Formulated Poly IC Destroys Other EGFR Over-Expressing Cancers In Vitro and In Vivo

We next examined whether our strategy could be implemented for the treatment of other EGFR over-expressing cancers. We used the A431 (vulval carcinoma, expressing ;2 3 106 EGF receptors per cell [18]) and MDA-MB-468 (breast cancer, similar EGFR expression and EGFR responsiveness to A-431 cells [19]), both non-engineered cell lines. Figure 7B shows targeted poly IC induced killing of A431 and MDA-MB- 468 cells, whereas the control U87MG cells remain unharmed. Figure 7B also demonstrates the enhanced efficacy of the allin- one vector MPPE as compared to PEI-PEG-EGFþPEI-Mel combination formula.

Encouraged by these results, we examined the efficiency of our strategy on models of these cancers in vivo. A431 cells were implanted subcutaneously into nude mice as described (Methods) and the tumors left to grow for 10 d. During this period, tumors of average size 10.1 mm³ were developed (Figure 7C). After tumor establishment, (poly IC)MPPE complexes were injected directly into the tumors twice per day for 6 d at 15 lg of poly IC/mouse/injection (30 lg/d) (Methods). On day 10 after treatment initiation, the tumors had disappeared completely from (poly IC)MPPE treated animals, while the tumors continued to grow in untreated animals, reaching up to139 mm³ (Figure 7C).

MDA-MB-468 cells were implanted over mammary fat pads of female SCID mice and the tumors left to grow for 14 d. During this period, tumors of average size 9.7 mm³ developed (Figure 7C). (poly IC)MPPE complexes were injected directly into the tumors for 6 d. On day 8 after treatment initiation, the tumors had disappeared completely from (poly IC)MPPE animals, while the tumors continued to grow in untreated animals, reaching 63 mm³ (Figure 7C).

We did not detect recurrence of any of the tumors in the follow up study (day 42þ after treatment initiation).

Discussion

Our results suggest that the EGFR-targeted delivery of poly IC can be implemented in the clinical treatment of GBM, for which current therapies are essentially ineffective. The therapeutic approach described here incorporates rapid and efficient killing of tumor cells by multiple mechanisms, a strong bystander effect, along with high selectivity. The localization of GBM in the central nervous system makes it an attractive candidate for local therapy by slow, constant, intratumoral delivery of the complexes described here (Figure 5). EGFR-targeted delivery of poly IC should be especially effective for the small tumors remaining after surgery, but could in principle be the first line of treatment, considering the results reported here. We have previously described an approach for GBM treatment using viral vectors to induce the in situ production of cancer-specific dsRNA within the cancer cell [8]. The EGFR-targeted synthetic dsRNA approach we describe now seems to be far superior, probably because we have achieved rapid delivery of a large dose of longer dsRNA into the cancer cell. This leads to a very fast response and induction of a bystander effect so that the tumor is demolished more rapidly than it can re-grow. Although many of the immune cells are still present in the nude mice we used, no significant immune reaction against the poly IC treated tumor was observed (Figure 6A). Most likely absence of the immune response is due to the fast elimination of the tumor. An immune reaction might be induced in patients following treatment that will take significantly longer periods of time. Nevertheless, expression of immunoactive IFN-a and T-cell chemokines, IP-10 and Gro-a selectively in the tumor (Figure 4) should drive the immune response specifically against GBM, reducing potential toxic effects on surrounding brain tissue.

To the best of our knowledge, the strategy described in this study has yielded the most effective treatment of EGFR overexpressing GBM reported so far, in an animal model. The therapeutic strategy described here differs significantly from other EGFR targeted agents aimed at GBM, such as erlotinib, gefitinib, [20] and anti-EGFR antibodies [21]. These agents inhibit the activity of the receptor and its downstream signaling. The response of GBM to gefitinib and erlotinib in GBM is weak [22]. Furthermore, the response of non-smallcell- lung cancer to either drug as a single agent is limited to tumors in which the EGFR harbor mutations in the kinase domain, where the EGFR is a survival element, and is usually transient [23]. Mutations (extracellular) in the EGFR receptor or frequent heterogeneity of GBM where only part of the cells over-express EGFR will result in only partial response to gefitinib and erlotinib as indeed observed in the recent clinical trials [22]. In the case of anti-EGFR therapy, the response seems not to be related to whether the EGFR is mutated but these drugs are also not life saving in patients with tumors involving EGFR [24]. The strategy described here is different from those currently implemented therapies because the EGFR is utilized to target a vehicle that delivers a strong pro-apoptotic molecule, namely dsRNA. Due to the strong bystander effect induced by the massive amount of dsRNA on neighboring cancer cells that were not targeted themselves, the therapy described here can eradicate EGFR over-expressing tumors even when only half of the cells overexpress wt EGFR (Figure 5C). The bystander effect is mediated at least partially by IFN-a (Figure 3B). This cytokine is clinically used against various cancers including GBM. Since IFN-a kills mainly fast proliferating cells, normal cells should not be affected. Indeed, we did not detect any pathology in normal brain tissue during or after the therapy (Figure 6A).

Another important advantages of the therapy described here are the fast cell killing (Figure 1) and activation of several growth inhibitory pathways simultaneously (Figures 3A and 4). This form of therapy is likely to prevent and/or overcome a potential mutation in one of the anti-proliferative proteins and development of resistance to the therapy. This is especially relevant for cancers with deficient PKR activity [25], because poly IC can still induce cell death in a PKR independent manner (Figure 3A). We would like to suggest that if this strategy is translated successfully to the clinical setting it may indeed help patients with GBM. In fact, several neurosurgery teams are interested in implementing this therapy to clinic once all preclinical studies will be completed.

The elimination of EGFR over-expressing adenocarcinoma and breast cancer in vivo (Figure 7C) suggests that in principle this strategy can be applied to treat other cancers that over-express EGFR. It should be emphasized that there are a number of scenarios, other than GBM in which the local application of the EGFR targeted poly IC is advantageous, like head and neck cancer in which EGFR is over-expressed and local therapies are often used.

Finally, we would like to suggest that the strategy of ligandguided delivery of dsRNA described here, can in principle, be applied to other cancers in which a particular receptor is over-expressed and undergoes endocytosis. Receptors that are over-expressed in many tumors and qualify as candidates for targeting poly(IC) include the transferrin receptor [26], the PDGF receptor [27] and the IGF-1 receptor [28].