Caveolae-deficient endothelial cells show defects in the uptake and transport of albumin in vivo.

The role of endothelial cell caveolae in the uptake and transport of macromolecules from the blood-space to the tissue-space remains controversial. To address this issue directly, we employed caveolin-1 gene knock-out mice that lack caveolin-1 protein expression and caveolae organelles. Here, we show that endothelial cell caveolae are required for the efficient uptake and transport of a known caveolar ligand, i.e. albumin, in vivo. Caveolin-1-null mice were perfused with 5-nm gold-conjugated albumin, and its uptake was followed by transmission electron microscopy. Our results indicate that gold-conjugated albumin is not endocytosed by Cav-1-deficient lung endothelial cells and remains in the blood vessel lumen; in contrast, gold-conjugated albumin was concentrated and internalized by lung endothelial cell caveolae in wild-type mice, as expected. To quantitate this defect in uptake, we next studied the endocytosis of radioiodinated albumin using aortic ring segments from wild-type and Cav-1-null mice. Interestingly, little or no uptake of radioiodinated albumin was observed in the aortic segments from Cav-1-deficient mice, whereas aortic segments from wild-type mice showed robust uptake that was time- and temperature-dependent and competed by unlabeled albumin. We conclude that endothelial cell caveolae are required for the efficient uptake and transport of albumin from the blood to the interstitium.

Caveolae are vesicular organelles located near or attached to the plasma membrane. They represent an appendage of the plasma membrane. Caveolae are most abundant in endothelial cells, adipocytes, smooth muscle cells, and fibroblasts, although they are thought to exist in most cell types (reviewed in Refs. [1][2][3][4][5]. The exact function of caveolae remains largely unknown; however, they are thought to function in both cellular transport processes, such as endothelial cell transcytosis, and signal transduction. In the early 1950s (6,7), caveolae were first implicated in the fluid phase transcytosis of both large and small molecules across capillary endothelial cells (8). In addition, albumin undergoes receptor-mediated transcytosis across endothelia via caveolae (9,10). As albumin is an abundant serum protein that functions as a carrier for fatty acids, steroids, and thyroid hormones, the caveolae-mediated transcytosis of albumin may represent the mechanism by which these important molecules are distributed from the vascular space to surrounding tissues (11,12). The uptake of albumin may be related to its ability to bind fatty acids as fatty acid-carrying albumin is endocytosed and transcytosed by endothelial cells at a rate 1.5-2.4-fold higher than defatted albumin (13).
In support of these morphological studies, albumin was shown to be a major protein component of caveolae purified from lung tissue (an endothelial-rich tissue source), providing a biochemical correlate to these earlier observations (14). The finding that a putative luminal content molecule is retained within these membrane domains could also imply that either albumin is receptor-bound or that a fraction of these vesicular structures remains sealed throughout the isolation procedure. However, over the years, many investigators have challenged the idea that caveolae function in the transport of macromolecules across the endothelial barrier (15)(16)(17). In this regard, the role of caveolae in endothelial uptake and transport still remains uncertain (11, 18 -23).
Caveolin-1 appears to be an essential component of caveolae (40). For example, caveolin-1 protein expression directly parallels caveolae formation during adipocyte differentiation (36,41,42). Conversely, caveolin-1 mRNA and protein expression are lost or reduced during cell transformation and caveolae are absent from these cell lines (43). Recombinant overexpression of caveolin-1 in caveolin-deficient cell lines results in (i) the correct biochemical targeting of caveolin-1 to caveolae-enriched membrane fractions (44) and (ii) the ability to drive the de novo formation of recombinant caveolae vesicles in mammalian cells (45,46) and Sf21 insect cells (40). These results provide direct evidence that caveolin family members participate in caveolae formation.
The mechanism of how caveolin-1 expression induces caveolae formation remains unknown. This may be related to the self-assembly properties of caveolin-1. Caveolin-1 undergoes two stages of oligomerization. First, in the endoplasmic reticulum, caveolin-1 monomers assemble into discrete multivalent homo-oligomers containing ϳ14 -16 monomers per oligomer (35,47). Subsequently, these individual caveolin-1 homo-oligomers (4 -6-nm spherical particles) can interact with each other to form clusters of particles that are ϳ25-50 nm in diameter (35). Also caveolin-1 homo-oligomers interact specifically with sphingolipids (48) and cholesterol (30,31) and require a high cholesterol content to insert into model lipid membranes (30,31). Thus, we envisage that through the interaction of caveolin-1 with itself and the caveolin-mediated selection of endogenous lipid components, a caveolae-sized vesicle is generated (40).
