcare USA, Inc.ISSN: 0191-3123 print/1521-0758 onlineDOI: 10.3109/01913121003725721

Fibrin Mimics Neutrophil ExtracellularTraps in SEM

ABSTRACT

Neutrophil extracellular traps (NETs) are extracellular web-like structures produced by activatedpolymorphonuclear neutrophils. NETs kill bacteria extracellularly, but their role in humanpathology remains largely unclear. One possible way of studying NETs is through the SEMapproach. However, web-like structures observed with SEM in sites of inflammation have beeninterpreted either as NETs or as fibrin. Thus, the question arises whether a reliable SEM dis-crimination between NETs and fibrin is at all possible. NET samples were collected as purulentcrevicular exudate from periodontal pockets. DNase-digested controls for SEM were employed todemonstrate the DNA backbone and immuno-staining for confocal laser scanning microscopy wasused to show the citrullinated histones of NETs. Blood clot samples were treated in the same wayas the exudate samples to demonstrate that fibrin and fibrinolysis can mimic NETs and DNAdigestion, respectively. No discrimination between fibrin and NETs based on morphologicalcriteria in SEM was possible. Furthermore, only a vague distinction between DNA digestion andfibrinolysis could be made. These findings unambiguously indicate that the discriminationbetween NETs and fibrin by means of SEM is untrustworthy for samples of inflammatory exudate.

Keywords: DNase, extracellular web-like structures, fibrinolysis, histone citrullination

Neutrophil extracellular traps (NETs) are extracellularweb-like structures produced by activated polymor-phonuclear neutrophils (PMNs) [1]. NETs kill gram-positive and gram-negative bacteria extracellularly[1, 2], prevent them from spreading and colonizing

new host surfaces [1, 3], and ensure a high localconcentration of antimicrobial agents able to degradevirulence factors [1, 2]. NETs mainly contain DNA,histones, and neutrophil elastase. The antimicrobialactivity of DNA [4] and histones [5] is well established.

Wolf Dietrich Krautgartner, PhDand Michaela KlappacherDepartment of Light & ElectronMicroscopy, Organismic Biology,University of Salzburg, Salzburg,Austria

Matthias Hannig, DMD, PhDDepartment of OperativeDentistry & Periodontology,Saarland University, Homburg,Germany

Astrid Obermayer, MSDepartment of Light & ElectronMicroscopy, Organismic Biology,University of Salzburg, Salzburg,Austria

Dominik Hartl, MD, PhD andVeronica Marcos, MDDFG Emmy-Noether Group,Immunodeficiency section and Cysticfibrosis section, Children’s Hospital

Ljubomir Vitkov, DMDDepartment of OperativeDentistry & Periodontology,Saarland University, Homburg,Germany and Private Practice,Straßwalchen, Austria

Received 28 December 2009; accepted 23 February 2010.Address correspondence to Ljubomir Vitkov, Mayburgerplatz 7, A-5204 Straßwalchen, Austria. E-mail: lvitkov@yahoo.comThe authors thank Andreas Zankl for the image processing.

226

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.

A characteristic feature of the NET is the histonecitrullination [6, 7]. Neutrophil elastase belongs to theoxygen-independent group of antimicrobial proteinsand acts through the digestion of bacterial proteins [8].Further, NETs also contain bactericidal substancessuch as lysozyme, bactericidal permeability increasingprotein, human peptidoglycan recognition protein S,and other neutrophil proteins [1, 2, 9].

NETs have been reported in only a few human diseases:appendicitis [1], preeclampsia [10], glomerulonephritis[11], and periodontitis [3]. Therefore, the NET-relatedpathology and the clinical implications of NETsremain to be addressed in future studies. Whereasimmuno-staining for fluorescent microscopy [1, 3, 6, 7]and highly specific cytochemical staining for TEM [12]have been reported, the NET identification with SEM isoften based on morphological criteria only, as originallyapplied for in vitro NETs. As all in vitro samplecomponents are known, the occurrence of web-likestructures other than NETs can be excluded. In contrastto the in vitro samples, human biopsies may containweb-like structures like fibrin, fibronectin, collagen, andcytoskeleton remnants. For that reason, the NETidentification with SEM, based on morphologicalcriteria, has been supplemented with a DNasedigestion as a negative control [3], relying on theability of DNase to selectively digest the DNAbackbone of NETs [1, 2]. Recently, SEM-observed, web-like, biofilm-associated structures in an animalinflammation model have been suggested to be NETs[13]. However, these structures showed pronouncedsimilarity to earlier reported SEM-observed, biofilm-associated [14] or inflammation-associated [15]structures, which had been suggested to be fibrin.Consequently, the questions arise: How reliable is theSEM identification of NETs based on morphologiccriteria and how reliable is the DNase digestion as aNET negative control?

