MDA-MB-453, MDA-MB-468, MDA-MB-231, MEF 3.5^−/−, A549, SKBR3, HT-29 and SKOV3 cell lines were obtained from the American Type Culture Collection and maintained as described previously described.29

To generate the GrB(C210A)-Fc-IT4, human GrB(C210A) cDNA was fused to the N-terminus of the coding region of hinge CH2 and CH3 (Fc) of a human IgG1 heavy chain followed by the humanized single-chain VH-VL variable fragment of the anti-Fn14 (scFvIT4) antibody using three separate PCR steps. The individual PCR fragments were then subjected to two splice overlap extension PCR reactions (first Fc and scFvIT4; then GrB(C210A) with Fc-IT4) to complete the fusion gene construct.30 The complete GrB(C210A)-Fc-IT4 fusion gene was then cloned into the mammalian expression vector pSecTag (Life Technologies, Waltham, Massachusetts, USA), expressed by transfection of HEK-293E cells and purified by immobilized metal affinity chromatography as previously described.7 The leader sequence containing the c-Myc epitope tag, (His)6 tag and the enterokinase cleavage site was cleaved by incubating the purified protein with recombinant enterokinase (New England Biolabs, Ipswich, Massachusetts, USA) overnight (20 U/mg) at room temperature (RT).31

GrB enzymatic activity was measured using a continuous colorimetric assay with Ac-IEPD-pNA (N-acetyl-Ile-Glu-Pro-Asp-p-nitroanilide) (Merck, Darmstadt, Germany) as a specific substrate as previously described.7 32

Cells were seeded onto plates overnight. After 24 hours incubation at 37°C, medium was replaced with complete media containing various concentrations of GrB or GrB(C210A)-Fc-IT4 and incubated at 37°C in 5% CO₂ for 72 hours. Cell viability was determined using the crystal violet staining method as described previously.33

A549 and MDA-MB-231 cell lines were treated with 30 nmol/L GrB or GrB(C210A)-Fc-IT4 for various time periods. Internalization was assessed with an immunofluorescence assay as previously described.7 Studies to identify potential delivery of the GrB(C210A)-Fc-IT4 construct in the endosomal/lysosomal compartment were performed using the Lysotracker DND-22 kit (Life Technologies, Carlsbad, California, USA). A549 cells were treated with 30 nmol/L GrB(C210A)-Fc-IT4 labeled with Alexa Fluor 594 (Molecular Probes, Eugene, Oregon, USA) for 2 hours at 37°C, then media was replaced with fresh media containing the Lysotracker Blue DND-22 for 2 hours. Then cells were washed with phosphate buffered saline (PBS) and visualized with a Nikon A1Rsi multiphoton/confocal microscope using 405 nm and 561 nm lasers for excitation.

Apoptosis was assessed using the Annexin V-FITC Kit (Invitrogen, Eugene, Oregon, USA).34 Mitochondrial membrane depolarization was studied using the cationic dye JC-1 (MitoProbe JC-1 Assay Kit, ThermoFisher Scientific, Eugene, Oregon, USA) as previously described.31

Western blot studies were performed using standard procedures.7 The primary antibodies used were anti-GrB (Santa Cruz Biotechnology, Santa Cruz, California, USA), anti-Caspase-3, anti-Caspase-7, anti-Caspase-9, anti-PARP-1 (Cell Signaling Technology, Danvers, Massachusetts, USA) and anti-β-actin (Sigma, St Louis, Missouri, USA).

The antibody dependent cell cytotoxicity (ADCC) response was evaluated using the ADCC Reporter Bioassay kit (Promega, Madison, Wisconsin, USA). Briefly, SK-BR-3 cells were seeded onto plates and incubated overnight at 37°C. Serial dilutions of Trastuzumab (Roche, San Francisco, California, USA), an anti-CD20 antibody and GrB(C210A)-Fc-IT4 were prepared in RPMI-1640 media and added to SK-BR-3 cells. Jurkat/FcγRIIIa/NFAT-Luc cells were suspended in RPMI-1640 with low IgG serum and added to the cells, followed by incubation at 37°C. After 24 hours, luciferase activity was measured by ONE-Glo Luciferase Assay Reagent.

