Our data clearly demonstrate an increase in ALCAM expression in osteosarcoma though the biologic consequences of this are difficult to gauge. The normal physiologic roles of ALCAM are still coming to light, but its molecular structure and clustering at tight junctions suggest that it could be involved in cell adhesion and migration [23
]. In this context, it is tempting to think that modulating ALCAM expression could potentiate the invasive and metastatic behaviors found in high-grade malignancies such as osteosarcoma. However, there is no consistent correlation between ALCAM expression level and patient survival across all cancers. For example, an increase in ALCAM expression is found in higher-stage more aggressive malignant melanoma [24
]. By contrast, high ALCAM is correlated with low-grade less aggressive cases of prostate cancer [11
]. Considering the high frequency of elevated ALCAM expression in even our small cohort of osteosarcomas, it may not be able to discriminate between high- and low-risk patients with this disease.
Though ALCAM may be a limited prognostic biomarker in osteosarcoma, it has potential to serve as a molecule through which to therapeutically target this tumor. Fluorescent nanoparticles coated with anti-ALCAM diabodies preferentially bind to osteosarcoma cell lines, even those that express ALCAM at relatively low levels. As seen in prostate cancer cells, ALCAM-targeted nanoparticles were rapidly internalized by osteosarcoma cells suggesting a strategy for intracellular delivery of anticancer agents.
The use of diacetylene containing lipids to create polymerizable films and vesicles has been intensively studied for creating biosensors [25
] and have been explored as cancer diagnostic and delivery vehicles [26
]. When these membranes are treated with ultraviolet irradiation the resulting intralipid cross-links form an intensely blue chromophore. When exposed to physiochemical perturbations such as heat, shear, or pH stress, these membranes shift from a blue nonfluorescent state to a red fluorescent state [17
]. A distinct advantage to the HPLN fluorescence is that little or no photobleaching occurs. Taking advantage of these properties, we were able to track binding and internalization of red, fluorescent ALCAM-targeted PLNs (α
-AL-PLN) that had been treated with UV irradiation and heat. Interestingly, we obtained the same results using a similar preparation ofα
-AL-PLN that received only UV irradiation and were therefore blue and nonfluorescent in solution (data not shown). It appears that the interaction between the coupled diabody molecules and the cell surface ALCAM proteins exerted sufficient stress to shift the boundα
-AL-PLN into a fluorescent state.
Though vesicles composed entirely of diacetylene containing lipids had excellent detection properties, they had limited capability as therapeutic delivery vehicles. We were unable to stably load these liposomes with doxorubicin either by passive encapsulation during vesicle formation or actively across ion gradients in formed vesicles. Others have been able to passively load hybrid liposomes composed of a 1
1 mixture of a phosphatidylcholine derivative with a dichain diacetylene lipid and another phospholipid [27
]. However, loading efficiencies were low and this strategy may be limited to hydrophobic payloads. We have found that for amphiphilic molecules such as doxorubicin, in HPLNs, with single-chain, neutral PCDA lipids the polymerizable lipid concentration needs to be 20 mole percent or less for efficient loading to occur (data not shown).
Though our HPLNs were initially formulated for their stable drug loading characteristics, they surprisingly also proved to be more therapeutically potent in in vitro testing. The IC50 concentrations of untargeted doxorubicin-loaded hybrid PLNs in three independent osteosarcoma tumor-derived cell lines were at least 6-fold lower than conventional liposomal doxorubicin composed of PEGylated saturated phospholipid. This boost in potency appears to depend on PCDA lipid content since it is progressively lost as the PCDA concentration is titrated down from an optimum of 15–20 mole percent (data not shown). From this point, the lower the PCDA lipid concentration is in our HPLNs, the higher the IC50 becomes in our osteosarcoma model. Recently, others have used mixtures of diacetylene lipids and phospholipids to create liposomes that could be selectively destabilized either by photochemical means or by thermal shock [28
]. The goal here was to create a therapeutic vehicle that would release its payload in a temporal spatially controlled fashion.
We have found that even without applying an external destabilizing stimulus, HPLNs can be more effective therapeutic delivery vehicles than standard liposomal formulations. The mechanisms underlying this effect are unclear and require further investigation. The presence of PCDA in our hybrid formulations could be having an effect at multiple steps in our in vitro assay from (i) nanoparticle binding to cells to (ii) cellular uptake to (iii) intracellular release of cytotoxic payload. This last step in particular may be rate limiting. The roughly 50-fold difference in IC50 between free doxorubicin and conventional liposomal doxorubicin seen in our osteosarcoma cell lines is consistent with that found in previously published model systems [22
]. Others have shown that this is primarily due to delayed release of free drug from the endocytic compartment of cells that have taken up liposomal doxorubicin [30
Evaluating the stability of doxorubicin drug containment within the HPLN versus conventional PEG-liposomes showed that there was a statistically significant increase in doxorubicin release from the HPLN over time. Furthermore, this enhanced drug release was accentuated under acidic conditions mimicking the receptor-mediated endocytic environmental conditions of late endosomes and lysosomes. It is tempting to hypothesize that the PCDA lipids may enhance the release of doxorubicin from HPLNs that have been taken up by osteosarcoma cells. Given their differences in molecular structure, it is highly likely that microsegregation occurs between PCDA lipids and phospholipid molecules on the surface of HPLNs. Evidence found in published studies with similar mixtures of longer chain diacetylene lipids and shorter chain phosphatidylcholine lipids suggests that a phase separation occurs between the lipid types [31
]. It is possible that these PCDA lipid islands could serve as destabilization points that could enhance drug release when exposed to intracellular environments.
The creation of an osteosarcoma-targeted doxorubicin loaded HPLN (α
-AL-HPLN/Dox) resulted in a 2-fold increase in cytotoxicity over the untargeted HPLN/Dox and a 12-fold increase in cytotoxicity over the conventional PEG-liposomal formulation in the HOS and KHOS240s osteosarcoma cell lines. These results suggest that ALCAM targeting in osteosarcoma adds an incremental therapeutic effect. Interestingly, in the MNNG-HOS cell line theα
-AL-HPLN/Dox had an even greater increase (23-fold) in cytotoxicity over the PEG-liposomal formulation. The MNNG-HOS cell line has high expression levels of the multidrug resistant protein 1 (MDR1) conferring chemotherapeutic resistance to doxorubicin [32
]. The increased sensitivity of the MNNG-HOS chemoresistant cell line to theα
-AL-HPLN/Dox formulation over the conventional formulation points to a therapeutic effect that may overcome multidrug resistance. We can hypothesize that the targeting and improved sustained drug release characteristics of ourα
-AL-HPLN/Dox formulation may help to bypass or overwhelm the drug efflux proteins mediating chemoresistance thereby improving cytotoxicity.
In conclusion, we have found a novel surface marker in human osteosarcoma, ALCAM, which we have used to specifically target osteosarcoma cells with a novel engineered drug-loaded hybrid PLN formulation anti-ALCAM immunoconjugate. Theseα
-AL-HPLN/Dox particles show improved cytotoxicity over a conventional untargeted PEG-liposomal doxorubicin formulation and show promise as a potential therapeutic delivery platform in osteosarcoma. This new liposomal nanoparticle formulation is particularly attractive for its potential therapeutic application in resistant, refractory, and metastatic osteosarcoma where current standard systemic untargeted chemotherapy is generally not efficacious and prognosis is dismal. Furthermore, the bystander and dose-limiting side effects of systemic chemotherapy are substantial. Thus far this formulation has only been tested in tissue culture based assays, so further assessment in tumorigenic animal models is a crucial next step to validate these findings. These experiments are currently under way.