The aim of this work was to examine the reliability ofthe differentiation between NETs and fibrin by meansof SEM.

METHODS AND MATERIALSPurulent crevicular exudate was collected from 8untreated diseased sites in 8 consecutive patients(4 male and 4 female) with periodontitis. Mean ageof the patients was 54 years, range 35–71 years. Allsites chosen for sampling had radiological evidence ofperiodontitis. The exclusion criteria were nectrotizingperiodontal diseases, abscesses of periodontium,periodontitis associated with endodontic lesions,systemic diseases, antibiotic therapy in the last6 months, and steroid and radiotherapy. The studyprotocol was approved by the local ethics committee

and informed written consent was obtained from eachpatient. The samples were collected with Aclar(Honeywell International, Morristown, NJ) strips cutto dimensions of 1.5/15 mm as earlier reported [16].Prior to sampling, the area was isolated with cottonrolls and gently dried using an air syringe. A strip wasinserted into the gingival crevice until mild resistancewas felt and left in place for 30 s. The strips werethen immediately transferred to either a plastic tubewith fixative or DNase. The first samples wereprocessed for scanning electron microscopy (SEM).The second ones were incubated in 300 mL of2000 U/mL DNase (DNase I recombinant, grade I,Roche Diagnostics, Vienna, Austria) at 378C for 30 minand then processed for SEM. DNase was dissolvedin 20 mM Tris–HCl buffer at pH 7.5 supplementedwith 1 mM MgCl2. The third samples were directlyprocessed for confocal laser scanning microscopy(CLSM).

Clinical examinations were performed prior to thesampling. Parameters evaluated were gingival redness,gingival swelling, and suppuration. Suppuration wasrecorded as purulent exudate either evident during theclinical inspection or apparent after the clinical test fordetecting the suppuration in periodontal inflammation[17], i.e., gently applying digital pressure (in a coronaldirection) to the gingival surface.

Blood clot samples were produced by swiping theAclar strips across the capillary blood droplets fromthe fingertips of the patients. Three samples werecollected from each patient. After 30 s adhesiontime the first samples were fixed; the second oneswere incubated in DNase and processed exactlyas described above; and the third samples wereincubated in the DNase buffer, but withoutDNase, and processed in the same way. Finally, allblood clot samples were processed for SEM asdescribed above.

SEM Analysis

After a fixation for 2 h with Karnovsky, postfixation ofthe samples was performed with 1% osmium tetroxide(buffered at pH 6.5 with 0.1 M sodium cacodylate)for a further 2 h. The postfixed samples weredehydrated in an ascending series of ethyl alcohol,critical-point-dried, and subsequently sputteredwith gold (circa 5 nm). The specimens were examinedin a scanning electron microscope ESEM XL30 (FEI,Philips, Eindhoven, Netherlands) operating at 20 kV.

CLSM Analysis

The samples were fixed with 4% paraformaldehyde inphosphate-buffered saline at pH 7.4 (PBS) for 2 h. The

Fibrin Mimics NETs in SEM

227

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.

fixed samples were washed with PBS and blocked(5% normal goat serum in PBS containing 0.5% bovineserum albumin and 0.5% Triton X-100). Subsequently,the samples were incubated with the primary antibodyanti-human citrullinated histone H3 (citrulline218116) antibody [CitH3] (ab5103, Abcam, Cam-bridge, UK), as recommended by the manufacturer.The primary antibody was detected with a secondaryantibody coupled to FITC (ab6717, Abcam). Finally,DNA was stained with the molecular probe pro-pidium iodide (PI) (Sigma, Schnelldorf, Germany).

The specimens were analyzed with a CLSM (ZeissLSM 510 meta UV, Carl Zeiss, Vienna, Austria).

