As an important component of an expanding vascular network the neovasculature can be viewed as a biomarker of inflammation associated with tumor aggressiveness or atherosclerotic plaque progression. 1-3 To date, noninvasive imaging of angiogenesis has been essentially a preclinical endeavor, although imaging candidates have reached the clinical testing level for MRI and nuclear medicine. Nuclear medicine probes, although exquisitely specific for neovascular biosignatures, such as the α_νβ₃-integrin, are not specific to angiogenesis, since this adhesion molecule is activated on numerous cell types present in tumors or atherosclerotic lesions. 4-6 By comparison, MRI paramagnetic nanoparticle agents are often vascular-constrained by size and very specific for neoendothelial expression of α_νβ₃-integrin, but such agents have experienced other roadblocks in the clinic. 7 Photoacoustic molecular imaging offers noninvasive vascular-constrained detection of angiogenic biomarkers with very high resolution imaging of the neovasculature.

Different imaging modalities have different characteristic resolution limits, with microscopic images having the highest resolution for thin specimens. For biomedical instrumentation, ultra-high spatial resolution MRI at 3.0T and high resolution CT provide in vivo blood vessel images but with a poorer resolution, in the range of 200 to 500 μm, which is inadequate for discrete neovessel imaging. 8-12 Photoacoustic (PA) imaging combines the advantages of ultrasound imaging and optical imaging and offers high depth-to-resolution ratios at different spatial scales, ranging from organelles to organs, 13 as well as unique and important physiological assessments, such as differential circulating blood oxygenation levels.

Previous studies have confirmed that PA imaging's utility can be markedly extended with exogenous contrast agents. 14-16 In many instances, these contrast agents have been based on gold cubes, spheres, or rods with particle sizes greater than the renal clearance threshold (<8 to 10 nm), creating difficult translational challenges when proving long-term safety in man. Gold nanobeacons (120-160 nm) were developed to achieve the large absorption cross-sections of bigger solid gold particles by suspending a multitude of small gold nanoparticles (3-5 nm) in a hydrophobic matrix encapsulated with phospholipid (160 nm). 17 Moreover, the optical absorption wavelength of nanoparticles was tunable well into the NIR optical window, minimizing endogenous absorption and maximizing PA imaging depth.

However, the pressure of healthcare economics has placed an emphasis on driving the costs of imaging studies down, and we have thus sought lower cost biocompatible alternatives to gold. We hypothesized that nanoparticles entrapping high densities of organically soluble small molecular weight copper complexes could provide an inexpensive and effective approach to neovascular targeted PA imaging. The choice of copper as a cheaper, biocompatible alternative to gold was independently supported by a recent report describing nontargeted crystalline copper sulfide nanoparticles as PA contrast agents for deep tissue imaging at 1064 nm. 18 In this research, a soft noncrystalline α_νβ₃-targeted copper-oleate nanoparticle (α_νβ₃-CuNP) was developed and its PA imaging signal with respect to blood and the previously reported gold nanobeacons was characterized 17 as a function of NIR wavelength and with varying nanoparticle concentrations. The in vivo efficacy of α_νβ₃-CuNPs for high-resolution imaging of mouse neovasculature was examined in a Matrigel™ model of angiogenesis. Furthermore, the theranostic antiangiogenic potential of α_νβ₃-CuNPs incorporating a novel Sn 2 lipase-labile fumagillin prodrug (Fum-PD) was also studied. 19

Given the inherently strong PA signal of hemoglobin, the efficacy of PA contrast agents for targeted molecular imaging of neovascular integrin expression is dependent on the strength and differentiation of its signal from the blood background. The PA signal of CuNPs (20 vol%) was compared with whole bovine blood (45% hct.) at 767 nm. As seen in Figure 2, the time-aligned PA signal of the copper NP emulsion was markedly greater than that of the bovine blood at 767. The peak-to-peak signal of the CuNPs (2.3 a.u.) was 4-fold greater than that of blood (0.58 a.u.) at 767 nm, suggesting that the PA signal of neovascular bound CuNPs could be readily detectable in vivo.

