An open-label, dose-escalating, non-randomized, single-center phase I study was conducted to study the intrapleural administration of attenuated vaccinia virus (Olvi-vec) as a bolus. Olvi-vec was administered either as a single dose or as three consecutive daily doses to patients with a histologically or cytologically documented diagnosis of MPE, as detailed in the study protocol ( Supplementary Material ).

The study protocol and amendments were approved by the Memorial Sloan Kettering Cancer Center Institutional Review Board (IRB# 12-169, NCT01766739). All patients provided written informed consent to participate in the study, and all response and toxicity outcomes were documented. Patients were enrolled in groups of three and individually assessed for safety and dose-limiting toxicity. Inclusion and exclusion criteria are listed in the protocol ( Supplementary Material ). Patients were treated following the diagnosis of histologically or cytologically documented MPEs (due to primary non-small-cell lung carcinoma, MPM, and other histologies) and had free pleural space (partial or total) that permitted intrapleural drug instillation.

A genetically engineered vaccinia virus, designated as GLV-1h68, was used in preclinical investigation. GLV-1h68 was derived from the LIVP strain by inserting RUC-GFP (a fusion gene of Renilla luciferase and green fluorescent protein), LacZ (beta-galactosidase), and gusA (beta-glucuronidase) expression cassettes into F14.5L (located between F14L and F15L), thymidine kinase (TK), and hemagglutinin loci, respectively. Disruption of these non-essential genes and expression of the foreign gene expression cassettes not only attenuated the virus but also enhanced its tumor-specific targeting. The GMP-derived material of this same virus is called Olvi-vec. Olvi-vec has been used primarily for all safety pharmacology and toxicological experiments, as well as for in vitro potency comparisons (in cell cultures) and in vivo potency comparisons (in tumorous animals). Details of the virus manufacturing process and analyses are described in the study protocol ( Supplementary Material ).

Eligible patients were admitted into the hospital for treatment on protocol. The pleural effusion was drained via insertion of a chest tube or pleural catheter (PleurX™ Catheter, Becton, Dickinson and Company, Franklin Lakes, NJ). A chest CT scan was performed to document drainage of the effusion and to assess the extent of pleural disease. Within 72 hours of the CT scan, the virus was instilled as a bolus into the pleural space via the chest tube or pleural catheter. Up to 150 ml of additional saline was used to flush the chest tube or pleural catheter to ensure that all the treatment drug was instilled into the pleural space. The chest tube or pleural catheter was left clamped for 4 hours (+/- 1 hour), after which it was reopened and placed to drainage in order to drain the pleural space. As dictated by the patient’s clinical status, the chest tube was either left inserted or removed until the surgical procedure (video-assisted thoracoscopic surgery, VATS) was performed 2-7 days after treatment to collect MPE and obtain pleural biopsy.

The primary objective of this study was to determine a recommended dose of Olvi-vec. The secondary objectives included the assessment of feasibility, safety and tolerability, evaluation of viral presence in the tumor, pleural fluid, serum, sputum, and urine, and evaluation of anti-vaccinia virus immune response. All patients were included in the reporting of adverse events (AEs). The safety of Olvi-vec was assessed by the evaluation of the type, frequency, and severity of AEs, changes in clinical laboratory tests (hematological and chemistry), immunogenicity, and physical examination. All AEs and laboratory toxicities were graded using the Common Terminology Criteria for Adverse Events (National Cancer Institute, version 4.0). Laboratory testing was performed at baseline (i.e., within 14 days before treatment), daily during the first 3 days after treatment, and at termination of study (day 60 ± 5).

Hematoxylin and eosin staining was performed on FFPE blocks of tumor biopsies collected before (pre-treatment) and 2-5 days after (post-treatment) Olvi-vec therapy. Semi-quantitative scoring of tumor cell density and immune cell density (0: very low density; 1: low density; 2: moderate density; 3: high density) was performed by a primary and secondary pathologist who were blinded to sample identity.

Multiplex immunofluorescence (mIF) staining was performed on tumor biopsies collected before (pre-treatment) and 2-5 days after (post-treatment) Olvi-vec therapy. Formalin-fixed and paraffin-embedded (FFPE) blocks were cut into sections of 5 µm thickness. Sections from each biopsy were stained with antibodies ( Supplementary Table 1 ) using the Opal™ 7-Color Kit for Multiplex Immunohistochemistry (Akoya Biosciences, Marlborough, MA). After mIF staining, slides were scanned using the Vectra^® 3.0 Automated Quantitative Pathology Imaging System (PerkinElmer Inc., Hopkinton, MA). Quantitative assessment of cell markers was performed using inForm^® software (version 2.2.1, PerkinElmer Inc., Hopkinton, MA). Cell segmentation and phenotyping algorithms were reviewed and confirmed by study pathologists.