RESULTSNETs were visualized in CLSM by staining their basiccomponents: (1) the DNA backbone and (2) thecitrullinated histone H3 (Figure 1). The quantities ofNETs varied considerably between the GCF samples.NETs also differed in density within each sample.Solitary bacteria or moderate numbers of dispersed

Figure 1. CLSM optical slices of 1.7 mm. Purulent crevicular exudate. Red: PI; Green: CitH3. (A) PI channel. Spread NETswith ensnared bacteria. (B) CitH3 channel. The NETs from (A). (C) Phase-contrast image. Most NETs and bacteria arediscernible. (D) Merged image of (A) and (B).

W. D. Krautgartner et al.

228

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.

bacteria were ensnared by NETs. PMNs at the stage ofinitial activation, i.e., only with CitH3 in the cytoplasm,were not observed in the purulent crevicular exudate.The second stage of neutrophil activation, i.e., with thetransition of CitH3 into nuclei, was evident in all PMNsin the purulent crevicular exudate. The stage of initialNETosis characterized by swollen nuclei and a lack ofdifference between eu- and heterochromatin wasfrequent in both GCF and purulent crevicular exudate.Spreading NETs were evident in all samples. Multi-tudes of dispersed bacteria were trapped by NETs.

Low magnification demonstrated the distribution of thepurulent crevicular exudate on the Aclar (Figure 2A).Fibrin from blood samples frequently showed verythick trunks, which forked into less thick branches(Figure 2B). In contrast, NETs reveal a characteristicweb-like appearance (Figure 2C). Frequently, bacteriawere ensnared by NETs. The DNase treatment causedthe complete digestion of NETs. However, some areas

of the blood clots revealed similar web-like fibrin fibers(Figure 2D). Higher magnification revealed thatnonoverlapping NET fibers had a diameter of approxi-mately 20 nm (Figure 3A). At similar magnification,some fibrin fibers showed a diameter of even lessthen 35 nm and strongly mimicked the appearanceof NET fibers densely covered by globular domains(Figure 3B).

The DNase treatment caused the completedisintegration of all fibers in the purulent crevicularexudate samples. Partial fibrinolysis, varying betweenmoderate and advanced, was observed in all bloodclot samples incubated for 30 min in DNase or inDNase buffer. Due to the fibrinolysis, the DNase-treated and the DNase buffer-treated blood clotsamples strongly mimicked the DNase-treatedpurulent crevicular exudate samples. However, insome areas of the blood clot samples incubated for30 min in DNase or in DNase buffer, fibrin remnants of

Figure 2. SEM. (A) Characteristic appearance of NETs in a sample of purulent crevicular exudate. (B) Blood clot sample fixedafter 30-s adhesion. Fibrin trunks with a diameter of 100 nm or above dominate the appearance. Characteristically, the fibrintrunks fork into less thick branches. (C) Purulent crevicular exudate. The NET web consists of relatively uniforms fibers.Some bacteria are trapped by NETs. (D) A blood clot sample. The fibrin fibers form web-like structures, which mimic theappearance of the NETs.

Fibrin Mimics NETs in SEM

229

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.

partially lysed fibrin fibers reveal a characteristicstump-like appearance (Figure 3C). Frequently, areasof blood fibrin clots with NET-like appearance and amultitude of adhering leukocytes were evident.Erythrocytes were rarely observed (Figure 3D).

DISCUSSIONWhereas the immuno-staining of the citrullinatedhistone H3 (CitH3) [7, 6] or the immuno-staining ofhuman neutrophil elastase [1–3] verifies on theconfocal microscopy level that the observed web-likestructures are NETs, web-like structures in theinflammatory exudate observed with SEM might alsobe fibrin [14, 15]. Fibrin formation is common ininflamed sites [18, 19] and fibrin shows a web-likeappearance in SEM [20, 21]. Fibrin has also been

reported in inflammatory gingival exudate [22, 23].The lack of fibrin in the purulent crevicular exudatemight be due to excessive quantities of the NETelastase [3], which efficiently digests fibrin [24, 25].However, bleeding on probing is a main clinical signof periodontitis [17]. This indicates that the possibilityof contaminating the gingival purulent exudate sam-ples with blood, i.e., with fibrin, is high. Consequently,the occurrence of erythrocytes might be an indicationof blood extravasation and thus also of fibrin. NETshave been morphologically characterized in SEM byapproximately uniform fibers and meshes as well asby the presence of globular domains, whereas thefibrin was mainly of a dendritic appearance. A similarbranching appearance has also been reported forNETs [1, 3, 26]. The web-like appearance of both NETsand fibrin considerably varied between samples andcannot be a reliable criterion for discrimination on

