1.

Introduction

Vulnerable atherosclerotic plaques in the coronary arteries are a leading cause of death and disease due to their propensity to rupture and cause thrombosis and myocardial infarction.¹ These rupture-prone plaques are characterized by large necrotic lipid cores, thin fibrous caps, and dense macrophage infiltration.² Because a large proportion of these plaques are nonstenotic (less than 60% narrowing of the lumen), they are frequently nonsymptomatic. In recent years, a plethora of methods have been devised in an attempt to detect them before a catastrophic rupture occurs.³ These detection methods include both invasive modalities that require cardiac catheterization and noninvasive imaging methods. Examples of the former invasive techniques include intravascular ultrasound,⁴ optical coherence tomography,⁵ intravascular thermography,⁶ and near-infrared spectroscopy.⁷ Examples of noninvasive techniques include multislice x-ray computed tomography,⁸ magnetic resonance imaging,⁹ electron-beam imaging of calcium deposits,¹⁰ and 2-[18F] fluoro-2-deoxy-D-glucose mediated positron emission tomography.¹¹

These detection methods rely on localization of various abnormalities associated with lesions in the vessel wall such as inflammation, fibrous cap thickness, lipid content, calcifications, and characteristic luminal appearance of the lesions. Activated macrophages are thought to be the chief culprit cells in causing plaque vulnerability by expressing proinflammatory cytokines,¹² secreting matrix metalloproteinases that further degrade the thin collagen cap¹³ and also producing the highly prothrombotic tissue factor that is released on plaque rupture.¹⁴ For these reasons, localizing and quantifying activated macrophages may be the best detection target in assessing vulnerability in coronary arterial lesions.

Photodynamic therapy (PDT) uses nontoxic photosensitizers (PS) that are photoactived on the absorption of light.^{15, 16} Exposure of the PS to the appropriate wavelength of light raises the PS molecule to an excited state that, depending on the intensity and wavelength of light used, can either lead to 1. the generation of cytotoxic reactive oxygen species, or 2. the emission of fluorescent light. PDT is currently approved for the treatment of esophageal and lung cancer and age-related macular degeneration.¹⁷ Its application to atherosclerotic disease is currently under investigation.¹⁸ There have been several reports that PS developed for treatment of other diseases such as cancer can localize in atherosclerotic lesions. ^{18, 19, 20, 21, 22} The reason for this selective localization is poorly understood, but may be related to their chemical structure being lipophilic and the plaques having high amounts of lipid accumulation. The fact that most PS that have been investigated for PDT are also fluorescent when excited with visible light, together with their ability to selectively accumulate in diseased tissue, has led to the administration of various PS being used for fluorescent diagnosis or detection of disease. ^{23, 24, 25, 26}

We have previously shown that covalent conjugates (MA-ce6) between the PS chlorin(e6) (ce6) and maleylated albumin (MA)²⁷ are taken up by macrophages with high specificity, and that subsequent photoactivation results in macrophage-specific phototoxicity.²⁸ MA-ce6 molecules are recognized by macrophage class A scavenger receptors (SR-As) that are high-capacity membrane glycoproteins that are confined mainly to tissue macrophages and related cell types.²⁹ The SR-As play an important role in the pathophysiology of atherosclerosis³⁰ by mediating the uptake of modified low-density lipoprotein uptake by human monocyte-derived macrophages,³¹ therefore encouraging their transformation into lipid-laden foam cells.³²

In this report, we investigated the use of MA-ce6 in an animal model of macrophage-rich atherosclerotic plaque in the rabbit aorta. This model, involving endothelial injury to the abdominal aorta followed by four months of an atherogenic diet, has been shown to lead to the development of inflamed atherosclerotic plaques with many characteristics of human vulnerable plaques including high levels of macro phage infiltration.^{33, 34} While unconjugated molecules closely related to ce6 have been reported to localize in atherosclerotic plaques, ^{19, 20, 21, 22} it was not known if MA-ce6 would localize in the same manner as free ce6 under the same conditions. We also decided to investigate the relative fluorescence accumulation at two different time points. Based on the fact that the pharmacokinetics of free ce6 are likely to be faster than that of the MA-ce6, we decided to compare accumulation 6 and $24\mathrm{h}$ postinjection.

