Monoclonal antibody-defined human endothelial antigens as vascular markers.

A review is given of human endothelial antigens recognized by monoclonal antibodies and used as vascular markers. These antigens can be classified tentatively into two categories that partly overlap: 1) differentiation markers and 2) antigens involved in specific cellular functions. Monoclonal antibodies recognizing endothelial differentiation markers reacting with all types of human endothelium can be regarded as constitutive endothelial markers. Other differentiation markers have a restricted distribution that is associated with a subtype of endothelium. Although sensitivity of the markers is high in general, specificity for endothelium is not absolute, based on distribution studies in tissues or in cell lines. With the exception of PAL-E and EN-3/EN-4, it is not clear from the literature whether the antibodies also react with lymphatic endothelium. Immunohistochemical examination of other species indicate that only BW 200 is restricted to humans. Immunoelectron microscopy of microvascular cells in tissue specimens has revealed that the monoclonal antibodies recognizing differentiation antigens show different subcellular distribution patterns. PAL-E and BW 200 react with the luminal endothelial surface, in a local and diffuse pattern, respectively. Anti-Von Willebrand factor (i.e., Factor VIII-related ag) antibodies react with Weibel-Palade bodies but also with subendothelial structures. Applications of immunohistochemistry using monoclonal antibodies in diagnostic pathology include assessment of vascular invasion by cancer cells, and identification of endothelial neoplasms and related disorders. Because anti-Factor VIII-related antigen and BW 200 are applicable on formaldehyde-fixed and paraplast-embedded tissue, they are most suitable for histodiagnostic application. Immunohistochemistry using monoclonal antibodies recognizing endothelial antigens involved in specific cellular functions also may contribute to pathobiologic research on the characterization of blood-tissue barriers, e.g., in the tumor vascular bed.

Phenotype of normal cutaneous microvasculature. Immunoelectron microscopic observations with emphasis on the differences between blood vessels and lymphatics.

The lymphatic system has been poorly characterized in comparison to the blood vessels. We investigated the expression of microvasculature markers in cutaneous lymphatics and blood microvessels in normal skin. Scrotal skin was chosen because of its high density of both types of microvessels. A pre-embedding peroxidase-conjugated immunoelectron microscopy technique was used, allowing both the visualization of the lymph and blood vessels and their immunohistochemical staining. The markers studied included endothelial antigens (recognized by PAL-E, EN-4, and von Willebrand factor/factor VIII-related antigen), structural molecules of the vascular wall (alpha-smooth muscle actin, heparan sulfate proteoglycan, collagen type IV), and adhesion molecules (endothelial leukocyte adhesion molecule-1 [E-selectin], intercellular adhesion molecule-1 [ICAM-1], platelet endothelial adhesion molecule-1 [PECAM-1], vascular cell adhesion molecule-1 [VCAM-1]). It is shown that lymphatics of normal skin are phenotypically different from blood microvasculature, only weakly expressing endothelial markers (EN-4+, von Willebrand factor/factor VIII-related antigen +/-, PAL-E-), mural markers (alpha-smooth muscle actin-, heparan sulfate proteoglycan-, collagen type IV+) and do not express the studied adhesion molecules except PE-CAM-1 (E-selectin-, ICAM-1-, PECAM-1+, VCAM-1-). The results were substantiated by a double-labeling immunoelectron microscopic technique, which facilitates detection and assessment of microvascular segments. By this technique, collagen type IV, recognized by a peroxidase-labeled 2nd antibody, stains the basal lamina by a linear pattern, whereas a second optional epitope is visualized as grains by a silver-enhanced ultra-small gold-conjugated antibody. Our study shows that not only morphology but also antigenic phenotype of lymphatics differs significantly from blood vessels.

Early and extensive contribution of pericytes/vascular smooth muscle cells to microvascular proliferation in glioblastoma multiforme: an immuno-light and immuno-electron microscopic study.