GNBs have been previously shown to produce strong PA contrast in vivo when applied to sentinel lymph node imaging or when targeted to neovasculature in vivo. 14, 17, 22, 23 The PA signal enhancement of CuNPs relative to blood (45% hct.) was compared to the signal of GNBs relative to blood across a NIR frequency range of 730 nm to 830 nm (Figure 3) on an equal nanoparticle number basis. Over almost the entire spectrum, the CuNPs signal strength relative to blood was around 5-fold greater with a noticeable decline beyond 820 nm. In general, the PA ratios of GNBs and CuNPs were very similar except between 760 nm and 790 nm, with the greatest divergence between GNBs and CuNPs at 770 nm, at which point the gold contrast result was nearly twice that of copper. GNB particles experienced the same declining results as CuNPs from 800 nm and beyond.

The concentration dependence of CuNPs' signals as a function of frequency was examined. CuNPs were serially titrated from 6.0 µM to 0.17 µM and studied at NIR wavelengths between 750 nm and 790 nm (Figure 4 (A)). PA amplitude decreased with reductions in CuNP concentration level, but the magnitude of signal diminishment was generally less than the percent change in CuNP level. This was particularly notable for the higher concentrations, between 6.0 µM and 1.5 µM. The magnitude of the CuNP's PA signal was greatest at 750 nm and 760 nm. At higher wavelengths, the magnitude of the CuNP signal trended lower, with the greatest contrast loss seen at the lower CuNP concentrations. It can be calculated from the extinction coefficient spectrum shown in Figure 4 (B) that the extinction coefficient of CuNPs at 700 nm was 1.3 times as strong as that at around 770 nm. However, the laser energy at 767 nm (the peak of laser energy) is an order of magnitude stronger than that at 700 nm. Therefore, neovascular imaging signal peaks around 767 nm.

To provide robust assessment of PA signal enhancement due to the CuNPs, in vivo neovascular imaging was studied in the Matrigel plug model in mice. Mice implanted with Matrigel™ 18 days previously received either α_νβ₃-CuNPs, nontargeted CuNPs, or α_νβ₃-CuNP preceded by 10 minutes with a competitive dose α_νβ₃-oil only NPs (1:1). As seen in Figure 5, at 0 min, forming neovessels were observed by the inherent PA contrast imparted by erythrocyte hemoglobin. The magnitude of the signal varied among animals with the extent of neovascular proliferation. Following α_νβ₃-CuNP injection, the magnitude of the PA signal in these immature tubules clearly increased, and numerous incomplete vascular sprouts appeared, as we have previously reported for GNBs. 14 The formation of the neovessel shoots depicted an early evolution stage of a dense and irregular microvasculature. In contradistinction to the animals receiving α_νβ₃-CuNPs, those given nontargeted-CuNPs had very little increase in the vessel-integrated PA amplitude. As shown, a minimal passive accumulation of the CuNPs in the forming bridges and sprouts combined with a modest amount of blood pool enhancement relative to the neovessel tubules at 0 min. We have previously reported that blood pool enhancement for GNBs in the intact microvasculature surrounding the Matrigel implant was negligible, 14 but within the plug, the slow moving blood may be associated with very slow washout. In the competition group at 170 min, very little change in the PA signal from the forming vascular tubules or sprouts was observed. In this group, the pretreatment binding of the α_νβ₃-oil NPs appeared to have blocked most of the receptors for the α_νβ₃-CuNP and even precluded significant passive accumulation. Figure 6 presents the average change in the vessel-integrated PA amplitude between the three treatment groups, showing that the α_νβ₃-CuNPs increased PA contrast, while the nontargeted-CuNP did not (p<0.05). Moreover, the specificity of targeting the neovasculature was again demonstrated by the low vessel-integrated PA amplitude measured in the competition group versus the α_νβ₃-CuNP group (p<0.05), and even relative to the nontargeted control. Dynamic accumulation of PA contrast measured in the α_νβ₃-CuNP group increased progressively over the 230 minutes post injection interval (Figure 7).