2.

Materials and Methods

3.

Results

3.1.

Preparation and Fluorescence of MA-ce6 Conjugate

A schematic representation of the molecular structure of the conjugate is given in Fig. 1 . There were an average of two ce6 molecules per BSA molecule, and the remaining lysine epsilon amino groups were approximately 80% maleylated as determined by analysis of the absorption spectrum. The fluorescence emission of the conjugate was centered at $665\mathrm{nm}$ . The fluorescence emission of free ce6 was centered at $660\mathrm{nm}$ (data not shown).

Fig. 1

Chemical structure of MA-ce6 together with absorption and emission spectrum.

3.2.

Rabbit Model of Atherosclerotic Plaque

After the balloon injury and 12 to 16 weeks of atherogenic diet, the rabbits exhibited extensive atherosclerosis in the excised aorta and iliac arteries. We numbered the 12 segments that we analyzed as depicted in Fig. 2 . The segments most affected by the balloon injury were numbers 6, 7, and 8, and they had the worst levels of atherosclerotic plaque. Segments at the borders of the injured area (numbers 4, 5, and 9) had somewhat less severe plaque. In addition, we found that the uninjured segments of the aorta at the aortic arch (numbers 1 and 2) also developed severe atherosclerosis. Segments 3 and 10 had the least pronounced atherosclerosis in the aorta. The introduction of the deflated balloon catheter through the right femoral and iliac arteries also led to injury to the endothelium of the right iliac artery and consequent development of severe atherosclerosis not found in the uninjured left iliac artery. There were no completely normal arterial segments in the atherosclerotic rabbits. By contrast, histology of the normal rabbits showed unremarkable arteries with no evidence of plaque or intimal thickening (not shown).

Fig. 2

Photograph of excised en bloc atherosclerotic aorta and iliac arteries showing numbering of segments. 1. ascending aorta; 2. aortic arch; 3., 4., and 5., thoracic aorta; 6. aorta at diaphragm; 7., 8., and 9. abdominal aorta; 10. aortic bifurcation; inj: right iliac artery through which balloon was introduced; and uninj: normal left iliac artery. H and E sections of the typical segments obtained from the (a) aortic arch, (b) mid-thoracic noninjured site, (c) injured mid-abdominal, (d) aorta above bifurcation, (e) uninjured iliac artery, and (f) injured iliac artery.

3.3.

Intravascular Fluorescence in Blood-Filled Arteries

The approved animal protocol specified a terminal procedure to analyze the distribution of ce6 in the rabbits. Before administering a fatal dose of pentobarbital, we anesthetized two rabbits with ketamine/xylazine to determine whether it was possible to capture intravascular fluorescence in the aortas of living rabbits in the presence of flowing arterial blood. The fiber was inserted through the noninjured femoral and iliac artery and advanced into the aorta. However, probably due to the construction of the fiber optic probe that had a flat polished end, signals were nonsignificant until the fiber reached the aortic arch where the tip came into contact with the arterial wall as the artery curved around (see Fig. 2; data not shown). Then we sacrificed the animals and opened the body cavity with a longitudinal incision through the thoracic and abdominal walls to expose the aorta. This enabled gentle pressure with a gloved fingertip to be applied to compress the artery against the fiber tip and allow optical contact with the arterial wall. Figure 3 illustrates examples of intravascular fluorescence emission spectra obtained from the mid-abdominal aorta of atherosclerotic and normal rabbits $24\mathrm{h}$ after injection of MA-ce6. The intravascular measurements were taken in triplicate and at ten locations at 1-cm intervals throughout the length of the aorta, plus one set in the uninjured and one set in the injured iliac arteries. There were five conditions (MA-ce6 at $6\mathrm{h}$ and $24\mathrm{h}$ ; free ce6 at 6 and $24\mathrm{h}$ and no injection) and two sets of rabbits for each condition (normal and atherosclerotic). The values are presented in Fig. 4a for six atherosclerotic rabbits and in Fig. 4b for six normal rabbits.

