Photodynamic therapy has emerged as an efficacious adjuvant treatment modality for several types of cancer.In PDT, light is used to locally excite a photosensitizer to generate reactive oxygen species. The resulting oxidative stress disrupts organelle functions, promotes cell apoptosis, and damages the tumor vasculature that supply oxygen and nutrients required for the tumor to survive.While a few PDT therapies have received FDA approval , efficient delivery of the PS to the target site remains challenging. Tumor accumulation of the PS is generally poor due to the physicochemical properties of the PS.Therefore, large doses are administered to compensate for the poor drug accumulation at the target site. This is particularly unfavorable because most PS suffer from slow in vivo clearance, which increases toxicity. For example, as skin is highly vascularized and easily exposed to light, the long circulation time of PS promotes skin phototoxicity. As a result, patients are required to limit their exposure to the sun several weeks post-treatment. Therefore, there is a critical need to develop delivery systems with enhanced clearance that promote the accumulation of the PS in the tumor site. To this end, I turned toward the development of plant virus-based nanoparticles for the delivery of PS. VNPs have been developed as carriers for the delivery of contrast agents, chemotherapeutics, protein therapies, epitopes, agro-pesticides, as well as PS . Plant VNPs have several attributes that are favorable for nanomedicine delivery and in particular PS delivery. Bio-manufacturing is well established and the biologic platform offers well-defined,hydroponic channel monodisperse structures that can be tailored with molecular precision.
Plant VNPs are non-infectious toward mammals, and most importantly the proteinaceous nanoparticles are cleared rapidly from circulation and from tissue,thus making this a particularly attractive platform for PS delivery. Plant VNPs as well as bacteriophage-derived nanoparticles have been developed for PS delivery;in most instances PS agents are covalently coupled to viral carriers. However, covalent binding of the PS to nanoparticles may impair their photoactivity due to quenching and reduced molecular freedom, and in turn limit their intracellular activity. Therefore, non-covalent drug delivery may be advantageous to enhance and control steady release of the PS within the tumor environment. This strategy relies on hydrophobic-hydrophilic and electron charge interactions between the PS and its carrier. In this work, I utilized two high aspect ratio, soft matter tubular nanostructures for PS delivery, namely tobacco mosaic virus and tobacco mild green mosaic virus . TMV and TMGMV were selected as carrier platforms based on their well-established surface chemistry and elongated shape. Elongated nanoparticles have enhanced blood margination, transport across tissue membrane, cell adherence, and macrophage avoidance, promoting their accumulation in the tumor tissue.TMV and TMGMV self assemble helically around a single-stranded RNA genome to form a 300 x 18 nm rod with a 4 nm wide hollow interior channel . As described in chapter II, both particles are made of 2,130 identical copies of coat protein units; TMV and TMGMV share 86% sequence homology.Of particular interest, the interior channels of TMV and TMGMV are covered with solvent exposed glutamic acids that are readily available for electrostatic loading of positively charged guest molecules .While TMV has been extensively studied for clinical applications, including the delivery of PS,this is the first study investigating TMGMV for medical applications. To probe drug loading and release, I studied the monocationic, dicationic, tricationic and tetracationic version of a zinc porphyrin photosensitizer. Lastly, we selected one formulation and developed a cancer cell targeting strategy to further enhance treatment efficacy.The TMV results can be attributed to the combined effect of electrostatic and hydrophobic/hydrophilic interactions; the greater the positive charge the better stabilization inside the TMV interior channel.
