Abstract
It is generally accepted that angiogenesis is delayed in aging. To define the effects of age on the neovascular response, polyvinyl alcohol sponges were implanted SC in young (6–8 months old, n = 11) and aged (23-25 months old, n = 13) mice and sampled at 14 and 19 days. Angiogenic invasion was significantly delayed in aged mice at 14d relative to young at 14d (% area of invasion 9.0 ± 3.7 vs 19.0 ± 5.6; p = 0.02). Although microvessel morphology and basement membrane composition were similar between the age groups, a significant decrease in capillary density was noted in aged tissues at 14d (7.5 ± 4.1) and 19d (12.1 ± 2.8) relative to young at 14d (18.7 ± 2.3) (p<0.01 A14d vs Y14d). In comparison to young at 14d, the inflammatory response was decreased by 43 ± 2.9% and 36 ± 7.8% in aged mice at 14d and 19d, respectively. Tissues of aged mice showed less newly deposited collagen. There was a lack of expression of transforming growth factor-β1 (TGF-β1) and vascular endothelial growth factor (VEGF) in aged mice at 14d (0.63 ± 0.3) and 19d (1.14 ± 0.5) vs young at 14d (1.92 ± 0.5) (p≤0.01 A14d vs Y14d for VEGF). However, similar production of VEGF receptor2 was observed. In contrast to young mice, there was significantly increased expression of thrombospondin-2 (TSP-2) in aged mice from 14d (14.6 × 103 ± 7.3 × 103) to 19d (34.9 × 103 ± 17 × 103). We conclude that angiogenesis in aging is not merely delayed, but is altered due to multiple impairments.
A
Although much is known about angiogenesis in general, the changes that occur during angiogenesis in aging are not well defined. Delayed neovascularization in aged tissues has been noted and proposed to contribute to slowed wound repair (Puolakkainen et al. 1995), neurodegeneration (Kalaria 1996; Lee et al. 2000), and renal disease (Kang et al. 2001). In contrast, decreased angiogenesis may be of benefit in slowing tumor growth in aged animals (Pili et al. 1994). It is not clear whether delayed neovascularization in aged tissues is analogous to a slower version of angiogenesis in young tissues or whether it reflects the confluence of multiple age-related impairments. Previous studies of angiogenesis in aging have been confounded by the use of pathological models such as ischemia in the rabbit hindlimb model (Rivard et al. 1999), glomerulonecrosis in the rat model (Kang et al. 2001), and excisional wounds in mice (Swift et al. 1999). Alterations in angiogenesis that were noted in these models included endothelial cell dysfunction, reduced expression of VEGF and FGF (Rivard et al. 1999; Swift et al. 1999), and increased expression of TSP-1 (Kang et al. 2001) in the injured or diseased aged animal relative to young controls. A defective inflammatory response has also been reported to modulate angiogenesis during cutaneous wound healing in aged mice and humans. The change in the immune response was associated with delayed infiltration of monocytes/macrophages and lymphocytes, as well as deficient expression of endothelial cell adhesion molecules (Ashcroft et al. 1997,1998; Swift et al. 2001).
Concurrent studies in vitro have confirmed that the age-related delay in angiogenesis is associated with several deficits in cell functions. The latter include slowed migration of aged microvascular endothelial cells (Reed et al. 2000) and fibroblasts (Reed et al. 2001; Mogford et al. 2002). Aged endothelial cells also produce less nitric oxide (NO) and demonstrate increased sensitivity to apoptotic stimuli (Vasa et al. 2000; Hoffmann et al. 2001; Dimmeler and Chavakis 2002). Others have reported that there is a defective response in aged cells to the mitogenic effects of VEGF (Rivard et al. 1999) and the anti-proliferative effects of TGF-β1 (McCaffrey and Falcone 1993).
