First-in-man study of western reserve strain oncolytic vaccinia virus: safety, systemic spread, and antitumor activity

Abstract

Oncolytic viral therapy utilizes a tumor-selective replicating virus which preferentially infects and destroys cancer cells and triggers antitumor immunity. The Western Reserve strain of vaccinia virus (VV) is the most virulent strain of VV in animal models and has been engineered for tumor selectivity through two targeted gene deletions (vvDD). We performed the first-in-human phase 1, intratumoral dose escalation clinical trial of vvDD in 16 patients with advanced solid tumors. In addition to safety, we evaluated signs of vvDD replication and spread to distant tumors, pharmacokinetics and pharmacodynamics, clinical and immune responses to vvDD. Dose escalation proceeded without dose-limiting toxicities to a maximum feasible dose of 3 × 10(9) pfu. vvDD replication in tumors was reproducible. vvDD genomes and/or infectious particles were recovered from injected (n = 5 patients) and noninjected (n = 2 patients) tumors. At the two highest doses, vvDD genomes were detected acutely in blood in all patients while delayed re-emergence of vvDD genomes in blood was detected in two patients. Fifteen of 16 patients exhibited late symptoms, consistent with ongoing vvDD replication. In summary, intratumoral injection of the oncolytic vaccinia vvDD was well-tolerated in patients and resulted in selective infection of injected and noninjected tumors and antitumor activity.

Figure 1

Toxicity of vvDD administration and acute pharmacokinetics. Blood samples were collected at various time points before and after virus administration. The enzyme activities in the serum and lymphocyte numbers in peripheral blood mononuclear cells were quantified. (a) Peak alkaline phosphatase levels (IU/l) relative to the baseline values before virus administration. (b) Median absolute lymphocyte count over time. (c) Lactate dehydrogenase peak levels (IU/l) relative to the baseline. (d) Recovery of serum vvDD DNA by qPCR at 15 minutes, 30 minutes, and 4 hours postinjection. The level of detection was 666 copies/ml and the level of quantification was 3333 copies/ml. At the 30-minute time point, viral DNA was detectable in patients treated at dose levels of 1.00E+08 and 3.00E+08 but below the level of quantification (at least 666 copies/ml but less than 3,333 copies/ml). This was also true at the 4-hour time point for patients treated at the two highest dose levels (1.00E+09 and 3.00E+09). The values displayed graphically represent serum vvDD DNA in copies/ml only in cases in which the detected level was greater than 3,333 copies/ml (the level of quantification). (e) Antibody response to viral administration demonstrating baseline and day 22 post-vvDD administration levels of anti-VV antibodies as measured by enzyme-linked immunosorbent assay in 12 of 17 patients. A 1.0 OD value represents a >50-fold induction of antibody.

Figure 2

Tumor selective infection, systemic spread, and clinical activity of vvDD in patients. (a) Evolution of response in patient #5 with breast cancer metastatic to the skin. vvDD replication leads to an ulcerated region encompassing only the tumor, which is resolving and scabbed over by day 28. (b) Another example of antitumor activity in patient #3 with breast cancer metastases to skin. A noninjected lesion 12 cm from the injected lesion became pustular and virus was recovered by plaque assay(*) on day 8. A similar lesion became pustular and resolved by day 28. (c) Patient #16 with metastatic colon cancer to the subcutaneous tissue and skin demonstrates significant erythema on day 8 at the two injected sites, consistent with active vvDD replication, and vvDD was recovered from the biopsy by plaque assay(*). A noninjected cutaneous lesion on the shoulder also became pustular and vvDD was recovered(*) on day 8. (D) In patient # 8, an omental mass of metastatic colon cancer was injected and a CT scan at day 30 revealed a markedly necrotic tumor and surrounding inflammation.

Figure 3

Immunologic effects of vvDD. (a) By phosflow analysis, pERK, pS6, and Ki67 were upregulated in circulating CD4 and CD8 T cells after vvDD injection in a dose-dependent manner. (b) Comparison of injected and noninjected tumor biopsies in compassionate use patient with melanoma on day 8 by qPCR for inflammatory markers. All proinflammatory markers are upregulated in the injected tumor, but the anti-inflammatory marker CCL22 is unchanged.

Figure 4

Antitumor activity in melanoma (compassionate use patient). (a) Evolution of tumor response demonstrating ulceration caused by vvDD replication specific for the injected tumors, with intervening skin unaffected. The tumor scabs over by day 22 and completely heals by day 75. (b) A larger adjacent uninjected lesion also responded. The larger lesion was reinjected on day 42, and surgically resected 5 months after the beginning of treatment. (c) A representative hematoxalyin-eosin stained slide from tissue obtained at the time of resection (5 months after the first vaccinia injection) demonstrating karyorrhectic as well as coagulative necrosis and visible tumor nest surrounded by hyalinized fibrous tissue.