vascular wall von willebrand factor in human diabetic...

Vascular Wall von Willebrand Factorin Human Diabetic Retinopathy

Daria Boeri, Enrico Cagliero, Francesca Podestd, and Mara Lorenzi

Purpose. To reconstruct the role played by vascular endothelium in the elevation of circulatingvon Willebrand factor (vWf) in diabetic patients with microangiopathy and, specifically, todetermine whether storage and synthesis of vWf is altered in diabetic retinal vessels.

Methods. Trypsin digests were prepared form retinas obtained post mortem from 11 patients(age 62 ± 9 years, mean ± SD) with 9 ± 5 years of diabetes and 12 nondiabetic control subjectsmatched for age and sex. Trypsin digests were inspected for the presence of lesions of diabeticretinopathy; vWf protein was localized by indirect immunofluorescence; and vWf mRNA levelswere studied by in situ hybridization.

Results. vWf immunofluorescence was present in vessels of all sizes. The granular fluorescencewas localized to the endothelial cell cytoplasm. Pattern and intensity of staining in diabeticmicrovessels and large vessels were similar to those observed in the vessels of nondiabeticsubjects. The amount of vWf mRNA detected by in situ hybridization in retinal endothelialcells was similar in diabetic (0.92 ± 0.32 grains/cell) and control (0.91 ± 0.42 grains/cell)microvessels. Likewise, no differences were observed in vWf mRNA levels in the large vesselsof diabetic (0.073 ± 0.034% grain area) and control (0.069 ± 0.018 grain area) subjects.

Conclusions. These observations are compatible with the occurrence in diabetes of the slowrelease of endothelial vWf through the pathway of vWf secretion not linked to synthesis, ie, theregulated pathway. Invest Ophthalmol Vis Sci. 1994;35:600-607.

Several features of diabetic microangiopathy indicateor suggest that vascular endothelium is affected in dia-betes mellitus,1 but very little is known of the cellularand molecular events that underlie the demonstratedor inferred abnormalities. As the task of uncoveringcell behavior in situ is made more approachable by thecombination of traditional and new techniques, wehave begun to investigate the type of abnormalitiesthat diabetes induces in the endothelial cells of humanretinal vessels.

A finding widely interpreted as an index of endo-thelial cell dysfunction or damage in diabetes is theelevated plasma concentration of von Willebrand fac-tor (vWf) encountered in patients with retinopathy23

From the Schepeiis Eye Research Institute and the Department of Ophthalmology,Harvard Medical School, Boston, Massachusetts.Supported by National Institutes of Health grant EY09I22, a Feasibility grantfrom the American Diabetes Association, and the George and Frances LevinEndowment.Submitted for publication April 27, 1993; revised July 6, 1993; accepted July 28,1993.Proprietary interest category: N.Reprint requests: Mara Lorenzi, Schepens Eye Research Institute, 20 StanifordStreet, Boston, MA 02114.

and other clinical signs of microangiopathy.4 6 vWf isan adhesive multimeric glycoprotein synthesized onlyby vascular endothelial cells and megakaryocytes andreleased to the plasma and the subendothelial matrixby endothelial cells and activated platelets.7 The in-creased circulating levels of vWf in diabetic patientshave been attributed to enhanced endothelial cell re-lease from a potentially greater intracellular storagepool8'9 and to increased endothelial synthesis9; the pos-sibility of impaired clearance has not been addressed.The studies performed to date on endothelial cells ex-posed in vitro to glucose concentrations mimicking dia-betic hyperglycemia have not contributed to the mech-anistic interpretations of the clinical data insofar asthey have shown increased intracellular levels of vWfwithout evidence of enhanced release to the me-dium,10 and no changes in the levels of vWf mRNA(Cagliero E, Lorenzi M, unpublished observations).

To assess whether indeed storage or synthesis ofvWf is altered in diabetic vessels and thus to recon-struct the role played by vascular endothelium in theelevated circulating levels of vWf, we examined thedistribution of the glycoprotein by immunofluores-

600Investigative Ophthalmology & Visual Science, February 1994, Vol. 35, No. 2Copyright © Association for Research in Vision and Ophthalmology

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933179/ on 07/13/2018

Vascular Wall vWf in Human Diabetic Retinopathy 601

cence and its synthesis by in situ hybridization in tryp-sin digests of retinas obtained from patients with dia-betic retinopathy and age- and sex-matched nondia-betic controls.