Recently, we and others have generated Cav-1-deficient mice by using a targeted gene disruption approach (49,50). Surprisingly, these mice are viable and fertile. However, Cav-1-null mice clearly lack caveolae organelles in the cell types where the caveolin-1 protein is normally expressed, such as fibroblasts and endothelial cells. The loss of caveolae organelles in Cav-1deficient mice makes them a useful experimental system to study the putative functions of caveolae in an in vivo setting. Given that the role of caveolae in endothelial cell-mediated transcytosis has been a hotly debated topic for many years, we decided to assess whether caveolae organelles were required in vivo to mediate the uptake and transport of albumin by endothelial cells. Here, we provide conclusive evidence that Cav-1deficient endothelial cells are defective in the uptake and transport of albumin.

EXPERIMENTAL PROCEDURES
Generation and Maintenance of Cav-1-deficient Mice-The strategy used to target the caveolin-1 locus and generate Cav-1-null mice was as previously described (49). All animals used in these studies (mice homozygous null for the caveolin-1 gene and their wild-type littermates) were in a C57BL/6 ϫ sv129 genetic background and were genotyped by PCR, as previously described (49). Housing and maintenance was provided by the Albert Einstein College of Medicine (AECOM) barrier facility; mice were kept on a 12-h light/dark cycle and had ad libitum access to chow (Picolab 20, PMI Nutrition International) and water. All animal protocols used in this study were pre-approved by the AECOM Institute for Animal Studies.
Gold-conjugated BSA Uptake Studies-0.5 ml of BSA 1 coupled to 5-nm gold particles (A 520 ϳ5.0, Sigma) was diluted into 5 ml of aerated serum-free DMEM (Invitrogen) supplemented with 14 mM glucose and prewarmed to 37°C. The circulatory system was washed with 3 ml of PBS after a laparatomy was performed. The gold solution was injected into the right ventricle of anesthetized aged-matched wild-type and Cav-1-null mice using a 5-cc syringe and a 30-gauge needle. The gold solution was injected at a rate of 0.5 ml/min. After 15 min from the beginning of perfusion, the lung tissue was removed from the mouse, rinsed in PBS, and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer. The tissue was then processed for transmission electron microscopy, as previously described (43,49,51).
Light Microscopy-Aortic segments from wild-type and agedmatched Cav-1-null mice were fixed in 4% paraformaldehyde. Samples were then placed in 30% sucrose overnight and frozen in OCT (Tissue-Tek). Five micron sections were cut using a Leica CM 3050 cryostat and placed on super frost plus slides (Fisher). The slides were then hematoxylin-and eosin-stained according to standard laboratory protocols. 125 I-BSA Uptake-Uptake studies with radioiodinated albumin were performed essentially as described previously, with minor modifications (52). Aortic segments from three wild-type and three Cav-1-null littermates (all 6-week-old females) were generated as previously described (49) and cut into 2-4-mm segments and placed into 48-well plates. Five segments were used per time point and were incubated with 25 l of 125 I-BSA (0.5 mCi/ml, PerkinElmer Life Sciences) in 0.5 ml of aerated serum-free DMEM (Invitrogen) for 15, 30, and 45 min. One set of aortic segments from both genotypes was also incubated with an excess of cold BSA at 5 mg/ml for 45 min. All the above time points and conditions were then incubated at 37°C. An additional set of segments from both genotypes was incubated for 15 min at 4°C to block any internalization via endocytosis. All samples were then washed four times in 0.2 M acetic acid and 0.5 M NaCl buffer, pH 2.5, to remove any 125 I-BSA that remained attached to the cell surface. The amount of radioactivity was determined for each aortic segment individually using an LKB 1282 CompuGamma scintillation counter. The samples were then dried and weighed. The final value for each sample was determined by dividing the cpm value by the dry weight. The individual segments adjusted cpm values for each time point and condition and were then averaged for that condition to yield the final cpm value.