GrB(C210A)-Fc-IT4 (50 µg/mL) was incubated with commercial mouse serum (BioIVT, Westbury, New York, USA) for different time periods (0, 24, 48, 72, 96 hours) at 37°C. Samples were harvested on the same day and spun through microtainer vials. The anti-GrBhorseradish peroxidase (HRP) conjugated antibody (Santa Cruz Biotechnology) was used to detect the target protein and the signal intensity was quantitated by Adobe Photoshop with GrB(C210A)-Fc-IT4 as standard. The signal was normalized with a serum loading control (Transferrin, Genetex, Irvine, California, USA). Stability experiments were performed in duplicate.

BALB/c mice (Experimental Radiation Oncology Breeding Core, MD Anderson Cancer Center) at 4–6 weeks of age were injected (intravenous, tail vein) with 200 µg GrB(C210A)-Fc-IT4. At various time points (0–48 hours after fusion protein administration), mice were sacrificed and blood was extracted from the chest cavity. Plasma samples were analyzed using the Western blot method previously described as well as a quantitative ELISA assay. Briefly, ELISA microtiter plates (Santa Cruz Biotech) were coated (18 hours, 4°C) with 2 µg/mL of the Fn14 extracellular domain (Cell Signaling) in coating buffer (0.05 M Na₂CO₃, 0.05 M NaHCO₃, pH 9.6) and blocked with 2% bovine serum albumin in PBS at RT for 2 hours. Samples (100 µL) were loaded in triplicate and incubated for 2 hours at RT, followed by the addition of HRP-conjugated goat anti-human IgG (KLP, Gaithersburg, Maryland, USA) in blocking buffer (1 hour, RT). Plates were washed five times after each step. The reaction was visualized by the addition of chromogenic substrate (ABTS) for 30 min and stopped with 10% sodium dodecyl sulfate (SDS). Absorbance at 450 nm was measured on a Thermomax plate reader. As a reference for quantification, a standard curve was established by a serial dilution of recombinant GrB(C210A)-Fc-IT4 in commercial mouse serum. Pharmacokinetic parameters were calculated by compartmental analysis of plasma data using an add-in program for pharmacokinetic and pharmacodynamic data analysis in Microsoft Excel the PKSolver (V.2.0).

Groups of female BALB/c nude mice (MDACC Experimental Radiation Oncology Breeding Core) (7 weeks old, five mice per group) were injected with 5×10⁶ human MDA-MD-231/Luc cells into the mammary fat pad (1:1 v/v with Matrigel). Once tumor volumes reached between 40 and 60 mm³, animals were randomized and treated (iv, tail vein) every other day for five total injections with vehicle or GrB(C210A)-Fc-IT4 (100 mg/kg total dose). Animals were monitored and tumor growth was measured every 3–4 days for an additional 90 days.

Human A549 lung cancer cells (7×10⁶ cells) were subcutaneously injected on the right flank of 6-week-old female infant BALB/c nude mice. Once tumors were established (~25 mm³), animals were randomized and treated (intraperitoneal) every day for 10 total injections with vehicle, enavatuzumab (50 mg/kg) or GrB(C210A)-Fc-IT4 (50 mg/kg or 100 mg/kg total dose). Animals were monitored and tumor growth was measured every 3–4 days for 60 days.

Animal experiments were conducted following approval by the Institutional Animal Care and Use Committee guidelines of St. Joseph’s Hospital and Medical Center. Fresh lung tumors resected from a patient-derived xenograft (PDX) were subcutaneously (s.c.) implanted into the flank of 6-week-old nu/nu athymic mice (Jackson Labs). Tumors from the same patient xenografts were allowed to grow to approximately 200 mm³. Animals were grouped randomly (six mice per group) and treated with vehicle or GrB(C210A)-Fc-IT4 (20 mg/kg) every other day over a 10-day period for a total dose of 100 mg/kg via tail vein injection. Tumor size was measured 2–3 times per week for 30 days.

Female BALB/c mice (4–6 weeks old) were injected intravenously every other day for five total injections of vehicle or GrB(C210A)-Fc-IT4 (100 mg/kg total dose). One day after the last injection (Acute group, Day 10) or 4 weeks after the last injection (Recovery group, Day 38) animals in the study were euthanized and terminally bled for hematological and tissue harvest/histological analyzes. The studies were performed by the MDACC Department of Veterinary Medicine and Surgery.

All in vivo efficacy experiments, statistical analysis was performed using Prism V.8.0.0 (GraphPad Software). Comparisons of continuous variables between two groups were performed using unpaired t-test. Comparisons of continuous variables between three or more groups were performed using a one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparisons test. Differences were considered to be statistically significant when p values were <0.05.