METHODS

Human EyesHuman eyes were provided by certified Eye Banks; thedonors remained anonymous. Anonymous collectionof human eyes was approved by the institutional re-view board. The clinical characteristics of the 11 dia-betic patients and 12 nondiabetic control subjectsstudied are detailed in Table 1. Preferential criteriafor inclusion in the study were age less than 70 years,duration of diabetes less than 15 years to addressmostly background retinopathy, and the fewest possi-ble chronic pathologic conditions other that diabetes.Criteria for exclusion were the presence of retinal orhematologic diseases, uremia, and administration of

chemotherapy or life support measures. The eyes werefixed in 10% buffered formalin as soon as possible, butno more than 36 hours after death (Table 1). Formalinfixation was found compatible with preservation of sev-eral antigens and with in situ hybridization and there-fore was chosen to ensure consistency between eyesfixed in the laboratory and those fixed by Eye Banks,which do not routinely have available less common fix-atives. Fixation was for 48 to 168 hours as fixation for24 hours or less provided insufficient crosslinking forrecovery of large areas of intact capillary network, andfixation for longer than 1 week appeared to interferewith complete peeling of the vitreous.

Trypsin Digests

Trypsin digestion of the retinas was performed ac-cording to the method of Kuwabara and Cogan11 withsome modifications. Each fixed globe or pole was cutin half, and washed overnight at 4°C in 3 changes of

TABLE i. Characteristics of Subjects and Specimens Used in the Study of vWF

Sex

Subjects with DM

123456789

1011

MeanSD

MMMFFMFFFMM

6M/5F

Control Subjects

123456789

101112MeanSD

MFMMMMFFMFMM

8M/4F

Age(Years)

6765477644646363626666

629

656752644444515757626762

588

DMDuration(Years)

1265

154

154

17109

10

95

DMTherapy

InsSuSuDietInsInsInsInsSuInsIns

Cause ofDeath

CHF, MIMICPACHFMIMI, Pneum.CHFPulm. edemaCVACPAMI

CVACPAMIMIRupture AAMIMICVAMIOvarian CaRupture AAMI

Other KnownPathology

CataractsCVAPDV, HTNCVA

Lung CaCardiomyopathyCardiomyopathy

PVD, HTNDiverticulosis

HTN

HTN, COPD

SeizuresAsthmaCardiomyopathy

Eyes: Timeto Fixation(Hours)

182024241219101462

15

157

362336

56

192623271422

3

2011

Eyes: Timein Formalin(Hours)

727248484872

12048964896

7025

48484848487296

168168484872

7646

DM, diabetes mellitus; Su, sulfonylurea; Ins, insulin; CVA, cerebrovascular accident; CPA, cardiopulmonary arrest; MI, myocardialinfarction; CHF, congestive heart failure; Ca, carcinoma; AA, aortic aneurysm; PVD, peripheral vascular disease; HTN, hypertension;COPD, chronic obstructive pulmonary disease.


602 Investigative Ophthalmology & Visual Science, February 1994, Vol. 35, No. 2

glycine 0.15 M in 50% H2O, 50% phosphate-bufferedsaline (PBS), pH 7.4. The retina was gently separatedfrom the retinal pigmented epithelium, cut in threepieces (such that six preparations were made fromeach retina), and incubated at 37°C in 3% Difco Tryp-sin (Trypsin 1:250, Difco Laboratories, Detroit, MI) in0.1 M Tris buffer, pH 7.8, in a shaking H2O bath.After 1 hour of incubation, when the internal limitingmembrane and the vitreous began to separate fromthe retina, the retina was transferred to PBS at roomtemperature and briefly agitated to complete detach-ment of the vitreous. Retinas were then returned totrypsin incubation at 37°C until they began to clear(approximately 30 to 60 minutes), and the vasculartree was washed free of remaining neural tissue in dis-tilled H2O under microscopic observation. The prepa-rations were set onto 3-aminopropyltriethoxysilane(Silane, Sigma, St. Louis, MO)-coated glass slides, air-dried, and stored at -20°C.