Caveolin-1-deficient Mice, A New Model System to Study in
Vivo Caveolae Functioning-Using targeted gene disruption, we recently generated caveolin-1-deficient mice (49). As these animals also lack caveolae organelles, these mice represent a new model system in which to study the in vivo function of caveolae. Our preliminary experiments with fibroblasts isolated from Cav-1-null mouse embryos indicated that these cells were defective in the in vitro uptake of fluorescein isothiocyanate-albumin, a known caveolar ligand, as seen by fluorescence microscopy (49). This initial finding prompted us to study the endocytic behavior of Cav-1-null endothelial cells in vivo. Transport of ligands from the blood-space to the tissue-space via endothelial cell caveolae has remained a controversial topic in cell biology for over 40 years. We reasoned that the use of caveolin-1-deficient mice should allow us to finally resolve this controversy.

Gold-conjugated Albumin Is Not Endocytosed by Cav-1-deficient Lung Endothelial Cells in Vivo, as Seen by Transmission
Electron Microscopy-Albumin has been used previously to examine caveolar endocytosis in the lung in vivo (9,53). Therefore, we examined the in situ ability of lung endothelial cells from Cav-1-null mice to endocytose gold-labeled BSA. Animals were perfused with gold-labeled BSA, and lung tissue samples were collected and processed for transmission electron microscopy (See "Experimental Procedures").

FIG. 1. Cav-1-deficient lung endothelial cells fail to endocytose albumin in vivo, as seen by transmission electron microscopy.
BSA coupled to 5-nm gold particles was perfused into the lungs of Cav-1-null mice and their wild-type littermates and followed by transmission electron microscopy. Panels A and B are micrographs taken of wild-type lung endothelial cells. Both panels show caveolae still open to the lumen (L) of the blood vessel and contain gold particles (small arrows). In addition, internalized caveolae (plasmalemmal vesicles) that are no longer open to the lumen and contain gold particles are also observed (large arrows). Some gold particles can also be seen in the lumen of the blood vessel (arrowheads). In contrast, the lung samples taken from Cav-1-deficient mice (panels C and D) show a loss of caveolae, and the gold particles remain within the lumen (L) of the blood vessel (arrowheads). Scale bar is 100 nm and applies to panels A-D.
Cav-1-deficient Mice: An Organelle Knock-out 48620 Fig. 1 shows that lung endothelial cells from Cav-1-deficient mice do not have any caveolae and cannot endocytose albumin. After 15 min, the gold-labeled BSA still remains within the lumen of the blood vessels (panels C and D). In contrast, lung endothelial cells from wild-type littermates (panels A and B), taken at the same time point, show a large portion of the gold-labeled BSA within caveolae. Fig. 2 presents higher magnification views of albumin endocytosis via caveolae in the lung. These micrographs clearly show multiple gold particles within caveolae both open to the lumen (panel A), as well as already detached from the plasma membrane and internalized (panel C). The internalized caveolae (plasmalemmal vesicles) support the idea that caveolae are indeed responsible for albumin endocytosis. However, micrographs of the Cav-1-deficient mouse endothelium show no caveolae and, as a result, the gold-labeled BSA remains in the lumen (panel B) of the blood vessel even at a time when endocytosis in the wild-type endothelium is well underway. These results directly support the idea that caveolae participate in the endocytosis of ligands, such as albumin.
Uptake of Radioiodinated Albumin Is Quantitatively Disrupted in Aortic Ring Segments from Cav-1-deficient Mice-In an attempt to more accurately quantify the endocytosis defect observed in the Cav-1-deficient mice, we used aortic ring segments generated from both wild-type and Cav-1-null mice, a system previously used for physiological studies (49,54). Prior to these studies, we needed to ensure there were no major anatomical defects in the Cav-1-null aorta that might affect our uptake studies. Cross-sections of wild-type and Cav-1-null aorta were generated and hematoxylin-and eosin-stained. These sections were examined and no major morphological differences could be detected. A low magnification view of the Cav-1-null aorta (Fig. 3C) shows that the overall size and shape is similar to the wild-type aorta (Fig. 3A). A higher magnification view of the Cav-1-null aorta (Fig. 3D) shows that all three layers of the aorta (tunica intima, media, and adventitia) appear intact and unchanged in their composition, compared with wild-type aorta (Fig. 3B).
