This study will test whether it is possible to introduce new genetic material into a small portion of a tumor and have the product of the new gene not only kill those tumor cells that were infected initially, but also the surrounding tumor cells as well with limited or no harm to the patient. The desired effects of this approach are achieved by focusing potent chemotherapies directly within the tumor itself and, as a result, avoiding injury to the remainder of the body. In this study, we will use two components, the first of which is a virus, known as an adenovirus, that has been crippled (i.e., it cannot make more of itself) and loaded with a bacterial gene called E. coli purine nucleoside phosphorylase (PNP). Adenoviruses are considered to be relatively safe vehicles for gene delivery and are presently being used in numerous human trials and therapies worldwide, including a head and neck cancer therapy approved for use outside the United States. The loaded adenovirus will be used to deliver the PNP gene directly into a tumor in patients. This gene is not expected to have an effect itself. However, the gene produces PNP inside the tumor and this protein will activate the second component of the therapy, a drug called fludarabine phosphate, which is approved by the FDA for certain types of blood-cell cancers, but has not been shown to be effective against most solid tumors. The proposed therapy gives the patient several infusions of fludarabine following the injection of the virus carrying the PNP gene and, as the fludarabine enters the tumor, it will be converted by PNP into a second compound, fluoroadenine. Numerous studies in mice and rats have shown that fluoroadenine is a very potent anti-cancer agent and that it will kill the tumor cells where it is made as well as those in the immediately surrounding area.

For this first study, we will inject the PNP-loaded adenovirus into the tumors of patients with cancers primarily in the throat and neck and then give them the drug. This study is designed with two goals in mind: 1) assessing the overall safety of this approach for the patient; and 2) observing the effects of this anti-cancer strategy on the tumor itself. This will be accomplished in two parts. First, we will introduce a modest, fixed amount of the gene-carrying adenovirus into the tumors of three separate groups of patients and then administer small, increasingly strong amounts of the fludarabine phosphate to each successive group over a three-day period. Even in the group that will receive the highest amount of fludarabine, the total amount given to any individual patient over those three days will be significantly less than the dose approved by the FDA for patients with non-solid tumors. Finally, a more concentrated amount of the adenovirus (approximately 10 times more viruses) will be given to a fourth group of patients who will also receive the highest dose of the drug that was shown to be well tolerated in the prior three groups (the highest dose at which no serious problems were observed).

Inclusion Criteria:

• Biopsy confirmed diagnosis of a solid tumor
• Failed or exhausted all standard or approved treatment options that would provide substantive palliation
• Have at least one measurable primary or metastatic tumor on imaging studies or physical exam whose potential reduction could provide relief of symptoms or benefit
• Tumor is accessible for direct intratumoral injection

Exclusion Criteria:

• Diagnosis of leukemia
• Have previously received any gene therapy products or oncolytic viral therapy
• Receiving treatment with allopurinol
• Received radiation treatment < 4 wks prior to first injection of Ad/PNP
• Received chemotherapy < 4 wks prior to first injection of Ad/PNP
• Have signs or symptoms of active infection
• Receiving chronic systemic corticosteroids or any chronic immunosuppressive medications within 14 days prior to first injection of Ad/PNP. Subjects receiving short courses of corticosteroids are considered eligible.

56 days