Study Design

The study was approved by the institutional Animal Research Committee. Growth factor–enriched matrigel was placed subcutaneously to study vasculogenesis in 18 wild-type C57Bl/6 male 8- to 10-week-old mice. Contrast-enhanced ultrasound (CEU) perfusion imaging of the matrigel plug and targeted imaging for α₅-integrins, VCAM-1, or neutrophil complement receptors were performed on day 3, 6, or 10 (n=6 for each time interval, 1 follow-up interval per animal). To study arteriogenesis, unilateral hind-limb ischemia was produced by interruption of iliac artery inflow. CEU perfusion imaging and targeted imaging of the proximal adductor muscles in the ischemic and control hind limb were performed 2, 4, or 7 days after ligation (n=5 for each time interval). Late recovery of perfusion was assessed at day 21 in an additional 5 mice. Molecular imaging and perfusion imaging also were assessed in 5 MCP-1^−/− mice (kindly provided by Dr Barrett Rollins, Dana-Farber Cancer Institute, Boston, Mass) 4 days after ligation. An additional 6 wild-type mice were used to validate the molecular imaging analysis method.

Matrigel Model

Matrigel (BD Biosciences, San Jose, Calif), composed of basement membrane extract from the Engelbreth-Holm-Swarm mouse sarcoma, was thawed at 4°C and enriched with fibroblast growth factor-2 (500 ng/mL; Sigma-Aldrich, St Louis, Mo) and heparin (64 U/mL). Enriched matrigel (1 mL) at 4°C was injected subcutaneously to form discrete ellipsoid plugs in the ventral abdominal wall. Imaging was performed in anesthetized mice at day 3, 6, or 10 (n=4 for each). In an additional 4 mice, the spatial extent of perfusion at either day 6 or 10 was characterized by the presence of intravascular fluorescent nanospheres. Approximately 1×10¹⁰ fluorescently labeled polystyrene spheres (Duke Scientific Corp, Fremont, Calif) with a mean diameter of 500 nm were injected intravenously. After 5 minutes, the plugs were excised, flash-frozen in liquid nitrogen, sectioned, and observed by microscopy (Axioskop2-FS, Carl Zeiss, Inc, Thornburg, NY) with fluorescent epi-illumination (530- to 560-nm excitation filter).

Hind-Limb Ischemia

Mice were anesthetized with an intraperitoneal injection (12.5 μL · g⁻¹) of a solution containing ketamine hydrochloride (10 mg · mL⁻¹), xylazine (1 mg · mL⁻¹), and atropine (0.02 mg · mL⁻¹). Body temperature was maintained at 37°C with a heating pad. The inguinal fold was prepped and draped in a sterile fashion. A short segment of the distal common iliac artery encompassing the origin of the epigastric vessel was isolated and removed. The skin incision was closed, and animals recovered in a heated environment. Imaging studies were performed at day 2, 4, 7, or 21 after arterial ligation.

Microbubbles

For perfusion imaging, lipid-shelled decafluorobutane microbubbles were prepared by sonication of a gas-saturated aqueous suspension of 2 mg · mL⁻¹ distearoylphosphatidylcholine and 1 mg · mL⁻¹ polyoxyethylene-40-stearate. For in vivo imaging of intravascular recruitment of activated neutrophils, microbubbles were targeted to complement receptors on activated leukocytes by addition of distearoylphosphatidylserine (0.3 mg · mL⁻¹) to the aqueous suspension before sonication.^16,17 Microbubbles targeted to α₅-integrins or VCAM-1 were prepared from biotinylated microbubbles containing distearoylphosphatidylethanolamine-PEG(2000)biotin in the shell. Biotinylated rat monoclonal antibody against murine α₅-integrin (5H10-27, BD Biosciences), VCAM-1 (MK2.7 purified from hybridoma), or isotype control antibody (R3-34, BD Biosciences) was conjugated to the surface of microbubbles as previously described.¹⁹ The mean diameters for targeted and control microbubbles were 2.7±0.3 and 2.9±0.4 μm, respectively.