Fig. 3

Representative intravascular fluorescence emission spectra from balloon-injured segment (mid-abdominal) of atherosclerotic rabbit aorta and corresponding section from normal rabbit. Both animals were injected with MA-ce6 $24\mathrm{h}$ before sacrifice.

Fig. 4

Mean intravascular fluorescence signal (in arbitrary units) for arterial segments gathered by fiber pullback in (a) atherosclerotic rabbits and (b) normal rabbits. Values are means of triplicate determinations at each segment from six rabbits, and bars are SEM.

The intensities of the intravascular fluorescence measurements in atherosclerotic animals decrease in the order MA-ce6 $24\mathrm{h}⪢$ ce6 $6\mathrm{h}$ $(\mathrm{p}<0.001)>$ MA-ce6 $6\mathrm{h}$ $(\mathrm{p}<0.05)>$ ce6 $24\mathrm{h}$ $(\mathrm{p}<0.05)>$ nothing $(\mathrm{p}<0.001)$ . Since the values obtained from rabbits injected with nothing can be assumed to be background autofluorescence from the artery wall, this number (approximately 100 arbitrary units) could be subtracted from the values, thus further increasing the advantage shown by MA-ce6 at $24\mathrm{h}$ . The highest values are in segments 1, 2, 6, 7, 8, 9, and the injured iliac artery. Segments 6 through 9 and the injured iliac artery develop the worst atherosclerosis due to the endothelial injury, while segments 1 and 2 near the aortic arch are the natural site for development of lesions in rabbits subjected to atherogenic diets. There is a very large selectivity for atherosclerotic rabbits compared to normal rabbits $(\mathrm{p}<0.0001)$ [Fig. 4b], where the values for both free ce6 and MA-ce6 at both time points do not differ significantly from values obtained from normal rabbits injected with nothing. This implies that none of these fluorophores accumulate in normal arteries. The autofluorescence (no injection) values from atherosclerotic rabbits were higher than those from normal rabbits (but not significantly).

3.4.

Fluorescence Localization ex-Vivo by Surface Spectrofluorimetry

Figure 5a shows the values obtained from emission spectro fluorimetry of the excised aortic segments opened up to expose the luminal surfaces after the intravascular experiments were completed. The order of intensities of the signals is MA-ce6 $24\mathrm{h}>\mathrm{MA}\text{-}\mathrm{ce}6$ $6\mathrm{h}$ $(\mathrm{p}<0.0005)>$ ce6 $6\mathrm{h}$ $(\mathrm{p}<0.0001)\cong$ ce6 $24\mathrm{h}\cong$ nothing. The signals from free ce6 at both 6 and $24\mathrm{h}$ did not significantly differ from background levels obtained in noninjected rabbits. The atherosclerotic signals were higher from segments 1, 4 through 8, and the injured iliac artery. The difference between fluorescence from MA-ce6 at 6 and $24\mathrm{h}$ was much smaller in segments when the maximum fluorescence was low (segments 2, 3, 9, 10, and uninjured iliac) than in the segments mentioned before where it was high. In the case of conjugate at 6 and $24\mathrm{h}$ the values obtained from segments excised from atherosclerotic rabbits were much higher $(\mathrm{p}<0.0001)$ than the corresponding segments taken from normal rabbits [Fig. 5b].

Fig. 5

Mean intimal surface fluorescence signal (in arbitrary units) for arterial segments removed and opened from (a) atherosclerotic rabbits and (b) normal rabbits. Values are means of 6 to 24 determinations at each segment from six rabbits, and bars are SEM.

3.5.

Extraction and Quantification of ce6 in Aortic Segments

Figures 6a and 6b show the values obtained by extraction of the ce6 from the arterial segments and quantification of the ce6 concentration in mol/g tissue by fluorescence spectrophotometry of the dissolved tissue. The concentration of MA-ce6 in atherosclerotic arteries is significantly higher at $24\mathrm{h}$ compared to $6\mathrm{h}$ $(\mathrm{p}<0.0001)$ . The values for MA-ce6 at $24\mathrm{h}$ are significantly larger in atherosclerotic rabbits compared to those from normal rabbits $(\mathrm{p}<0.0001)$ . The values of free ce6 at both 6 and $24\mathrm{h}$ in atherosclerotic rabbits are not different from background (no injection).

