2.1.

Functional Building Blocks and Polymer Synthesis

For the photocleavable linker system, which attaches the therapeutic drug 5-fluorouracil (5FU) to the polymer backbone, a coumarin molecule is chosen (Fig. 3 ). The acrylic polymer serves two functions; first it is the IOL material, and second it is the drug reservoir. For this purpose, a copolymer (PC) from a nonmodified monomer $\left(P_{\mathrm{mono}}\right)$ and a monomer with a coumarin side group $\left(C_{\text{link}}\right)$ is used. The drug employed is 5FU $\left(D_{\text{drug}}\right)$ , which is effective for secondary cataract treatment, but has a very low solubility and diffusion constant in the used PC polymer. To overcome this problem, a side group $\left(D_{\mathrm{diff}}\right)$ is attached that improves both parameters. The formed linkage in the formed drug-diffusion modulator complex (DD) is hydrolysable, and upon contact with chamber water the 5FU is released in its original form. In a photochemical reaction, the DD complex is attached through a $[2+2]$ cycloaddition reaction to the coumarin side group, and a polymer with photocontrolled drug delivery (PCDD) capabilities is obtained.

Fig. 3

Functional building blocks for photocontrolled drug delivery intraocular lenses (PCDD-IOL). The polymer backbone is a copolymer (PC) obtained from a polymer monomer $\left(P_{\mathrm{mono}}\right)$ and a polymer monomer with a photosensitive linker molecule attached $\left(C_{\text{link}}\right)$ at a molar ratio of about 9:1. The drug complex (DD) comprises the drug molecule $\left(D_{\text{drug}}\right)$ and an optional diffusion modulator molecule $\left(D_{\mathrm{diff}}\right)$ , which is attached through an easily hydrolysable linkage. $D_{\mathrm{diff}}$ is required, as the solubility and diffusion constant of the nonmodified drug in the polymer is often very poor. The photocontrolled drug delivery polymer (PCDD) is finally obtained by coupling the drug complex (DD) to the copolymer (PC) by means of a photocleavable linkage.

The photoreversibility of cyclobutane systems is well known. ^{53, 54, 55, 56, 57, 58} UV photons as a trigger cannot be employed here due to the strong UV absorption of the cornea, which limits the penetration depth of the light. We employ a TPA-induced process that enables a cleavage of the UV-active linkage behind the cornea as a photochemical barrier. The required intensities are achieved by focusing a pulsed laser system to a spot inside the PCDD material. Since the probability for a TPA-dependent reaction is relatively low, the process requires considerably high photon densities. This is because the device is not affected by ordinary light.^{59, 60, 61, 62} Compared with conventional sustained-release intraocular drug delivery devices based on biodegradable polymers, this novel photomodulated delivery device offers temporal and spatial control over the release of the drug. The required or desired drug levels can be easily controlled by dosing the irradiation energy.

The material properties should be equivalent or superior to polymethylmethacrylate (PMMA), which is currently one of the standard materials for IOL manufacturing. As a first example, we have chosen n-butylmethacrylate (NBMA) as a polymer backbone, which provides the unique mechanical and optical properties of the acrylic polymers. The fabrication of the device consists of three different steps, as sketched in Fig. 2. The copolymer matrix bearing photoreactive moieties $\left({\mathrm{C}}_{\text{link}}\right)$ was obtained by copolymerization of NBMA $\left(P_{\mathrm{mono}}\right)$ and monomeric methylmethacrylate modified with coumarin $\left(C_{\text{link}}\right)$ . The coumarin side group acts as a molecular linker due to its capability for reversible $[2+2]$ photocyclizations/photoreversions in dependence on the wavelengths of light used.^{63, 64, 65} We have shown that coumarin photodimers used in these experiments can be cleaved by two-photon absorption.⁶⁶ The model drug, 5-fluorouracil (5FU), was modified with heptanoic acid $\left(D_{\mathrm{diff}}\right)$ resulting in the drug precursor (DD), 1-heptanoyl-5-fluorouracil (H5FU). While possibilities of direct photochemical crossdimerization of coumarin derivatives and 5FU were reported in several studies,^{67, 68, 69} the utilization of alkylated 5FU provides some advantages such as improved solubility, miscibility, and diffusion properties in organic solvents and hydrophobic polymers, respectively. The diffusion modulator $\left(D_{\mathrm{diff}}\right)$ is attached to the drug by a hydrolysis-labile ester bond that is cleaved on exposure to water at the target site. Finally, the polymer-drug conjugate (PCDD) was obtained using photochemical $[2+2]$ cycloaddion reaction. Analysis by UV/VIS spectroscopy shows that the typical absorption band of coumarin at $314\mathrm{nm}$ corresponding to the double bond besides the carbonyl group,⁷⁰ is missing in the absorption spectrum of the obtained polymer-drug conjugate. This indicates a complete cyclodimerization of all coumarin moieties to $[2+2]$ -photoproducts. In the spectrum, no absorbance in the visible region is observed, which guarantees the excellent transparency of the polymer for its use in IOLs. Because the temperature at which 5% of the polymer decomposed $(244°\mathrm{C})$ is much higher than the glass-transition temperature $T_g=83°\mathrm{C}$ , it is possible to fabricate IOLs via compression molding. The high value of $T_g$ guarantees that the IOL is stable after implantation. With the data from elemental analysis, we can calculate a drug loading in the polymer of about 6% (wt/wt).