Ultrasound

CEU imaging was performed with 2 different imaging modalities to optimize balance between sensitivity and resolution for the 2 target tissues. For matrigel imaging in which sensitivity for detecting very low levels of flow was paramount, power Doppler imaging (Sonos 5500, Philips Ultrasound, Bothell, Wash) was performed at 7.1 MHz with a mechanical index of 0.8, a pulse-repetition frequency of 2.4 KHz, and a medium packet size. This imaging modality did not provide sufficient spatial resolution to study regional hind-limb muscle imaging. Accordingly, in vivo limb imaging was performed with Contrast-Pulse Sequencing (Sequoia, Siemens Medical Systems, Mountain View, Calif), which balances microbubble sensitivity and spatial resolution. Imaging was performed at a centerline frequency of 7 MHz, a mechanical index of 1.6, and a dynamic range of 60 dB. The proximal hind-limb adductor muscles in the ischemic and contralateral control nonischemic limb were imaged in a transaxial plane midway between the inguinal fold and the knee.

Perfusion Imaging

Perfusion was assessed during continuous intravenous infusion of nontargeted lipid microbubbles at 1×10⁶ min⁻¹ for matrigel and 2×10⁶ min⁻¹ for hind-limb imaging (total volume ≤50 μL). Background images were acquired at 30 Hz. Intermittent imaging was performed by progressive prolongation of the pulsing interval from 0.2 to 12 seconds with an internal timer. Several frames were acquired at each pulsing interval. Averaged background frames were digitally subtracted from averaged contrast-enhanced frames at each pulsing interval. Pulsing interval versus video intensity (VI) data were fit to the following function: y=A(1−e^−βt), where y is the VI at the pulsing interval t; A is the plateau VI, which is an index of microvascular blood volume; and β is the rate constant, which provides a measure of microvascular blood velocity.²⁰ Microvascular blood flow was calculated by the product of A and β. For display purposes, parametric images of microvascular blood flow were created in which the value of A · β was displayed on a pixel-by-pixel basis. Because neovascularization of matrigel plugs progresses centripetally (from the outer surface toward the center),²¹ perfusion data were analyzed separately for the outer, middle, and central regions of the plug defined by a central and 2 surrounding concentric ellipsoid rings, all with equal radial width.

Targeted Imaging

Targeted imaging was performed as previously described after intravenous injection of control microbubbles or microbubbles targeted to VCAM-1, α₅-integrin, or neutrophil complement receptors (5×10⁶ for matrigel and 1×10⁷ for hind-limb imaging).¹⁹ The order of injection was randomized. A single image reflecting only retained microbubbles was created by capturing the initial frame obtained 5 minutes after injection and subtracting averaged frames at a pulsing interval of 10 seconds obtained after destroying all microbubbles within the sector with several high–mechanical-index frames. Because the number of microbubbles retained is influenced by microbubble influx, the image representing retained microbubbles was normalized to microvascular blood flow. Data for matrigel were analyzed according to radial position as described above. Flow-normalized targeted imaging data were validated by comparison with the microbubble retention fraction derived from modeling time-intensity curves during continuous low–mechanical-index CEU from 6 additional animals undergoing iliac occlusion (see the online Data Supplement). These studies demonstrated a good correlation between techniques in the ischemic hind-limb model.

Histology

Staining was performed on perfusion-fixed and paraffin-embedded sections. Hematoxylin and eosin staining was performed to evaluate inflammatory cell infiltration. For immunohistochemistry, an antigen retrieval protocol was performed, and primary antibody labeling was performed with sc1504 (VCAM-1), sc1506 (CD31), sc315 (flk-1), sc6593 (α₅-integrin), and sc-6999 (CD14) (Santa Cruz Biotechnology, Santa Cruz, Calif). Neutrophils were labeled with 7/4 (AbD Serotec, Raleigh, NC). Biotinylated polyclonal secondary antibodies (Vector Laboratories, Burlingame, Calif) were used for VCAM-1 and CD31 detection, and staining was performed with a peroxidase kit (Vector Laboratories) and 3,3′-diaminobenzidine chromagen. For detection of the remaining antigens, fluorescent labeling was performed with secondary antibodies conjugated with Alexa Fluor 488 or 555 (Molecular Probes, Carlsbad, Calif). Capillary density and ratio to myocyte number in skeletal muscle were determined from CD31-stained tissue sectioned in the transaxial fiber plane. Analysis was blinded to animal identity.

Statistical Analysis

Unless otherwise specified, parametric data are expressed as mean±SD. Comparisons between targeted and control agent were performed with ANOVA. Follow-up comparisons were made for each animal group by use of a paired t test with Bonferroni adjustment for multiple comparisons for bubble type. A Wilcoxon signed-rank test was used for VCAM-1 in the hind limb because these data violated assumptions for the paired t test. Comparisons between MCP-1^−/− and wild-type mice were made with an unpaired t test; data for α₅-integrin imaging were compared with an unequal variance version of the test. Blood flow data in each region of matrigel were assessed by an unpaired t test with Bonferroni adjustment for multiple comparisons for time interval. Tests for trends in hind-limb blood flow or microvascular flow components according to duration after arterial occlusion were made with Spearman’s rank correlation. Differences were considered significant at P<0.05 (2 sided).

