2.1.

Materials

The photosensitizers aluminum (III) phthalocyanine chloride tetrasulfonic acid (Al-4) and aluminum (III) phthalocyanine disulfonate chloride adjacent isomer (Al-2) were purchased from Frontier Scientific, Incorporated, UT, USA. The photosensitizer 2-(1-hexyloxyethyl)-2-devinyl pyropheophorbide (HPPH) was kindly provided by Roswell Park Cancer Institute. Liposomes were prepared using 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and cholesterol (CHOL, Northern Lipids Incorporated, B.C., Canada). Solvents used in bulk solution and liposome studies were either phosphate-buffered saline (PBS, pH 7.4) or 20:80 fetal bovine serum (FBS)/PBS (Sigma) dissolved in Milli-Q ultrapure water (UV-treated $18.2\mathrm{M}\Omega \cdot \mathrm{cm}$ resistivity). Solutions were filtered through a Nalgene $0.45\text{-}\mu \mathrm{M}$ filter to remove scattering contaminants.

2.2.

Liposome Preparation with Photosensitizers and Spectral Characterization

Liposomes were prepared by dissolving $50\mathrm{mg}$ of DOPC and $8.6\mathrm{mg}$ $\left(35\%\mathrm{mol}\right)$ of CHOL in chloroform. The chloroform was subsequently roto-evaporated until a uniform thin lipid film was formed. This lipid cake was hydrated by drop-wise additions of PS $\left(5.0\mathrm{mL}\right)$ at $40°\mathrm{C}$ , and swirling, over $30\min$ . The photosensitizer concentrations used were $50\text{-}\mu \mathrm{M}$ Al-4 in PBS, $25\text{-}\mu \mathrm{M}$ Al-2 in PBS (presonicated to aid solubility), and $100\text{-}\mu \mathrm{M}$ HPPH in FBS/PBS. The hydrated phospholipids were agitated and resuspended by rotating on a roto-evaporator $\left(180\mathrm{rpm}\right)$ without applying vacuum for $15\min$ at $65°\mathrm{C}$ . The resulting liposomes were subjected to 10 freeze-thaw cycles to induce stability (at $-80°\mathrm{C}$ and $60°\mathrm{C}$ , respectively). The liposomes were extruded 11 times through a $100\text{-}\mathrm{nm}$ pore-size polycarbonate membrane to ensure a narrow size distribution. Vesicle size and distribution were determined for both liposomes incorporating PSs (L-PS) and blank liposomes containing only PBS (L-blank). Particle sizing was based on dynamic light scattering measurements (DLS) acquired at varying angles between $90\text{to}150\mathrm{deg}$ (BI9000 apparatus from Brookhaven Instruments at $532\mathrm{nm}$ using BI200SM goniometer and B2FBK/RFI PMT). The vesicle sizes from Gaussian/Nicomp distribution analysis (intensity weighted) were tightly grouped around a single mean value $\left(135\mathrm{nm}\right)$ . Free PSs were removed from the solution containing encapsulated PSs using Sephadex G-25 prepacked size exclusion columns (NAP 5, GE Healthcare). Images of the L-PSs were acquired with an epifluorescence inverted microscope (Eclipse TE2000-U, Nikon) equipped with a charge-coupled device (CCD) camera (QICAM fast 1394, QImaging) and $100\times$ objective. The absorption spectra of the bulk PS solutions used for liposomal preparation and the confined L-PS solutions corresponding to the highest concentration (sample A) were recorded with a spectrophotometer (Cary500, Varian). Fluorescence measurements of a range of concentrations of bulk PS solution and L-PS were performed with a spectrofluorometer (Cary Eclipse, Varian) using $661\text{-}\mathrm{nm}$ excitation. The emission was collected with a bandpass filter $\left(680\text{to}710\mathrm{nm}\right)$ identical to that used for fluorescence lifetime studies (Ref. 36), allowing direct comparison between spectral and temporal studies.

2.3.