Although florid microvascular proliferation (MVP) in glioblastoma multiforme (GBM) has long been considered as proliferation of endothelial cells (EC), recent immuno-light microscopic studies demonstrated many alpha-smooth muscle actin (alpha-sm actin)-positive cells in this MVP, suggesting a major contribution of pericytes and/or vascular smooth muscle cells (VSMC). Under certain culture conditions, however, alpha-sm actin expression has also been described in EC. In order to further investigate to what extent pericytes/VSMC participate in MVP in GBM, we performed an immunohistochemical study at both the light and electron microscopic levels with anti-alpha-sm actin, with an antibody against EC (EN-4) and with an antibody recently described to react with "activated" pericytes in conditions with neovascularization (anti-high molecular weight-melanoma associated antigen). In this detailed study of MVP in GBM, two distinct cell types could be recognized on the basis of a consistent ultrastructural localization and immunophenotype: EC and pericytes/VSMC; no transitional forms were found between these two cell types. The contribution of pericytes/VSMC to MVP in GBM was extensive and already present in many delicate tumor capillaries, suggesting not only an essential but also an early role of these cells in this type of tumor angiogenesis.

A new surgical technique for deep stromal, anterior lamellar keratoplasty.

AIMS: To describe a new surgical technique for deep stromal anterior lamellar keratoplasty. METHODS: In eye bank eyes and sighted human eyes, aqueous was exchanged by air, to visualise the posterior corneal surface--that is, the "air to endothelium" interface. Through a 5.0 mm scleral incision, a deep stromal pocket was created across the cornea, using the air to endothelium interface as a reference plane for dissection depth. The pocket was filled with viscoelastic, and an anterior corneal lamella was excised. A full thickness donor button was sutured into the recipient bed after stripping its Descemet's membrane. RESULTS: In 25 consecutive human eye bank eyes, a 12% microperforation rate was found. Corneal dissection depth averaged 95.4% (SD 2.7%). Six patient eyes had uneventful surgeries; in a seventh eye, perforation of the lamellar bed occurred. All transplants cleared. Central pachymetry ranged from 0.62 to 0.73 mm. CONCLUSION: With this technique a deep stromal anterior lamellar keratoplasty can be performed with the donor to recipient interface just anterior to the posterior corneal surface. The technique has the advantage that the dissection can be completed in the event of inadvertent microperforation, or that the procedure can be aborted to perform a planned penetrating keratoplasty.

Transition of horizontal to vertical growth phase melanoma is accompanied by induction of vascular endothelial growth factor expression and angiogenesis.

Melanoma progression in general is characterized by an increase in both metastatic frequency and the vascular density of the tumour tissue. Although a direct correlation between these two parameters in individual cases seems to be lacking, it is clear that metastasis is invariably preceded by angiogenesis. One of the angiogenic factors that is produced by human melanoma cells is vascular endothelial growth factor (VEGF). To investigate the role of this factor in the angiogenic process in primary cutaneous melanoma we determined the mean vascular density and the presence of VEGF protein in biopsies of human lesions. The results were compared with those found in normal skin or uninvolved skin from melanoma patients. In addition, we studied morphological and antigenic features of the proliferating neovasculature. We show that (1) the mean vascular density gradually rises along with melanoma progression, (2) transition of horizontal to vertical growth phase melanoma is accompanied by induction of VEGF protein expression and accumulation of this factor in the stroma, (3) vertical growth phase melanoma is often organized in nodules separated by septa containing blood vessels, but without lymphatics, and (4) blood vessel lumina in vertical growth phase melanoma are separated from tumour nodules by two basal lamina containing collagen type IV and the endothelium shows activated morphology and focal expression of the adhesion molecule E-selectin. Our findings indicate that VEGF is a prominent angiogenic factor in melanoma angiogenesis. Although its expression is induced during progression, the effect of VEGF on the incidence of metastasis is probably indirect.

Healing of reopened-and-sutured radial keratotomy wounds.