In a previous study of the angiogenic response in healthy aged mice, our laboratory demonstrated a delay in fibrovascular invasion in aged mice at 14 d relative to young mice at 14 d that was coincident with decreased levels of type 1 collagen mRNA and deficient expression of TGF-β1 protein (Reed et al. 1998). However, it is probable that age-related alterations in angiogenesis result from additional factors, including cellular dysfunction (e.g., migration, proliferation, apoptosis), upregulation or downregulation of growth factors and matrix proteins, and a decrease in the inflammatory response. The latter processes have not been examined during the neovascular response in healthy aged animals. In this study we wished to characterize in vivo the mechanisms that influence delayed angiogenesis in healthy aged mice relative to young mice. We used implanted SC PVA sponges, an accepted animal model of angiogenesis (Davidson et al. 1985; Fajardo et al. 1988; Reed et al. 1998). Sponges were placed distant from the point of initial subdermal entry to minimize the injury stimulus. Although others have termed the sponge a “wound,” this model differs significantly from wound repair in that there is minimal involvement of the clotting cascade and there are no epithelial cells, thereby permitting one to focus on the endothelial network and fibroblasts. Features studied included quantification of neovascular invasion and cell proliferation, blood vessel morphology, density, and basement membrane composition, collagen deposition and distribution, the inflammatory response, expression of the angiogenic factor VEGF and its receptor, VEGFR2, and levels of thrombospondin-2 (TSP-2), an inhibitor of vascularization. We report the novel finding that angiogenesis in healthy aged mice is not only slower than in young mice but is altered as a result of impairments at multiple steps in the neovascular response.
Materials and Methods
Animal Model
An established model of angiogenesis in mice using SC PVA sponges was applied (Davidson et al. 1985; Fajardo et al. 1988; Reed et al. 1998). In this study, male mice (6–8 months of age for young mice and 23-25 months for aged mice) of the {C57BL/6 × DBA/2} F1 “B6D2F1” and {C57B1/6 × C3H} F1 “B6C3F1” strains were used. All animals were received from the NIA colony maintained by Harlan (Indianapolis, IN). Mice were maintained in the modified pathogen-free environment of the Department of Comparative Medicine at Harborview Medical Center. Two circular Clinicel PVA sponges (10 × 2 mm) (M-Pact; Eudora, KS) were implanted SC in the dorsum of each mouse. Sponges were tunneled at least 2 cm away from the incision site to minimize the response to injury. Animals were sacrificed and sponges were recovered at 9, 14, and 19d. Implants were fixed in neutral-buffered formalin, embedded in paraffin, and sectioned.
Histology and Invasion Index
Five-μm sections of paraffin-embedded sponges were dewaxed in xylene, hydrated in a graded series (100%-70%) of ethanol solutions, and stained with Masson's Trichrome. For quantification of invasion, magnified images were viewed on a Leica microscope and were captured by a spot camera. Images were imported into NIH Image software program (US National Institutes of Health; http://rsb.info.nih.gov/nih-image/). Fibrovascular invasion into the sponges was clearly visible on the imported images and was defined by the presence of vessels that contained red blood cells (Reed et al. 1998). All vessels were then labeled with an antibody to VEGFR2 (a gift from Dr. Rolf Brekken, UT-Southwestern Medical Center) to confirm the presence of a lining of endothelial cells. The width of each sponge was approximately three fields at X25; each field was analyzed separately. The area of invasion was digitally outlined and the designated area of the sponge was determined by the NIH Image program. The percent area of invasion was calculated by dividing the amount of invasion by the area available for invasion (the total area minus the area occupied by PVA sponge material).
BrdU Injection
A subset of mice (n = 4–7 in each group) were injected IP with BrdU (2 μg/g body weight) 8 hr before sacrifice. Sponges were then removed, fixed in neutral buffered formalin, and embedded in paraffin. Tissue sections were examined for the incorporation of BrdU as previously described (Reed et al. 1996). Cells with dark nuclei, representing positive immunostaining, were quantified by two different observers.