Immunofluorescence

Each experiment included retinal trypsin digests fromtwo or three diabetic subjects and from an equal num-ber of age-matched nondiabetic controls. The digestswere soaked in glycine buffer for 30 minutes at roomtemperature and covered with 50 nl of the permeabili-zation buffer (0.5% Triton X-100 vol/vol and 2% bo-vine serum albumin in PBS) containing the monoclo-nal mouse antihuman vWf antibody (Dakopatts, Glos-trup, Denmark) at a dilution of 1:50. After overnightincubation at 4°C in a moist chamber, the slides werewashed in PBS (15 minutes X 3) and exposed for 1hour at room temperature in the dark to FITC-conju-gated polyclonal goat antimouse antibody (Coulter Im-munology, Hialeah, FL) diluted 1:40 in PBS-2% bo-vine serum albumin. The slides were washed in PBS asabove, mounted in p-phenylendiamine, and immedi-ately observed under a Zeiss (Thornwood, NY) Axio-phot fluorescence microscope. Negative controls (in-cluded in each experiment) were prepared by omittingthe first antibody or by replacing it with nonimmuneserum.

Photographs (Tri-X pan 400 film, Eastman KodakCo., Rochester, NY) were taken of three or four repre-sentative fields of microvasculature and three or fourdifferent large vessels in each preparation. The sameenlargements (XI00 and X200 for details) and thesame time of exposure (60 seconds) were usedthroughout all experiments. Two independent ob-servers experienced in immunofluorescence and blindto the nature of the study and identity of the subjectsscored each picture for prevalence and intensity offluorescence on a scale of 0 to 4. Zero was assigned tothe negative control and 4 to the picture with the mosthomogeneous bright fluorescence in the film. Micro-vessels and large vessels were graded independently,the fluorescence of microaneurysms was not taken

into account. The mean of scores assigned to the pic-tures of any given sample by the two observers wasused in the summary and statistical analysis of the data.

In Situ Hybridization

We have previously established that nondegradedRNA can be extracted from human retinal vessels iso-lated within 36 hours of death.12 To test whether thetrypsin digestion procedure is compatible with preser-vation of cytoplasmic RNA, human retinal trypsin di-gests were stained with acridine orange (0.5 mg/ml) inphosphate buffer according to an established proto-col.13 Under the fluorescence microscope the cell nu-clei emitted bright yellow-green fluorescence, whereasthe cytoplasm showed diffuse orange fluorescence(not shown). The latter was abolished by pretreatmentof the digests with ribonuclease A (20 /ig/ml) for 30minutes, indicating retention of cytoplasmic RNA intrypsin digests.

For the in situ hybridization experiments, the for-malin-fixed retinal trypsin digests set onto Silane-coated ribonuclease-free slides were incubated withproteinase K (100 Mg/ml) at 37°C for 15 minutes toincrease probe accessibility. The slides were then ace-tylated in triethanolamine buffer 0.1 M, pH 8.0, con-taining acetic anhydride (0.25%, vol/vol) for 10 min-utes, rinsed in 2 X sodium chloride-sodium citratebuffer (SSC) and incubated in 0.1 M Tris-HCl, pH 7.0,0.1 M glycine for 30 minutes. Immediately before hy-bridization the slides were immersed for 5 minutes in50% formamide-2 X SSC at 55°C. Sense and anti-sense 3)S-labeled cRNA probes were synthesized in vi-tro from the cDNAs coding for human vWf14 and 7-actinlr> and fragmented to 150 bp by alkaline hydroly-sis. To confirm the specificity of the vWf probe, theantisense vWf cRNA was used to hybridize a Northernblot containing RNA extracted from human umbilicalvein endothelial cells16; the expected 8.8-kb band wasclearly visualized. The trypsin digests were hybridizedin a solution containing formamide 50% vol/vol; dex-tran sulfate 10% wt/vol; dithiothreitol 10 mM; 2 XSSC; tRNA 1 mg/ml; salmon sperm DNA 1 mg/ml;nuclease-free bovine serum albumin 2 mg/ml; and 1 X106 cpm of probe per specimen. The hybridizationtook place in a moist chamber at 50°C for 4 hours.After hybridization the slides were washed in 50%formamide-2 X SSC at 52°C, and treated with ribonu-clease A (100 Mg/ml) at 37°C for 30 minutes. Autoradi-ography was performed with NTB-2 nuclear trackemulsion (Kodak), the slides were exposed for 14 to 21days at 4°C. The retinal digests were stained with peri-odic acid-Schiff hematoxylin, and because the acidtreatment damages the photographic emulsion, thestaining was performed before coating the slides withthe autoradiographic emulsion.