We designed a study to evaluate healing in reopened-and-sutured (RAS) keratotomy wounds to determine the efficacy of reoperations in treating radial keratotomy overcorrections. Using light and transmission electron microscopy, we compared stromal scar tissue organization (transverse fibroblast orientation and collagen fiber continuity across the wound) in RAS wounds and in sutured and unsutured control wounds in 18 monkey eyes one to nine weeks after surgery. Wound healing morphology of RAS wounds varied with the interval between reoperation and termination of the experiment. Scar tissue organization was sagittal at one week postoperatively, transverse in the anterior and mid regions after four weeks, and transverse over the entire wound after nine weeks. Sutured wounds showed a similar pattern of healing, although transverse scar tissue organization was restricted to the anterior and mid regions in the late healing phases. In contrast, unsutured wounds showed a temporary, transverse scar tissue organization over the entire wound depth at two to four weeks and a progressive reorientation of the mid and posterior scar tissue sagittal to the wound at later intervals. The results suggest that reopening and suturing keratotomy incisions to treat radial keratotomy overcorrections may be effective through a myopic shift induced by sutured wound apposition and long-term wound remodeling, contraction, or both.

A technique to visualize corneal incision and lamellar dissection depth during surgery.

PURPOSE: To describe a surgical technique to visualize the depth of corneal incisions and lamellar stromal dissections during surgery. METHODS: In porcine cadaver eyes, the aqueous was exchanged by air. Thus an air-to-endothelium interface (i.e., a useful optical surface) was created at the posterior corneal surface. The air-to-endothelium interface was used as a reference plane to visualize the corneal thickness and the relative depth of corneal incisions and dissections. Freehand peripheral corneal incisions, tangential keratotomy incisions, and lamellar stromal dissections were made at an intended corneal depth of 60, 80, and 99%. Light microscopy was used to measure the relative depth of the incisions and dissections. RESULTS: Achieved depth for peripheral corneal incisions averaged 65.2+/-5.3%, 78.8+/-5.1%, and 93.4+/-6.0%, respectively (p<0.05); and for tangential keratotomy incisions, 68.2+/-7.3%, 83.2+/-4.4%, and 95.8+/-3.6%, respectively (p<0.05). Achieved depth for lamellar stromal dissections averaged 58.3+/-9.4%, 81.1+/-3.4%, and 94.4+/-1.5%, respectively (p<0.05). Microperforations occurred with three incisions made at 99% intended depth. CONCLUSION: During surgery, the depth of incisions and lamellar dissections relative to the corneal thickness can be visualized by filling the anterior chamber with air (i.e., by creating an optical interface at the posterior corneal surface).

Depth predictability of stromal pockets in the posterior cornea.

PURPOSE: To evaluate the predictability of the depth of stromal pockets made in the posterior cornea for the excision of anterior or posterior lamellar corneal buttons with a planned thickness. METHODS: Stromal corneal pocket dissections were created in human eye bank eyes by making a peripheral arcuate keratotomy incision at 60, 80, or 95% of central pachymetry and creating a pocket from the bottom of the incision across the cornea. Pocket depth was measured by pachymetry immediately after surgery and by light microscopy. RESULTS: Mean achieved central pocket depth differed by 0.03+/-0.03 mm from the intended depth. Variation in depth across the pocket decreased from 0.07+/-0.02 mm for pockets made at 60% of the intended depth to 0.05+/-0.01 mm for pockets made at 80% depth, and 0.04+/-0.02 mm for pockets made at 95% depth (p < 0.01). Pachymetric and histological measurements of relative pocket depth averaged 64+/-9% and 73+/-7%, respectively, for pockets made at 60% of the intended depth, 82+/-7% and 86+/-3% for pockets made at 80% depth, and 91+/-7% and 92+/-3% for pockets made at 95% depth. The difference between pachymetric and histological relative pocket depth measurements decreased with deeper pocket depth (p < 0.01). CONCLUSIONS: In the posterior cornea, stromal pockets can be created to within 30 microm from the intended depth. Variation in depth throughout the pocket decreases with deeper pocket depth. Pachymetry is a reliable method to check the achieved pocket depth during surgery; the accuracy of pachymetry readings improves with deeper pocket depth.