GrB(C210A)-Fc-IT4 (figure 1A) was transiently expressed by HEK-293E cells and purified to homogeneity. Analysis by SDS-PAGE of GrB(C210A)-Fc-IT4 under non-reducing conditions confirmed the purity of the construct (figure 1B). The theoretical molecular weight is 160 kDa for the homodimer but the presence of several glycosylation sites in the primary sequence increased the apparent molecular weight to 220 kDa. GrB(C210A)-Fc-IT4 had a similar enzymatic activity to commercial native GrB (264 U/nmol and 159 U/nmol, respectively), demonstrating the mutation has no impact on the enzymatic activity of GrB (figure 1C).

Next, we assessed the cytotoxic effects of GrB(C210A)-Fc-IT4 on a panel of tumor cell lines (table 1). The IC₅₀ of GrB(C210A)-Fc-IT4 ranged in the low nanomolar range, showing that the construct is able to induce cell death on Fn14 positive cell lines in a targeted manner. As previously published for the original GrB-Fc-IT4,7 we found no correlation between cell surface Fn14 levels and the sensitivity to the construct (data not shown). The fusion protein did not affect the growth of the mouse embryonic fibroblast MEF3.5^−/− cell line, derived from Fn14 Knockout mice, indicating that the cytotoxicity of the construct on the tumor cell lines was dependent on the presence of the Fn14 receptor.

We also examined the effect of GrB(C210A)-Fc-IT4 on MDA-MB-231 viability over time. While GrB(C210A)-Fc-IT4 treatment for 12 hours showed moderate levels of cytotoxicity, the lowest IC₅₀ was obtained after 72 hours of continuous exposure (online supplemental table 1S). These results suggested cytotoxicity was due to receptor-mediated internalization and subsequent release of GrB(C210A)-Fc-IT4 into the cytosol.

To investigate the cytotoxic effect of the GrB(C210A)-Fc-IT4 on Fn14 positive cells, we first assessed the internalization process in MDA-MB-231 and A549 cells by immunofluorescence microscopy. We detected GrB(C210A)-Fc-IT4 in the cytoplasm within 30 min after exposure (figure 2A). In contrast, no staining was detected in cells treated with GrB alone, suggesting that internalization of GrB(C210A)-Fc-IT4 was driven by Fn14 receptor-mediated endocytosis. The intracellular delivery of the GrB portion and the metabolic fate of GrB(C210A)-Fc-IT4 was assessed at various times by Western blot analysis. Intact GrB(C210A)-Fc-IT4 was detected within the cell in less than 1 hour after exposure (figure 2B). Intracellular claeveage to release free GrB into the cytosol was observed at approximately 4 hours of treatment. Surprisingly, neither the intact GrB(C210A)-Fc-IT4 nor free GrB were observed after 16 hours, suggesting either complete degradation of the construct and its components or epitope masking of the GrB portion of the molecule.

Because accumulation of cell-targeted fusion proteins can occur in the endosomal and lysosomal compartments and can result in significant reduction in the efficiency of certain targeted therapeutics, we next evaluated the impact of chloroquine, a lysosomotropic agent that allows endosomal escape, on construct activity against MDA-MB-231 cells. Treatment of cells with different (non-cytotoxic) concentrations of chloroquine had no impact on the IC₅₀ of GrB(C210A)-Fc-IT4 suggesting that endosomal or lysosomal trapping does not play a role in GrB(C210A)-Fc-IT4 internalization (figure 2C). We also studied the localization of GrB(C210A)-Fc-IT4 protein within the endosomes or lysosomes of A549 cells using Lysotracker Blue DND-22, a fluorescent dye that stains acidic compartments in live cells. We did not detect specific colocalization of the fusion protein in endosomes or lysosomes after 2 hours of treatment (figure 2D). Together, these studies confirm that GrB(C210A)-Fc-IT4 is internalized rapidly and efficiently escapes into the cytosol of target cells without appreciable accumulation in the lysosomal or endosomal compartments.

GrB is a well-known potent initiator of apoptosis once delivered to the cytosol.35 To confirm that the cytotoxicity of GrB(C210A)-Fc-IT4 is mediated by apoptosis, we first assessed the Annexin V levels on MDA-MB-231 and A549 cell lines in the presence of different concentrations of either GrB(C210A)-Fc-IT4 or GrB. The construct GrB(C210A)-Fc-IT4 was able to increase Annexin V levels whereas GrB treatment alone showed no impact over 24 hours compared with non-treated cells (figure 3A). Similarly, treatment with GrB(C210A)-Fc-IT4 rapidly induced mitochondrial depolarization in treated cells over 24 hours (figure 3B).