Each experiment included its own negative con-trol (hybridization with sense probe) to determine non-



specific background, and positive control (hybridiza-tion with cytoplasmic g-actin antisense probe). Quan-titation of signal in microvessels was performed bycounting the silver grains in ten randomly selected mi-croscopic fields, and the results are expressed (aftersubtracting the nonspecific background) as grains perendothelial cell nucleus per 109 cpm specific activity ofthe probe (grains per cell). Quantitation of signal inlarge vessels was performed with the aid of an imageanalysis system (Biological Detection Systems, Inc.,Pittsburgh, PA) in ten randomly selected microscopicfields (five arterioles and five venules). Results are ex-pressed as area occupied by grains in total tissue area(percent grain area) after subtraction of nonspecificbackground and normalization for 109-cpm specificactivity of the probe.

All results are summarized with the mean ± SDexcept the immunofluorescence scores because theywere measured on an ordinal (rather then interval)scale. The distribution of scores in the diabetic andcontrol groups was compared with the Mann-Whitneyrank sum test.

RESULTS

Human retinal trypsin digests studied with indirect im-munofluorescence showed the presence of vWf in ves-sels of all sizes (Fig. 1 A). Staining intensity was variableamong and within vessels, and the granular patternmost likely reflected reactivity of vWf stored in thediscrete organelles known as Weibel-Palade bodies;staining was absent upon omission of the vWf antibody(Fig. IB). The granular fluorescence was localized tothe endothelial cell cytoplasm (Figs. 1C, ID); the endo-thelial cell nuclei appeared as negative images, and nofluorescence was visible where only pericyte nucleiwere present (Figs. 1C, ID). Digest preparations fromthe same retina were tested after different duration offormalin fixation (2, 5, or 12 days) or storage at—20°C. Neither these variables nor the age of the do-nors appeared to affect the pattern and intensity ofvWf fluorescence.

Of the 11 diabetic patients studied, 8 had retinopa-thy as documented by the presence of microaneurysmsin retinal trypsin digests. vWf immunostaining was sim-

FIGURE l. Localization of vWf by immunofluorescence in human retinal vessels. (A, X200)Granular fluorescence of variable intensity is present in small and large vessels. (B, X200)Fluorescence is absent upon omission of vWf antibody (negative control). (C, X400) Granu-lar fluorescence is localized to the endothelial cell cytoplasm, as evidenced by comparisonwith same field (D, X400) stained with periodic acid-Schiff hematoxylin. The endothelial cellnuclei appear as negative images (thick arrows in C and D), and no fluorescence is visible wereonly pericyte nuclei are present (thin arrows in C and D).


604 Investigative Ophthalmology &; Visual Science, February 1994, Vol. 35, No. 2

ilar in control (Figs. 2A, 2B) and diabetic microvessels(Figs. 2C, 2D), irrespective of whether the capillariesappeared morphologically intact (Fig. 2C) or showedirregular contours and microaneurysms (Fig. 2D). Inthe specimens from 2 diabetic patients (9 and 11 inTables 1 and 2) the pattern of fluorescence was diffuserather than granular and more intense than that pre-vailing in the specimens from the other diabetic pa-tients or the controls. The quantitative scores attrib-uted to the fluorescence of microvessels as well aslarge vessels in individual specimens are reported inTable 2; the distribution of scores in diabetic patientswas not different from that in control subjects. Theinterassay coefficient of variation for immunofluores-cence signals in trypsin digests was 17%.

