Ultrastructural Analysis of Vascular Calcifications in Uremia

Ultrastructural Analysis of Vascular Calcifications inUremia

Georg Schlieper,* Anke Aretz,† Steven C. Verberckmoes,‡ Thilo Kruger,* Geert J. Behets,‡

Reza Ghadimi,† Thomas E. Weirich,† Dorothea Rohrmann,§ Stephan Langer,�

Jan H. Tordoir,¶ Kerstin Amann,** Ralf Westenfeld,*†† Vincent M. Brandenburg,*††

Patrick C. D’Haese,‡ Joachim Mayer,† Markus Ketteler,‡‡ Marc D. McKee,§§ andJurgen Floege*

Departments of *Nephrology and Clinical Immunology, §Urology, �Vascular Surgery, and ††Cardiology and †CentralFacility for Electron Microscopy, Rheinisch-Westfalische Technische Hochschule Aachen, Aachen, Germany;‡Laboratory of Pathophysiology, University of Antwerp, Antwerp, Belgium; ¶Department of Vascular Surgery,Maastricht University Hospital, Maastricht, Netherlands; **Department of Pathology, University of Erlangen,Erlangen, Germany; ‡‡Clinic for Nephrology, Coburg, Germany; and §§Faculty of Dentistry and Department ofAnatomy and Cell Biology, Faculty of Medicine, McGill University, Montreal, Quebec, Canada

ABSTRACTAccelerated intimal and medial calcification and sclerosis accompany the increased cardiovascular mortalityof dialysis patients, but the pathomechanisms initiating microcalcifications of the media are largely unknown.In this study, we systematically investigated the ultrastructural properties of medial calcifications frompatients with uremia. We collected iliac artery segments from 30 dialysis patients before kidney transplan-tation and studied them by radiography, microcomputed tomography, light microscopy, and transmissionelectron microscopy including electron energy loss spectrometry, energy dispersive spectroscopy, andelectron diffraction. In addition, we performed synchrotron x-ray analyses and immunogold labeling to detectinhibitors of calcification. Von Kossa staining revealed calcification of 53% of the arteries. The diameter ofthese microcalcifications ranged from 20 to 500 nm, with a core-shell structure consisting of up to three layers(subshells). Many of the calcifications consisted of 2- to 10-nm nanocrystals and showed a hydroxyapatite andwhitlockite crystalline structure and mineral phase. Immunogold labeling of calcification foci revealed thecalcification inhibitors fetuin-A, osteopontin, and matrix gla protein. These observations suggest that uremicmicrocalcifications originate from nanocrystals, are chemically diverse, and intimately associate with protein-aceous inhibitors of calcification. Furthermore, considering the core-shell structure of the calcifications,apoptotic bodies or matrix vesicles may serve as a calcification nidus.

J Am Soc Nephrol 21: 689–696, 2010. doi: 10.1681/ASN.2009080829

Cardiovascular mortality in patients with ESRD isdramatically increased when compared with thegeneral population.1 Cardiovascular calcificationsare major predictors of mortality in patients withESRD.2,3 Ectopic calcification is a tightly regulatedprocess that results from an imbalance betweenproteinaceous and small-molecule inhibitors (e.g.,fetuin, matrix gla protein, pyrophosphate) and in-ducers of mineralization (e.g., an elevated serumcalcium-phosphate product).4,5 In cell culturemodels, hyperphosphatemia leads to a transforma-

tion of vascular smooth muscle cells to osteoblast-like cells, which then express bone-specific pro-

Received August 15, 2009. Accepted December 23, 2009.

Published online ahead of print. Publication date available atwww.jasn.org.

Correspondence: Dr. Georg Schlieper, Department of Nephrologyand Clinical Immunology, RWTH University of Aachen, Pauwelsstrasse30, 52074 Aachen, Germany. Phone: ��49-241-8089530; Fax: ��49-241-8082446; E-mail: [email protected]

Copyright � 2010 by the American Society of Nephrology

CLINICAL RESEARCH www.jasn.org

J Am Soc Nephrol 21: 689–696, 2010 ISSN : 1046-6673/2104-689 689

teins.6 Thus, vascular wall calcifications in particular do notseem to result from passive precipitation of calcium and phos-phate but rather involve active cellular processes.

So far, it is generally assumed that the mineral deposited in thevascular wall has the physicochemical properties of hydroxyapa-tite, the main mineral phase of bone; however, for the ultrastruc-ture and systematic characterization of uremic cardiovascular cal-cifications, limited information is available in the literature. Theseprevious studies used different tissues and yielded different re-sults: Hydroxyapatite [Ca10(PO4)6OH2]—as seen in the skele-ton—was described as the sole mineral phase of uremic arterialcalcifications in two studies,7,8 whereas another study found bothbrushite and hydroxyapatite in calcifications of stenotic arterio-venous fistulas.9 In a rodent study of uremic calcification, whit-lockite, a magnesium-containing crystal [(Ca,Mg)3(PO4)2], wasdetected in addition to hydroxyapatite.10

Different mineral phases arising during uremic medial cal-cification may potentially be the result of different patho-mechanisms ultimately affecting dissolution of the calcifica-tions and the choice of therapy. We therefore performed asystematic ultrastructural analysis of uremic media calcifica-tions in patients with ESRD.

