Journal of Cell Science partnership with Dryad

Journal of Cell Science makes data accessibility easy with Dryad

Endothelial expression of the α6β4 integrin is negatively regulated during angiogenesis
Tejindervir S. Hiran, Joseph E. Mazurkiewicz, Paul Kreienberg, Frank L. Rice, Susan E. LaFlamme


Development and homeostasis of the vascular system requires integrin-facilitated cellular adhesion, migration, proliferation and survival. A specific role for the α6β4 integrin in the vasculature, however, has not been identified. Using immunohistochemistry, we observed α6β4 expression on the dermal microvasculature of human foreskin. Analysis of individual cells isolated from trypsin-disrupted foreskin tissue indicated that α6β4 was expressed by a subset of epithelial and endothelial cells, and not by smooth muscle cells. Expression of α6β4 was also analyzed during new vessel growth using explants of human saphenous vein cultured in fibrinogen gels. The results indicate that α6β4 is not expressed by outgrowing endothelial cells, and is downregulated by the original α6β4-positive endothelial cells of the explant. To determine whether α6β4 is expressed during angiogenesis in vivo, the expression of the β4 subunit was analyzed during the development of the mouse mystacial (whisker) pad. Immunohistochemical staining of the whisker pad indicates that β4 is expressed by the adult vasculature. To identify when and where β4 is turned on in the vasculature, we examined the whisker pads from the developing embryo (E19.5 pc), and from postnatal days zero (P0), three (P3) and seven (P7) pups. The expression of α6β4 was found to be turned on spatially and temporally from caudal to rostral regions and from the deep to superficial vasculature, correlating with the maturation of the whisker pad and its corresponding vasculature. Together, these findings suggest a potential role for α6β4 as a negative component of the angiogenic switch, whereas expression of α6β4 on the adult vasculature may indicate regions requiring additional adhesive mechanisms.


The vascular system is responsible for the nutrient and oxygen supply of an organism. It develops initially via vasculogenesis, and subsequently through angiogenic remodeling (Flamme et al., 1997). Angiogenesis involves the sprouting of endothelial cells from the existing vasculature, by a process that requires the migration and proliferation of endothelial cells, resulting in new vessel formation (Carmeliet, 2000). Stromal cells destined to become smooth muscle cells are recruited, ensheathing vessels except the capillaries. The end product is a fully developed vasculature, in which the cells are differentiated and quiescent, with endothelial cells stably adhering to their underlying basement membrane (Carmeliet, 2000). Throughout these processes, integrins play important roles in the development and maintenance of the vasculature (Hynes et al., 1999), by regulating cell adhesion, migration, proliferation and survival (Albelda et al., 1991; Giancotti and Ruoslahti, 1999; Pinter et al., 1997).

Integrins are α/β heterodimeric transmembrane proteins, and the specific combination of α and β subunits determines ligand-binding specificity (Hynes, 1992). Endothelial cells express several different integrins, whose roles in regulating endothelial cell adhesion, migration, proliferation and apoptosis have been examined (Bazzoni et al., 1999; Hynes et al., 1999). The αv integrins have received most attention, because they are upregulated during angiogenesis (Brooks et al., 1994; Clark et al., 1996) and blocking their function with antagonists can inhibit angiogenesis (Clark et al., 1996; Eliceiri and Cheresh, 1999). However, the role of αvβ3 and αvβ5 is more complex than earlier thought, since mice lacking α3 and/or β5 integrins show normal vascular development and enhanced pathological angiogenesis, suggesting that the αvβ3 and αvβ5 integrins may, in fact, negatively regulate the angiogenic process (Reynolds et al., 2002). The α5β1 integrin may play an important role in regulating new vessel growth, as demonstrated by α5-null mice that die during embryogenesis as a result of defects in vascular development (Yang et al., 1993). Blocking the function of α5 integrins can also inhibit angiogenesis (Kim et al., 2000). Roles for other integrin heterodimers in new vessel growth have also been described (Bazzoni et al., 1999; Hynes et al., 1999; Senger et al., 1997; Senger et al., 2002). Additionally, several studies have reported the expression of α6β4 in the vasculature (Enenstein and Kramer, 1994; Kennel et al., 1992; Koukoulis et al., 1991; Ryynanen et al., 1991); however, there have been conflicting reports as to whether it is expressed by endothelial cells or smooth muscle cells (Cremona et al., 1994).

Although the role of the α6β4 integrin in the vasculature is unclear, its function in epithelia is well characterized. In keratinocytes, α6β4 is required for maintaining firm epithelial adhesion to the underlying dermis (Dowling et al., 1996; van der Neut et al., 1996) by connecting the laminin-containing basement membrane with the intracellular keratin intermediate filaments (Borradori and Sonnenberg, 1999). In some carcinoma cells, however, α6β4 promotes migration by activating specific signaling pathways and interacting with the actin cytoskeleton (Mercurio et al., 2001; Trusolino et al., 2001). The function of α6β4 in the vasculature may be analogous to its function in epithelial and carcinoma cells; α6β4 may confer a promigratory phenotype in response to angiogenic stimuli, and also contribute to stable adhesion in the mature vasculature.