Figure 3. SEM. (A) Purulent crevicular exudate. When NET fibers lie apart, they show a diameter of approximately 20 nm.Some of them carry globular domains. (B) A blood clot sample with the same magnification. Some fibrin fibers show a dia-meter of even less than 35 nm and strongly mimic the appearance of NET fibers densely covered by globular domains. (C)A blood clot sample incubated for 30 min in DNase. Partial fibrinolysis, varying between moderate and advanced, wasobserved in all samples. The samples incubated in the DNase buffer showed no differences to the DNase-treated samples.Characteristically, the partially lysed fibrin fibers reveal a stump-like appearance. (D) Overview of blood fibrin clot withNET-like appearance and a multitude of leukocytes adhering to it. Erythrocytes are not evident.

W. D. Krautgartner et al.

230

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.

an ultrastructural level. NETs in purulent exudate areregularly accompanied by a multitude of PMNs [3].However, the co-occurrence of leukocytes in thesample does not facilitate the assessment of the web-like structures, because the blood leukocytes adhere tofibrin clots, whereas erythrocytes remain on thesample only if they are entangled in the fibrin clot.

The partial fibrinolysis in blood clot samples treatedeither with DNase or the DNase buffer was due toplasmin and erythrocyte-coupled fibrinolytics [27]. Thisphenomenon strongly indicates that DNAse digestionused as a negative control in vitro [1] is not reliablyapplicable to exudate, since the fibrin lysis could bemisinterpreted as DNA digestion. The only suitablemorphological criterion for the NET identificationappears to be the occurrence of fibers with a diameterof 20 nm [3]. In contrast, the thinnest fibrin fibers,which we frequently observed by SEM, had a diameterof 35 nm, and previous SEM investigations reported40 nm [20]. However, considering sample shrinkagedue to dehydration and alterations of dimension due tosputtering, the fiber diameter cannot be used solely as asufficient criterion for discriminating between NETsand fibrin web-like structures. Furthermore, thedistribution of NET fibers within one sample variesconsiderably. Thus, no overall assessment of thesample can be made by the observation of NET fiberswithin a random sample section.

CONCLUSIONSDiscriminating between NETs and fibrin based onmorphological criteria is not possible with SEM.Further, the DNase digestion is not a reliable SEMnegative control for the occurrence of NETs in exudate,as fibrin occurrence and subsequent fibrinolysis canmimic the DNA digestion. Consequently, both criteriaconsidered sufficient for an in vitro examination arenot applicable for samples of inflammatory exudate.

Declaration of interest: The authors report no conflictsof interest. The authors alone are responsible for thecontent and writing of the paper.

REFERENCES

1. Brinkmann V, Reichard U, Goosmann C, et al. Neutrophilextracellular traps kill bacteria. Science. 2004;303:1532–1535.

2. Fuchs TA, Abed U, Goosmann C, et al. Novel cell death programleads to neutrophil extracellular traps. J Cell Biol. 2007;176:231–241.

3. Vitkov L, Klappacher M, Hannig M, Krautgartner WD. Extra-cellular neutrophil traps in periodontitis. J Periodontal Res.2009;44:664–672.

4. Mulcahy H, Charron-Mazenod L, Lewenza S. Extracellular DNAchelates cations and induces antibiotic resistance in Pseudomonasaeruginosa biofilms. PLoS Pathog. 2008;4:e1000213.

5. Kawasaki H, Iwamuro S. Potential roles of histones in host defenseas antimicrobial agents. Infect Disord Drug Targets. 2008;8:195–205.

6. Neeli I, Khan SN, Radic M. Histone deimination as a response toinflammatory stimuli in neutrophils. J Immunol. 2008;180:1895–1902.

7. Wang Y, Li M, Stadler S, et al. Histone hypercitrullinationmediates chromatin decondensation and neutrophil extracellulartrap formation. J Cell Biol. 2009;184:205–213.