In addition, the increased hydrophobic nature of the monocationic and dicationic Zn-Por formulations in combination with their electrostatic properties led to the formation of more aggregates compared to their tricationic and tetracationic counterparts, thereby reducing the loading efficiency. Several factors may explain the differential loading results between TMGMV and TMV. In chapter II, I have previously compared the amino acids sequences of TMV and TMGMV and analyzed their distribution of charged residues on both the inner and outer surfaces of the virus.While it has been shown that only two glutamic acid residues are chemically available on TMV , our analysis revealed that in addition to the Glu 95 and Glu 106 in the interior channel of TMGMV, Glu 145 and aspartic acid 66 were also exposed on the outer surface and could be available for electrostatic charge interactions. The difference in the amino acid sequences of TMV and TMGMV could also play a role in the difference in loading by changing the charge and hydrophobicity surrounding the glutamic acid residues. Furthermore, the virus coat proteins are not rigid structures, and therefore small molecules could diffuse in between coat proteins.Based on the above studies, I prepared drug-loaded VNPs using the 2000:1 Zn-Por:VNP ratio. I studied whether changing the pH of the 10 mM KP buffer solution would influence the loading efficiency of Zn-Por into TMV and TMGMV . At pH 3, VNPs aggregated and disassembled, which led to lower yields. The corresponding loading efficiency was low due to the protonation of carboxylate groups, resulting in weak electrostatic interaction. At pH 5, the reaction yields and loading efficiency were improved compared to pH 3, and reached their maximum at pH 7.8. Increasing the pH to 10 did not increase the loading yield, but rather just slightly decreased loading efficiency and reaction yields. While ~60−75% of starting materials were recovered at pH 7.8, the yield dropped to ~40% at pH 10.
Based on the findings of the pH studies, I conducted the remaining experiments at pH 7.8 due to the relatively high loading efficiency and recovery observed at this pH. Next, I analyzed the drug release profile of each Zn-Por:VNP formulation . 1 mg of particles was resuspended in 300 μL PBS and loaded in triplicate in 10,000 MW cutoff Slide-A-Lyzer MINI dialysis units for 72 hrs. To mimic physiological conditions, samples were dialyzed against 3 L of PBS adjusted to pH 7.4 as well as pH 5 at 37°C. At time t = 0, 1, 3, 6, 18, 24, 48, and 72 h, 10 μL was extracted from each dialysis unit and the remaining Zn-Por entrapment was measured by UV/Visible spectroscopy. The half-life t1/2, defined as the time required for 50% of the drug to be released from the VNPs, decreases as the electropositivity of Zn-Por increases. At pH 7.4, TMV: Zn-Por4+ and TMGMV: Zn-Por4+ formulations had the lowest t1/2 . In contrast, only 20% and 25% of ZnPor1+ was released from TMV and TMGMV respectively within 72 hrs. The release profiles of Zn-Por3+ were similar to that of Zn-Por 4+, while the release rates of Zn-Por2+ were in between those of Zn-Por4+/3+ and Zn-Por1+. While the t1/2 values of each Zn-Por:VNP formulation were slightly lower at pH 5, the trend remained the same. These results indicate that the dominant force of interaction between TMV/TMGMV and Zn-Por is not electrostatic,hydroponic dutch buckets but rather hydrophobic/hydrophilic interactions. Since Zn-Por becomes more hydrophobic as its electropositivity is reduced, its ability to solubilize in PBS surrounding the VNP is impaired, thereby decreasing the rate of drug release. I also tested stability of the Zn-Por:VNP formulations under storage conditions , and observed a slow and constant release of Zn-Por from TMV and TMGMV over a period of 6 weeks Less than 45% of Zn-Por was released from the other formulation within 6 weeks. To evaluate in vitro efficacy of Zn-Por:VNP formulations, I first compared TMV and TMGMV’s uptake by B16F10 melanoma cells. Melanoma was chosen as a model because PDT has shown promise in melanoma.While most melanomas are removed by surgery supplemented with adjuvant chemotherapy and/or immunotherapy, some melanomas remain unresponsive to these therapies. A growing body of data indicates that PDT could be applied as an adjuvant therapy for those melanomas not responsive to traditional therapies.For cell uptake studies, TMV and TMGMV were conjugated with the fluorophore Cyanine 5 using solvent exposed tyrosine side chains click chemistry,followed by the purification of the reaction mixture as previously described. The covalent attachment of Cy5 was confirmed by UV-vis and denaturing SDSNuPAGE gel electrophoresis . We have previously demonstrated that a minimum conjugation of Cy5 to ∼8% of TMV coat proteins is sufficient to yield maximum fluorescence intensity.237 TMV and TMGMV particles displayed ~160 and ~490 dyes respectively. The higher dye conjugation efficiency in TMGMV could be due to differences in the chemical micro-environment and greater surface exposure of the tyrosine side chain.