To examine the synthesis of vWf we used the insitu hybridization method. In experiments performedto optimize the procedure, the signal obtained in tryp-sin digests was compared to that obtained in cytospinsof cultured human umbilical vein endothelial cells.The tatter showed 3.3 ± 1.0 versus 0.37 ± 0.1 grains/cell in human retinal capillaries. In the retinal digestslabeling was clearly present in both small and largerretinal vessels (Fig. 3A) and minimal in control speci-mens hybridized with the sense vWf probe (Fig. 3B). In

retinal microvessels of diabetic patients (Figs. 3C, 3D)labeling intensity was similar to that observed in con-trol microvessels (Fig. 3A), both when vessel morphol-ogy was still intact (Fig. 3C) and when irregular con-tours and pericyte ghosts had become apparent (Fig.3D). Quantitation of signal in individual specimens isreported in Table 2. Microvessels showed 0.92 ± 0.32grains/cell in the diabetic group and 0.9] ± 0.42grains/cell in the nondiabetic control group. Labelingof larger vessels was also similar in the diabetic andcontrol group (0.073 ± 0.034 and 0.068 ± 0.018%grain area, respectively). The interassay coefficientof variation for the in situ hybridization methodwas 28%.

DISCUSSION

With the goal of ascertaining if the diabetic state altersthe storage and synthesis of vWf in vascular endothe-lium, we studied the protein and its cognate mRNA inhuman retinal vessels. The characteristics of vWf im-munofluorescence in retinal trypsin digests —pres-ence in all size vessels with variable intensity and pref-erential localization to the perinuclear cytoplasm ofendothelial cells—were in perfect agreement with

FIGURE 2. vWf iiTiniunofluorescence in control and diabetic retinal microvessels. (A, B) Con-trol microvessels. (C) Diabetic microvessels showing intact morphology. (D) Diabetic micro-vessels showing irregular diameter and microaneurysm. (Magnification X200.)



TABLE 2. vWF Immunofluorescence (IF) and In Situ Hybridization(ISH) in Human Retinal Vessels

Microvessels

Subjects with DM

123456789

1011

MeanSD

Control Subjects

123456789

101112

MeanSD

1.31.32.32.00.51.32.11.13.11.43.0

1.82.52.1—1.02.02.02.12.01.41.8

vWFIF

Large Vessels(Score)

2.02.01.93.01.51.5——3.13.02.9

3.02.04.02.8—2.53.02.02.83.33.03.5

Microvessels(Grains/Cell)

1.121.200.46———

0.721.10———

0.920.32

0.540.29—

1.151.40——

1.011.42———

0.910.42

vWF ISH

Large Vessels(% Grain Area)

0.0700.0440.038

———

0.0950.119

———

0.0730.034

0.0920.076

—0.0470.084

——

0.0510.062

———

0.0690.018

DM, diabetes mellitus. Numbering of diabetic and control subjects corresponds to that in Table 1.

those observed in frozen biopsy specimens of severalhuman tissues.17 Preservation and accurate localiza-tion of vWf in trypsin digests can provide a usefulmarker of endothelial cell cytoplasm in studies requir-ing precise cellular attribution of specific mRNAs orproteins in these vascular preparations.

To our knowledge, application of the in situ hy-bridization method to retinal trypsin digests has notbeen previously reported. We had established that in-tact RNA can be extracted from human retinal micro-vessels up to 36 hours from death,12 and we have nowobserved that cytoplasmic RNA is retained in trypsindigests. The amount of vWf mRNA detected by in situhybridization in the retinal vascular endothelial cellswas one order of magnitude lower than that detectedby the same method in cultured umbilical vein endo-thelial cells. This result is in agreement with that ob-tained using a liquid hybridization assay,12 indicatingthat application of the in situ hybridization techniqueto retinal trypsin digests can provide reliable, if semi-

quantitative, information. Our previous studies haveshown that not only the vWf transcript, but also theplasminogen activator-inhibitor 1 transcript is muchless prevalent in endothelial cells of retinal vessels thanin cultured endothelial cells.12 Among the many rea-sons that may account for the difference,12 we favor amajor role for the different degrees of biosyntheticactivity expected of two endothelial cell populationswith vastly dissimilar turnover rates.