Immunohistochemistry
For immunohistochemistry (IHC), sections were dewaxed in xylene and hydrated in a graded series (100%-70%) of ethanol solutions. Slides were blocked overnight in PBS with 5% normal goat serum and incubated with specified primary antibodies (at 5–25μg/ml) for 1 hr at room temperature. The following primary antibodies were used: a rabbit polyclonal antibody against laminin (Sigma Chemical; St Louis, MO), a mouse monoclonal antibody against smooth muscle α-actin (Dako; Carpinteria, CA), an affinity-purified rabbit polyclonal antibody against VEGF-165 (Santa Cruz Biotechnology; Santa Cruz, CA), a rabbit polyclonal antibody against VEGR receptor2 (a gift from Dr. Rolf Brekken), and a rabbit polyclonal antibody against mouse TSP-2 (a gift from Dr. Paul Bornstein, U. of Washington). For assessment of basement membrane components, slides were dewaxed, hydrated, and then incubated with Jones silver stain. For immunostaining, slides were exposed to the primary antibody, incubated with the appropriate secondary antibodies conjugated to horseradish peroxidase (1-5 μg/ml) (Jackson Immuno Research; West Grove, PA) for 1 hr at RT, followed by uniformly timed detection with 3,3-diaminobenzidine (Vector; Burlingame, CA). In some cases, an amplification with strepavidin/biotin was performed (Dako). Immuno-stained sections were counterstained with toluidine blue, dehydrated with a graded series of ethanol, cleaned with xylene, mounted, and visualized by light microscopy. In all experiments, pre-immune serum and secondary antibody alone served as negative controls.
Assessment of Vessel Morphology and Capillary Density
Slides were prepared as indicated previously and stained with Masson's Trichrome. Vessels were identified by the presence of hematopoetic cells in their lumens and positive staining for VEGFR2 (to confirm the presence of a lining of endothelial cells). Vessels were counted by three different investigators at X200 magnification in at least six randomly selected fields per section.
Picrosirius Red Staining for Collagen
Slides were dewaxed, hydrated, and stained with Picrosirius Red (PSR). PSR is an anionic dye that differentiates collagen fiber thickness and density by the color emitted under polarized light. Whereas thin, loosely packed fibers are green-yellow, thicker, tightly packed fibers emit longer wavelength colors, such as orange and red (Kesler et al. 2000).
Quantification of the Inflammatory Response
Measurement of the presence of macrophages/monocytes was performed by staining tissue with an antibody against F4/80 antigen (10 μg/ml), a specific marker for these cell types (Serotec; Raleigh, NC). The number of macrophages/ monocytes was counted and analyzed in five to eight random fields by two different observers.
Computer-assisted Morphometric Analysis
TSP-2-immunostained sections were captured with a spot digital camera (Diagnostic Instruments; Sterling Heights, MI). Morphometric analysis of digital images was done by applying Metamorph Software (Universal Image; Westchester, PA). For quantification, a threshold value was set representing the maximal background intensity observed in control tissues (no primary antibody). Relative values above background readings were measured in young and aged tissues. Only sections immunostained in the same experiment were evaluated to avoid interexperiment variability (Kyriakides et al. 2001).
Statistical Analysis
All comparisons between groups were analyzed for significance by a two-tailed unpaired t-test. All data are presented as the mean ± SD.
Results
The Angiogenic Response is Delayed in Aged Mice
To quantify the delay in angiogenesis in the aged mice, we measured the amount of fibrovascular invasion into a PVA sponge by analysis of digital images. There was minimal invasion at 9d in either age group. There was maximal difference in the young and aged mice at 14d. Whereas sponges in aged mice at 19d were still undergoing the angiogenic response, fibrovascular invasion into the sponges was completed by 19d in the young mice. Therefore, the data to be presented will use young mice at 14d (n = 7) as a reference point and will focus on the response of aged mice at 14d (n = 6) and 19d (n = 7).
The angiogenic response in aged mice was delayed at 14d (Figure 1B) but reached the value of young mice at 14d (Figure 1A) by 19d (Figure 1C). Figure 4A shows a bar graph of the percent area of invasion into the sponge obtained from young mice at 14d and aged mice at 14d and 19d. The percent area of angiogenic invasion was significantly delayed in aged mice at 14d (9.0 ± 3.7) relative to the young at 14d (19 ± 5.6). By 19d, sponges from the aged mice showed a similar amount of angiogenesis (20 ± 7.6) as sponges from young mice at 14d.