In a second aspect of the Matrigel^™ angiogenesis experiment, fumagillin-prodrug (Fum-PD), 24, 25, was incorporated into the phospholipid surfactant at 2.28 mol%. The chemical synthesis, characterization, and in vivo effectiveness of Fum-PD incorporated into perfluorocarbon nanoparticles has been previously reported. 19, 26 α_νβ₃-CuNP (2 µl/g) incorporating Fum-PD was administered on days 11 and 15 post Matrigel^™ implant, and the vessel-integrated PA amplitude in these animals was determined on day 18. At baseline, the amount of neovasculature observed by PA imaging was sparse, but no more so than that in the other groups. Moreover, following α_νβ₃-CuNP injection there was little sprouting enhancement at 170 minutes. As shown collectively in Figures 5 and 6, the vessel-integrated PA amplitude change was similar to that measured in the nontargeted-CuNP group, and appeared related to modest particle accumulation in the nascent tubules, with less signal related to sprout formation. Although drug delivery and imaging were not performed simultaneously to minimize the differential animal handling between the groups, these results illustrated the potential of CuNPs as vehicles for drug delivery and diagnosis.

Photoacoustic imaging has demonstrated broad applications in the imaging and physiological characterization of the vasculature, particularly the microvasculature 13, 27, and microvascular expansion driven by brain tumor hypoxia 28. The vascular endothelium provides numerous “road-signs” reflecting its developmental stage, as seen for the neovasculature, or its state of activation, as seen with inflammation. In each instance, specific PA characterization depends on molecular imaging through the recognition of unique endothelial biomarkers. 15, 16, 22, 29-33 In the present study, sparse activated α_νβ₃-integrin was selected as a biomarker of angiogenesis. While we have previously reported dynamic imaging of the forming neovasculature with gold nanobeacons, the high and unpredictable cost of gold, and the unclear long-term safety of larger nonmetabolizable nanoparticles, such as carbon fullerenes or nanotubes, led to the design and development of CuNPs. While free copper can elicit systemic toxicity, CuNPs are comprised of high densities of small organometallic complexes that can be eliminated through the bile or urine, minimizing the risk potential for this agent. 34 In the present study, CuNPs offered comparable PA contrast enhancement, on an equal particle basis, to GNBs both in vitro and in vivo.

In vivo PA imaging was performed at 767 nm, which was optimal for the previously targeted GNB agent, but the spectral assessments of CuNP relative to blood and as a function of particle concentration suggest 740 or 750 nm may be preferred for targeting sparse epitopes like adhesion molecules. Also, the time for maximum neovascular signal was found to be nearly 4 hours in the Matrigel model, but similar particles administered in vivo for deeper-tissue angiogenesis imaging with MRI were shown to provide minimal targeting benefit after 2 hours. 35, 36 The difference likely relates to vascular vasoconstriction with reduced blood flow to the skin in the subcutaneous Matrigel implant model, which is aggravated by the cooling effects of acoustically coupling the PA probe to the body during anesthesia, despite provision of supplemental heating to maintain core body temperatures. Lastly, CuNPs were designed to be vascular constrained agents to reduce nonspecific signal generated by interactions with extravascular cells expressing the α_νβ₃-integrin. Close examination of the 170-minute PA images in Figure 5 showed negligible extravasation of the particles. Although much attention has been attributed to the leakiness of angiogenesis, even in the sprouting regions where the neovasculature is essentially open-ended, no evidence of nanoparticle accumulation beyond the tips of the forming neovessel branches was noted. Moreover, this was a consistent observation, regardless of whether the nanoparticles were integrin-targeted or not.

Fumagillin prodrug was incorporated into the phospholipid surfactant of CuNPs and demonstrated to elicit anti-angiogenic effects in this Matrigel model. Microanatomically, the treatment virtually eliminated the rich web of sprouts in the mice receiving α_νβ₃-CuNPs without Fum-PD. “Pruning” of neovasculature was first suggested by Jain in a visionary manner, but in this study, the concept was visually apparent. 37-39 Despite all of the extensive clinical research on the effects and benefits of anti-angiogenesis therapy in cancer and selected other pathologies, virtually all of the work has centered on vascular endothelial growth factor (VEGF) inhibition. While VEGF is a clear driver of angiogenesis, it elicits pleotropic effects and is not solely responsible for neovascularization. 40 Moreover, VEGF is produced by a variety of cell types, making effective blockade of the growth factor challenging. Typically, relatively high dosages of medicants are used clinically and these treatments have well known side-effects. 41