The corresponding average distances between fluorophores are equal to 2.7 nm and 1.6 nm for TMV and TMGMV respectively, which are large enough to prevent quenching due to energy transfer between dye molecules and trapping by dimers.Therefore these Cy5-TMV and Cy5-TMGMV constructs are suitable for imaging experiments. To assess VNP–cell interactions, B16F10 melanoma cells were incubated with 100,000 VNPs per cell at 37 °C and 5% CO2 for 1 h and 8 h in Dulbecco’s modified Eagle’s media supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. Cells were washed thoroughly with FACS buffer 0.5 M EDTA, 1% FBS and 2.5% 1 M HEPES pH 7.0 in DPBS and fixed with 2% paraformaldehyde. Cells were then analyzed using a BD Accuri C6 Plus flow cytometer and 1 x 104 events were recorded. Data were analyzed using FlowJo v8.6.3 software. After 1 h of incubation, 85% and 100% of TMV and TMGMV were taken up by B16F10 cells, respectively . This is reflected by an increase in mean fluorescence intensity compared to cells only . The slightly higher uptake of TMGMV may be attributed to greater particle instability during viral production and purification, which causes some of the particles to be broken218; a shorter TMGMV rod would have a faster rate of cell penetration. Nonetheless, the cellular uptake of TMV and TMGMV reached 93% and 100%, respectively, after 8 h of incubation. This time point was selected to allow VNPs to traffic through the cells before proceeding with the photodynamic treatment. I evaluated efficacy of the drug delivery approach against B16F10 cells using previously established white light therapy.The following samples were tested: Drug-free VNPs, free Zn-Por, and Zn-Por-loaded VNPs, and a dark control for each sample was included. Cells were incubated with 0.001, 0.01, 0.1, 1, 5, and 10 μM of Zn-Por, Zn-Por:VNP, or controls for 8 h at 37 °C and 5% CO2. Cells were washed to remove any Zn-Por that was not endocytosed and samples were illuminated under white light for 30 min . Control samples were kept in the dark at 37 °C and 5% CO2. In all experiments, neither dark controls nor any of the VNP only controls showed significant cell toxicity. Free Zn-Por1+ was 1.8 to 2.8-fold more effective compared its TMV/TMGMV formulation; this reduced efficacy was even more dramatic for the Zn-Por3+ loaded particles which showed a 30-50 fold decrease in efficacy. The decreased drug activity of VNPs loaded with Zn-Por vs. free Zn-Por is expected. The reactive oxygen species produced by PS drugs have a very short half-life and act locally from their generation site. Therefore, the subcellular localization of the PS greatly influences its phototoxicity. Like most nanoparticles, TMV and TMGMV are internalized by endocytosis and follow the endosomal-lysosomal pathway. Previous data suggest the phototoxicity of PS localized in lysosomes is significantly reduced compared to PS localized in other organelles, in particular in mitochondria.On the other hand, hydrophobic PS with cationic charges such as free Zn-Por is likely to localize in mitochondria.Nonetheless, TMV and TMGMV are here used to improve the bio-availability and tumor accumulation of Zn-Por while reducing non-specific tissue toxicity. TMV and TMGMV can be further chemically or genetically modified to display moieties such as cancer cell targeting ligands, cell penetrating ligands, and chemotherapeutics for combined therapy, which would further improve the treatment efficacy. As a proof of concept, we set out to develop a targeted Zn-Por delivery system. We chose Zn-Por3+ and TMV, in particular we used the well-established and characterized Lys-added mutant denoted as TMVlys. While TMGMV showed greater toxicity than TMGMV, the genetic engineering of TMGMVlys mutant has yet to be established in the future. TMVlys offers amine functional groups for bio-conjugation: targeting ligands synthesized with a terminal Cys side chain can be conjugated using heterobifunctional NHS-maleimide linkers. Here we chose the F3 peptide as the ligand.