In this study, we were interested in determining whether the α6β4 integrin functions in the vasculature to promote angiogenesis. In initial studies, we confirmed the expression of α6β4 on dermal microvascular endothelial cells in situ. We did not observe expression of α6β4 on smooth muscle cells. Additional experiments examined whether the expression of α6β4 was regulated during the formation of new vessels. Explant angiogenesis assays, using segments of human saphenous vein, demonstrated that endothelial cells downregulate α6β4 in the explant and that outgrowing endothelial cells also do not express α6β4. Additionally, we did not observe expression of α6β4 during early vascular development of the murine whisker pad, but α6β4 was expressed by the same vasculature of the adult animal. The developmental expression of α6β4 correlated spatially and temporally with the maturation of the whisker pad and its corresponding vasculature. Taken together, our studies indicate that α6β4 is not expressed during new vessel growth in our assays. This result implies the α6β4 integrin may be a negative component of the angiogenic switch.

Materials and Methods


The following antibodies were used: mouse monoclonal antibody (mAb) 3E1 (Chemicon International, Temecula, CA) and rat mAb 439-9B (BD Biosciences, San Diego, CA) to the human β4 subunit; rat mAb 346-11A to the mouse β4 subunit (Kennel et al., 1986) (a gift from Dr Stephen Kennel, Oak Ridge National Laboratory, TN); rabbit polyclonal antibody (pAb) to human von Willebrand Factor (Accurate Chemical & Scientific Corporation, Westbury, NY); rabbit pAb to human PECAM-1 (Albelda et al., 1991) (a gift from Dr Albelda, University of Pennsylvania Medical Center, PA); rabbit anti-mouse PECAM-1 (Pinter et al., 1997) (a gift from Dr J. Madri, Yale University School of Medicine); rat mAb MEC13.3 to mouse PECAM-1 (BD Biosciences); mouse mAb 1AF to human smooth muscle actin (Sigma, St Louis, MO); mouse mAb AE1/AE3 to human cytokeratins (Sigma); and rabbit pAb to human s100A and s100B (Serotec, Raleigh, NC). The secondary antibodies used were Alexa 488- or Alexa 546-conjugated goat anti-mouse, anti-rat and anti-rabbit IgG, CY3-conjugated donkey anti-rat IgG, FITC-conjugated goat anti-rat IgG adsorbed against mouse, TRITC-conjugated goat anti-mouse IgG adsorbed against rat (Jackson ImmunoResearch Laboratories, West Grove, PA), and the biotin-conjugated antibodies goat anti-rat IgG (Jackson ImmunoResearch Laboratories), goat anti-mouse IgG (Molecular Probes) and goat anti-rabbit IgG (Santa Cruz Biotechnology, Santa Cruz, CA). When double staining with primary antibodies 3E1 (mouse IgG1) together with 1AF (mouse IgG2a) the isotype-specific secondary antibodies used were Alexa 488- and Alexa 546-conjugated goat anti-mouse IgG1 and IgG2a, respectively (Molecular Probes, Eugene, OR).

Isolation of primary cells from human neonatal foreskin

Human neonatal foreskin, obtained under Albany Medical College IRB approved protocols, was cut into small pieces (∼1×1 mm3) and placed in 0.25% Trypsin in PBS for 16-20 hours at 4°C. The supernatant was filtered through cheesecloth and centrifuged at 435 g at 4°C for 5 minutes. The cells were resuspended in PBS and plated into 6-well plates at 5-10×104 cells/well. Each well contained coverslips precoated with poly-l-lysine (10 μg/ml in PBS for 2 hours at 37°C, washed 3× with PBS). The 6-well plates were centrifuged at 560 g for 5 minutes at 4°C, and the attached cells were fixed with 4% paraformaldehyde in PBS for 30 minutes. The cells were washed 3× with PBS, then immunostained.

Explant cultures

Explant cultures were prepared with slight modifications from a previously published method (Nicosia and Madri, 1987; Nicosia and Ottinetti, 1990). A segment of human saphenous vein, obtained under Albany Medical College IRB approved protocols, was cut into cross sections approximately 1 mm thick. The sections were washed 10× with PBS, and then placed onto a solidified gel of fibrinogen (Sigma) in 24-well plates. The gel was prepared by adding 3 mg of fibrinogen to 1 ml of EBM (Clonetics, San Diego, CA), and gelled by the addition of 1 unit/ml of thrombin (Sigma). The vein was positioned in the center of the well on top of the fibrinogen gel, and overlaid with additional fibrinogen to which thrombin had been added immediately prior to use. The gel was allowed to solidify, and EBM containing 15% serum, 100 units/ml penicillin/streptomycin, 1 μg/ml hydrocortisone (Sigma), 10 ng/ml EGF (Collaborative Biomedical Products, Bedford, MA) was added on top. The culture medium was changed every other day with additions of the fibrinolytic inhibitor ϵ-amino-n-caproic acid (Sigma) for the first 4 days at 300 μg/ml, then subsequently at 50 μg/ml. Human neonatal foreskin explants were placed directly onto coverslips in 6-well plates, and medium changed as above. Saphenous vein explants were fixed (as above) for 4 hours at 4°C then washed 3× for 5 minutes with PBS, cryoprotected in 30% sucrose in PBS overnight at 4°C, dehydrated by passage through 75% and 100% OCT for 24 hours each, and then cut into 10-12 μm sections. Human saphenous vein, tissue and explant, exhibited high intensity broad-spectrum autofluorescence and therefore we used immunoperoxidase staining, yielding colored reaction products. Human foreskin explants were fixed for 30 minutes, then immunostained.