A surgical technique for posterior lamellar keratoplasty.

PURPOSE: To design a surgical technique for transplantation of posterior corneal tissue, while leaving the recipient anterior cornea intact. METHODS: In human cadaver eyes, and in a cat and monkey model, recipient eyes had an 8.0-mm limbal incision made with a diamond blade set to 50% of central pachymetry. A stromal pocket was created across the cornea, and a 6.0-mm diameter posterior lamellar disc was excised. A donor posterior disc was implanted into the recipient opening, and the limbal incision was sutured. The procedure was evaluated with keratometry, biomicroscopy, endothelial (supra)vital staining, and light microscopy. RESULTS: In human cadaver eyes, post-operative astigmatism averaged 1.2 D (SD, +/- 0.6 D). Posterior transplants showed an intact endothelial cell layer with 1.0% (SD, +/- 1.2%) of cell death. In the animals, six (75%) eyes had clear transplants 2 weeks after surgery; one of these eyes later developed an allograft rejection. Two (25%) eyes showed corneal decompensation, because of inverted implantation of the donor disc. Microscopy showed minimal scarring at the donor-to-host interface and a normal wound-healing response at the posterior stromal wound edges. CONCLUSION: In experimental models, posterior lamellar keratoplasty can be performed through a limbal incision and a mid-stromal pocket. The procedure may be a potential alternative in the surgical management of corneal endothelial disorders.

Aminopeptidase a is a constituent of activated pericytes in angiogenesis.

Monoclonal antibody (MAb) RC38 recognizes a human renal antigen of 160 kD recently identified as human aminopeptidase A (APA; EC 3.4.11.7). This ectoenzyme is able to hydrolyse selectively N-terminal glutamyl and aspartyl residues from oligopeptides. By enzyme histochemistry, APA activity has also been localized in the microvessels of all organs in animals and man. The purpose of this study was to investigate the distribution of human APA as recognized by MAb RC38 in the microvasculature of normal human tissues and pathological conditions associated with neovascularization. Unexpectedly, in normal tissues vascular staining with MAb RC38 was generally weak and often absent, while in tumours, granulation tissue, and chronic synovitis, marked microvascular staining was demonstrated. By immuno-electron microscopy, the antigen was found on the cell membrane of activated pericytes and their processes in the tumour vasculature. RC38 expression could not be detected on cultured human endothelial cells or pericytes. These observations suggest that pericyte expression of a subtype of APA (as recognized by MAb RC38) is markedly enhanced in the vasculature of tumours and wound healing tissue as compared with normal resting tissues. This provides further evidence of the altered state of pericytes in these conditions. Pericyte APA may be involved in the metabolism of biologically active oligopeptides during neovascularization, supporting a regulatory role of pericytes in this process. In addition, MAb RC38 may be useful as a marker of pericyte activation in tissue sections.

Encephalocytozoon intestinalis-specific monoclonal antibodies for laboratory diagnosis of microsporidiosis.

Two monoclonal antibodies which can be used for the unambiguous identification by fluorescence microscopy of Encephalitozoon intestinalis spores in clinical specimens are described. Monoclonal antibody Si91 is specific for the extruded polar filament, and Si13 recognizes the surfaces of E. intestinalis spores. No cross-reaction with spores of Encephalitozoon hellem was observed. Immunogold electron microscopy confirmed the specific reactivities of both antibodies. Combined in an indirect immunofluorescence assay, these antibodies are used to identify spores in feces. Although there was some cross-reaction with fecal bacteria and fungi, the typical morphology of the extruded polar filaments enabled proper identification of the E. intestinalis spores. Parasites could also be demonstrated to be present in urine, nasal swabs, lung brush biopsy specimens, and bronchoalveolar lavage fluid from a patient with disseminated infection with E. intestinalis. The use of these monoclonal antibodies facilitates the detection and species determination of E. intestinalis in clinical specimens.