Flow Cytometry

The liposome vesicle concentration of each sample was determined by flow cytometry using the fluorescence channel, since a portion of the liposome’s distribution size was too small to be adequately detected by forward scattering. The fluorescence channel was then used directly as a gate. The liposomes were analyzed using an EPICS Elite ESP (Beckman Coulter) device equipped with a $633\text{-}\mathrm{nm}$ He-Ne laser at $10\mathrm{mW}$ . The fluorescence emission was separated from the excitation light by using a $650\text{-}\mathrm{nm}$ dichroic mirror and a bandpass filter centered at $670\mathrm{nm}$ . Fluorescent calibration beads were used to align the laser to minimize the coefficient of variation induced by the device itself.

The liposome concentration (quantity of liposomes per milliliter) were subsequently adjusted to the same value for all L-PS samples. The blank liposome concentration was determined by bilayer labeling with Nile Red (Sigma).³⁷ No separation of unbound Nile Red was performed, as this dye selectively fluoresces when attached to lipids.^{38, 39} The concentration of L-blank (unlabeled) liposomes were subsequently matched to the L-PS concentrations.

3.1.

Physical Characterization of Liposomes

PSs Al-2, Al-4, and HPPH were incorporated in liposomes separately, and L-blank was prepared for background comparison. After the extrusion step, the average liposome size was determined, summarized in Table 1 . It was observed that the mean vesicle diameter was similar for all preparations, $135\pm 45\mathrm{nm}$ . Once the PSs were incorporated into the liposomes, the unencapsulated PSs were removed from the media by size exclusion separation (not performed for L-blank). The variable retention of liposomes in the size exclusion column is a dynamic and reversible known process.⁴⁰ Therefore, the exact amount of liposomes lost due to separation was quantified by flow cytometry, and found to depend on the specific PS incorporated. The resulting liposome concentrations were $6.6\times {10}^7$ , $8.8\times {10}^7$ , and $7.4\times {10}^7$ liposomes ${\mathrm{mL}}^{-1}$ for L-Al-2, L-AL-4, and L-HPPH, respectively. These results indicate that the liposome retention in the column was not the same for every L-PS, and may be related to the spatial localization of PS within the liposome particles traveling through the Sephadex material. Using the flow cytometry data, linear dilutions were performed to achieve a liposome concentration series ranging from $2.7\times {10}^7$ liposomes ${\mathrm{mL}}^{-1}$ (concentration A) to $1.1\times {10}^5$ liposomes ${\mathrm{mL}}^{-1}$ (concentration F).

Table 1

Average liposome sizes and concentrations of L-Ps used after separation.

3.2.

Microscopic Characterization of Confined Photosensitizers

Examination of the three selected PSs chemical structures, particularly the acid-base properties of the substituents, revealed very different hydrophilic-hydrophobic properties (Fig. 1 ). Consequently, it is expected that there will be structure-dependent, preferential localization of PS within a liposome.¹³ Figure 1 presents images of the confined PSs in liposomes obtained by epifluorescence microscopy. To aid visualization, these pictures were derived from highly concentrated solutions containing large, polydispersive liposomes. Accompanying Fig. 1 are schematic representations of the PSs spatial confinement in liposomes suggested by the fluorescence micrographs.

Fig. 1

Epifluorescence microscope images of confined PSs (in polydisperse liposomes) with their chemical structure and suggested schematic of the PS liposomal localization of (a) L-Al-2, (b) L-Al-4, and (c) L-HPPH.

Figure 1a shows fluorescence emitted both from the liposome membrane and the core of L-Al-2. In contrast, the fluorescence observed in L-HPPH [Fig. 1c] arises solely from the bilayer membrane. Previous studies of $\mathrm{Al}\mathrm{Pc}\mathrm{S}n$ in cationic reversed micelles suggest that amphiphilic molecules, such as Al-2, tend to localize at the interface of the lipid/core construct.¹³ Hence, the localization of Al-2 is believed to occur at both the internal and external interface of the phospholipid bilayer, explaining the diffuse pattern of the fluorescence observed in Fig. 1a. The highly hydrophilic nature of Al-4, arising from the tetrasulfonate substituents, is believed to drive this PS primarily to the aqueous core of the liposome. Indeed, Fig. 1b confirms the hydrophilicity argument, and Al-4 fluorescence is mainly observed inside the central cavity of the vesicle. As HPPH is highly lipophilic,⁴¹ fluorescence is only detected from hydrophobic domains of the liposome, defined strictly by the phospholipid boundary [Fig. 1c].