Preparation of mouse mystacial (whisker) pads

Swiss Webster mice were purchased from Taconic. Adult, timed pregnancy mice and neonates were euthanized and the mystacial pads were removed and post-fixed and cryoprotected as for the explants. All procedures were in accordance with Albany Medical College IACUC approved protocols.


Sectioned tissue was rehydrated in PBS, permeabilized with 0.5% Triton X-100 in PBS, blocked with 3% BSA-glycine, and then preincubated with 10% normal goat serum (Pierce, Rockford, IL) in PBS. Sections were incubated overnight at 4°C with primary antibodies diluted in PBS containing 0.5% BSA, 10% goat serum and 0.1% Tween 20. Sections were washed and incubated with the appropriately labeled fluorescent secondary antibodies, and then washed and mounted in Anti-Fade (Molecular Probes). Cells on coverslips were permeabilized, blocked with BSA-glycine, then incubated in primary antibody for 1 hour at room temperature, but otherwise treated as above. For immunoperoxidase staining, sections were rehydrated and endogenous peroxidase quenched by incubating in 3% H2O2 in H2O for 10 minutes, then rinsed 2× for 1 minute in distilled H2O. Sections were permeabilized and incubated with primary antibody and washed as above, and then incubated with the appropriate biotinylated secondary antibodies for 30 minutes, followed by incubation with Vectastain ABC Elite reagents, and developed with the peroxidase substrate DAB reagents (Vector Laboratories, Burlingame, CA) as described by the manufacturer. Sections were dehydrated in a graded series of ethanol washes, cleared in xylene and mounted with Permount (Fisher, Pittsburgh, PA). Controls were as above. In some instances, tissue sections were rehydrated in PBS, incubated for 1 minute in Gills Hematoxylin No. 2 (Polysciences, Warrington, PA), followed by 3 minutes in tap water, 3 minutes in Scotts solution (24 mM NaHCO3, 166 mM MgSO4 in H2O), 3 minutes in tap water, 10 seconds in Eosin-Y (Richard Allen Scientific, Kalamazoo, MI), 3 minutes in tap water, dehydrated and mounted as above. Immunostained material was visualized with either an Olympus BX60 microscope with an attached Spot camera and associated software, or with an Olympus AX70 with a Sony DKCST4 Digital Photo Camera using Northern Eclipse software. Immunofluorescently labeled sections were also visualized on a Noran Oz confocal laser scanning microscope interfaced with a Nikon Diaphot 200 inverted microscope equipped with a PlanApo ×60, 1.4 NA oil-immersion objective. Controls for all immunofluorescent staining in the absence of primary antibody showed no significant fluorescence over tissue or cellular autofluorescence.


The α6β4 integrin is expressed by microvascular endothelial cells

Reports from several laboratories have described the expression of the α6β4 integrin on a subset of blood vessels; however, there have been conflicting reports as to whether the α6β4 integrin is expressed by endothelial or vascular smooth muscle cells (Cremona et al., 1994; Enenstein and Kramer, 1994; Kennel et al., 1992; Koukoulis et al., 1991; Mechtersheimer et al., 1994; Ryynanen et al., 1991). In our initial studies, we used sections of human neonatal foreskin to examine the expression of α6β4 in the dermal vasculature by fluorescent immunohistochemistry. As shown in Fig. 1, the β4 subunit colocalized with the endothelial markers von Willebrand Factor (vWF) (Fig. 1C, yellow) and the panendothelial marker PECAM-1 (data not shown). Colocalization was also observed between β4 and its only known integrin subunit partner α6 (data not shown) (Hynes, 1992). As expected, expression of α6β4 was detected on the basal keratinocytes, which served as a positive internal control. In addition, the analysis of β4 expression on cross sections of vessels, viewed at high magnification by confocal microscopy, suggested that α6β4 localized at the basal surface of the endothelial cells, in contrast to vWF, which is mostly intracellular (Fig. 1D).

Fig. 1.