The observations that in diabetic retinal digestsimmunostaining for vWf was unchanged from controllevels in both microvessels and large vessels, and thatvWf mRNA levels were not different from control indi-cate that diabetes does not substantially alter storageor synthesis of endothelial vWf. Insofar as we studiedpatients with documented lesions of diabetic retinopa-thy and thus likely to have had elevated plasma levelsof vWf,23 our findings can be applied to interpretationof the latter abnormality.

The observations made in retinal endothelial cells


Investigative Ophthalmology &; Visual Science, February 1994, Vol. 35, No. 2

k.

A•4

FIGURE 3. Localization of vWf mRNA by in situ hybridization in control and diabetic retinalvessels. (A) Silver grains are present on small and large vessels after hybridization with anti-sense vWf probe. (B) Silver grains are absent after hybridization with sense vWf probe (nega-tive control). (C) Diabetic microvessels showing intact morphology. (D) Diabetic microvesselsshowing irregular diameter and a pericyte ghost (arrow). (All specimens counterstained withperiodic acid-Schiff hematoxylin; magnification X600.)

are compatible both with decreased clearance or in-creased release of vWf to account for the elevatedplasma levels encountered in diabetic patients. The ca-tabolism of vWf probably involves rapid proteolysiswithin the circulation,18 and if there are no informa-tion on whether the process is impaired in diabetes,there are also no reasons to anticipate that it might be.If the increased plasma levels of vWf reflect increasedrelease of the glycoprotein, the main contributorwould be vascular endothelium because the plasmaticcompartment is only minimally replenished by the vWffrom platelets and megakaryocytes.19 Within the likelypossibility of increased release of vWf by endothelialcells, our observations help define the pathway in-volved. The unchanged synthesis of vWf in diabeticvessels strongly suggests that increased release shouldnot occur through the constitutive pathway insofar asthis pathway is directly linked to biosynthesis.7 Stimu-lated release of vWf through the regulated pathway(which involves the vWf secreted from the Weibel-Pa-lade bodies) has instead no significant effect on vWfbiosynthesis,20 and could therefore be a potentialmechanism for increased vWf in the circulation of dia-

betic patients. Although massive stimulation of the reg-ulated pathway would be expected to drastically de-crease the fluorescence of Weibel-Palade bodies,20'21

subtler and more chronic stimuli may not result inappreciable differences in number and distribution ofthe storage organelles. This paradigm has been used toexplain increased release of vWf to the medium with-out depletion of the Weibel-Palade bodies in culturedendothelial cells exposed to radiation,22 and may ac-count for the analogous set of findings in diabetic pa-tients.

Goals for future studies should be to confirm theoccurrence in diabetes of excess release of endothelialcell products, and to define whether the increased se-cretion reflects alterations intrinsic to endothelial cells(eg, increased membrane permeability to Ca2+) or en-hanced availability of physiologic secretagogues suchas thrombin and histamine.7

Key Words

endothelial cells, von Willebrand factor, diabetic angiop-athy, retinal vessels, in situ hybridization


Vascular Wall vWf in Human Diabetic Retinopathy

Acknowledgments

The authors thank Dr. Dennis C. Lynch for providing thevWf probe and Dr. M. Rosario Hernandez and Dr. Nancy C.Joyce for scoring the immunofluorescence pictures.

References

1. Lorenzi M. Glucose toxicity in the vascular complica-tions of diabetes: The cellular perspective. DiabetesMetabRev. 1992;8:85-103.

2. Coller BS, Frank RN, Milton RC, Gralnick HR.Plasma cofactors of platelet function: Correlationwith diabetic retinopathy and hemoglobins Ala-c.Studies in diabetic patients and normal persons. AnnIntern Med. 1978;88:311-316.

3. Porta M, La Selva M, Molinatti PA. von Willebrandfactor and endothelial abnormalities in diabetic mi-croangiopathy. Diabetes Care. 1991 ;14 (suppl 1):167-172.

4. Jensen T. Increased plasma concentration of von Wil-lebrand factor in insulin dependent diabetics with in-cipient nephropathy. BrMedJ. 1989;298:27-28.