3.3.

Liposomal Confinement of Amphiphilic Photosensitizers

Figure 2 presents the absorption and fluorescence emission spectra for different concentrations of free Al-2 in bulk solution, as well as confined in liposomes. The characteristics of the absorbance spectra of monomeric phthalocyanines in bulk solution are well documented. A strong Soret peak around $350\mathrm{nm}$ , a weak maximum around $600\mathrm{nm}$ , and a narrow, very strong Q-band absorption peak in the far-red region of the visible spectra around $670\mathrm{nm}$ have been reported previously.^{4, 13, 14, 15, 42} However, phthalocyanines in water tend to form dimers and higher order aggregates due to the propensity of the large hydrophobic skeleton to avoid contact with an aqueous medium.¹³ Dimer formation can be observed spectroscopically, as the Q-band exhibits a decrease in the extinction coefficient and a blueshift to $\sim 640\mathrm{nm}$ for the staggered dimers, or a redshift toward $\sim 700\mathrm{nm}$ for the stacked dimers.¹³ Extended dimerization causes a large decrease of the extinction coefficient of the Q-band with broadening of the band over a spectral range from $550\text{to}800\mathrm{nm}$ . The Soret absorption peak is also broadened and shifted toward a shorter wavelength.^{14, 15} Less spectral information is available concerning the higher order aggregates of phthalocyanines in solution.

Fig. 2

(a) Normalized absorbance spectra of bulk solution Al-2 ( $25\mu \mathrm{M}$ , solid line) versus confined L-Al-2 (dashed line). (b) Fluorescence emission spectra of bulk Al-2, and (c) liposome L-Al-2 as a function of PS concentration and liposome concentration (defined in Table 1), respectively.

The absorption spectrum [Fig. 2a] of Al-2 shows a Soret peak at $354\mathrm{nm}$ , a weak maximum at $605\mathrm{nm}$ , a strong Q-band at $672\mathrm{nm}$ , and the weaker dimer band at $642\mathrm{nm}$ .^{4, 13} Al-2 is considered less hydrophilic (i.e., less soluble in water) than Al-4, because it has two fewer sulfonic acid groups, as shown in Fig. 1a. As such, a smaller concentration of Al-2 $\left(25\mu \mathrm{M}\right)$ was prepared with the aid of presonication, while complete solubility was achieved at $50\text{-}\mu \mathrm{M}$ Al-4. Al-2 is thus considered an amphiphilic PS exhibiting both hydrophobic and hydrophilic properties.⁸ The liposomal confinement of Al-2 produced a slight redshift in the absorption Soret and Q-bands, as shown in Table 2 .

Table 2

Summary of maximum peaks in absorbance and fluorescence spectra of bulk and liposomal-containing PS.

The excitation of Al-2 at $661\mathrm{nm}$ yielded a fluorescence emission peak at $684\mathrm{nm}$ . Note that the expected peaks above $710\mathrm{nm}$ with different PSs were not detected due to the use of a bandpass filter $\left(680\text{to}710\mathrm{nm}\right)$ . The intensity of the fluorescent peak at $684\mathrm{nm}$ increased with Al-2 concentration in bulk solution up to $10\mu \mathrm{M}$ , after which the intensity decreased concomitant with a redshift in the maximum peak emission to 687 and $691\mathrm{nm}$ for 25 and $50\mu \mathrm{M}$ , respectively [Fig. 2b]. The same trend was observed with increasing concentrations of liposomes containing Al-2 [Fig. 2c]. At the smallest concentration of liposomes (F), there was negligible fluorescence. The increase of L-Al-2 concentration caused an increase of peak fluorescence intensity until concentration B was reached, after which fluorescence saturation occurred as observed by decrease in the fluorescence signal concomitant with redshifting of concentration A. At concentration A, the fluorescence intensity decrease was similar to what was observed in bulk solution, with a small associated redshift $\left(685\text{to}688\mathrm{nm}\right)$ . It should be noted here that for all experiments, each liposome contained approximately the same amount of PS; it was the increase of liposome concentration that caused the increase in the effective concentration of Al-2 in the sample. All PS concentrations introduced to the lipid cake were kept high (and approximately identical at $\sim 50\mu \mathrm{M}$ , within solubility limit), to allow for comparison at maximum drug incorporation for each L-PS. However, the absolute encapsulation efficiency of the PS into liposomes was unknown. The confinement of Al-2 provided a unique lipophilic/aqueous interface facilitating greater solubility of Al-2 molecules.