The α6β4 integrin is expressed by human dermal microvascular endothelial cells. (A-D) Sections of human neonatal foreskin were double immunostained with antibodies to the endothelial marker von Willebrand Factor (vWF) (A), and with 3E1 antibodies to the β4 subunit (B). The white dotted line indicates the epidermal-dermal interface. (C) The colocalization of β4 and vWF is seen in the merged image as yellow, with a representative β4-positive vessel indicated by the arrow. (D) A microvessel in cross section examined at higher magnification using confocal laser scanning microscopy, showing vWF (green) and β4 (red). (E-J) Cells isolated from human neonatal foreskin were double immunostained to detect β4 and endothelial cell-, smooth muscle cell-, and keratinocyte-specific markers: (E) cells stained with pAb to PECAM-1 (red) and 3E1 to β4 (green); (G) cells stained with 1A4 to smooth muscle actin (red) and 3E1 to β4 (green); and (I) cells stained with AE1/AE3 to human epidermal keratins (red) to detect keratinocytes and 439-9B to β4 (green). (E,G,I) The corresponding phase contrast images are shown in F, H and J, respectively. The cells that stained brightly with only β4 in E and G are probably keratinocytes. Scale bars: (A-C) 50 μm; (D-J) 10 μm.

To more precisely characterize the expression of β4 integrins on vascular cells, individual cells isolated from trypsin-disrupted human neonatal foreskin tissue were adhered to polylysine-coated coverslips and used for immunocytochemistry. The cell-type expression of α6β4 was examined by determining whether β4 was co-expressed with epithelial-, endothelial-, and/or smooth muscle cell-specific markers. These studies indicated that β4 integrin was expressed by endothelial cells as identified either by PECAM-1 (Fig. 1E) or vWF expression (data not shown). Some endothelial cells did not show β4 expression (not shown), which is consistent with published findings that α6β4 is expressed by a subset of vessels (Mechtersheimer et al., 1994). We did not observe the co-expression of β4 with smooth muscle actin (Fig. 1G), suggesting that β4 may not be expressed by smooth muscle cells as previously reported (Cremona et al., 1994). As expected only a subset of cytokeratin-positive cells expressed β4 (Fig. 1I), since α6β4 is known to be expressed by basal keratinocytes and not by cells of the suprabasal layers (De Luca et al., 1990). Interestingly, the expression of β4 remained polarized in both endothelial cells and keratinocytes dissociated from foreskin tissue (Fig. 1E,I).

The α6β4 integrin is negatively regulated by angiogenic endothelial cells in explant culture

To begin to understand the function of α6β4 in the vasculature, we investigated whether the expression of α6β4 was regulated during new vessel growth. To determine this, we performed explant angiogenesis assays using segments of human saphenous vein (Kruger et al., 2000; Slomp et al., 1996) which have been successfully used in these types of assays. When we examined β4 expression in cross sections of saphenous vein, we found it was expressed by endothelial cells of the vasa vasorum, but not by endothelial cells lining the main lumen of the saphenous vein (Fig. 2A-E). To determine whether the expression of β4 integrins is regulated by outgrowing endothelial cells, human saphenous vein explants were cultured in fibrinogen gels as described by others (Kruger et al., 2000; Nicosia and Ottinetti, 1990; Slomp et al., 1996). Endothelial cell outgrowth was observed from the explant after 7 days with robust growth observed after 14 days. At day 14, most of the cells were arranged in tube-like structures that resembled vessels. Day-14 explants were immunostained to confirm endothelial outgrowth with vWF (Fig. 3A,B) and PECAM-1 (data not shown). Virtually all cells in the outgrowth stained positive with vWF and PECAM-1, with almost none of the outgrowth positive for smooth muscle actin (Fig. 3E,F). Interestingly, β4 expression was not observed on outgrowing endothelial cells, or on endothelial cells remaining in the vasa vasorum of the explant at day 14 (Fig. 3C,D). This was also true of α6 (data not shown).

Fig. 2.

β4 integrin is expressed by endothelial cells of the vasa vasorum but not by endothelial cells lining the lumen of human saphenous vein. (A) Frozen section of human saphenous vein stained with Hematoxylin and Eosin; the asterisk indicates the lumen of the saphenous vein. The dashed box shows an example of the region chosen for higher magnification immunohistochemical imaging in BE. Cross sections of saphenous vein were stained with antibodies to vWF (B), marking endothelial cells lining the main lumen (arrow) and the vasa vasorum (arrowheads) and 3E1 to β4 (D), selectively staining endothelial cells of the vasa vasorum (arrowheads). (C,E) Control immunostaining in the absence of primary antibody, for B and D, respectively. Scale bars: (A) 500 μm; (B-E) 200 μm.

Fig. 3.