5. Stehouwer CDA, Stroes ESG, Hackeng WHL, MulderPGH, den Ottolander GJH. Von Willebrand factorand development of diabetic nephropathy in IDDM.Diabetes. 1991 ;40:971-976.

6. Stehouwer CDA, Nauta JJP, Zeldenrust GC, HackengWHL, Donker AJM, den Ottolander GJH. Urinary al-bumin excretion, cardiovascular disease, and endothe-lial dysfunction in non-insulin-dependent diabetesmellitus. Lancet. 1992;340:319-323.

7. Wagner DD. Cell biology of von Willebrand factor.Annu Rev Cell Biol. 1990;6:217-246.

8. Porta M. Availability of endothelial von Willebrandfactor and platelet function in diabetic patients in-fused with a vasopressin analogue. Diabetologia.1982;23:452-455.

9. Giustolisi R, Musso R, Russo M, Catania N, LombardoT, Cacciola E. Possible evidence for an increased fac-tor VIII antigen synthesis in vascular endothelium ofdiabetic subjects. Thromb Haemostas. 1982;47:293.

10. Mordes DB, Lazarchick J, Colwell JA, Sens DA.Elevated glucose concentrations increase factorVIIIR:Ag levels in human umbilical vein endothelialcells. Diabetes. 1983;32:876-878.

11. Kuwabara T, Cogan DG. Studies of retinal vascular

607

patterns: I. Normal architecture. Arch Ophthalmol.1960;64:904-911.

12. Cagliero E, Grant MB, Lorenzi M. Measurement ofgene expression in human retinal microvessels by so-lution hybridization. Invest Ophthalmol Vis Sci.1991;32:1439-1445.

13. Berger CN. In situ hybridization of immunoglobulin-specific RNA in single cells of the B lymphocyte lin-eage with radiolabelled DNA probes. EMBO J.1986;5:85-93.

14. Bonthron D, Orr EC, Mitsock LM, Ginsburg D, Han-din RI, Orkin SH. Nucleotide sequence of pre-pro-von Willebrand factor cDNA. Nucleic Acids Res.1986;14:7125-7127.

15. Gunning P, Ponte P, Okayama H, Engel J, Blau H,Kedes L. Isolation and characterization of full-lengthcDNA clones for human a-, /?-, and 7-actin mRNAs:Skeletal but not cytoplasmic actins have an amino-ter-minal cysteine that is subsequently removed. Mol CellBiol. 1983:3:787-795.

16. Cagliero E, Maiello M, Boeri D, Roy S, Lorenzi M.Increased expression of basement membrane compo-nents in human endothelial cells cultured in high glu-cose. 7 Clin Invest. 1988;82:735-738.

17. Hoyer LW, De Los Santos RP, Hoyer JR. Antihemo-philic factor antigen: Localization in endothelial cellsby immunofluorescent microscopy. / Clin Invest.1973:52:2737-2744.

18. Sadler JE. vonWillebrand disease. In: Scriver CR,Beaudet AL, Sly WS, Valle D, eds. The Metabolic Basisof Inherited Disease. 6th ed. New York: McGraw-Hill;1989:2171-2187.

19. Bowie EJW, Solberg LA, Fass DN, et al. Transplanta-tion of normal bone marrow into a pig with severe vonWillebrand's disease, f Clin Invest. 1986;78:26-30.

20. Mayadas T, Wagner DD, Simpson PJ. von Willebrandfactor biosynthesis and partitioning between constitu-tive and regulated pathways of secretion after throm-bin stimulation. Blood. 1989;73:706-7ll.

21. Reinders JH, Vervoorn RC, Verweij CL, van MourikJA, de Groot PG. Perturbation of cultured human vas-cular endothelial cells by phorbol esters or thrombinalters the cellular von Willebrand factor distribution./CellPhysiol. 1987;133:79-87.

22. Sporn LA, Rubin P, Marder VJ, Wagner DD. Irradia-tion induces release of von Willebrand protein fromendothelial cells in culture. Blood. 1984;64:567-570.


vascular wall von willebrand factor in human diabetic...

Documents