The angiogenic response is delayed in aged mice. Sections were stained with Masson's s Trichrome. Fibrovascular invasion into the sponges was identified by the presence of vessels that contained red blood cells (arrows). The angiogenic response was delayed in aged mice at 14d (
The Proliferative Response Is Not Significantly Decreased in Aged Animals
Angiogenesis requires the proliferation and migration of endothelial cells and their supporting cells. To assess the relative contribution of proliferation to neovascular invasion into the sponge, we measured the incorporation of BrdU by all cells in the sponges from animals injected with BrdU 8 hr before sacrifice. The absolute number of cells stained for BrdU in each section, as well as the number of cells divided by the percent area of fibrovascular invasion, was determined. The number of BrdU-positive cells in the tissues was low (Figure 2). Comparison of the number of BrdU-positive cells in sponges in young 14d and aged mice at 14d and 19d showed that fewer than 15% of the cells were proliferating at either time point. Note that the absolute number of BrdU-positive cells in the sponges, when corrected for the percent area of fibrovascular invasion, was similar in both groups of mice at 14d.
The Appearance of Large Vessels Is Delayed in the Sponges of Aged Mice
The size of newly formed vessels in the area of fibrovascular invasion was measured via a micrometer. As shown in Figures 3A–3C, whereas capillaries (<30 μm wide; small arrows) were present at all time points in young and aged sponges (Figures 3A–3C), only the sponges from aged at 19d (Figure 3C) showed the larger blood vessels (>40 μm wide; large arrows) that were present in the tissues of young mice at 14d (Figure 3A).
The proliferative response is not decreased in aged animals. The contribution of proliferation to fibrovascular invasion was assessed by immunostaining for BrdU. As shown in the scatter plot, when corrected for the percent area of fibrovascular invasion, the numbers of proliferating cells in young mice at 14d and in aged mice at 14d and 19d were similar.
Assessment of vessel morphology. Sponge sections were stained with Masson's Trichrome. Whereas capillaries (small arrows) were present at all time points in young and aged sponges (
Vessel Density Is Significantly Decreased in Aged Mice
Although sponges from aged mice reached a comparable percent area of invasion at 19d as those of young mice at 14d, tissues from aged mice never attained the vessel density of sponges from young mice. Both 14d (7.5 ± 4.1) and 19d (12.1 ± 2.8) sponges from aged mice showed significantly decreased vessel density relative to that of young tissue at 14d (18.7 ± 2.3) (Figure 4B). The endothelial cell marker VEGFR2 was applied to confirm that the vessels were lined with endothelial cells (Figure 8A inset).
Basement Membrane Composition Is Similar in Young and Aged Mice
Jones silver stain and an antibody against laminin showed no detectable difference in the distribution of basement membrane proteins in the vessels of young and aged mice at any time point (data not shown). However, immunofluorescence for laminin deposition supported our finding that there was a significant decrease in vessel density in the sponges from aged mice relative to those from young mice. Immunostaining for actin, representing the presence of mural cells (smooth muscle cells and pericytes) (Hellstrom et al. 2001), was similar for a given number of vessels in the sponges of young and aged mice at the 14 and 19d time points (data not shown).
Sponges from aged mice demonstrate a delay in percent area of invasion and a persistent decrease in vessel density. (
Newly Deposited Collagen Is Decreased in Sponges of Aged Mice
The amount of collagen mRNA in the PVA sponge is decreased during angiogenesis in aged mice (Reed et al. 1998). To determine the quality of the collagen protein that is deposited, sponges were examined after PSR staining. As shown in Figure 5, the tightly packed mature collagen, as visualized by red staining on polarized microscopy, was decreased in aged mice at 14d (Figure 5B) compared to young mice at 14d (Figure 5A). Moreover, although there were significant amounts of mature collagen in aged mice at 19d (Figure 5C), a greater amount of newly deposited collagen (green-yellow) was observed in young mice at 14d (Figure 5A) than in aged mice at any time point.
Sponges from Aged Mice Have Decreased Infiltration of Macrophages/Monocytes
It is well known that the immune response is decreased in aged tissues during infection and wound repair (Ashcroft et al. 1997,1998; Swift et al. 2001). However, inflammatory infiltration into the PVA sponge has not been previously reported in aged mice. To quantify the number of macrophages/monocytes in the areas of fibrovascular invasion, young and aged mouse tissues were immunostained with F4/80, a specific marker for monocytes and macrophages. A decrease in the infiltration of macrophages/monocytes was observed in aged tissues at 14d and 19d (Figures 6B and 6C) compared to young tissues at 14d (Figure 6A). The number of macrophages and monocytes was decreased by 43 ± 2.9% in 14d aged mice and decreased by 36 ± 7.8% in 19d aged mice compared to young mice at 14d (Figure 7).