Fig. 6

Mean values of ce6 content (in mol ce6 equivalent per gram tissue) from arterial segments that were removed and opened from (a) atherosclerotic rabbits and (b) normal rabbits, and subsequently dissolved and quantified by fluorimetry. Values are means of 1 to 3 determinations at each segment from six rabbits, and bars are SEM.

3.6.

Biodistribution in Nontarget Organs

Figure 7 depicts the values of ce6 concentration in the tissue of nontarget organs obtained by extraction. We only measured the values for atherosclerotic rabbits injected with MA-ce6 at 6 and $24\mathrm{h}$ . Skin and muscle had low levels, followed by heart, fat, plasma, and bile with intermediate levels, and liver, spleen, bladder, lung, and gall bladder had the highest levels. Most of the organs had lower levels at $24\mathrm{h}$ compared to $6\mathrm{h}$ , except the liver and bile, which increased at $24\mathrm{h}$ .

Fig. 7

Mean values of ce6 (in mol equivalent per gram tissue) in nontarget organs from atherosclerotic rabbits injected with MA-ce6 and sacrifced either 6 or $24\mathrm{h}$ later. Values are means of 1 to 3 samples from each organ from six rabbits, and bars are SEM.

4.

Discussion

There have been many reports of cell-type specific targeting using a conjugate between a reporter or therapeutic molecule, joined to a macromolecular ligand, recognized by a receptor, expressed or overexpressed by a certain cell type involved in disease. To validate this claim, an essential control is the administration of the unconjugated molecule on its own. This is even more necessary in the case of targeted delivery of a PS to atherosclerotic plaques, as there have been several reports of free PS alone localizing in various atherosclerotic lesions in diverse animal models of atherosclerosis. ^{19, 20, 21, 22}

The findings in the present study are highly supportive of the hypothesis that scavenger-receptor targeting of ce6 can indeed provide an advantage over the accumulation of free ce6 in plaques that appears to occur by an unknown mechanism. When MA-ce6 is used with a 24-h time interval, there is significantly more ce6 in plaques than free ce6 at either time point, without a corresponding increase in accumulation in normal rabbit arteries. This applies to all segments of the diseased aorta but particularly to the segments that have more severe disease either by being located near the aortic arch (segments 1 and 2) or by suffering endothelial denudation (segments 6, 7, 8, and injured iliac). It was interesting to observe that the MA-ce6 conjugate at $24\mathrm{h}$ had higher values in the plaques than the same conjugate at the same dose at $6\mathrm{h}$ . The explanation for this observation is presumably that the conjugate is actively taken up by macrophages in the plaques by the time-dependent process of endocytosis, and therefore the cells can continue to accumulate fluorescence as long as there is conjugate still in circulation. This time difference in fluorescence accumulation is not observed in arterial segments from normal rabbits, as these nondiseased arteries are devoid of macrophages. The data from the nontarget organs agree with the hypothesis of active accumulation. The liver has the highest concentration of tissue macrophages, and this was the only organ that had much higher amounts at $24\mathrm{h}$ compared to $6\mathrm{h}$ . In addition, the amount in bile was higher at $24\mathrm{h}$ , in agreement with the function of the liver to excrete lipophilic molecules in bile.

As expected, the free ce6 did show a somewhat selective localization in plaques both at 6 and $24\mathrm{h}$ (especially by intravascular fluorescence). This was more pronounced at the 6-h time point. This can be explained by the fact that as ce6 is a small molecule, the pharmacokinetics may be expected to be faster than the macromolecular MA-ce6 conjugate. When free ce6 is used as a PS for cancer treatment, illumination is usually carried out relatively soon after injection into the bloodstream.³⁵ The present study has not shed any light on the mechanism of plaque localization by free ce6, but plausible hypotheses include accumulation in lipid-rich areas of plaques, penetration through leaky endothelial barriers, and even uptake by the macrophages present in the lesions.

