|Home | About | Journals | Submit | Contact Us | Français|
Hyaluronan (HA) is a glycosaminoglycan. Despite its high molecular weight, Brown et al. has shown that topically applied HA rapidly disseminated throughout the epidermis and reached the dermis in both mouse and human skin . Importantly, the majority of cutaneous malignant melanomas are positive for the HA receptor, CD44 . We envisioned that HA conjugated to anti-cancer drugs would be useful for: 1) Transport of the drug into the skin after topical application, and 2) Targeted delivery of the drug to CD44 positive malignant melanoma cells. To test this concept, we conjugated doxorubicin (DOX) to HA and tested its impact on B16-F10 melanoma cells in vitro and in vivo. We used DOX because of its intrinsic fluorescence and the availability of anti-DOX antibodies. Thus, DOX-HA could be evaluated in vitro and in vivo.
B16-F10 melanoma cells expressed CD44 as assessed by FACS (Fig. 1A). The binding of CD44 to HA and DOX-HA was evaluated using cell adhesion assays. As shown in Fig. 1B, panel 1, B16-F10 melanoma cells showed significantly better binding to wells coated with HA as compared to the wells lacking the HA substrate. An anti-CD44 monoclonal antibody significantly (P<0.05) blocked the adhesion of B16-F10 cells to HA showing that their binding to HA is CD44 mediated. Similarly, B16-F10 melanoma cells showed significantly better binding to wells coated with the DOX-HA as compared to wells lacking the DOX-HA (Fig. 1B, panel 2). Again, an anti-CD44 monoclonal antibody significantly (P<0.05) blocked the adhesion of B16-F10 cells to DOX-HA showing that their binding to DOX-HA is CD44 mediated.
To assess the uptake of DOX and DOX-HA by B16-F10 melanoma cells, we performed FACS. DOX and DOX-HA showed time dependent uptake in B16-F10 cells (Fig. 1C). However, the midpoints of the curves showed that the rate of DOX uptake was greater than the uptake for DOX-HA. DOX and DOX-HA both showed dose-dependent uptake by the B16-F10 cells but DOX accumulated to higher concentrations inside of the melanoma cells as compared with the DOX-HA (Fig. 1D).
Intracellular distributions of DOX and DOX-HA in melanoma cells were assessed by confocal microscopy. DOX and DOX-HA fluorescent patterns in B16-F10 cells were similar with cytoplasmic and punctate nuclear staining (Fig. 1E).
Both DOX and DOX-HA inhibited the proliferation of melanoma cells in a dose dependent fashion (Fig. 1F). However, DOX was more potent than DOX-HA. We failed to detect any effect of HA alone on the proliferation of B16-F10 melanoma cells (Fig. 1G) showing that DOX-HA inhibited the proliferation of B16-F10 cells due to the attached DOX moieties. Finally, primary human keratinocytes were sensitive to DOX but were insensitive to DOX-HA (Fig. 1H). Decreased cytotoxicity to normal keratinocytes suggests that DOX-HA has improved selectivity as compared with DOX.
Next, we tested the penetration of DOX and DOX-HA into mouse skin using a transdermal delivery assay . DOX and DOX-HA were recovered in the receptor solutions of Franz diffusion cells (Fig. 2A). The amount of recovered DOX was only modestly more than the amount of recovered DOX-HA but the difference between the two was significant (P<0.01).
To characterize the translocated DOX-HA, we assessed the recovered receptor solution using Western blotting and size exclusion chromatography. Results showed that an anti-DOX monoclonal antibody detected a single band. Treatment of the sample with hyaluronidase resulted in disappearance of the band, showing that the DOX detected in the sample recovered from the Franz cell was linked to HA (Fig. 2B). Therefore, DOX remained linked to the HA molecule as it transited the skin. On the other hand, the DOX-HA that crossed the skin was reduced in molecular weight (from ~1 × 106 Da to ~0.5 × 106 Da; Fig. 2C).
To assess the potential utility of DOX-HA for the local treatment of malignant melanomas, we applied it to tumors of graded size using DOX and vehicle as comparators. DOX-HA significantly reduced the growth rates of B16-F10 melanoma tumors compared with vehicle alone (P<0.01; Fig. 2D). By contrast, DOX had an insignificant effect relative to the vehicle treated mice (P>0.05). Comparisons of the tumor sizes in the DOX treated groups versus the DOX-HA treated groups showed that the bioconjugate more significantly (P<0.05) impaired tumor growth. Thus, DOX-HA improved the pharmacological activity of DOX. Because in vitro experiments showed that B16-F10 melanoma cells expressed CD44, bound the DOX-HA biopolymer in a CD44 dependent fashion and cells actively endocytosed the DOX-HA, we concluded that DOX-HA is targeted to the B16-F10 melanoma cells after topical applications most likely due to molecular interaction with CD44.
Next, we compared the impact of topically applied DOX and DOX-HA on normal mouse skin. We failed to detect macroscopic (induration, erythema and ulceration, data not shown) and histological alterations in the skin of mice after applications of either DOX or DOX-HA (cumulative dose of 42 μg; data not shown) suggesting that the dose and/or route of administration of DOX and DOX-HA did not induce local toxicities.
How can we explain the improved pharmacological activity of DOX-HA compared to DOX given that DOX showed higher cytotoxicity in vitro and that DOX showed better transdermal delivery? Previous in vitro tumor models have predicted that the rate of drug uptake into solid tumors is due to both protein binding and the diffusion gradient in the tumor interstitium . Thus, the cytotoxicity of drugs on tumor cell monolayers may not correlate with their effects in the context of the tumor. We propose that a combination of CD44 targeting and interstitial diffusion gradients can account for the improved anti-tumor activity of DOX-HA as compared with DOX. In fact, as shown in Fig. 2E, we recovered higher concentrations of DOX-HA from B16-F10 tumors in vivo as compared with DOX which showed that HA improved drug localization.
In conclusion, we have described the preparation of a DOX-HA bioconjugate and have evaluated its impact on tumor growth after topical applications in vivo. Results described in this study provide the conceptual and technical foundations for further developing HA formulations for the topical treatment of cutaneous malignant melanoma.
We thank Neema Lakshman for her assistance with confocal microscopy and Marguerite Starr for her secretarial assistance. This work was supported by NIH grants RO1 AR48840 (M.E.M.) and R24 EYO16664 (W.M.P.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.