VEGF Protein, but Not Its Receptor, Is Decreased in Aged Tissues
Others have shown that the expression of VEGF is decreased during the response to ischemic injury in aged muscle (Rivard et al. 1999). In this study, IHC was performed with an antibody against VEGF and VEGF receptor2. The immunostaining per high-power field was assessed by two investigators and was corrected for the degree of fibrovascular invasion as defined previously. In Figures 8A–8C), the immunostaining for VEGF protein is decreased in aged mice at 14d (Figure 8B) and 19d (Figure 8C) relative to the young mice at 14d (Figure 8A). Figure 9A is a graphic depiction of the differences in the deposition of VEGF protein. A significant lack of VEGF expression was demonstrated in aged 14d (0.63 ± 0.30) mice and aged 19d (1.14 ± 0.50) mice relative to young mice at 14d (1.92 ± 0.5). In contrast, the expression of VEGF receptor2 was similar in both young and aged tissue at all time points (Figure 8A inset).
Aged mice have less total and newly deposited collagen. Sections were stained with PSR and analyzed by polarized microscopy. Although there was mature collagen in aged mice at 19d (
Inflammatory influx into the fibrovascular response is decreased in aged mice. Sections from young and aged mice were stained with an antibody to F4/80. Immunostaining, specifically representing macrophages/monocytes (arrows), is depicted in sponge implants from young mice at 14d (
Aged Tissues Express High Levels of TSP-2
Although it is accepted that aged tissues have less access to both circulating and locally secreted angiogenic growth factors, little is known about their expression of inhibitors of angiogenesis. In this context, we wished to examine the expression of TSP-2, a potent inhibitor of angiogenesis (Bornstein 2001; Hawighorst et al. 2001; Kyriakides et al. 2001). Sections from young and aged mice were stained with a polyclonal antibody to TSP-2 and counterstained with 1% toluidine blue (Figures 8D–8F). There was greater deposition of TSP-2 around the cells and in the extracellular matrix in aged mice at 14d (Figure 8E) and 19d (Figure 8F) relative to young 14d mice (Figure 8D). Images from stained tissues were digitally captured and analyzed using Metamorphic Software. As shown in the graph (Figure 9B), there was significantly increased expression of TSP-2 in aged tissues from 14d to 19d (14.6 × 103 ± 7.3 × 103 vs 34.9 × 103 ± 17 × 103). This was in contrast to young mice at 14d (25 × 103 ± 8.9 × 103), which did not show a significant increase in expression of TSP-2 from 14d to 19d (data not shown).
The number of macrophages/monocytes in the areas of angiogenesis is decreased in aged mice. Graph shows that the inflammatory response, relative to young mice at 14d, was decreased significantly in aged 14d mice by 43 ± 2.9% and aged 19d mice by 36 ± 7.8% (∗p = 0.04 A14d vs Y14d).
Aged tissues express less VEGF protein and more TSP-2. Sections were stained with a polyclonal antibody against VEGF and with a polyclonal antibody to TSP-2. The amount of immunostaining for VEGF protein is decreased in aged mice at 14d (
The expression of VEGF is decreased, whereas that of TSP-2 is increased, in aging. (
Discussion
Angiogenesis is requisite for many physiological and pathological processes. With few exceptions (Ashcroft et al. 1997), most studies have documented an age-related delay in the angiogenic response (Yamaura and Matsuzawa 1980; Kreisle et al. 1990; Cohn 1994; Pili et al. 1994; Marinho et al. 1997; Reed et al. 1998; Rivard et al. 1999; Swift et al. 1999; Kitlinska et al. 2002). Whereas slowed neovascularization in aged tissues is detrimental to the re-perfusion of ischemic tissue, it can be of benefit in inhibiting the growth of tumors (Kreisle et al. 1990; Cohn 1994; Marinho et al. 1997). The delay in angiogenesis in aging may reflect changes in multiple components of the neovascular response: endothelial proliferation (Nissen et al. 1998, Rivard et al. 1999), migration (Reed et al. 2000, 2001; Mogford et al. 2002), cell death (Vasa et al. 2000), growth factors such as VEGF (Arthur et al. 1998; Nissen et al. 1998; Rivard et al. 1999; Ferrara and Gerber 2001), b-FGF (Augustin-Voss et al. 1993; Sartippour et al. 2002), IGF-1 (Simmons et al. 2002); and extracellular matrix proteins such as type 1 collagen (Reed et al. 1998), SPARC (Bradshaw et al. 2001), fibronectin (Ashcroft et al. 1997), and TSP-1 and 2 (Bornstein 2001; Hawighorst et al. 2001; Kyriakides et al. 2001). Recent studies have also noted the contribution of an altered inflammatory response