RESULTS

Prevalence and Predictors of Vascular CalcificationsMicrocalcifications in the media of iliac arteries were detected byvon Kossa staining in 53% of the samples. They were often dif-fusely dispersed in the tissue, but sometimes numerous cells werefound nearby (Figure 1). The von Kossa and hematoxylin andeosin stains of our iliac arteries did not show intimal calcificationor major atherosclerotic plaques. Moreover, samples with micro-calcifications did not exhibit major histologic differences com-pared with samples without calcification (data not shown).

Neither the presence nor the degree of medial calcificationsin iliac arteries was related to clinical or serum parameters(Table 1), but sample size limited the ability to detect differ-ences. Of note, only one patient had a history of diabetes. Plainpelvic x-rays for the assessment of iliac calcifications in 19 ofthe 30 patients revealed that eight (73%) of 11 patients with apositive von Kossa stain for calcification exhibited calcifica-tions on the plain pelvic x-ray, whereas only two (25%) of eightpatients with a negative von Kossa stain had calcifications byx-ray (P � 0.039). When comparing laboratory results of the10 patients who had iliac calcifications on plain x-rays with thenine patients who did not have iliac calcifications, serum cal-cium (2.56 � 0.16 versus 2.39 � 0.19 mmol/L; P � 0.049) andhemoglobin (129 � 13 versus 108 � 18 g/L; P � 0.008) wereincreased, whereas the other parameters indicated in Table 1did not differ significantly (data not shown).

Radiography and Microcomputed Tomography of IliacArtery CalcificationFaxitron radiography and a three-dimensional reconstruc-tion of mineral deposits in a uremic artery are shown inFigure 2. Calcification could be detected in three of fourarteries and seemed extensive within certain regions of theartery, but closer examination revealed that all of these cal-cification sites were not continuous and contained smallermineralized areas and microcalcifications, the majority ofwhich seemed to be extracellular by microscopy of serialsections (Figures 1 and 3).

Electron Microscopic Morphology of Iliac ArteryCalcificationsMicrocalcifications were found in the vicinity of both healthyand damaged vascular smooth muscle cells. Moreover, thesemineral deposits often occurred in areas rich in collagen and

A B E

C

D

Figure 1. Microcalcifications can be noticed in approximately half of the uremic iliac artery segments as visualized by light microscopy.(A and B) Sample without detectable calcification. (C through E) The von Kossa stain shows microcalcifications in the media. Some cellsseem to be completely surrounded by microcalcifications (arrows in E).


690 Journal of the American Society of Nephrology J Am Soc Nephrol 21: 689–696, 2010

vesicles (Figure 3), whereas an association with elastinfibrils or the inner elastic lamina could not be detected.

Transmission electron microscopy (TEM) revealed variousmorphologies of the microcalcifications (Figure 4). The diameterof the microcalcifications was mainly in the range of 20 to 500 nm,with only 12% of the microcalcifications exceeding 500 nm. Ap-

proximately one third of the microcalcifications exhibited a core-shell layered structure (32%), some with more electron-dense cal-cification in the core than in the shell (22%), whereas mostmicrocalcifications with core-shell layered structure lacked theelectron-dense material in their core (78%). In some samples, adouble-layered shell could be observed (Figure 4B). More rarely, amulti-shell structure with up to three subshells could be observed.When a core was present, the shells were arranged around a com-pact electron-dense spherical particle in its center. The spheruliticand layered nature was verified by a three-dimensional recon-struction from a tilt series (Figure 4E). Many of the calcified ag-glomerates consisted of spherical nanocrystals with grain sizes be-tween 2 to 10 nm (Figure 4F).

Chemical Composition of Iliac Artery CalcificationsUsing TEM techniques such as energy-dispersive x-ray spec-troscopy (EDX) and electron energy loss spectroscopy (EELS),all spectra demonstrated the presence of the elements phos-phorus, calcium, oxygen, and carbon in the calcified regions(Figure 5A). Control spectra taken from adjacent, noncalcifiedtissue always showed only amorphous carbon from the embed-ding material and/or biologic matrix, whereas different spectrawere recorded from the electron-dense particles (Figure 5A).