Angiogenic endothelial cells do not express the α6β4 integrin in explant cultures. Segments of human saphenous vein were grown as explant cultures in fibrin gels for 14 days. The plus sign indicates the explant tissue, with the outer surface of the original explant tissue outlined with a dotted blue line in A,C,E. (A,B) Cross sections of explant and outgrowth stained with antibodies to vWF, which stained outgrowing endothelial cells as well as endothelial cells remaining in the vasa vasorum and lining the saphenous vein lumen. A higher magnification of a region of outgrowth is shown in B. (C,E) Cross sections stained with 3E1 to β4 (C) showing no β4 expression on either the explant saphenous vein or the outgrowing endothelial cells, and 1A4 to smooth muscle actin (E) which stained the original explant, but not outgrowing cells. (D,F) Phase contrast images of C and E, respectively. Arrows indicate the direction of outgrowth. Controls with secondary antibody alone showed insignificant staining similar to the controls for the immunostainings in Fig. 2 (data not shown). Scale bars: (A,C-F) 200 μm; (B) 20 μm.

To exclude the possibility that the outgrowing endothelial cells were originating only from the lumenal endothelial cells of the explant as opposed to the β4-positive endothelial cells present in the vasa vasorum, the endothelial lining of the lumen was stripped by collagenase digestion, leaving the endothelial cells of the vasa vasorum intact as determined by PECAM-1 staining (Fig. 4A) and vWF (data not shown). Collagenase digestion did not affect the expression of β4 in the vasa vasorum (Fig. 4B). Explant assays were performed using collagenase-treated and untreated saphenous vein segments. A phase image of day-14 explants that had not been treated with collagenase showed normal outgrowth in the vessel lumen (Fig. 4C). In contrast, there was significantly diminished outgrowth into the lumen of the collagenase-treated explant (Fig. 4D), whereas the outgrowth of endothelial cells from the vasa vasorum of the vessel was unaffected (Fig. 4E). Immunostaining of collagenase-treated explants showed endothelial-positive staining for PECAM-1 (Fig. 4F,G) and vWF (data not shown). Similar to the outgrowth from untreated tissue, β4 expression was also downregulated both in the endothelial outgrowth and the original explant at day 14 (Fig. 4H,I). This was also true of α6 expression (data not shown). These results indicate that endothelial cells downregulate α6β4 expression during new vessel growth in culture.

Fig. 4.

α6β4-integrin-positive endothelial cells of the vasa vasorum contribute to angiogenesis in explant cultures of human saphenous vein. Human saphenous vein was treated with collagenase to strip the endothelial lining from the main lumen. (A,B) Cross sections of collagenase-treated explants were stained with pAb to PECAM-1 (A) and 3E1 to β4 (B). Endothelial cells of the vasa vasorum maintained PECAM-1 and β4 expression (black arrowheads); however, expression of PECAM-1 was not observed on the lumenal surface (open arrowhead), indicating the endothelial cells had been successfully removed. (C,D) Phase contrast images of endothelial outgrowth from collagenase-treated (D) or untreated (C) segments of saphenous vein after 14 days in culture in fibrinogen gels. Little angiogenic outgrowth occurred in the collagenase-treated lumen (D); however, robust endothelial outgrowth was observed from the vasa vasorum (E). (F-I) Cross sections of collagenase-treated saphenous vein explants were immunostained after 14 days in culture with either pAb to PECAM-1 (F,G), or with 3E1 to β4 (H), indicating the outgrowing endothelial cells (arrow) do not express β4. A region of outgrowth stained for PECAM-1 is shown at higher magnification in G. (I) A phase contrast image of H. (F,H) The perimeter of the explant is outlined by the dotted blue line. Controls with secondary antibody alone showed insignificant staining similar to the controls for immunostainings in Fig. 2 (data not shown). (J) Day-12 explant of human neonatal foreskin double stained with AE1/AE3 to human epidermal keratins (red) and 439-9B to β4 (green), demonstrating that keratinocytes do not downregulate β4 in explant cultures. Asterisks indicate vessel lumen. Plus signs indicate original explanted tissue. Arrows indicate outgrowing endothelial cells or keratinocytes. Scale bars: (A,B,F,H,I) 100 μm; (G) 20 μm; (J) 50 μm.

To confirm that the loss of endothelial expression of β4 integrins was due to cell-type specific regulation and was not an artifact of our outgrowth assay, explants of human neonatal foreskin tissue were similarly cultured, since migrating and proliferating basal keratinocytes are known to maintain β4 expression (Gipson et al., 1993; Larjava et al., 1993; Mercurio et al., 2001; Nguyen et al., 2000). The outgrowing cells maintained an epithelial appearance, and their identity as keratinocytes was confirmed by immunohistochemical staining with an epithelial cytokeratin marker (Fig. 4J). Dual labeling showed that the outgrowing keratinocytes maintained their expression of β4 integrins, which appeared in a punctate pattern typical of hemidesmosomes. This result indicates that outgrowing endothelial cells in explant culture down-regulate β4 expression in a cell-type specific manner.