3.4.

Liposomal Confinement of Hydrophilic Photosensitizers

The absorption spectrum of tetrasulfonated phthalocyanine, Al-4, also presents a Soret $\left(350\mathrm{nm}\right)$ and Q-bands (674 and $606\mathrm{nm}$ ), as shown in Fig. 3a . Containing four sulfonic acid substituents, Al-4 is the most hydrophilic of the three PSs. Unlike Al-2, Al-4 did not exhibit an aggregate absorption band at $\sim 640\mathrm{nm}$ . It was expected that an increase of the number of sulfonate moieties in the phthalocyanine should produce more intensive repulsion between rings, and therefore, minimize aggregation.¹³ As evidenced by similar curves obtained for AL-4 and L-Al-4 [Fig. 3a], the absorption properties of Al-4 were not affected by the confinement in liposomes.

Fig. 3

(a) Normalized absorbance spectra of bulk solution Al-4 ( $50\mu \mathrm{M}$ , solid line) versus confined L-Al-4 (dashed line). (b) Fluorescence emission spectra of bulk Al-4, and (c) liposome L-Al-4 as a function of PS concentration and liposome concentration, respectively.

Excitation of free solution Al-4 in the low concentration range $\left(1\text{to}10\mu \mathrm{M}\right)$ resulted in a strong fluorescence emission peak centered at $685\mathrm{nm}$ . In the higher concentration regime, a pronounced redshift from $691\text{to}696\mathrm{nm}$ was observed when incrementing the concentration from $25\text{to}50\mu \mathrm{M}$ , accompanied by a significant decrease in the relative fluorescence intensity of 50%. While both free solutions of Al-2 and Al-4 were redshifted at higher concentrations, the magnitude of this effect was larger in the case of hydrophilic Al-4. In the case of liposome-confined Al-4, the primary fluorescence emission peak at $685\mathrm{nm}$ was not affected by changes in liposome concentration, as seen across the range shown in Fig. 3c. The fluorescence intensity increased with concentration up to concentration B, but decreased at the maximum L-Al-4 concentration A, which was associated with a $1\text{-}\mathrm{nm}$ redshift.

3.5.

Liposomal Confinement of Lipophilic Photosensitizer

The third PS, HPPH, is a new drug that is currently undergoing clinical testing. Information on this PS is limited, particularly in the context of liposomal confinement, but there is some spectral information available reporting two main absorption peaks (408 and $665\mathrm{nm}$ ) and fluorescence emission located approximately at $675\mathrm{nm}$ with shoulder above $700\mathrm{nm}$ .^{43, 44} The two main absorption peaks for the bulk solution were observed at 398 and $667\mathrm{nm}$ , with the presence of small peaks located approximately at 508, 540, and $607\mathrm{nm}$ . Those small peaks, with the one located at $667\mathrm{nm}$ , are called Q-bands and were also present on the absorbance spectra in many other studies.^{43, 45, 46, 47, 48} Even though those peaks are reported in the literature, little information is available on the link between the chemical structure of HPPH and the presence of those peaks, or the impact of aggregate states on the molar extinction and wavelength dependence of these Q-bands.

HPPH has been described as lipophilic, as its carboxylic groups are ionized at physiological pH. This PS has been found to be localized in intracellular membranes, namely in endoplasmic reticulum, Golgi apparatus, lysosomes, and mitochondria.^{47, 49} The liposomal confinement of this lipophilic PS induced a change in the shape of the Soret absorption peak, which becomes narrower and better defined, shifting the maximum from $398\text{to}421\mathrm{nm}$ and decreasing the intensity of the far-red band. The inclusion of HPPH in the phospholipids/CHOL membrane also modifies the distribution of the Q-bands. The hydrophobic tail of the phospholipid provides a microenvironment that favors solubilization of the PS in the membrane of the liposome and thus reduces the aggregation.