EDX and EELS spectra of the electron-dense particles andvesicles identified elements consistent with hydroxyapatite as achemical compound in microcalcifications of 10 samples,which was confirmed by crystallographic analysis using elec-tron diffraction (Figure 5B). In addition to the TEM analyses,synchrotron radiation analysis was applied to six iliac arteries,which allowed for a more systematic scanning of the samples.Interestingly, synchrotron radiation analysis revealed the pres-ence of whitlockite in addition to hydroxyapatite (Figure 5D).The analysis of six iliac arteries showed that two samples hadonly hydroxyapatite and one sample only whitlockite, whereasthree samples contained both hydroxyapatite and whitlockite.Because of the limited resolution of the synchrotron radiation

Table 1. Demographic data of patients undergoing kidney transplantation

ParameterAll Patients

(n � 30)

No Iliac ArteryCalcification

(n � 14)

Iliac ArteryCalcifications

(n � 16)P

Age (years; mean � SD) 49 � 10 50 � 9 49 � 12 0.709Male/female 17/13 7/7 10/6 0.491Diabetes 1 0 1 0.341Hypertension 26 13 13 0.351BMI (kg/m2; mean � SD) 24.3 � 3.2 24.5 � 2.4 24.1 � 3.9 0.692Dialysis mode (HD/PD) 26/4 12/2 14/2 0.886Dialysis vintage (years; mean � SD) 5.9 � 3.1 5.1 � 3.1 6.7 � 3.0 0.175Hemoglobin (g/L; mean � SD) 119 � 17 119 � 16 119 � 18 0.733Protein (g/L; mean � SD) 68.7 � 8.8 67.8 � 8.3 69.4 � 9.4 0.739Calcium (mmol/L; mean � SD) 2.47 � 0.23 2.41 � 0.22 2.51 � 0.23 0.291Phosphate (mmol/L; mean � SD) 1.53 � 0.58 1.51 � 0.38 1.54 � 0.71 0.759Magnesium (mmol/L; mean � SD) 0.91 � 0.18 0.86 � 0.16 0.96 � 0.19 0.149CRP (mg/L; mean � SD) 14.2 � 16.1 11.9 � 14.9 16.0 � 17.2 0.455Calcification of the iliac artery was investigated by von Kossa stain. Blood was drawn before kidney transplantation. BMI, body mass index; CRP, C-reactiveprotein; HD, hemodialysis; PD, peritoneal dialysis.

A

B

Figure 2. Mineral deposits of variable size can be found in auremic artery (longitudinal view) by use of x-ray imaging. (A)Radiograph showing a thin plate of mineralization (*) in the bloodvessel wall, with occasional nearby mineralization foci (arrows). (B)Three-dimensional reconstruction of mineral deposits after micro-computed tomography of a uremic artery. A large mineralizedarea is apparent in addition to small mineralization foci (arrows).The insert shows a single x-ray microcomputed tomographic“slice” from this region used in the reconstructions.

CLINICAL RESEARCHwww.jasn.org

J Am Soc Nephrol 21: 689–696, 2010 Ultrastructural Analysis 691

analysis, we could not analyze whether whitlockite was presentonly in certain types of calcification. Serum calcium, phos-phate, and magnesium of the four patients with whitlockitewere 2.42 � 0.29, 0.99 � 0.39, and 0.81 � 0.15 mmol/L, re-spectively, whereas the mean serum concentration of these pa-rameters of the two patients with hydroxyapatite were 2.25,2.28, and 1.04 mmol/L, respectively.

Analyses of Brachial and Coronary Artery SpecimensIn addition to analyses of iliac arteries, we investigated twobrachial and two coronary arteries. Both brachial and coronaryarteries showed a similar pattern of microcalcifications withintheir media. Using synchrotron radiation for diffraction anal-ysis, the coronary arteries again showed both hydroxyapatiteand whitlockite, whereas the brachial arteries contained onlywhitlockite.

Immunoelectron Microscopy Studiesin Uremic Artery CalcificationsIn addition to ultrastructural analysis, weexamined five uremic arteries (two iliacand three brachial arteries) for circulatingand matrix mineral-binding proteins us-ing immunoelectron microscopy. In allarteries investigated, immunogold label-ing was consistently observed for fe-tuin-A, osteopontin, and matrix gla pro-tein, showing a strong to moderatelabeling of the microcalcifications (i.e., aclose spatial relationship) but no labelingof the surrounding tissue for these pro-teins (Figure 6).