Angiogenesis in wound healing and tumor metastasis.

Formation of new blood vessels is essential for several physiological and pathological events, e.g. embryogenesis, wound healing and tumor growth and metastasis. In order to increase the insight into the mechanisms of angiogenesis we have visualized the different components of the microvasculature in human wounds and tumors by immunohistochemistry on the light and electronmicroscopic level. For this purpose, antibodies recognizing distinct markers for human endothelial cells, pericytes and basal lamina were used on freshly frozen or paraformaldehyde-fixed tissue samples. In terms of efficacy, the PAL-E antigen is highly specific for blood vessel endothelium. Its sensitivity is less than other endothelial markers, such as von Willebrand factor and CD 31, as it is not expressed in arterioles. Within the context of the microvasculature alpha-smooth muscle actin and the HMW-MAA chondroitin sulphate proteoglycan are useful markers for pericytes. Type IV Collagen and Laminin can be visualized consistently in the microvascular basal lamina. During the formation of granulation tissue in wound healing a heterogeneity of the expression of endothelial and pericyte markers is found. In the least matured zone in granulation tissue of decubitus lesions and experimental skin wounds microvessels already contained both endothelial cells and pericytes, suggesting a role for both cell types in the early steps of angiogenesis. Regarding the tumor microvasculature, antibodies to von Willebrand factor often failed to stain capillaries, that did show expression of the other endothelial markers studied. Broad staining in pericytes was found for the HMW-MAA chondroitin sulphate proteoglycan. In contrast, these cells only locally expressed alpha-smooth muscle actin. Staining of the basal lamina components Type IV Collagen and Laminin within tumors was not restricted to the microvasculature. Therefore, antibodies recognizing endothelial markers, particularly PAL-E and BMA 120, are preferable as tools to visualize the tumor microvasculature. In accordance with the situation in granulation tissue of wound healing the broad presence of pericytes in the microvasculature of human tumor suggests an involvement of this cell type in tumor angiogenesis. Recent immunohistochemical studies on human tumor lesions indicated that a high number of microvessels adjacent to the tumor as a measure of tumor angiogenesis is an unfavorable prognostic factor in cutaneous melanoma, mammary carcinoma and non-small cell pulmonary carcinoma. This new application of immunohistochemistry represents a valuable, clinically relevant adjunct to the repertoire of the surgical pathologist.

Expression of the high molecular weight melanoma-associated antigen by pericytes during angiogenesis in tumors and in healing wounds.

In the course of immunohistochemical characterization of murine monoclonal antibodies recognizing the human high molecular weight-melanoma associated antigen (HMW-MAA), a striking reactivity with blood vessels in the tumor stroma was noted. Immunocytochemical analysis by light and electronmicroscopy of a panel of tissues and cell lines showed that the staining of microvessels by anti-HMW-MAA monoclonal antibodies was restricted to pericytes. Correspondingly, anti-HMW-MAA monoclonal antibodies were found to react with cultured pericytes from human brain, but not with endothelial cells in serologic assays, and to immunoprecipitate from biosynthetically labeled pericytes an antigen with the characteristic structural profile of HMW-MAA. At the subcellular level, the expression of HMW-MAA in cultured pericytes was mainly restricted to microspikes that are localized in clusters on the cellular membrane. Staining by anti-HMW-MAA monoclonal antibodies of pericytes was not only found in the tumor stroma, but also in other lesions associated with angiogenesis, such as granulation tissue of wound healing and synovitis. In these lesions, microvascular staining for another marker of pericytes, ie, alpha-smooth muscle actin, also was observed. These results suggest that, in conditions associated with vascular proliferation, 1) pericytes acquire HMW-MAA and 2) the number of pericytes may be increased as compared with normal tissues.