In a bulk solution of PBS, HPPH has a very low fluorescence quantum yield. To achieve higher fluorescence intensities, different solvents were used to solubilize this PS (Fig. 4 ). Addition of FBS produced an increase in the fluorescence emission of HPPH. Specifically, changing the FBS fraction from 5 to 20% in the PBS solvent produced a significant increase, greater than two fold, in the fluorescence intensity at identical concentrations. A significant gain in intensity was obtained when HPPH was solubilized in 20:80 FBS/PBS, and this optimized solvent mix was used in all subsequent experiments. The absorbance contribution of FBS/PBS was negligible, as shown in Fig. 5a , and no fluorescence was emitted from this solvent when excited at $661\mathrm{nm}$ . The intensity of the HPPH fluorescence increased with increasing concentrations up to $25\mu \mathrm{M}$ [Fig. 5b], with a prominent emission peak at $685\mathrm{nm}$ and a shoulder around $702\mathrm{nm}$ . At concentrations above $25\mu \mathrm{M}$ , a redshifting occurs in which the relative contribution of the shoulder increases. At the highest concentration of $100\mu \mathrm{M}$ , one broad peak is identified concomitant with redshift to $695\mathrm{nm}$ . When the PS was confined in liposomes, the fluorescence emission signal of L-HPPH was found to increase with increasing concentration of liposomes from F to A. The shoulder observed in the bulk solution was also present in L-HPPH. In contrast to the free solution of HPPH, and unlike the other PS molecules investigated, no decrease in the intensity or redshifting was observed through the entire concentration series of L-HPPH.

Fig. 4

Fluorescence emission spectra of bulk solution HPPH $\left(100\mu \mathrm{M}\right)$ as a function of relative FBS content in PBS solvent.

Fig. 5

(a) Normalized absorbance spectra of bulk solution HPPH ( $100\mu \mathrm{M}$ , solid line) versus confined L-HPPH (dashed line) and background FBS solvent (dotted line). Fluorescence emission spectra of (b) bulk Al-4, and (c) liposome L-Al-4 as a function of PS concentration and liposome concentration, respectively.

4.1.

Photosensitizer Behavior in Bulk Solution

We compared the absorbance spectra of bulk PS solutions (identical to those used in the L-PS formulation) to that of the highest L-PS concentration prepared (concentration set A) [Figs. 2a, 3a, 5a]. Comparison of the absorbance spectra provides insight on the aggregation behavior of PSs in liposomes and bulk solution. From Fig. 3a, it is clear that even at high concentrations $\left(50\mu \mathrm{M}\right)$ , bulk Al-4 presents no aggregation products that absorb in the visible range, also observed by Juzenas ⁴² Liposomal encapsulation to L-Al-4 did not change the absorbance features. However, the absorbance spectrum of Al-2 (at the maximum concentration of $25\mu \mathrm{M}$ ) exhibited some aggregation characteristics. Specifically, the Soret peak and Q-band were redshifted from $349\text{to}353\mathrm{nm}$ and from $669\text{to}672\mathrm{nm}$ , respectively, and the overall absorbance was higher in the red region of the spectrum for the bulk Al-2 compared to the L-Al-2.

In the case of the hydrophobic porphyrin molecule HPPH, two main absorption peaks at 408 and $665\mathrm{nm}$ , with three small peaks located approximately at 508, 540, and $607\mathrm{nm}$ , were observed, as reported previously.^{43, 44} In bulk solution, the photophysical properties of porphyrins with negatively charged peripheral substituents change due to the aggregation process.⁵⁰ Dimers and higher aggregates modify the absorption and fluorescence spectra: dimers cause shifting toward shorter wavelength and an increase in the width of the Soret band at full width at half-maximum (FWHM), with redshifting of the Q-band,^{30, 45} and higher order aggregates typically shift the Soret peak to $360\mathrm{nm}$ and cause a large redshift of the principle Q-band.⁴⁶ The spectra of nonaggregated porphyrins typically show a linear decrease in the intensity of the Q-bands with increasing wavelength, a behavior known as etio-type spectrum.²⁵