DISCUSSION

Our first major finding is that we discov-ered a core-shell layered structure inmany of the microcalcifications and thatsome of these microcalcifications exhib-ited at least a double layer in their shell. Invitro calcification has been shown to beinitiated by the release of membrane-bound matrix vesicles from vascularsmooth muscle cells and/or by apoptoticbodies from dying cells11 acting as crystalnucleation points. Given the size of thecore-shell structures in our study, thelarger objects could represent calcifica-tion of apoptotic bodies and the smallerones matrix vesicles, as has been de-scribed in a cell culture model of calcifi-cation.11 In nonuremic arteries (i.e., me-dia of aorta and temporal arteries),objects with similar appearance to themicrocalcifications detected in our study

had been described.12,13 A recent study of human arteriesshowed that medial calcification was in part triggered bysmooth muscle cell apoptosis.14 Whether the calcification pro-cess starts within the core or the shell structure still remains anopen question. According to in vitro models, calcification canstart both within vesicles and on the surface of vesicles.11,15

Alternative explanations are also possible, where a nidus ofcalcification (by whatever means) is followed by sequentialbouts of growth and inhibition, leading to a layered structure.

Our second major finding was that some microcalcificationswere composed of nanocrystals with a small size of 2 to 10 nm. Tothe best of our knowledge, such nano-sized calcifications have notyet been visualized in human media calcification. Becker et al.7

calculated a size of 20 to 25 nm for nanocrystals in atheroscleroticplaques. Recent data on calcification in an in vitro model indicatedthat spherical particles were composed of nanocrystalline needles

A B

C

D E F

Figure 3. Microcalcifications in uremic arteries are located extracellularly in the vicinity ofcollagen and vesicular structures as shown by light and electron micrographs. (A) In unde-calcified arterial samples embedded in plastic, sectioned, and examined by light microscopyfor mineral deposits by von Kossa staining, numerous small mineralization foci of variable size(black arrows) near collagen fibrils (light blue staining, white arrows) are observed throughoutthe arterial media. The inset shows a higher magnification of the collagen matrix and mineraldeposits. (B) Transmission electron micrograph showing a vascular smooth muscle cell(VSMC) surrounded by multiple small mineralization foci (black arrows) in the extracellularmatrix. The white arrows indicate collagen fibrils. The inset shows a higher magnification ofmineralization foci and collagen. (C) Mineralization foci (black solid arrows) in the vicinity ofvesicular structures (black dashed arrows) and collagen fibrils (white arrows). (D through F)Transmission electron micrographs of mineralization foci in the extracellular matrix showingvarious morphologies. In many cases, there is evidence of a concentric lamellar pattern of themineral deposits. The white arrow in F indicates cross-sectioned collagen fibrils.



of approximately 10 � 100 nm.16 In addition, it has been pro-posed that self-assembly of nano-apatite particles constitutes asubunit mechanism for biologic mineral crystals.17 Thus, ourfinding suggests that the nanocrystals play an early and essentialrole in the process of media calcification; however, a temporalsequence of microcrystals progressing to overt medial calcifica-tion cannot be proved by our cross-sectional study design. Thesemicrocalcifications could also be a different form of calcificationthat does not progress and thus are not necessarily indicative ofthe early stages of overt medial calcification. This could explain

the discrepancy between the lack of calcification in elastin in thisstudy and the calcification of elastin noted in other studies.

Our third major finding is the novel observation of whitlockitein human uremic arterial calcifications. Whitlockite is a magne-sium-substituted crystal that has rarely been reported to be a com-ponent of human calcifications of the mesenteries, lymph nodes,dental calculus, or dystrophic gouty calcification of the aorticvalve.18–21 In patients with uremia, whitlockite has been describedas a preliminary stage in soft tissue calcifications but not in vascu-lar calcifications.8,22,23 Most studies so far have exclusively de-tected hydroxyapatite or carbonated apatite in uremic vascularcalcifications7,8—except for one study, which reported brushitein arteriovenous fistulas.9 Whitlockite has been described onlyonce as a chemical compound in the nonuremic human aorta24;however, in this context, it has to be mentioned that hydroxyap-atite rarely exists in its pure form in biologic systems but occursoften as carbonated apatite, which may also contain magnesium.

Our novel finding of whitlockite is at variance with two previ-ous reports that described hydroxyapatite as the sole mineralphase present in human uremic vascular calcifications.7,8

Whereas in the study by Contiguglia et al.8 the vascular location ofcalcifications (i.e., intimal versus medial) was not reported, it isnoteworthy that in the study by Becker et al.,7 intimal but notmedial uremic calcifications were investigated. Atherosclerosisand calcification of the arterial media (i.e., arteriosclerosis) mayhave overlapping mechanisms but differ in their pathogenesis.More important, in our study, whitlockite could be found onlywhen using synchrotron radiation analysis but not with TEMtechniques. With TEM techniques, only small parts of the tissuecan be examined as a compromise for the available ultrahigh res-olution. With synchrotron radiation analysis, however, because ofthe combination of micrometer resolution and short acquisitiontimes (a few seconds), a more systematic scanning of the samplecovering a larger area can be obtained, which is a major advantageof this technique—albeit at the cost of a lower resolution whencompared with high-resolution TEM techniques. To the best ofour knowledge, only Verberckmoes et al.10 applied synchrotronradiation analysis to examine vascular calcifications in rats withuremia, in which they also identified whitlockite as an importantcomponent of the mineral phase, particularly in vitaminD–treated animals. No studies using synchrotron radiation fordiffraction analysis of human vascular calcification have been re-ported so far. This might be one of the reasons that in this studywhitlockite was found as a novel compound in vascular calcifica-tion of human origin.