Interfacial phenomena: an in vitro study of the effect of calcium phosphate (Ca-P) ceramic on bone formation.

In previous studies we developed a RF magnetron sputter technique for the production of thin Ca-P coatings. With this technique coatings can be produced that vary in Ca/P ratio as well as in structural appearance. The aim of this investigation was to obtain more understanding of the biological behavior of these coatings by way of in vitro experiments. The effect of noncoated titanium (Ti) and three different Ca-P-sputtered surfaces on the proliferation and differentiation (morphology and matrix production) of osteoblast-like cells was studied. Proliferation was determined using counting procedures; morphology was studied by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Fluorescent markers and energy-dispersive X-ray microanalysis (EDX) were used to obtain quantitative and compositional information about the resultant calcified extracellular matrix (ECM). Results demonstrated that proliferation of the osteoblast-like cells was significantly (p < 0.05) higher on noncoated than on Ca-P-coated samples. On the other hand, more mineralized ECM was formed on the coated surfaces. In addition, TEM confirmed that the cells on the coated substrates were surrounded by ECM with collagen fibers embedded in crystallized, needle-shaped structures. On the basis of these findings, we concluded that: (1) the investigated Ca-P sputter coatings possess the capacity to activate the differentiation and expression of osteogenic cells, and (2) bone formation proceeds faster on Ca-P surfaces than on Ti substrates. Further, this bone-inductive effect appeared to be dependent on the Ca-P ratio of the deposited coatings.

Novel proline substitution mutations in keratin 16 in two cases of pachyonychia congenita type 1.

Pachyonychia congenita (PC) is a group of inherited ectodermal dysplasias, the characteristic phenotype being hypertrophic nail dystrophy. Two main clinical subtypes, PC-1 and PC-2, are inherited as autosomal dominant disorders, but other less well characterized clinical forms also exist. The PC-1 phenotype may be distinguished by the absence of the epidermal cysts found in PC-2, and it has been shown to be caused by mutations in either keratin K16 or its expression partner, the K6a isoform of K6. Mutations in K16 have also been shown to cause a milder related phenotype, focal non-epidermolytic palmoplantar keratoderma. Recently, we have developed a long-range polymerase chain reaction (PCR) strategy which allows specific amplification of the entire functional K16 gene (KRT16A), without amplification of the two K16 pseudogenes (psiKRT16B and psiKRT16C), enabling mutation analysis based on genomic DNA. Here, using this methodology, we describe novel mutations R127P and Q122P in the helix 1A domain of K16 in two families presenting with PC-1. Both mutations were excluded from 50 normal unrelated individuals by restriction enzyme analysis of K16 PCR fragments. In one family, ultrastructural analysis was performed, revealing distinctive tonofilament abnormalities. Specifically, keratin filament bundles were greatly condensed, but did not form the dense amorphous aggregates seen in a number of other keratin disorders. In the second kindred, autosomal dominant cataract was present in some but not all members affected by PC. As the cataract phenotype did not fully cosegregate with the K16 mutation, and given that K16 is not expressed in the lens, these two phenotypes may be coincidental.

Differential expression of markers for endothelial cells, pericytes, and basal lamina in the microvasculature of tumors and granulation tissue.