Liposome formulations with highly concentrated solutions (i.e., where PSs are in dimer conformation) displace the equilibria toward monomerization and lipid localization.^{27, 51} Redshifts in the Soret and Q-bands were also observed with liposomal confinement but were typically small. Kuciauskas ³³ reported that changing the solvent of the porphyrin [5,10,15,20 - tetrakis - (4-hydroxyphenyl) -21,23H-porphyrin] from acetone to a liposome without cholesterol caused a redshift of the Soret and Q-bands of $1\text{to}4\mathrm{nm}$ . Encapsulation of the PS in the ordered lipid bilayer can also change the symmetry of the PS molecule and simplify its absorption spectrum into a two-band system. Changes in symmetry due to membrane incorporation, and its consequences on absorption, are analogous in mechanism to the effect of peripheral substituents on the inductive polarization of the chromophore.⁴⁵ Hence, the structure of a porphyrin, the local microenvironment, including the solvent used and the dye concentration, will determine the contribution from monomers, dimers, and higher order aggregates to the net spectroscopic properties. It has been reported that even in liposomes, where the equilibrium favors monomerization, some contributions from aggregate species are always present.⁵⁰

Based on the observation of the Soret peak position and broad FWHM of the HPPH absorbance spectrum [Fig. 5a], we can state that aggregates are present in the highly concentrated bulk solution. Also, the Q-bands do not appear as an etio-type spectrum. In fact, the three low absorbance Q-bands (508, 540, and $607\mathrm{nm}$ ) had approximately the same intensity, while the red band $\left(665\mathrm{nm}\right)$ had a relatively higher intensity, similar to previous results obtained with $2.5\mu \mathrm{M}$ of three different chlorine derivatives in tetrahydrofuran (THF).²⁴ Following liposomal confinement, the Soret band was redshifted and the FWHM was narrower, suggestive of monomerization of the HPPH. The Q-bands were simplified to a two band system, a consequence of PS incorporation into a highly ordered bilayer of phospholipids. The absorption spectra obtained in liposomes were similar to those obtained by Kuciauskas ³³ for other hydrophobic porphyrins in acetone solvent and in liposomes. In a manner analogous to the effect of acetone, the polarity of the solvent influenced the effective solubilization of HPPH in a phospholipid environment. It has been suggested that the hydrophobic core structure of porphyrins can be effectively embedded deep in the aliphatic chain of the phospholipids, while its carboxylic substituent(s) may either interact with the polar head of the phospholipid or with the cholesterol situated between the phospholipids.^{27, 33, 51}

From the absorbance spectra, it was observed that at high concentrations, HPPH and Al-2 tend to form dimers or higher order aggregates, as opposed to hydrophilic Al-4, which is highly soluble and not aggregated, even at high concentrations. The fluorescence spectra of the three PSs were acquired over a broad concentration range for both the free solution and the corresponding liposome-confined analogs. For all the bulk solutions, results show that the fluorescence intensity increased with concentration until a maximum was reached, after which the intensity decreased concomitant with a redshift in the emission peak maximum. In the case of phthalocyanines, there is a debate in the literature about the decrease of fluorescence intensity at high PS concentration and the redshift associated with this phenomenon. The decrease in emission intensity is often attributed to the formation of the commonly encountered cofacial phthalocyanine dimers that are reported to be nonfluorescent and photodynamically inactive.^{14, 45, 52} In their studies, redshifting of fluorescence peaks has been identified and associated with dimer emission.^{53, 54} However, among the nonfluorescent dimers, some fluorescent species have been observed but emission is restricted to particular species.^{55, 56} At room temperature and $\mathrm{pH}=7.2$ , the decreases in intensity observed for Al-2 and HPPH may in part be explained by the presence of dimers or higher order aggregates that are present at high concentrations and that do not fluoresce. This is supported by the absorbance spectra obtained for Al-2 and HPPH at high concentrations [25 and $100\mu \mathrm{M}$ , respectively, Figs. 2a and 5a]. However, the contribution of aggregate quenching to the fluorescence signal may not entirely account for the total decrease in the fluorescence intensity observed and the concomitant redshift, especially at high concentrations. For example, the contribution of dimers or higher order aggregates was not evident from the absorbance spectrum of Al-4 at $50\mu \mathrm{M}$ (Fig. 3), although a reduction in fluorescence intensity and redshift were observed with increasing PS concentration. Hence, the decrease in fluorescence intensity with increasing concentration may in part be associated with self-absorption, a common optical phenomenon known to occur particularly at higher fluorophore concentrations.^{15, 57, 58}