2.2.

Phototriggered Drug Release

Both single-photon absorption (SPA) and TPA-induced processes were investigated. Polymer films for testing were fabricated from a blend of the PCDD polymer-drug conjugate and PMMA (for SPA 3:1 wt/wt and for TPA 1:1 wt/wt, respectively) by the solvent casting technique. During the irradiation of the obtained films with $254\text{-}\mathrm{nm}$ UV light, the absorption band with a maximum at $312\mathrm{nm}$ , which corresponds to the coumarin moieties, increased [Fig. 4a ], which indicates the cleavage of the cyclobutane ring and the related drug release. TPA-induced photocleavage of the PCDD material by intense $532\text{-}\mathrm{nm}$ pulses is observed at much higher energies. The spectral changes during the course of irradiation with an average pulse intensity of $43.5\mathrm{mJ}$ /pulse are shown in Fig. 4b. The observed spectral changes of SPA- and TPA-induced processes are similar.

Fig. 4

Single-photon and two-photon absorption-dependent drug release from PCDD-IOL material. Difference absorption spectra of a PBMAOCH5FU film during photocontrolled drug release. (a) Single-photon absorption: UV light $(\lambda =254\mathrm{nm})$ is used to trigger release. (b) Two-photon absorption: a frequency-doubled Nd:YAG-laser ( $\lambda =532\mathrm{nm}$ , $43.5\mathrm{mJ}$ /pulse, $20\mathrm{Hz}$ ) is employed. (c) The initial photocleavage rate is plotted as a function of the incident intensity at $532\mathrm{nm}$ . A slope of about 2 is derived, which corresponds to a two-photon reaction.

The dependence of the initial photocleavage rate on the incident intensity was determined to confirm the TPA nature of the process induced by the $532\text{-}\mathrm{nm}$ pulses. The initial rates of photocleavage in the model IOL were derived from the changes in the absorption at $312\mathrm{nm}$ , and the incident intensities range from $14\text{to}43.5\mathrm{mJ}$ /pulse at a repetition rate of $20\mathrm{Hz}$ were employed. The experimentally obtained slope of 1.93 in the log-log plot, shown in Fig. 4c, corresponds nicely to the theoretically expected value of 2,^{59, 61, 62} and indicates that the cleavage of the PCDD conjugate is solely induced by TPA.

We analyzed the identity and quantity of the delivered H5FU by HPLC analysis (Fig. 5 ). From solutions of PCDD in chloroform, samples were analyzed before and after irradiation with the energies given. In both cases, the obtained HPLC profiles are very similar and show only two products, which are H5FU and its hydrolysis product 5FU, released during photocleavage without any side products. The total amount of delivered H5FU and 5FU, respectively, correlates exactly with the absorption changes at $312\mathrm{nm}$ measured by UV/VIS spectroscopy.

Fig. 5

HPLC analysis of phototriggered released compounds. (a) PBMAOCH5FU solution on irradiation with UV light $(\lambda =254\mathrm{nm})$ . (b) PBMAOCH5FU solution on irradiation using frequency-doubled Nd:YAG-laser ( $\lambda =532\mathrm{nm}$ , $67\mathrm{mJ}$ /pulse, $20\mathrm{Hz}$ ). The solvent peak (chloroform) is marked.

2.3.

Multidose Drug Delivery

Secondary cataract is observed many months even years after IOL implantation. Even after treatment, secondary cataract may reappear. Due to this, a multidose capability of the PCDD IOL is desired. To demonstrate an external triggered repeated drug release, the model IOL was irradiated in a stepwise manner, and the release pattern of 5FU was monitored using UV/VIS spectroscopy. The result depicted in Fig. 6 shows the cumulative amount of the discharged 5FU after three consecutive steps of irradiation with a constant energy dose of $112\mathrm{mJ}$ . During each irradiation step, approximately $2\mu \mathrm{g}$ of 5FU were released from approximately $30\text{-}\mathrm{mg}$ PCDD material. The amount of released drug was strictly proportional to the applied energy dose. This was also confirmed by HPLC analysis. In view of the low volume inside the capsular back, the LD50 of 5-fluorouracil, which was reported for rabbit lens epithelial cells (RLEC) to be $0.58\mu \mathrm{g}∕\mathrm{ml}$ ,⁷¹ can easily be obtained.