8. Weinrauch Y, Drujan D, Shapiro SD, Weiss J, Zychlinsky A.Neutrophil elastase targets virulence factors of enterobacteria.Nature. 2002;417:91–94.

9. Cho JH, Fraser IP, Fukase K, et al. Human peptidoglycanrecognition protein S is an effector of neutrophil-mediated innateimmunity. Blood. 2005;106:2551–2558.

10. Gupta AK, Hasler P, Holzgreve W, Gebhardt S, Hahn S.Induction of neutrophil extracellular DNA lattices by placentalmicroparticles and IL-8 and their presence in preeclampsia. HumImmunol. 2005;66:1146–1154.

11. Kessenbrock K, Krumbholz M, Schonermarck U, et al. Nettingneutrophils in autoimmune small-vessel vasculitis. Nat Med.2009;15:623–625.

12. Krautgartner WD, Vitkov L. Visualization of neutrophil extra-cellular traps by TEM. Micron. 2008;39:367–372.

13. Reid SD, Hong W, Dew KE, et al. Streptococcus pneumoniae formssurface-attached communities in the middle ear of experimen-tally infected chinchillas. J Infect Dis. 2009;199:786–794.

14. Marrie TJ, Costerton JW. Scanning and transmission electronmicroscopy of in situ bacterial colonization of intravenous andintraarterial catheters. J Clin Microbiol. 1984;19:687–693.

15. Wang E, Simard M, Ouellet N, Bergeron Y, Beauchamp D,Bergeron MG. Pathogenesis of pneumococcal pneumonia incyclophosphamide-induced leukopenia in mice. Infect Immun.2002;70:4226–4238.

16. Vitkov L, Klappacher M, Hannig M, Krautgartner WD.Neutrophil fate in gingival crevicular fluid. Ultrastruct Pathol.2010;34:25–30.

17. Armitage GC. Clinical evaluation of periodontal diseases.Periodontology 2000. 1995;7:39–53.

18. Chung CL, Chen YC, Chang SC. Effect of repeated thoracenteseson fluid characteristics, cytokines, and fibrinolytic activity inmalignant pleural effusion. Chest. 2003;123:1188–1195.

19. Wait MA, Sharma S, Hohn J, Dal Nogare A. A randomized trialof empyema therapy. Chest. 1997;111:1548–1551.

20. Veklich Y, Francis CW, White J, Weisel JW. Structural studies offibrinolysis by electron microscopy. Blood. 1998;92:4721–4729.

21. Langer BG, Weisel JW, Dinauer PA, Nagaswami C, Bell WR.Deglycosylation of fibrinogen accelerates polymerization andincreases lateral aggregation of fibrin fibers. J Biol Chem.1988;263:15056–15063.

22. Talonpoika J, Paunio K, Soderling E. Molecular forms andconcentration of fibronectin and fibrin in human gingivalcrevicular fluid before and after periodontal treatment. Scand JDent Res. 1993;101:375–381.

23. Talonpoika JT, Soderling E, Paunio K. Characterization offibronectin and fibrin(ogen) fragments in gingival crevicularfluid. Scand J Dent Res. 1993;101:26–32.

24. Weitz JI, Landman SL, Crowley KA, Birken S, Morgan FJ.Development of an assay for in vivo human neutrophil elastaseactivity: increased elastase activity in patients with alpha1-proteinase inhibitor deficiency. J Clin Invest. 1986;78:155–162.

25. Kohno I, Inuzuka K, Itoh Y, et al. A monoclonal antibody specificto the granulocyte-derived elastase-fragment D species of humanfibrinogen and fibrin: its application to the measurement ofgranulocyte-derived elastase digests in plasma. Blood. 2000;

95:1721–1728.26. Brinkmann V, Zychlinsky A. Beneficial suicide: why neutrophils

die to make NETs. Nat Rev Microbiol. 2007;5:577–582.27. Ganguly K, Krasik T, Medinilla S, et al. Blood clearance and

activity of erythrocyte-coupled fibrinolytics. J Pharmacol ExpTher. 2005;312:1106–1113.

Fibrin Mimics NETs in SEM

231

Ultr

astr

uct P

atho

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f W

ashi

ngto

n on

08/

27/1

4Fo

r pe

rson

al u

se o

nly.