Expression of α6β4 is temporally and spatially regulated during vascular development

Since the expression of some genes can be altered when cells are placed in culture (Antequera et al., 1990; St Croix et al., 2000) and this could potentially occur in a cell-type-dependent manner, we were interested in determining whether α6β4 was expressed by newly forming vessels in situ. To accomplish this, we analyzed the expression of α6β4 in newly forming vasculature using the developing murine mystacial (whisker) pad as a model. This model was chosen because the whisker pad has a well described and predictable vascular architecture shown schematically in Fig. 5A (Fundin et al., 1997b). Initial studies demonstrated the expression of β4 integrins in the vasculature of the adult whisker pad (Figs 5, 8). Because β4 may be expressed on endothelial cells, perineural cells and Schwann cells within the dermis (Niessen et al., 1994) and because of limitations in the availability of immunological reagents for murine tissue, we analyzed the expression of α6β4 by double-immunofluorescence staining of neighboring sections, using either antibodies to PECAM-1 and to s100, a marker for Schwann cells (Kligman and Hilt, 1988; Zimmer et al., 1995), or antibodies to β4 and s100. [Note that although s100 expression is concentrated in Schwann cells, it also labels perineurium, fat cells, and chondrocytes to varying degrees (Zimmer et al., 1995).] Endothelial staining with PECAM-1 antibodies was observed throughout the different vascular beds of the whisker follicle (Fig. 5B). This staining was distinct, compared to s100 expression. In the neighboring section, β4 staining was prominent on the basal keratinocytes and epithelial invaginations of both hair and whisker follicles, which served as a positive control (Fig. 5C). β4 was also expressed by the Schwann cells and perineurium of nerves that were identified by s100 labeling. β4 was expressed on endothelial cells of most vessels supporting the various whisker pad sites (Fig. 5C,E,I). This was confirmed by double-label immunofluorescence showing the colocalization of β4 and PECAM-1 (Fig. 8A-C). Interestingly, β4 was lacking on a set of vessels affiliated with papillary muscle slings (Fig. 5G).

Fig. 5.

α6β4 is expressed by endothelial cells in the adult mystacial pad. (A) Schematic representation of the vasculature associated with a whisker follicle (wf) in the adult mouse mystacial whisker pad. Small fur hairs (h) and a large whisker (w) are shown in dark gray. The epidermis (e), small hair follicles, and large whisker follicle are shown in light gray. The whisker follicle is surrounded by a vascular sinus (pink) which is enclosed by a dense collagen capsule (medium gray). Arteries, arterioles and capillary beds are red; veins and venules are blue; nerves are green; and piloerector muscles are brown (Fundin et al., 1997a; Fundin et al., 1997b; Rice et al., 1997). Dashed rectangles indicate areas comparable to sites in B,C, and in higher magnification in D-I. (B-I) Immunofluorescence staining of longitudinal sections of an adult whisker follicle. Adjacent sections were immunostained with anti-s100 (B,C: green) for Schwann cells on nerve axons. Double immunostaining with MEC 13.3 (B: red) labels PECAM-1 on the endothelium of all blood vessels. Double immunostaining with 346-11A (C: red) labels β4 on the perineurium of nerves and endothelium of many blood vessels. Regions of B and C (dotted boxes) are shown at higher magnification (D,F,H) and (E,G,I) respectively. Arrows indicate nerves. Arrowheads indicate examples of β4-positive vessels; dotted ovals indicate β4-negative vessels. β4 labeling is also present on basal keratinocytes (E, arrows) of the epidermis and hair follicles. Scale bars: 100 μm.

Fig. 8.

Confocal microscope images showing double immunostaining of microvascular endothelial cells for β4 (red) and PECAM-1 (green). Each image is a maximum intensity projection image of a sub-stack of six optical sections taken from a larger stack of images collected at 0.3 μm intervals. The images were collected in the region just below the epidermal-dermal junction from sections of skin from adult (A-C) and P3 (D-F) mice. In the adult, almost all the vessels in this region showed immunoreactivity for β4 and PECAM-1 (A-C, arrowheads). At P3, vessels expressing β4 were rare (D-E, solid arrowhead), most lacked β4 (open arrowheads). No vessels in this location showed β4 immunoreactivity at P0 (not shown). Arrows indicate β4 immunoreactivity on basal keratinocytes in the epidermis (e) and hair follicles (f). Scale bar: 25 μm.

To examine β4 expression during the development of the vasculature, the whisker pad from E19.5 embryos (where E20 corresponded with parturition) was analyzed similarly to the adult tissue shown in Fig. 5. In general, β4 expression was not detected in the embryonic microvasculature when neighboring sections were stained with antibodies to PECAM-1 and β4. However, some endothelial β4 expression was observed in the caudal regions (Fig. 6A,B) and deep vasculature (Fig. 6C,D). Double labeling with β4 and vWF supported this observation (data not shown). β4 was also not coexpressed with s100, indicating that β4 is not expressed by Schwann cells at this time during the development of the whisker pad (Fig. 6A,B).

Fig. 6.