Our fourth major finding was the close spatial relationshipbetween the microcalcifications and calcium-binding or calcifica-tion-inhibitory proteins. The proteins fetuin-A, matrix gla, andosteopontin are known calcification inhibitors, because their ab-sence is associated with spontaneous calcifications in vivo25,26 andenhanced susceptibility to calcification in vitro.27 Indeed, all ofthese proteins have been shown to co-localize with uremic calci-fications by immunohistologic methods.14,28 Until now, however,electron immunogold–labeling techniques for osteopontin andfetuin-A have been applied only to animal or in vitro models of

A B

C D

E F

Figure 4. The diameter of microcalcifications ranged from 20 to500 nm and a core-shell layered structure could be noticed in ap-proximately one third of the microcalcifications which suggest thatapoptotic bodies or matrix vesicles may serve as a calcification nidus.Microcalcifications seem to originate from 2- to 10-nm nanocrystals.(A) TEM showing multiple microcalcifications with diameters be-tween 20 and 500 nm. Microcalcifications show various morpholo-gies. A lamellar core-shell structure can be observed in many parti-cles. (B through D) TEM of microcalcifications with a core-shellstructure. A core was not present in all microcalcifications (D); how-ever, when present (B and C), it consisted of more electron-densematerial than the shell(s). (E) Three-dimensional reconstruction of themicrocalcification shown in B with visualization of a solid core sur-rounded by an inner, less dense and an outer, more dense shell. (F)High-resolution TEM of a uremic vascular microcalcification. Themicrocalcifications consist of nanocrystals with a size of 2 to 10 nm.



arterial calcification, respectively,15,29 but not to human arteries.Thus, our study provides further evidence that calcification inhib-itors are indeed intricately involved in the process of calcificationin humans. The localization of calcification inhibitors to areas ofcalcification is not yet completely understood.5 One possible ex-planation is an insufficient or exhausted defense mechanismagainst calcifications; however, other explanations are possible(e.g., removal of calcification complexes by vascular smooth mus-cle cells). Future studies are warranted for full understanding ofthe role of calcification inhibitors in the regulation of vascularcalcification.

In conclusion, arterial microcalcifications in uremia seemto originate from nanocrystals. These uremic microcalcifica-tions often exhibit a core-shell structure that may indicate acellular origin. Calcification inhibitors were closely related touremic calcifications. The novel observation of whitlockitewithin calcifications, in particular its potential clinical impli-cations, requires further study.

CONCISE METHODS

PatientsWe studied 30 dialysis patients who were under-

going renal transplantation. Patients were en-

rolled between March 2003 and March 2006.

Characteristics of the patients are depicted in Ta-

ble 1. None of the patients was on magnesium-

containing phosphate binders. The study proto-

col adhered to the Declaration of Helsinki and

was approved by the ethics committee of the

Rheinisch-Westfalische Technische Hochschule

Aachen, and each patient gave written informed

consent.

Calcification AssessmentFor 19 of the 30 patients, a plain pelvic x-ray

for assessment of vascular calcification was

performed. X-ray images were analyzed by one

experienced physician, who was blinded to the

patient’s condition.

Arteries and BiochemistrySmall pieces of iliac arteries (approximately 5

mm length), which normally would have been

discarded, were obtained during renal transplan-

tation. A standardized protocol was used for

sample collection. Surgeons performed sampling

at areas of the iliac artery that did not have any

obvious macroscopic calcifications. Samples

were fixed immediately in fixation buffer. For

von Kossa staining (in paraffin sections) and syn-

chrotron analysis, tissue was fixed in methacarn

solution and embedded in paraffin. Sections

were then stained using the von Kossa method

and counterstained with hematoxylin and eosin

as described previously.25

Of all artery segments, one section was used for the von Kossa stain

to assess whether microcalcifications were present. In 10 samples, we

obtained sections of two parts (one at the beginning and one in the

middle part of the arterial segment) and could confirm that von Kossa

stain yielded the same results for these two parts (either only positive

or only negative for calcification).

To assess the degree of calcification, we applied a semiquanti-

tative score according to the study of Gross et al.30: Grade 0, no

staining; grade 1, minimal positive staining; grade 2, positive stain-

ing involving up to 50% of the field of view; grade 3, positive

staining involving �50%; grade 4, positive staining of all struc-

tures within the field of view.