Some differences were observed between the various modalities used to quantify the ce6 accumulation in the atherosclerotic plaques. The MA-ce6 at $24\mathrm{h}$ was the highest by a considerable margin using all the different techniques. However, the MA-ce6 at $6\mathrm{h}$ gave significantly higher values in the surface fluorescence measurements than in the intravascular measurements, while the situation was reversed for free ce6 at both time points. The explanation for this observation is not obvious. It is possible that the detection of the fluorescence through the blood that filled the arteries was more efficient for the free ce6 that has a higher fluorescence quantum yield than the conjugate, where the fluorescence quantum yield of the dye is reduced by the covalent conjugation. The free ce6 may have been more loosely attached to the surface of the arteries, so it could have been detected in the intact aortas but was lost when the arterial segments were washed in saline before surface detection. This possibility would also explain the absence of significant amounts of free ce6 in the arterial segment extraction data at both 6 and $24\mathrm{h}$ [Fig. 6a].

The optical arrangement of the surface fluorescence measurements involves a 3-mm-diam window that sits in contact with the tissue, and means that the light enters the tissue at right angles. This configuration is in contrast with the intravascular fiber optic probe, where the light is launched parallel to the intimal surface and therefore will not penetrate as far into the tissue. A comparative estimate of the penetration depths would be in the order of 100 to $200\mathrm{\mu }\mathrm{m}$ for the surface fluorescence but only 10 to $60\mathrm{\mu }\mathrm{m}$ for the intravascular probe. It is possible that the macrophages that accumulate MA-ce6 are located deeper in the intimal lesions than the accumulation of free ce6 that may be shallower due to endo thelial leakiness. This would then explain the difference between the relative magnitudes of the MA-ce6 at $6\mathrm{h}$ and those of free ce6, as determined by intravascular and surface fluorescence measurements.

The highest relative signals for MA-ce6 at $24\mathrm{h}$ compared to $6\mathrm{h}$ were obtained using the tissue extraction procedure. The explanation for this observation is likely to be the continued active accumulation of MA-ce6 by receptor-mediated endocytosis in the plaque macrophages. The nature of the fluorescence measurements at the artery wall (both intravascular and surface) means that the measured signal may become saturated at certain levels of plaque accumulation, possibly because the ce6 is in a less fluorescent aggregated state when present at high levels, but on dissolution in a disaggregating solvent $(\mathrm{NaOH}∕\mathrm{SDS})$ , the fluorescence becomes linear with ce6 concentration. One question that has not been addressed by the present study is how the conjugate is transported to the macrophages in the plaques. It is possible that the conjugate passes through the endothelium lining the lumen of the artery, or alternatively, it could be delivered to the plaques by the neovasculature (vasa vasorum) that is typical of atherosclerotic plaques.

The intravascular fluorescence detection system requires optimization to function satisfactorily in flowing blood. This might involve redesign of the fiber tip to produce a side-firing optical configuration, and to ensure that the optical path length through blood is minimized by allowing the tip to maintain gentle contact with the artery wall. There have been considerable technical improvements in intravascular fiber catheter design to enable such techniques as optical coherence tomography³⁶ and near-infrared spectroscopy⁷ to proceed to clinical trials. Although the rabbits did not suffer any toxicity from injections of the scavenger-receptor targeted conjugate, it should be noted that such preparations might not be pharmacologically neutral. Binding of ligands to macrophage scavenger receptors may induce signaling cascades, leading to macrophage activation and possibly to inflammation,^{37, 38} but recent data has shown that this SRA-mediated signaling is unlikely.^{39, 40}

In summary, this study demonstrates the superior local ization of a scavenger-receptor-targeted PS conjugate in macrophage-rich atherosclerotic plaques $24\mathrm{h}$ after injection into rabbits. A rationally designed fiber-optic-based fluorescence detection system with the correct fiber-tip geometry may answer the question of whether intravascular detection of macrophage-rich plaques is feasible. Future work will correlate the accumulation of MA-ce6 in plaques to immunohistochemical localization of macrophages, and investigate the ability of the MA-ce6 conjugate to mediate therapeutically beneficial PDT of the atherosclerotic lesions after intravascular delivery of red light in living rabbits.