to the delay in wound repair in aging (Ashcroft et al. 1997,1998; Swift et al. 2001). These observations were made in vitro and in injury/disease models. In this study we wished to define angiogenesis in healthy aged mice by examination of the neovascular response into an inert sponge in vivo. The sponge model differs significantly from wound repair. There is minimal involvement of the clotting cascade, there are no epithelial cells/keratinocytes, there is no wound contraction, and there is no scar. Therefore, invasion into the sponge focuses on a fibrovascular response composed of endothelial cells and fibroblasts. Moreover, studies in transgenic mice have demonstrated contrasting responses during cutaneous wound repair and angiogenic invasion into the sponge, further highlighting the unique features of each of these models (Bradshaw et al. 2001). In this study, the following features were analyzed: neovascular invasion and cell proliferation, blood vessel morphology, density, and basement membrane composition, collagen deposition and distribution, the inflammatory response, the expression of the angiogenic factor VEGF and its receptor, VEGFR2, and the levels of thrombospondin-2 (TSP-2).
Angiogenesis was delayed in sponges obtained from aged mice at 14d compared to sponges from young mice at 14d. However, by 19d the percent area of fibrovascular invasion in the aged tissue was similar to that of their young counterparts at 14d. Of note, despite the similar area of invasion into the sponge by 19d in aged animals, the density of vessels in the sponge remained significantly decreased in the aged tissues relative to the young tissues. Previous reports have noted that capillary density in aged animals is less than that of young animals in the myocardium and brain (Kalaria 1996; Rivard et al. 1999; Lee et al. 2000). An increase in the density of vessels in aged myocardium in rats was achieved through the administration of growth hormone (Khan et al. 2001). The present study is the first to report that there is a significant decrease in the number of vessels per given area in healthy aged tissues at both the mid and later time points of the angiogenic response.
Neovascular invasion requires the proliferation of endothelial cells and their supporting cells (Rivard et al. 2000). We examined the relative contribution of cell proliferation to invasion into the sponge by measuring the incorporation of BrdU in the tissues. Despite the consensus that aged cells have reduced replicative capacity, when corrected for the area of fibrovascular invasion there were no significant differences in BrdU incorporation in the sponges from the young and aged cells. These data do not exclude a decrease in proliferation in the sponges of the aged mice at all timepoints. Nevertheless, this result supports our premise from prior observations that impaired migration contributes significantly to delayed angiogenesis in aging (Reed et al. 2000,2001).
To further define the features of fibrovascular invasion in aged mice, we evaluated sponge tissues from young and aged mice for vessel morphology, matrix composition, and the presence of mural cells. We utilized Jones silver stain and an antibody against laminin to visualize basement membrane composition, and an antibody against actin to define the presence of mural cells. The morphology of vessels in the aged mice at 19d was similar to that of young at 14d with respect to vessel size. In that context, larger vessels that may represent the coalescence of smaller vessels were present at 14d in the young and at 19d in the aged tissues. Others have examined the shape of blood vessels induced in normal adult tissues by VEGF and have noted that vessel size can change significantly within a few days during the neovascular response (Pettersson et al. 2000; Dor et al. 2002). These data demonstrate that microvessels undergo alterations in shape and size in a rapid fashion in any given tissue. As noted previously, vessels in the sponges from aged mice were significantly less dense than those in young tissues (independent of their size and shape) at all time points. However, once present, vessels showed similar amounts of basement membrane proteins surrounding the vessel. In addition, there were no differences in mural cell distribution in the vessels between the age groups. Future experiments will determine whether the similar appearance of basement membrane proteins and mural cells in the vessels of the aged mice are predictive of intact functional characteristics such as vessel integrity.