For TEM techniques, tissue was fixed with glutaraldehyde and

embedded in Epon (Serva, Heidelberg, Germany) or LR White acrylic

resin (London Resin Company; Berkshire, UK) as described previ-

ously.31 In addition to iliac arteries, two pieces of brachial arteries

were collected during arteriovenous fistula creation, and two coro-

nary arteries were obtained from autopsy samples. Blood was drawn

Figure 5. Uremic arterial calcifications are chemically more diverse than previouslythought with composition of hydroxyapatite and/or whitlockite. (A and B) Elemental andcrystallographic analysis of microcalcifications using electron microscopy techniques. (A)EELS showing the spectrum of hydroxyapatite. (B) Electron diffraction pattern of hy-droxyapatite. (C and D) Synchrotron radiation fluorescence and diffraction analysis. (C)Diffraction pattern for apatite (left) and whitlockite (right) standards. (D) The sample wasscanned, and the results are depicted. The top panel shows the scan for the intensity ofthe calcium signal. The middle bottom panel shows the diffraction mapping, where thediffraction pattern of each point of the sample is represented in the scan on a one-to-one basis. The intensity of the signal is depicted by different colors. Both side panelsshow the diffraction pattern of the two points with the highest and second-highest calciumsignal, respectively (as indicated by the white vertical line). The left spectrum reveals apatiteas the chemical compound, whereas the right spectrum shows whitlockite.



at admission to the hospital before kidney transplantation. Biochem-

ical analysis was performed by standard laboratory procedure using

an automated analyzer.

Radiography and Microcomputed TomographyHigh-resolution radiographic images of four artery segments were

taken under identical conditions by means of a Faxitron Model

MX-20 (Faxitron x-ray Corp., Wheeling, IL). Digital images were re-

corded at 26 kV and 0.3 mA over a 5-second exposure. For visualiza-

tion of arterial calcifications in three dimensions, an x-ray microcom-

puted tomograph (SkyScan model 1072, Kontich, Belgium) was used

to scan the arteries. The x-ray source was operated at maximum

power (45 KeV) and at 222 �A. Images were captured using a 12-bit

cooled charge-coupled device camera (1024 � 1024 pixels) coupled

by a fiber-optic taper to the scintillator. Using a rotation step of 0.9°,

total scanning time was 35 minutes for each rotated sample. Sections

were reconstructed using Skyscan tomography software (3D-Creator)

based on a triangular surface rendering a three-dimensional distribu-

tion of the calcified tissue.

TEM TechniquesFor high-resolution TEM, a 200-kV FEI Tecnai F20

with Gatan GIF 2000 imaging filter and EDX (EDAX)

was used to examine ultrathin sections of undecalci-

fied uremic arteries with areas of microcalcifications.

Chemical microanalysis was performed by recording

electron energy loss spectra of various calcium depos-

its in the scanning TEM mode. For mineral phase

identification, digitized selected-area electron dif-

fraction patterns from regions with a large number

of calcium deposits were quantified using the pro-

gram ELD (Calidris, Sollentuna, Sweden) and sub-

sequently processed by the program FULLPROF.32

Structural parameters for hexagonal hydroxyapa-

tite, Ca10(PO4)6OH2, were taken from ICSD-60521

and were partially refined.

Synchrotron AnalysisFor synchrotron radiation �-analysis (i.e., x-ray flu-

orescence and diffraction at the European Synchro-

tron Radiation Facility in Grenoble, France), we used

unstained (and undecalcified) 10-�m-thick sections

with both microcalcifications and areas of more

dense calcification sequential to those used for the

von Kossa staining as described already.10 X-ray flu-

orescence for calcium enabled localization of calcifi-

cations and indicated regions of interest to be further

investigated by x-ray microdiffraction. Synthetic hy-

droxyapatite [Ca10(PO4)6OH2] and whitlockite

[(Ca,Mg)3(PO4)2] embedded in paraffin served as

positive controls. A two-dimensional 135-mm MAR

charge-coupled device–based diffraction camera (Mar-

research, Norderstedt, Germany) was used to capture

the diffraction patterns. A microfocus monochromatic

x-ray beam (2 � 7 �m) with an energy of 14.4 keV was

used to scan the samples. X-ray fluorescence mappings

for calcium indicated calcified regions in the artery sections. Several line

�-fluorescence and diffraction scans of adequate length (30 to 100 �m;

2-�m steps) were recorded per sample. Spectra of investigated samples

were compared with synthetic mineral controls.