Fig. 6

Multidose drug delivery triggered by repeated irradiation. The cumulative amount of the 5FU released was determined from the absorbance changes observed at $267\mathrm{nm}$ . At the points A, B, and C, the film was irradiated with $254\text{-}\mathrm{nm}$ light for $10\min$ . The insert shows corresponding absorbance changes.

2.4.

In Vitro Test

In in-vitro cell tests, IOLs were tested to show that the drug-loaded material itself is nontoxic to the cells, but as soon as it is photochemically activated, the drug release is triggered and the cell count is reduced significantly. PCDD polymer disks of $8\mathrm{mm}$ radii with a thickness of $0.5\mathrm{mm}$ carrying approximately $800\mu \mathrm{g}$ of 5FU each were used for the tests. Such samples were incubated with pancreatic carcinoid cells (line Bon 1) and the proliferation of the cells was analyzed (Fig. 7 ).

Fig. 7

Cell tests. Proliferation of (a) control cells and (b) cells treated with the activated IOL, both after $7\text{days}$ in culture. The treated cells show a significantly lower cell density.

The data obtained clearly indicate that IOLs that were not photochemically activated have no significant influence on the cell proliferation. However, photoactivated IOLs cause a significant reduction of 29% of the cell mass compared to an untreated control group (Table 2 ). It should be mentioned that the in-vitro tests are for screening purposes, as in-vivo lens epithelial cells may show a different sensitivity toward 5FU.

Table 2

Cell test: The protein content after 7days was analyzed with the RC-DC Protein Assay. The statistic correlation performed with a Turkey multicomparison test shows a significant difference between nonactived and activated PCDD-IOLs (p less than 0.001).

2.5.

Summary and Conclusion

We have developed a photocontrolled multidose drug delivery device that may be applied in the form and having the function of an IOL. This combination we named PCDD-IOL. The drug delivery is photochemically triggered by TPA, which enables ambulant treatment of readily available laser sources. The multidose capability of the PCDD-IOLs was demonstrated in vitro. Cell tests confirmed that the PCDD-IOL material itself shows not cytotoxicity until photochemical stimulation. Such materials will be tested in rabbits in the near future. The presented materials and concept are a new improved route in secondary cataract treatment.

3.1.

Syntheses

3.2.

Preparation of Polymer Films

Polymers or polymer blends were dissolved in chloroform to the highest concentration possible by stirring overnight. The solutions were filtered through $0.45\text{-}\mu \mathrm{m}$ Teflon filters (P819.1, Roth, Karlsruhe, Germany). A few milliliters of the filtered chloroform solution were pipetted on a glass plate. By means of a coating knife, a wet film of 0.4 or $0.8\mathrm{mm}$ thickness was obtained ( $=\text{solvent}$ casting). After solvent evaporation (typically for $20\mathrm{h}$ ), a polymer film was obtained. Samples were prepared by either punching out or cutting out the desired pieces from the prepared polymer films.

3.3.

Phototriggered Drug Release: Single-Photon Absorption

Single-photon absorption (SPA)-induced drug release was analyzed in solutions. Samples were prepared by dissolving $32.8\mathrm{mg}$ of PBMAOCH5FU in $4.5\text{-}\mathrm{ml}$ $\mathrm{C}\mathrm{H}{\mathrm{Cl}}_3$ . The solution was through a $0.45\text{-}\mu \mathrm{m}$ Teflon filter to remove any scattering particles. Excitation light of $266\text{-}\mathrm{nm}$ wavelengths from a fluorescence spectrometer (RF-1502, Shimadzu, Duisberg, Germany) was used to induce SPA-dependent drug release in the test solutions directly in the spectrometer, where the resulting absorption spectra were recorded. The test samples were continuously stirred (Telemodul 20P and mini, $\mathrm{H}+\mathrm{P}$ Labortechnik, Oberschleißheim, Germany), but during recording of the absorption spectra, the stirring was switched off. The intensity of the $266\text{-}\mathrm{nm}$ light in the fluorescence spectrometer was $196\mu \mathrm{W}∕{\mathrm{cm}}^2$ .

Polymer films for SPA-induced drug release were made from a mixture of PBMAOCH5FU and PMMA $(3:1\mathrm{wt}∕\mathrm{wt})$ as described before. A wet thickness of $0.8\mathrm{mm}$ was used. Light of $254\text{-}\mathrm{nm}$ wavelength (MinUVIS, Desaga, Heidelburg, Germany, $187\mu \mathrm{W}∕{\mathrm{cm}}^2$ ) was used to induce the drug release by light-induced cycloreversion of the coumarin linker system.