Following the caudal to rostral and deep to superficial development of the whisker pad vasculature (E), embryonic microvascular endothelial cells of the whisker pad only express the β4 integrin subunit in the caudal and deep vasculature in the whisker pad from an E19.5 p.c. animal. Sections from the caudal region (A,B) are shown immunostained with s100 (green) and either MEC 13.3 to PECAM-1 (A: red) or with 346-11A to β4 (B: red). Neighboring sections from a more rostral region are shown immunostained with MEC 13.3 to PECAM-1 (C) or 346-11A to β4 (D). The majority of the vasculature was β4 negative, however, some larger caliber vessels deeper in the tissue were β4 positive (arrowheads). (E) Schematic representation of the whisker pad showing six whisker follicles along the caudal to rostral axis, where the filled arrowhead indicates a whisker and the open arrowhead indicates a hair follicle. Intense β4 labeling is present on basal keratinocytes of the epidermis (e), hair follicles (f) and whisker follicles (wf). Scale bars: 100 μm.

Induction of β4 expression in the whisker pad vasculature was analyzed at postnatal day (P), zero (P0, equivalent to E20), P3 and P7. At P0, β4 expression could be seen on the vasculature in the caudal most region, and also in deeper regions of the tissue (Fig. 7B). However, at this time point, large regions of the vasculature still remained negative for β4 expression, such as between all of the whisker follicles (Fig. 7A,B). At P3, β4 expression was observed in the capillaries that lie between the facial muscles, again caudal and deep to the developing whisker pad (Fig. 7C,D). Using double-label immunofluorescence for β4 and PECAM-1 expression, little or no detectable endothelial β4 expression was observed on vessels between the whisker follicles (not shown) or between developing hair follicles in the upper dermis (Fig. 8D,E,F). By P7, the vasculature associated with the follicles of the whisker pad had started to express β4 (Fig. 7E,F). It should also be noted that the expression of β4 in the peripheral nervous system (perineural sheaths and Schwann cells) and β4 expression in the vasculature showed a similar pattern of progression from caudal to rostral and deep to superficial regions during the development of the whisker pad (Fig. 7B). Together these data suggest that the temporal and spatial progression of the β4 integrin in the vasculature may correlate with vascular maturation in the whisker pad.

Fig. 7.

Microvascular endothelial cells turn on β4 expression during postnatal development. The progression of endothelial β4 expression was examined during postnatal development of the whisker pad. Neighboring sections were immunostained with s100 (A-F: green) and either MEC 13.3 to PECAM-1 (A,C,E: red) or with 346-11A to β4 (B,D,F: red). Caudal regions from P0 embryos (A,B) and P3 embryos (C,D), and rostral region from P7 embryos (E,F) are shown. Arrowheads indicate examples of vessels that express β4 (A-F). At P0, only vessels in the most caudal and deep regions of the whisker pad showed expression of β4. By P3, the same regions show more vascular β4 expression. By P7, β4 expression is observed throughout the vasculature of the whisker pad, as shown on the rostral-most whisker follicles. Dotted ovals indicate examples of regions where vessels do not express β4. Scale bars: 100 μm.


Our studies focused on understanding the expression of the α6β4 integrin in the vasculature. We demonstrated that α6β4 is expressed by endothelial cells of mature vessels, but not by developing vessels in murine tissue. Using an explant angiogenesis assay, we cultured human adult saphenous vein segments in fibrin gels, and observed angiogenic outgrowth of endothelial cells. Using this explant system, we further showed that in response to signals that promote angiogenesis, the expression of α6β4 is turned off by mature vessels and is not expressed by outgrowing endothelial cells. These same outgrowing endothelial cells did, however, express the α5 and αv integrin subunits (data not shown). Together our results suggest that the α6β4 integrin is not required for developmental angiogenesis, and may need to be downregulated in some instances for angiogenesis to proceed.

The mechanisms regulating the expression of α6β4 in endothelial cells are poorly understood. Our data suggests that signals known to drive the angiogenic process, such as hypoxia, basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF) signaling (Carmeliet, 2000), may negatively impact on the expression of α6β4 in endothelial cells. In contrast, the expression of the αv, α1 and α2 integrins is positively regulated by angiogenic signals (Brooks et al., 1994; Senger et al., 1997). This is likely to be physiologically important because endothelial cell migration and invasion associated with angiogenesis is believed to occur through a collagen I-rich matrix (Senger et al., 1997) or a provisional matrix rich in fibronectin and vitronectin, which are ligands for these integrins (Senger, 1996). Interestingly, bovine adrenal cortex endothelial cells upregulate the expression of α6β4 in culture when stimulated with bFGF, suggesting that α6β4 may also promote angiogenesis in a tissue-specific manner (Klein et al., 1993). This is consistent with the observation that matrices rich in laminin 1, a known ligand for α6β4, induce the morphological differentiation of endothelial cells into capillary-like structures in culture (Grant and Kleinman, 1997).

Our data further suggests that signals promoting vessel maturation may positively regulate the expression of α6β4 by endothelial cells. For example, angiogenic vessels are known to undergo stabilization and maturation by mechanisms that require interaction between endothelial and mural cells, and the localized activation of TGFβ, resulting in the production of extracellular matrix components by endothelial cells (Folkman and D'Amore, 1996; Hellstrom et al., 2001; Hirschi and D'Amore, 1996; Neubauer et al., 1999; Nicosia and Madri, 1987; Shanker et al., 1999). Future studies in our laboratory will address these questions.