Ultrastructural Colloidal Gold ImmunocytochemistryThe antibodies used for the immunocytochemistry included poly-

clonal anti-human fetuin-A (courtesy of Dr. W. Jahnen-Dechent,

Aachen, Germany), polyclonal anti-mouse osteopontin (R&D Sys-

tems, Minneapolis, MN), and polyclonal anti-mouse matrix gla pro-

tein (courtesy of Dr. L.W. Fisher, Bethesda, MD). Grid-mounted, LR

White tissue sections were processed for colloidal gold immunocyto-

chemistry by incubation of the sections containing microcalcifica-

tions with primary antibody (1:10 dilution), after which immunola-

beling patterns were visualized by incubation with protein

A– colloidal gold complex (14-nm gold particles; Dr. G. Posthuma,

University of Utrecht, Utrecht, Netherlands), followed by conven-

tional staining with uranyl acetate and lead citrate, as described pre-

viously.31

Figure 6. Calcification inhibitors are involved in the process of calcification inhuman arteries (high-resolution immunogold labeling of calcification foci revealedcirculating and matrix mineral-binding proteins). Immunolabeling for fetuin-A (Aand B), osteopontin (OPN; C), and matrix Gla protein (MGP; D) show strong tomoderate gold particle labeling of mineralization foci in the uremic vessel wall.Immunolabeling in each case mainly localizes to electron-dense lamellar struc-tures in the mineralization foci.



Statistical AnalysisContinuous variables were summarized by means and corresponding

SD. Comparisons of the values of continuous variables between two

groups and among three groups were made using an unpaired t test

and ANOVA, respectively. Categorical variables were summarized by

relative frequencies. �2 test was used for investigating associations

between various categorical variables.

ACKNOWLEDGMENTS

This work was funded by grants from the Rheinisch-Westfalische

Technische Hochschule Aachen (Interdisziplinares Zentrum fur Kli-

nische Forschung BioMAT and START).

We thank Katrin Harthe for excellent technical assistance.

DISCLOSURESNone.

REFERENCES

1. Patient mortality and survival. United States Renal Data System. Am JKidney Dis 32[Suppl 1]: S69–S80, 1998

2. Blacher J, Guerin AP, Pannier B, Marchais SJ, London GM: Arterialcalcifications, arterial stiffness, and cardiovascular risk in end-stagerenal disease. Hypertension 38: 938–942, 2001

3. Schlieper G, Kruger T, Djuric Z, Damjanovic T, Markovic N, Schurgers LJ,Brandenburg VM, Westenfeld R, Dimkovic S, Ketteler M, Grootendorst DC,Dekker FW, Floege J, Dimkovic N: Vascular access calcification predictsmortality in hemodialysis patients. Kidney Int 74: 1582–1587, 2008

4. Ketteler M, Westenfeld R, Schlieper G, Brandenburg V: Pathogenesisof vascular calcification in dialysis patients. Clin Exp Nephrol 9: 265–270, 2005

5. Moe SM, Reslerova M, Ketteler M, O’Neill K, Duan D, Koczman J,Westenfeld R, Jahnen-Dechent W, Chen NX: Role of calcificationinhibitors in the pathogenesis of vascular calcification in chronic kid-ney disease (CKD). Kidney Int 67: 2295–2304, 2005

6. Moe SM, Duan D, Doehle BP, O’Neill KD, Chen NX: Uremia inducesthe osteoblast differentiation factor Cbfa1 in human blood vessels.Kidney Int 63: 1003–1011, 2003

7. Becker A, Epple M, Muller KM, Schmitz I: A comparative study ofclinically well-characterized human atherosclerotic plaques with histo-logical, chemical, and ultrastructural methods. J Inorg Biochem 98:2032–2038, 2004

8. Contiguglia SR, Alfrey AC, Miller NL, Runnells DE, Le Geros RZ: Natureof soft tissue calcification in uremia. Kidney Int 4: 229–235, 1973

9. Olsson LF, Odselius R, Ribbe E, Hegbrant J: Evidence of calciumphosphate depositions in stenotic arteriovenous fistulas. Am J KidneyDis 38: 377–383, 2001

10. Verberckmoes SC, Persy V, Behets GJ, Neven E, Hufkens A, Zebger-Gong H, Muller D, Haffner D, Querfeld U, Bohic S, De Broe ME,D’Haese PC: Uremia-related vascular calcification: More than apatitedeposition. Kidney Int 71: 298–303, 2007

11. Reynolds JL, Joannides AJ, Skepper JN, McNair R, Schurgers LJ,Proudfoot D, Jahnen-Dechent W, Weissberg PL, Shanahan CM: Hu-man vascular smooth muscle cells undergo vesicle-mediated calcifi-cation in response to changes in extracellular calcium and phosphateconcentrations: A potential mechanism for accelerated vascular calci-fication in ESRD. J Am Soc Nephrol 15: 2857–2867, 2004

12. Mohr W, Gorz E: Granular media calcinosis of the aorta: Structural

findings, historical review and pathogenetic significance [in German].Z Kardiol 90: 916–928, 2001

13. Mohr W, Gorz E: Calcifications in temporal arteries: Their morphogen-esis in comparison to physiological osteogenesis [in German]. Z Kar-diol 92: 60–72, 2003