Self-absorption occurs when a portion of the fluorescence emission is reabsorbed by neighboring fluorophores due to the overlap of the absorption and emission bands of the fluorophore. This in itself has the effect of shifting the apparent peak emission of the observed fluorescence. Moreover, a fraction of these reabsorbed photons are re-emitted, thus increasing further the redshift in the emission peak wavelength. This phenomenon is typically observed at high fluorophore concentrations and/or when the “effective” light path length is large, as the result of scattering, for example.⁵⁹ The total fraction of photons that will be reabsorbed and re-emitted depends on several geometric and chemical factors, including the concentration. Increasing the fluorophore concentration improves the probability that photons are reabsorbed and re-emitted, with the redshift becoming more apparent, observed for all free PSs in aqueous solution over the concentration range studied [Figs. 2b, 3b, 5b]. The concentration at which the intensity starts to decrease and where the redshift becomes apparent is PS dependent and related to their respective Stokes shift and fluorescence quantum yield. The observed redshift, although similar for Al-2 and HPPH $(\sim 8\mathrm{nm})$ , was found to be higher for Al-4 $(\sim 12\mathrm{nm})$ . Some Al-2 and HPPH molecules are present as dimers or higher order aggregates when the concentration of PS in solution is high. This presence of nonfluorescent conformations decreases the effective concentration of fluorescent species that can re-emit redshifted photons. This may explain the smaller degree of redshift with Al-2 and HPPH and explains the large redshift observed for Al-4, which is negligibly aggregated.

4.2.

Photosensitizer Fluorescence in Liposomes

The fluorescence study of liposome-confined L-Al-2 showed a similar trend of redshifting at the highest liposome concentrations; increasing the liposome concentration from F to A had the same results on the measured fluorescence as increasing the bulk concentration. A redshift and decrease in fluorescence intensity were observed at the highest concentrations B and A. While the uptake efficiency of Al-2 in liposomes was not known, the spatial localization of the PS, shown in the epifluorescence image, and the amphiphilic properties of Al-2 suggested that it resides at the membrane interface (on the external and/or internal bilayer surface). Interfacial alignment of amphiphilic cis Al-2 molecules on similar membrane surfaces has been reported previously.^{13, 60} The trends observed in bulk Al-2 and L-Al-2 as a function of concentration were similar. This is because the liposomal localization of the amphiphilic Al-2 molecules did not result in encapsulation as in the case of L-Al-4, which has its local concentration of fluorescing species maintained within the core of the liposome. The localization of Al-2 on the surface of the liposome membrane means that the PS can interact with other surface-bound PS molecules. Consequently, as the liposome concentration is increased, we observe the corresponding red-shift that mirrors the fluorescence behavior observed in free solution Al-2.

The liposomal membrane is constituted of DOPC, an unsaturated phospholipid, and a significant quantity of cholesterol $(35\mathrm{mol}\%)$ to mimic cellular membrane chemistry. The lipid molecules are highly mobile in membranes. For example, the unsaturated fatty acids of membrane lipids rotate easily. Cholesterol is a key constituent in the cell membrane having a critical role on membrane organization, function, and sorting.⁶¹ It is also known to improve the tendency for stable vesicles to form by facilitating hydrophobic interactions in the bilayer.⁶² Lipids are known to undergo lateral diffusion, an effect that generally decreases with increasing cholesterol content. DOPC phospholipids incorporating cholesterol are relatively disordered, imparting more motion and fluidity into the membranes via transmembrane “flip-flop” movement.^{63, 64} Given that cholesterol has pronounced effects on the structural properties of membranes, PS molecules localized within or near the surface of the bilayer display high sensitivity to membrane dynamics. As such, it is not surprising that interfacial Al-2 molecules, located on the external interface of the liposomal membranes, are highly mobile, allowing intra- and interliposomal aggregation, consequently producing the reabsorption and re-emission effects observed here. Although the concentration of liposome is relatively low, interliposomal aggregation is seen with L-Al-2 [Fig. 1a] which facilitates interaction of interfacial Al-2 molecules. Conversely, no interliposome aggregation is present in the case of L-Al-4, as evidenced by Fig. 1b.