Aortic ring segments were then generated from Cav-1-null mice and wild-type littermates to quantify their ability to endocytose albumin. For this study, aortic rings were incubated with 125 I-BSA at 37°C for 15, 30, and 45 min, and then the amount of radioiodinated BSA incorporated into the tissue was measured. Control samples were incubated: (i) at 4°C to prevent endocytic uptake, or (ii) at 37°C with excess unlabeled cold BSA to demonstrate that the observed internalization was specific. All samples were washed extensively in 0.2 M acetic acid and 0.5 M NaCl buffer, pH 2.5, to remove any cell surface bound 125 I-BSA. As a consequence, only internalized albumin that remained in a protected endocytic compartment was detected. Fig. 4 shows that little or no uptake of radioiodinated BSA was observed in the aortic segments from Cav-1-deficient mice, while aortic segments from wild-type mice showed robust uptake that was time-and temperature-dependent and competed by unlabeled cold albumin. Taken together, our current results clearly demonstrate that endothelial cells from Cav-1 null mice are incapable of endocytosing albumin to any appreciable level, compared with wild-type animals. However, the fact that Cav-1-null mice remain viable suggests that other pathways must at least partially compensate for the absence of caveolae in vivo.
Conclusions-As mentioned above, the uptake and transport of macromolecules by endothelial cell caveolae (plasmalemmal Gold particles still within the lumen of the blood vessel are also present (arrowheads). Panel C shows three caveolae that have become internalized (large arrows) and are no longer in contact with the plasma membrane. Panel B is a high magnification micrograph of a lung endothelial cell taken from a Cav-1-deficient mouse. No caveolae are present in these Cav-1-deficient endothelial cells, and the gold-labeled BSA (large arrowheads) remains within the lumen (L) of the blood vessel. Scale bar is 100 nm and applies to panels A-C.
FIG. 3. The morphology of the Cav-1-deficient mouse aorta appears grossly normal. Panels A-B are micrographs of the aorta from a wild-type mouse. Panels C-D are micrographs of the aorta from a Cav-1-null mouse. The overall size, shape, and structure of the Cav-1-null aorta (C) appears to be similar to the wild-type aorta (A). At higher magnification, the various layers that make up the aorta (tunica intima (ti), tunica media (tm), and tunica adventitia (ta)) are all present in the Cav-1-null mouse (D) and are similar in appearance, structure, and composition to the wild-type aorta (B).
Cav-1-deficient Mice: An Organelle Knock-out 48621 vesicles) has remained controversial for more than four decades. To resolve this issue, we have now employed a new experimental model system (Cav-1-deficient mice) that allows the study of the in vivo function of caveolae. The power of this genetic model system is that the ablation of a single gene (CAV-1) results in the loss of morphological caveolae, i.e. an organelle knock-out. Using this model system, we show here that Cav-1-null lung endothelial cells lack caveolae and fail to endocytose gold-labeled BSA in vivo. Similarly, aortic ring segments from Cav-1-null mice are quantitatively defective in the internalization of radioiodinated BSA. This is also the first clear demonstration that caveolae represent an internalized membrane compartment that is sealed off from the external environment, because the internalized pool of radioiodinated BSA in wild-type cells was resistant to stringent washes (at acid pH and with high concentrations of NaCl) that remove surface-bound material. Undoubtedly, Cav-1-deficient mice will help resolve many of the other controversies that still exist in the caveolae field.
FIG. 4. Uptake of radioiodinated albumin is quantitatively disrupted in aortic ring segments from Cav-1-deficient mice. Aortic ring segments from wild-type and Cav-1-null mice were incubated with 125 I-BSA at 37°C for various lengths of time (15,30, and 45 min) to determine whether the ability of Cav-1-null mice to endocytose serum albumin was affected by a loss of caveolin-1 expression. Black bars represent wild-type aortic segments; gray bars represent Cav-1-null mouse aortic segments. Note that no significant increase in internalized 125 I-BSA could be detected in the Cav-1-null samples over time, suggesting that there is a defect in their ability to endocytose serum albumin (asterisks). As a negative control, aortic segments were incubated with 125 I-BSA for 15 min at 4°C to block endocytosis and prevent any albumin uptake. As an additional negative control, aortic segments were incubated with 125 I-BSA plus an excess of unlabeled cold albumin for 45 min at 37°C. Virtually identical results were obtained with both male and female mice.