14. Shroff RC, McNair R, Figg N, Skepper JN, Schurgers L, Gupta A,Hiorns M, Donald AE, Deanfield J, Rees L, Shanahan CM: Dialysisaccelerates medial vascular calcification in part by triggering smoothmuscle cell apoptosis. Circulation 118: 1748–1757, 2008

15. Reynolds JL, Skepper JN, McNair R, Kasama T, Gupta K, WeissbergPL, Jahnen-Dechent W, Shanahan CM: Multifunctional roles for serumprotein fetuin-a in inhibition of human vascular smooth muscle cellcalcification. J Am Soc Nephrol 16: 2920–2930, 2005

16. Sandin K, Kloo L, Nevsten P, Wallenberg RL, Olsson LF: Formation ofcarbonated apatite particles from a supersaturated inorganic bloodserum model. J Mater Sci Mater Med 20: 1677–1687, 2009

17. Robinson C: Self-oriented assembly of nano-apatite particles: A sub-unit mechanism for building biological mineral crystals. J Dent Res 86:677–679, 2007

18. Kodaka T, Mori R, Hirayama A, Sano T: Fine structure and mineralcomponents of fibrous stonelike masses obtained from the humanmesenteries. Med Electron Microsc 36: 272–281, 2003

19. Sakae T, Yamamoto H: Crystals and calcification patterns in two lymphnode calcifications. J Oral Pathol 16: 456–462, 1987

20. Kani T, Kani M, Moriwaki Y, Doi Y: Microbeam x-ray diffraction analysisof dental calculus. J Dent Res 62: 92–95, 1983

21. Gawoski JM, Balogh K, Landis WJ: Aortic valvular tophus: Identifica-tion by X-ray diffraction of urate and calcium phosphates. J Clin Pathol38: 873–876, 1985

22. Conger JD, Hammond WS, Alfrey AC, Contiguglia SR, Stanford RE,Huffer WE: Pulmonary calcification in chronic dialysis patients: Clinicaland pathologic studies. Ann Intern Med 83: 330–336, 1975

23. LeGeros RZ, Contiguglia SR, Alfrey AC: Pathological calcificationsassociated with uremia: Two types of calcium phosphate deposits.Calcif Tissue Res 13: 173–185, 1973

24. Reid JD, Andersen ME: Medial calcification (whitlockite) in the aorta.Atherosclerosis 101: 213–224, 1993

25. Schafer C, Heiss A, Schwarz A, Westenfeld R, Ketteler M, Floege J,Muller-Esterl W, Schinke T, Jahnen-Dechent W: The serum proteinalpha 2-Heremans-Schmid glycoprotein/fetuin-A is a systemically act-ing inhibitor of ectopic calcification. J Clin Invest 112: 357–366, 2003

26. Luo G, Ducy P, McKee MD, Pinero GJ, Loyer E, Behringer RR, KarsentyG: Spontaneous calcification of arteries and cartilage in mice lackingmatrix GLA protein. Nature 386: 78–81, 1997

27. Speer MY, Chien YC, Quan M, Yang HY, Vali H, McKee MD, GiachelliCM: Smooth muscle cells deficient in osteopontin have enhanced sus-ceptibility to calcification in vitro. Cardiovasc Res 66: 324–333, 2005

28. Moe SM, O’Neill KD, Duan D, Ahmed S, Chen NX, Leapman SB,Fineberg N, Kopecky K: Medial artery calcification in ESRD patients isassociated with deposition of bone matrix proteins. Kidney Int 61:638–647, 2002

29. Kaartinen MT, Murshed M, Karsenty G, McKee MD: Osteopontinupregulation and polymerization by transglutaminase 2 in calcifiedarteries of Matrix Gla protein-deficient mice. J Histochem Cytochem55: 375–386, 2007

30. Gross ML, Meyer HP, Ziebart H, Rieger P, Wenzel U, Amann K, Berger I,Adamczak M, Schirmacher P, Ritz E: Calcification of coronary intima andmedia: Immunohistochemistry, backscatter imaging, and x-ray analysis inrenal and nonrenal patients. Clin J Am Soc Nephrol 2: 121–134, 2007

31. McKee MD, Nanci A: Postembedding colloidal-gold immunocyto-chemistry of noncollagenous extracellular matrix proteins in mineral-ized tissues. Microsc Res Tech 31: 44–62, 1995

32. Rodriguez-Carvajal J: FULLPROF: A program for Rietveld refinementand pattern matching analysis [Abstract]. Abstracts of the SatelliteMeeting on Powder Diffraction of the XV Congress of the InternationalUnion of Crystallography, Toulouse, France, July 19–28, 1990, p 127.



Ultrastructural Analysis of Vascular Calcifications in Uremia

Documents