Viral plaque assays were performed on body fluid samples (blood, sputum, urine, pleural fluid) collected from patients immediately before and 24, 48, 72, and 96 hours after Olvi-vec treatment to assess for the presence of viral particles. Post-treatment tumor biopsies collected 2-5 days after treatment also underwent assessment for viral particles using viral plaque assays. In brief, patient samples were plated on confluent layers of CV-1 cells. Evaluation of virus infection was done by visual assessment of viral plaque in wells with both CV-1 cells and patient samples. Additionally, post-treatment serum samples obtained from patients 60 days after Olvi-vec treatment were assessed for the presence of Olvi-vec neutralizing antibodies via standard vaccinia virus neutralization assay and compared to corresponding pre-treatment serum samples.

Pleural fluid and serum samples were obtained from patients both pre-treatment (baseline) and post-treatment (at 24, 48, 72, and 96 hours, and on days 2, 3, and 60). All specimens available were assessed for a panel of effector and suppressive cytokines using a 41-plex MILLIPLEX^® MAP Human Cytokine/Chemokine kit (MilliporeSigma, Burlington, MA). The kit was run on a Luminex^® 100/200™ System (Luminex Corporation, Austin, TX).

Values represent the mean of the duplicate wells ± standard deviation. These data were analyzed using IS 2.3 software (Luminex Software, Inc., Riverside, CA), Microsoft Excel and GraphPad Prism. Additionally, RNA isolation was performed on FFPE sections of tumor biopsies taken pre- and post-treatment using the RNeasy^® FFPE Kit in accordance with the manufacturer’s protocol (Qiagen, Germantown, MD). RNA concentration and purity were measured using the NanoDrop™ 2000/2000c Spectrophotometer (Thermo Scientific, Waltham, MA). RNA profiling was performed using the nCounter^® PanCancer Immune Profiling Panel by NanoString Technologies, Inc. (Seattle, WA).

The sample size was based on a standard dose-escalation design. All statistical tests were two-sided, and statistical significance was defined as p<0.05. Data with normal distribution was assessed using paired t-test. Data without normal distribution was assessed using Wilcoxon matched-pairs signed-rank test. Analyses were conducted using R 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria).

From February 2013 to April 2015, 18 patients were enrolled who were treated in a dose-escalating fashion ( Table 1 ). Fifteen patients had MPM (13 epithelioid, 1 biphasic, and 1 sarcomatoid), 2 had non-small cell lung cancer (1 adenocarcinoma and 1 squamous cell carcinoma), and 1 had triple negative breast cancer. 5 out of 18 patients were female, and the mean age of all patients was 66 years. All 3 patients with NSCLC or breast cancer and 1 patient with MPM had received previous lines of therapy.

Attenuated vaccinia virus (Olvi-vec) was administered intrapleurally as a bolus through a pleural catheter after complete evacuation of pleural effusion in all patients ( Figure 1 ). The intrapleural administration of vaccinia virus was feasible, and there were no failures in administration of the agent. Table 2 lists the adverse events that occurred at any grade (1–4) in ≥15% of the total cohort (n=18), up to day 60 post-treatment.

There was 1 reversible grade-4 laboratory abnormality (hypocalcemia). The most frequent grade 3 adverse events were lymphopenia (3 patients, 17%), fatigue (2 patients, 11%), and hypophosphatemia (2 patients, 11%). The most frequent grade 2 adverse events were anemia (5 patients, 28%), hyperglycemia (5 patients, 28%), and fever (4 patients, 22%). There were no dose-limiting toxicities, and maximally tolerated dose was not reached. As a result, the primary objective of establishing a recommended dose was not reached.

Resected post-treatment samples from 14 patients were stained by immunohistochemistry with an antibody against Olvi-vec (A27L). Positive cytoplasmic expression was observed in 7 of 14 specimens (representative images shown in Figures 2A, B ). When the A27L antibody was tested in a multiplex immunofluorescence panel with anti-mesothelin antibody on pre- and post-treatment specimens from Patient #16, cytoplasmic expression of Olvi-vec was observed in the post-treatment specimen but not the pre-treatment specimen (representative images shown in Figure 2C ).