The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.

Matrigel Neovascularization

On histology, a dense inflammatory response was present at the interface between the matrigel surface and surrounding tissue 3 days after subcutaneous injection (Figure 1A). Linear penetrations of mononuclear cells were observed at the outer margin of the plug. These cellular penetrations were surrounded by focal clearing of matrigel material, consistent with early channel formation. At day 6, there was further penetration of the cellular projections toward the plug center (Figure 1B). The outer portions of the channels had enlarged and were partially lined with spindle-shaped cells, consistent with vascular channel maturation. At day 10, these channels were interconnected in a reticular network, and almost all were partially lined with spindle-shaped cells (Figure 1C). The centripetal growth of channels on histology correlated temporally with the inward advancement of the leading edge of perfusion on CEU (Figure 2). The peripheral or outer region of the matrigel was well perfused even at day 3. Flow to the middle region was initially low at day 3 but increased significantly by day 10. The late increase in perfusion in this segment was due primarily to an increase in microvascular blood volume (median A values of 18.7, 12.4, and 37.0 VI units [VI_u] on days 3, 6, and 10, respectively). Flow in this region was verified by the presence of intravenously injected nanospheres in this region on fluorescent microscopy at day 10. Flow to the central core region was minimal at all time intervals (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Examples of hematoxylin and eosin staining of matrigel plugs seen as the pink amorphous material at day 3 (A), 6 (B), or 10 (C) after subcutaneous placement under high- and low-power magnification. Linear projections of inflammatory cells into the matrigel surface were observed at day 3. At subsequent time intervals, there was inward growth, enlargement, and interconnection of channels. There also was time-dependent lining of the channels with spindle-shaped cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Microvascular blood flow in matrigel determined by CEU perfusion imaging. A, Examples of parametric images of blood flow in matrigel plugs illustrating inward progression of blood flow over time but a persistent avascular core at day 10. B, Mean (±SEM) microvascular blood flow in the concentric regions of the matrigel plug. *P<0.05 vs day 6.

Molecular Imaging of Matrigel Neovascularization

According to CEU perfusion imaging data, the middle region of the matrigel provided an opportunity to temporally study de novo appearance of functional microvessels. In this region, there was robust molecular imaging signal from microbubbles targeted to α₅-integrin on days 3 and 6 when flow was relatively low (Figure 3A). This signal subsequently diminished by day 10. Immunohistochemistry demonstrated α₅-integrin staining predominantly on a subset of mononuclear cells infiltrating the matrigel material. Occasionally, spindle-shaped cells also stained positive for α₅-integrin, although the intensity was weak and was localized primarily to the outer matrigel region. Signal enhancement for VCAM-1–targeted microbubbles in the middle region was initially low at day 3 but then increased on day 6 (Figure 3B). This signal originated primarily from the spindle-shaped cells that lined the vascular channels. Some of these cells also stained positive for flk-1 (vascular endothelial growth factor receptor-2) (Figure 3C), consistent with endothelial differentiation. Infiltrating monocytes positive for flk-1 were present but uncommon. Positive staining for CD14 was found not only on a portion of the monocytes in the matrigel but also on some spindle-shaped cells lining the vascular channels, suggesting a possible monocytic origin for some of these cells (Figure 3D). Signal enhancement from microbubbles selective for activated neutrophils was low (<3.0 VI_u) in the middle region at all time points. Likewise, polymorphonuclear cells that stained positive for anti-neutrophil monoclonal antibody 7/4 were sparse on histology. Signal for control nontargeted microbubbles was low (<2.0 U) at all time intervals, indicating that targeted microbubble signal was due to ligand-specific binding.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Targeted CEU imaging data and examples of immunohistochemistry from middle matrigel region. Mean (±SEM) signal enhancement during molecular imaging for α₅-integrin (A) and VCAM-1 (B) (bars) is displayed with perfusion data superimposed (triangles). For both, examples of targeted signal enhancement on CEU at day 6 are shown (color scale at bottom). Immunohistology with diaminobenzidine chromagen (brown) demonstrated α₅-integrin present on infiltrating mononuclear cells. VCAM-1 signal originated primarily from the spindle-shaped cells lining the matrigel channels. Immunohistology for flk-1 (C; green) and CD14 (D; red) at day 6 demonstrated positive staining for spindle-shaped cells during channel formation and subsets of mononuclear cells. Scale bar=25 μm. *P<0.05 vs other time points for targeted imaging data.