For L-Al-4, an increase in fluorescence intensity with liposome concentration was not accompanied by a redshift. Attributed to its hydrophilicity, the bulk absorbance spectrum of Al-4 was without aggregation bands, and the epifluorescence image showed the PS to be confined primarily to the core volume of the liposome. Previous reports suggested that due to its tetrasulfonate substitution with strong acids, Al-4 is highly soluble and stable in water, and does not form aggregates as observed in the visible spectral range.⁴² Hence, confinement into the aqueous core of the liposomes was not expected to cause a spectral shift in the fluorescence emission pattern. The liposomes prepared in this experiment had a mean diameter of $130\mathrm{nm}$ , which resulted in a liposome volume of $9.2\times {10}^{-21}{\mathrm{m}}^3$ . Considering that the highest liposome concentration used was $2.7\times {10}^7{\mathrm{mL}}^{-1}$ , we estimate that the PS-containing liposome volume fraction in solution was $2.5\times {10}^{-7}$ . Assuming that the liposome solution was a homogeneous colloidal suspension, we can estimate that the mean distance between liposomes in solution was approximately $33\mu \mathrm{m}$ . Since Al-4 molecules were encapsulated inside the liposome core, the molecules were restricted to interact intraliposome only. The small liposome volumic fraction of PS and the larger mean inter-liposome distance implies that there was a low chance that emitted photons from L-Al-4 in one liposome would be reabsorbed by L-Al-4 molecules residing in another liposome. Reabsorption and re-emission depends not only on the concentration of the fluorophores in solution, but also on the average optical pathlength of the fluorescence inside the solution itself. Based on the calculated average optical pathlength, reabsorption effects for core-confined L-Al-4 were believed to be negligible.^{58, 59} Therefore, the confinement of hydrophilic Al-4 PS inside the core volume of the liposome created a homogenous and protected microenvironment over the large range of liposome concentrations studied here. Hence, reabsorption and re-emission phenomena of PS molecules via interliposome interactions were not significant, as compared to those in bulk aqueous solution. Consequently, increasing the liposome concentration over 2 orders of magnitude also produced fluorescence redshifting like that observed in bulk solution with increasing concentration.

The increase of the L-HPPH concentration resulted in increased fluorescence intensity for all concentrations, with no redshifting or aggregation observed. Although the uptake efficiency of liposomes was unknown for HPPH, the lipophilic nature of the PS suggests that the molecule was driven to the inside of the liposomal membrane, where it is stabilized by the nonpolar domain of the phospholipid. The fact that no spectral modifications were observed at the highest concentration of L-HPPH can be explained by both optical and chemical arguments. Despite the relatively high concentrations of HPPH used for liposomes, no detection of re-emission due to reabsorption was detected. Also, the average optical path length of PS embedded in the liposome membrane was small enough to assess that the reabsorption of fluorescence was insignificant.^{58, 59} Chemical considerations of hydrophobic interaction initiated from membrane incorporation of HPPH suggest that the phospholipid bilayer of the liposome provided an environment that favored monomeric HPPH. Furthermore, the low liposomal volumic fraction indicated that the mean interliposomal distance was large compared to their size, minimizing interliposome HPPH interactions. The liposomal confinement of HPPH created a lipophilic microenvironment where the number of HPPH in close proximity was low, protected by the membrane from any interaction with other PS molecules despite aggregation of liposomes, as observed in Fig. 1c.

4.3.

Implication of Photosensitizer Localization

The localization of PS is important for efficient PDT. One reason is that singlet oxygen diffuses rapidly out of membranes, and thus efficient reactions with its target are necessary for phototoxicity generated in lipid membranes.³² The microenvironment of PSs in cells also affects aggregation which has a deleterious effect on photodynamic activity. Hence, PS drugs need to be introduced into the bloodstream in a form that is either monomeric, or easily dissociated by the plasma. For IV administration, hydrophobic PSs require formulations that minimize aggregation in vivo, like liposomal vehicles that favor the monomeric form of the PS. Amphiphilic PSs have been reported as being excellent candidates for liposomal formulation.⁵⁹ Studies are currently under way in our laboratory to corroborate the results observed in our model system with PS localization effects in cells.