As reported by others, fumagillin is an antiangiogenic agent which is specific for proliferating endothelial cells, but its clinical analogue, TNP-470, possessed only antidotal effectiveness in the clinic at the high doses required, which were accompanied by numerous toxicities, including neurocognitive dysfunction. 42-44 Fumagillin can be effectively incorporated into the phospholipid surfactant of nanoparticles for targeted delivery of anti-angiogenesis therapy through a mechanism referred to as “contact facilitated drug delivery” (CFDD) at a small fraction of the dosages previously used systemically. 2, 45, 46 Tethering of the lipid-encapsulated nanoparticle to the target cell surface facilitates the interaction and hemifusion of the two lipid membranes, which affords the passive transfer of the drug and phospholipids from the nanoparticle surface to the outer leaflet of the target cell membrane. The drug is then translocated to the inner leaflet through an ATP-dependent mechanism. 47, 48 CFDD eliminates the need for particle internalization with subsequent endosomal drug payload escape.

However, pharmacokinetic studies showed that fumagillin dissolved into the lipid membrane was substantially lost prematurely during circulation to the target neovasculature, despite its effectiveness in vivo. Fum-PD was developed to address this early release by coupling the drug through the Sn 2 acyl position (i.e., the stereospecific hydroxyl group of the second carbon of glycerol). Subsequent transfer of the monolayer components into the target cell membrane allows cytosolic phospholipases to enzymatically cleave the Sn 2 ester and release the drug, allowing it to diffuse into the cytosol for effect. Fum-PD as well as other similar prodrug compounds is incorporated stably into the hydrophobic aspects of the particle membrane. Importantly, Sn 2 lipase-labile prodrugs within the particle lipid membrane are stable in blood and plasma, even in the presence of excess exogenous lipase, and do not passively exchange to co-circulating RBC. 49 The results of the present research show the first example of a systemically targeted antiangiogenic drug delivery with a photoacoustic contrast nanoparticle. Given the rapid developments in hardware and software for PA imaging, one can envision numerous medical applications for low-cost CuNPs particles to target neovascular or alternative endothelial biosignatures for diagnostic and theranostic purposes.

Unlike current VEGF inhibitors, fumagillin-prodrug specifically targeted to angiogenic vessels suppresses angiogenesis, clinical disease, and inflammation in a preclinical model of rheumatoid arthritis (RA). 50 Enhanced endothelial nitric oxide (NO) modulates local macrophage inflammatory activity through NO activation of AMP-activated protein kinase (AMPK). In vivo, NO-induced AMPK activation increased autophagy by inhibiting mammalian target of rapamycin (mTOR) activity. Increased autophagy mediated the degradation of IkappaB kinase (IKK) and suppressed NF-κB. The suppression of NF-κB was associated with diminished inflammatory cytokine release. Importantly, the NO mediated inhibition of inflammation was reversed in vitro and in vivo by the co-administration of N(G)-nitro-L-arginine methyl ester (L-NAME), a nitric oxide synthase inhibitor. These unique anti-angiogenic and anti-inflammatory properties of Fum-PD nanotherapy may be applicable to the treatment of other angiogenesis-dependent diseases.

Soft copper oleate nanoparticles were synthesized and conceptually shown to offer effective photoacoustic contrast comparable to gold-based agents, but at a much lower cost and higher material availability. CuNPs possessed strong PA contrast relative to blood from 730 nm to 830 nm, which was similar to the PA signal response of gold nanobeacons relative to blood overall. In the Matrigel angiogenesis model, α_νβ₃-CuNPs markedly enhanced neovessel tubules and identified a myriad of incomplete emergent neovascular sprouts, which were not well detected or specifically differentiated as angiogenesis with inherent PA imaging. α_νβ₃-CuNPs effectively delivered fumagillin-prodrug, a potent anti-angiogenic therapy, in vivo, providing the first example of a systemically targeted drug delivery therapy with a photoacoustic contrast agent.