Ischemic Hind-Limb Perfusion

In mice undergoing arterial ligation, microvascular blood flow in the ischemic proximal hind-limb adductor muscle was reduced to 10% to 20% of that in the control nonischemic limb at 2 days and increased in a time-dependent fashion (Figure 4). The increase in ischemic limb perfusion was secondary to changes in both microvascular blood volume and blood velocity (Table). Immunohistology for CD31 failed to detect any significant change in muscle capillary density over time (Figure 4), indicating that the increase in microvascular blood volume was due to restoration of precapillary hemodynamic conditions as described previously.²² However, enlargement of intramuscular arterioles and venules between days 2 and 7 was particularly prominent in transverse arterioles between major muscle fibers.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Microvascular responses to hind-limb ischemia produced by iliac artery occlusion. A, Examples of perfusion imaging from a single animal at day 2 illustrating reduced flow in the ischemic vs control contralateral limb. Background-subtracted color-coded images are shown at incremental pulsing intervals in seconds. B, Mean blood flow in the ischemic limb relative to the control contralateral limb. The probability value for trend is shown (Spearman’s ρ=0.77). C, Capillary density determined by CD31 immunohistology in ischemic and control hind-limb adductor muscle. D, Examples of immunohistology for CD31 illustrating the proliferation of small to medium vessels at days 4 and 7 after interruption of arterial inflow. Scale bar=50 μm.

Molecular Imaging of Hind-Limb Skeletal Muscle

Molecular imaging with microbubbles targeted to activated neutrophils demonstrated selective signal enhancement in the ischemic limb at day 2 that rapidly declined thereafter (Figure 5A). This signal correlated temporally with an abundance of polymorphonuclear leukocytes (positive staining with monoclonal antibody 7/4) in the perivascular and interfibrillar space at day 2 that was not present on subsequent days. Also present on day 2 were abundant α₅-integrin–positive mononuclear cells that continued to be detected on days 4 and 7 (Figure 5B). Likewise, targeted imaging signal for α₅-integrin in the ischemic limb was greatest at day 2 but remained significantly higher than signal from control microbubbles at days 4 and 7. Molecular imaging of VCAM-1 showed a similar pattern of robust signal enhancement in the ischemic limb at day 2 that persisted, albeit at a lower level, on subsequent days and correlated with endothelial expression of VCAM-1 in the intramuscular arterioles and venules on immunohistochemistry (Figure 5C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Molecular imaging data for microbubbles targeted to activated neutrophils (A), α₅-integrin (B), and VCAM-1 (C). For each molecular target, mean (±SEM) signal enhancement in the ischemic limb is shown for targeted and control microbubbles. Data are normalized to muscle blood flow. Corresponding examples of molecular imaging with targeted microbubbles at day 2 and immunohistochemistry for neutrophils (green), α₅-integrin (red), and VCAM-1 (brown) at days 2 to 7 are shown to the right of each graph. *P<0.05 vs control microbubble signal.

Imaging was performed in MCP-1^−/− mice 4 days after interruption of arterial inflow. This time interval was chosen on the basis of studies showing that genetic deletion of the MCP-1 receptor (CCR2) suppresses monocyte recruitment at day 4, before there is any detectable impairment in flow recovery.¹³ In the present study, there was a nonsignificant trend toward lower blood flow in the adductor muscles of the ischemic hind limb at day 4 in MCP-1^−/− compared with wild-type mice (Figure 6). Signal enhancement from α₅-integrin–targeted microbubbles in the ischemic limb was severely attenuated in MCP-1^−/− mice, whereas signal enhancement with neutrophil and VCAM-1–targeted microbubbles was only mildly reduced (Figure 6).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Microvascular perfusion and targeted imaging data in MCP-1^−/− mice. Mean (±SEM) data are shown for microvascular blood flow in the ischemic limb adductor muscle relative to control limb (A) and targeted molecular imaging for neutrophil complement receptors (B), α₅-integrin (C), and VCAM-1 expression (B) in wild-type (WT) and MCP-1^−/− and mice. VI_u is normalized to blood flow. *P<0.05 vs wild type.