Resected samples from 13 patients were stained by multiplex immunofluorescence with a panel of antibodies against mesothelin (MSLN), CD3, CD4, CD8, and FoxP3. 4 patients had matched pre- and post-treatment specimens, and 9 patients had post-treatment specimens only. The density of MSLN+ tumor cells was qualitatively observed to decrease and the density of CD3+ immune cells was qualitatively observed to increase from pre-treatment to post-treatment specimens (representative images shown in Figure 2D ).

Resected samples from 16 patients were stained with hematoxylin and eosin and independently scored semi-quantitatively for tumor-cell density and immune-cell density by two pathologists ( Table 3 ). Four patients had matched pre- and post-treatment specimens, and 12 patients had post-treatment specimens only. When matched tumor specimens were compared (n=4), tumor cell density score decreased from pre-treatment to post-treatment in all patients ( Figure 3 ). Immune cell density score increased from pre-treatment to post-treatment in 3 of 4 patients. Among all post-treatment tumor specimens (n=16), the average score per high power field (1 mm²) was lower for tumor cell density compared to immune cell density.

When matched tumor specimens were stained using multiplex immunofluorescence and then compared (n=4), mean CD8+ cells per mm² increased in 3 of 4 patients ( Figure 4 ). Mean MSLN+ tumor cells per mm² decreased from pre-treatment to post-treatment in all patients. Comparing available pre-treatment specimens to all post-treatment specimens (n=13), mean CD8+ cells increased and mean MSLN+ tumor cells decreased, although neither achieved statistical significance ( Figure 5 ).

Gene expression analysis of pre- and post-treatment tumor specimens (n=16), using nCounter immune cell type scoring module, revealed increased scores (i.e., change in score by 1 unit, indicating twice the abundance of that cell type) for CD45+, Th1+, Tregs, CD8+, exhausted CD8+, NK+, cytotoxic cells, dendritic cells, macrophage, and neutrophil immune cell populations in post-treatment tumor specimens compared to pre-treatment specimens ( Figure 6 ). Similarly, scoring of individual protein mRNA levels pre- and post-treatment tumor specimens (n=13) revealed increased scores in immune effector proteins (IFN-γ, granzyme B, perforin), immune suppressive proteins (TGFβ, FoxP3), and immune checkpoint regulatory proteins (PD-1, PD-L1, PD-L2) following treatment ( Figure 7 ). The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (31) and are accessible through GEO Series accession number GSE223395 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE223395).

Vaccinia virus was detectable by viral plaque assay in the pleural fluid of 7 patients ( Table 4A ). Vaccinia was initially detected in pleural fluid at the dose 1.00E8 PFU (Cohort 2) and exhibited dose-dependent increase in PFU/mL in Cohorts 3, 4, 5, and 6. Viral plaque assay was also positive for vaccinia in the post-treatment tumor lysate of 4 patients. There was minimal viral shedding into other compartments, as shown by low positivity in the urine of 2 patients, blood lysate of 1 patient, and sputum of 1 patient ( Table 4B ). Of note, among patients who received multiple doses of Olvi-vec (Cohorts 5 and 6), significant increase in the number of plaque-forming units was observed in the pleural fluid of 2 patients.

Among 7 patients whose baseline serum was available to perform vaccinia virus neutralization assay, 4 patients had low levels of anti-vaccinia neutralizing antibodies pre-treatment, and 3 had no neutralizing antibodies ( Table 5 ). Five of the patients had high levels of neutralizing antibodies at day 60 post-treatment.

Luminex analysis of pleural fluid specimens from baseline to up to 96 hours following treatment indicated significant increase in the levels of the following cytokines by 48 hours: IFN-γ, TNF-α, VEGF, IL-1ra, IL-1β, and IP-10 ( Figure 8 ). In contrast, analysis of serum specimens showed an increase only in IL-8 levels from baseline to day 3 following treatment (p=0.0065; Figure 8 ). By day 60, only RANTES (CCL5) was found to be significantly elevated in serum compared to baseline (p=0.0276; Figure 9 ).

All patients received subsequent other therapies as determined by treating physician following participation in the trial. Among all patients, median overall survival (OS) was 19.5 months. The median OS among patients who had MPM was 22 months ( Figure 10 ). One patient with epithelioid MPM is alive; 87 months after treatment, the patient received other treatments that are standard of care for patients with MPM.