The expression of α6β4 by mature vessels, and its absence in newly developing vessels suggests that α6β4 may only be required after vessel maturation. This is supported by the fact that α6β4 integrin is a receptor for several laminin isoforms (Borradori and Sonnenberg, 1999; Lee et al., 1992), and although the isoforms expressed in the vasculature are not fully characterized, the endothelial basement membrane is known to be rich in laminins (Colognato and Yurchenco, 2000). Thus, α6β4 may be required for endothelial adhesion to the underlying basement membrane to promote the integrity of mature vessels. Consistent with this notion, previous studies from our laboratory suggest that α6β4 may uniquely contribute to endothelial cell adhesion by forming a transmembrane link between the basement membrane and the vimentin intermediate filament cytoskeleton (Homan et al., 1998).

The endothelial-specific downregulation of α6β4 during endothelial outgrowth contrasts with the known role of α6β4 in certain carcinoma cells where it promotes migration and invasion (Gambaletta et al., 2000; Mercurio et al., 2001; Trusolino et al., 2001). Basal keratinocytes also maintain expression of α6β4 during re-epithelialization (Gipson et al., 1993; Larjava et al., 1993; Mercurio et al., 2001; Nguyen et al., 2000). These differences in cell-type specific regulation of α6β4 may reflect the essential requirement of the epidermal layer to be firmly attached to the underlying basement membrane by the basal keratinocytes through their expression of α6β4. This is illustrated by the extensive detachment of the epidermal-dermal interface in neonatal β4 knockout mice (Dowling et al., 1996; van der Neut et al., 1996), and the microblistering phenotype observed in the skin of patients with Epidermolysis Bullosa, which is the result of mutations in β4 (Pulkkinen and Uitto, 1999). Together this strongly implies that even temporary downregulation of α6β4 expression by basal keratinocytes during re-epithelialization may compromise skin integrity, whereas remodeling vessels do not have this same requirement for maintained α6β4 expression.

Also, in contrast to keratinocytes, Schwann cells do not express β4 during in vitro migration except when they begin ensheathing/myelinating axons (Einheber et al., 1993; Niessen et al., 1994; Previtali et al., 2001). Our data indicates that β4 is not expressed by Schwann cells during embryonic development of the whisker pad, but is detected on Schwann cells at P0, as reported by others (Feltri et al., 2002). This expression pattern of β4 on perineural cells and Schwann cells, appears similar temporally and spatially to that observed on endothelial cells. Thus, similar signaling and/or transcriptional mechanisms may regulate β4 expression in Schwann cells and endothelial cells.

Our findings that α6β4 is not expressed on newly forming vessels is consistent with, and may in fact help explain, the lack of a gross pathological phenotype in the vasculature of the β4-null mice (Dowling et al., 1996; van der Neut et al., 1996). Our results also suggest that endothelial expression of α6β4 may be a negative component of angiogenesis, and that its expression needs to be downregulated at the onset of new vessel growth. This is consistent with recent reports that suggested that some integrins may negatively modulate angiogenic processes (Reynolds et al., 2002; Stupack et al., 2001), perhaps by inhibiting endothelial growth and even triggering apoptosis when the appropriate ligands are unavailable.

Additional studies are needed to fully understand the role of α6β4 in the vasculature. A vascular-specific conditional knockout would be valuable in this regard. Even with this in hand, if α6β4 functions together with other adhesion molecules to promote vascular integrity, a vascular phenotype may only be observed after vascular challenge, such as intense cardiovascular activity or an inflammatory response. Based on our findings, we propose a novel function for α6β4 as a potential negative regulator of angiogenesis in some instances. Future experiments will test this hypothesis by determining whether forced vascular expression of β4 integrins significantly inhibits angiogenesis and/or vascular development. Additionally, it will be important to determine whether α6β4 is also negatively regulated during angiogenesis associated with wound healing and tumor vascularization, since normal and pathological angiogenesis can involve different mechanisms (Carmeliet et al., 2001; Carmeliet and Jain, 2000; St Croix et al., 2000). Finally, if expression of β4 can be anti-angiogenic, it will become important to determine the molecular mechanisms involved, since such studies may have potential therapeutic value in controlling pathological angiogenesis in specific tissue environments.


We thank Drs S. Kennel, S. Albelda, and J. Madri for generously providing antibodies to β4 and PECAM-1, Drs J. Sottile, C. M. DiPersio, L. Van De Water and A. Berrier for critically reading this manuscript, and Debbie Moran for her assistance during the preparation of this manuscript. This work was funded by grants from the American Heart Association/New York State Affiliate and the National Institute of General Medical Sciences to S.E.L., from the National Institute of Neurosciences (NS-34692) and the Albany Medical College Strategic Research Plan to F.L.R., and from the National Center for Research Resources (RR-12894-01A1) to J.E.M.

  • Accepted May 23, 2003.


View Abstract