We have previously reported that the delay in angiogenesis was coincident with decreased transcripts for type 1 collagen (Reed et al. 1998). In this study, we applied polarized light microscopy and PSR staining to distinguish between mature and newly deposited collagen protein. We observed that the content of mature collagen was decreased in aged mice at 14d compared to young mice at the same time point. This difference was obviated at aged 19d. Moreover, consistent with our previous data using in situ hybridization, a greater amount of newly deposited collagen was observed in young mice at 14d compared to aged mice at any time point. These data show that altered collagen synthesis and organization not only are features of aged tissues such as skin (Gilchrest 1999) but also occur during the neovascular response in aging.
During wound repair in aged animals, an impaired inflammatory response precedes the delay in angiogenesis (Ashcroft et al. 1997,1998; Swift et al. 2001). Although the PVA sponge differs significantly from wound repair and does not elicit as large an inflammatory influx, the number of macrophages/monocytes was decreased in aged mice at 14d and 19d relative to young mice at 14d. It is likely that the altered appearance or function of inflammatory cells in aged tissues is a key mediator in slowed angiogenesis via both direct (lack of angiogenic chemokines) and indirect (deficient induction of angiogenic growth factors by inflammatory mediators) mechanisms. Therefore, modification of the immune response is likely to provide a key upstream pathway for altering the angiogenic response in aged tissues.
Others have documented that a deficiency in VEGF secretion by connective tissue cells contributes directly to altered angiogenesis in aging in ischemic vascular disease (Rivard et al. 1999). In this study we found that the expression of VEGF was decreased in sponges from the aged mice at 14d and 19d compared to young tissue at the same time points. A moderate increase in VEGF expression was observed in aged tissue from 14d to 19d, but the level of VEGF expression in aged tissue at 19d remained lower than that of young tissue at 14d. A previously reported study by Rivard et al. (1999) showed that the expression of VEGF receptor2 (VEGFR2), as assessed by Western blotting, was preserved in both young and aged animals. However, the same group noted that the ultimate level of recovery in old animals after recombinant VEGF therapy was inferior to that observed in young animals. In this study we found no differences in the expression of VEGFR2 in the areas of fibrovascular invasion in young and aged animals at any time point. Taken together, these data show that the presence of equivalent amounts of VEGFR in tissues does not ensure that their function is similar between young and aged animals.
An additional novel finding in this study was the increased expression of TSP-2 in the tissues of the aged mice. TSP-2, a matricellular protein, has been reported to regulate angiogenesis by direct influences on endothelial cell behavior as well as modulation of cell- matrix interactions (Bornstein 2001). The TSP-2-null mice demonstrate increased and prolonged neovascularization, providing support for TSP-2 as a potent inhibitor of angiogenesis (Lawler 2000; Hawighorst et al. 2001; Kyriakides et al. 2001). By applying IHC staining and computer-assisted histomorphometric analysis, we observed a significant increase in the expression of TSP-2 in aged tissues from 14d and 19d. Although the deposition of TSP-2 was higher in the young mice than in the aged mice at 14d, our analyses showed that the expression of TSP-2 in young animals did not increase in the later stages of neovascularization as was observed for aged tissues from 14d to 19d. The high amounts of TSP-2 in aged tissues at 19d points to a possible role for TSP-2 as a late antagonist of the angiogenic response in aging.
In summary, this study demonstrates that delayed angiogenesis in aged tissues does not merely reflect a slower model of angiogenesis in young tissues. Whereas the neovascular response in the aged mice approximates that of young mice in isolated features such as deposition of basement membrane proteins, other aspects of angiogenesis are truly impaired. The changes in aging include decreased vessel density, less newly deposited collagen, altered inflammatory response, reduced expression of proangiogenic factors, and increased expression of TSP-2, an inhibitor of angiogenesis. The many components of the neovascular response that are altered with age indicate that those interventions that affect multiple pathways will have the greatest impact on enhancing or inhibiting angiogenesis in aged tissues.
Footnotes
Acknowledgements
Supported by the National Institutes of Health AG015837 and the Paul Beeson Physician Faculty Scholars in Aging Research Program.
We wish to thank Drs E. Helene Sage, Paul Vernon, Pauli Puolakkainen, Teruhiko Koike, Paul Bornstein, Azin Agah, and Themis Kyriakides for thoughtful discussions, and Nancy Ferara for technical assistance.