The structure and function of the tumor microvasculature is of great interest for cancer biology, diagnosis, and therapy. The distribution of endothelial cells, pericytes, and basal lamina in tumors is not well documented. In this study, the authors investigated the distribution of markers for these different components in a series of malignant human tumors and in human granulation tissue, both situations with extensive angiogenesis. Their results show a striking heterogeneity in the expression of markers for pericytes and endothelial cells between different tumors, but also within a single tumor lesion. To be able to distinguish between these two adjacent cell types decisively, all marker studies were carried out both on the light and the electron microscopical level and compared with staining results in granulation tissue of cutaneous wounds in healthy volunteers and of decubitus lesions. In granulation tissue of decubitus lesions, well-defined zones with increasing levels of maturation can be delineated. It was found that antibodies recognizing von Willebrand factor often failed to stain the tumor capillaries. Of the pericyte markers, alpha-smooth muscle actin was only locally expressed by pericytes in the tumor vasculature, whereas the high-molecular-weight melanoma-associated antigen, a chondroitin sulfate proteoglycan, stained the microvasculature broadly. Staining of the basal lamina components collagen type IV and laminin was, within the tumor, not restricted to the microvasculature. From their findings the authors conclude that 1) for the visualization of the tumor vasculature, antibodies recognizing endothelial markers, especially monoclonal antibodies PAL-E and BMA 120, are preferable to those recognizing pericytes or basal lamina; 2) within the microvasculature of tumors and granulation tissue, a heterogeneity of expression of endothelial and pericyte markers is observed; 3) during the formation of granulation tissue, all three microvascular components can be demonstrated already in the histologically earliest stage, suggesting not only an involvement of endothelial cells but also of pericytes and basal lamina in the initial steps of angiogenesis in wound healing.

Comparison of immunohistochemistry with immunoassay (ELISA) for the detection of components of the plasminogen activation system in human tumour tissue.

Enzyme-linked immunosorbent assay (ELISA) methods and immunohistochemistry (IHC) are techniques that provide information on protein expression in tissue samples. Both methods have been used to investigate the impact of the plasminogen activation (PA) system in cancer. In the present paper we first compared the expression levels of uPA, tPA, PAI-1 and uPAR in a compound group consisting of 33 cancer lesions of various origin (breast, lung, colon, cervix and melanoma) as quantitated by ELISA and semi-quantitated by IHC. Secondly, the same kind of comparison was performed on a group of 23 melanoma lesions and a group of 28 breast carcinoma lesions. The two techniques were applied to adjacent parts of the same frozen tissue sample, enabling the comparison of results obtained on material of almost identical composition. Spearman correlation coefficients between IHC results and ELISA results for uPA, tPA, PAI-1 and uPAR varied between 0.41 and 0.78, and were higher for the compound group and the breast cancer group than for the melanoma group. Although a higher IHC score category was always associated with an increased median ELISA value, there was an overlap of ELISA values from different scoring classes. Hence, for the individual tumour cases the relation between ELISA and IHC is ambiguous. This indicates that the two techniques are not directly interchangeable and that their value for clinical purposes may be different.

Leukocyte antigen CD34 is expressed by a subset of cultured endothelial cells and on endothelial abluminal microprocesses in the tumor stroma.

It has been reported that the human haemopoietic progenitor cell antigen CD34 is also expressed by vascular structures. To investigate its precise vascular localization, we have studied the cellular and subcellular distribution of CD34 in normal tissues and pathologic tissues with neovascularization. In normal resting tissues, anti-CD34 antibodies, ICH3 and QBEND-10 predominantly stain the luminal endothelial membrane, whereas the abluminal membrane is negative or weakly positive. In contrast, a striking staining of endothelial abluminal microprocesses (EAM) was found in the tumor stroma. These structures, measuring up to 20 microns in length, could be observed in thick vibratome sections both at the tips of vascular sprouts and, also frequently, on fully formed microvessels. The number of vascular sprouts and EAM varied widely between different tumors. CD34-stained EAM were sparsely present in fetal tissue of 10 weeks gestation, but they could not be demonstrated in granulation tissue of wound healing. By immunoelectron microscopy, the EAM were continuous with the cytoplasm of endothelial cells showing an immature phenotype as seen in regeneration. In cultured human umbilical vein endothelium, CD34 was preferentially found on a small subset of cells with the morphologic appearance of migrating cells. These findings suggest that CD34 is an endothelial marker for EAM present during angiogenesis.