3.4.

Phototriggered Drug Release: Two-Photon Absorption

Solutions for two-photon absorption (TPA)-induced drug release tests were prepared by dissolving $51.7\text{-}\mathrm{mg}$ PBMAOCH5FU in $4\text{-}\mathrm{ml}$ $\mathrm{C}\mathrm{H}{\mathrm{Cl}}_3$ . The solutions were filtered through $0.45\text{-}\mu \mathrm{m}$ Teflon filters before use. An Infinity 40-100 mode-locked Nd:YAG laser (Coherent, Rödermark, Germany) emitting $3\text{-}\mathrm{ns}$ pulses at $532\mathrm{nm}$ at a repetition rate of $20\mathrm{Hz}$ was used to excite the samples. The pulse energy was $67\mathrm{mJ}$ and the beam diameter was $5.5\mathrm{mm}$ .

Polymer films for TPA-induced drug release were made from PBMAOCH5FU and PMMA (1:1 wt/wt) blends. A wet thickness of $0.4\mathrm{mm}$ was used. To induce the cycloreversion of the coumarin system via two photon absorption, the Infinity 40-100 system described earlier was used. Pulse energies ranging from $14\mathrm{mJ}$ /pulse to $67\mathrm{mJ}$ /pulse were used.

3.5.

Quantitative Analysis of Released Drug

HPLC analysis of the released drug was done on a Hewlett-Packard Model 1050 system equipped with a diode array detector. The $260\text{-}\mathrm{nm}$ trace was used for quantitative determinations. A reversed phase column (Nucleosil, $3\mu \mathrm{m}$ , RP18, $250\times 4\mathrm{mm}$ , Bischoff, Leonberg, Germany) equipped with a precolumn ( $20\times 4\mathrm{mm}$ , Bischoff) was employed for trapping the polymer content. A water / acetonitrile gradient was used.

3.6.

Multistep Drug Release

A polymer film comprising PBMAOCH5FU and PMMA at a ratio of 1:3 (wt/wt) was prepared from a chloroform solution as described earlier. A suitably cut piece of the polymer film was mounted on one side of the inner walls of a fluorescence cuvette (101 QS, Hellma, Müllheim/Baden). The cuvette was filled with $3\mathrm{ml}$ of water. The film was irradiated with $254\text{-}\mathrm{nm}$ light from a MinUVIS (Desaga, $187\mu \mathrm{W}∕{\mathrm{cm}}^2$ ) until the desired energies were reached. After light-dependent release of 5FU from the polymer matrix, the drug diffuses from the polymer film into the aqueous solution. The UV/VIS spectra of the aqueous solution were recorded in an UVIKON 922 (Kontron, Munich, Germany) spectrophotometer. Accompanying HPLC analyses of small samples taken from the cuvettes were done as described before.

3.7.

Cell Test

Polymer films for cell tests were prepared from 1:3 (wt/wt) mixtures of PBMAOCH5FU and PMMA dissolved in chloroform. A $0.8\text{-}\mathrm{mm}$ coating knife was used to prepare the films as described earlier. Test disks having $16\mathrm{mm}$ diam and a thickness of $0.5\mathrm{mm}$ were punched out of the polymer foil and used for the cell tests. The drug release was triggered by light with $254\text{-}\mathrm{nm}$ wavelength having a total energy of $0.504\mathrm{J}∕{\mathrm{cm}}^2$ .

2000 Bon 1 cells in $1.2\text{-}\mathrm{mL}$ Dulbecco's modified Eagle's medium and Ham's F-12 medium containing 10% fetal calf serum (PAA, Cölbe, Germany) and 1% penicillin/streptomycin were seeded in 24 well plates (Greiner Bio-One, Frickenhausen, Germany). The polymer disks were added after $4\mathrm{h}$ in a way that any contact of the polymer disks with the Bon cells was omitted. No polymer disks were added to the control samples. The cell cultures were incubated at $37°\mathrm{C}$ in a saturated ${\mathrm{H}}_2\mathrm{O}$ atmosphere containing 5% $\mathrm{C}{\mathrm{O}}_2$ (HS incubation cabin, Heraeus, Hanau, Germany). The cells were observed for $7\text{days}$ . Then the cells were washed three times with phosphate buffered saline (PBS). Then, $1\text{-}\mathrm{mL}$ $20\text{-}\mathrm{mM}$ 3-(N-Morpholino) propanesulfonic acid containing 0.1% Triton-X-100 was added to each well to homogenize the cells. $50\mu \mathrm{L}$ of the suspension was analyzed with the RC-DC Protein Assay (Bio-Rad, Hercules, California) to determine the protein content following a modified Lowry method.⁷⁵ The protein content correlates to the total cell count. The statistic correlation was performed with a Turkey multicomparison test.