Next, we used Western blot analysis to examine various cellular mediators of apoptosis known to be part of the GrB-induced apoptotic mechanism. Cells treated with GrB(C210A)-Fc-IT4 showed activation of the apoptotic cascade by cleavage of caspase-3, caspase-7, caspase-9 and PARP-1 at 8 hours after treatment (figure 3C). These data confirm that the mechanism of cytotoxic action appears to be similar to that of GrB itself.

An important mechanism of cytotoxic action of some conventional IgG surface-bound antibodies is the mediation of ADCC. To evaluate whether the Fc portion of the GrB(C210A)-Fc-IT4 is able to mediate ADCC on target cells, we performed an in vitro assay using the SK-BR-3 cell line which is known to express Fn14 and HER2 on the cell surface.36 As shown in figure 3D, treatment with the anti-HER2 antibody, trastuzumab, induced the activation of Jurkat cells whereas treatment with GrB(C210A)-Fc-IT4 and the irrelevant control antibody did not induce any response in the effector cells. This suggests that GrB(C210A)-Fc-IT4 is unable to induce ADCC likely due to the steric hindrance at the C-terminus of the Fc domain by the presence of the anti-Fn14 scFv domain. Therefore, all of the cytotoxic effects observed in vivo are likely to be solely due to the GrB component of the fusion construct.

In order to assess the stability of the GrB(C210A)-Fc-IT4 molecule in a physiological environment, we incubated GrB(C210A)-Fc-IT4 in commercial mouse serum for up to 96 hours at 37°C and conducted Western blot analysis. Intact or full length GrB(C210A)-Fc-IT4 was present in serum for up to 96 hours with an overall ex vivo half-life of 88 hours, indicating the construct is stable under physiological conditions (figure 4A). In addition, quantitative ELISA and Western blot analysis of GrB(C210A)-Fc-IT4 from plasma samples harvested at 24, 48 and 72 hours after administration to BALB/c mice showed similar quantitative results (data not shown).

Pharmacokinetic analysis of GrB(C210A)-Fc-IT4 revealed a biphasic clearance profile (figure 4B). Following intravenous administration, GrB(C210A)-Fc-IT4 demonstrated a rapid α-phase half-life of 0.36 hour, followed by a prolonged terminal β-phase clearance of 35.6 hours. The apparent small volume of distribution of 1.66 mL suggests limited distribution of GrB(C210A)-Fc-IT4 into peripheral sites outside the vascular space immediately after administration. The estimated initial (t=0) concentration in plasma was 120 µg/mL, similar to the projected initial dose concentration of 153 µg/mL of drug injected per mouse. These studies further confirm the stability of the molecule in vivo.

Next, we examined the in vivo toxicity observed in mice treated with total dose of 100 mg/kg of GrB(C210A)-Fc-IT4 using a QODX5 schedule. Body weight, tissue and serum samples were collected 24 hours (acute group) or 4 weeks (recovery group) after intravenous administration of the construct to BALB/c mice. There were no differences in body weight (figure 4C) or specific organ weight (data not shown) between the controls and the treated mice. The enzymatic activities of liver enzymes were determined and no differences in aminotransferase (AST), alanine transferase (ALT) and lactate dehydrogenase (LDH) levels between groups were found, indicating no evidence of hepatotoxicity (figure 4D). We performed a complete hematological and chemical analysis; we did not find significant differences between groups (data not shown). The histological analysis showed no moderate or marked lesions in any mice, and only a few rare and minor mild incidental lesions were present (online supplemental table 2S). Thus, GrB(C210A)-Fc-IT4 is a safe and well-tolerated drug at doses as high as 100 mg/kg in mice as there were no indications of severe toxicity.

Once the pharmacokinetics and toxicity of GrB(C210A)-Fc-IT4 were determined, we then evaluated the ability of GrB(C210A)-Fc-IT4 to inhibit the growth of established MDA-MB-231 tumors in BALB/c nude mice. Three out of five mice with tumors treated every other day with GrB(C210A)-Fc-IT4 for 10 days showed complete tumor growth inhibition whereas the other two mice showed a significant tumor growth inhibition (p<0.001) compared with control (vehicle) tumors 80 days after implantation (figure 5A). Furthermore, mice bearing MDA-MB-231 tumors treated via intravenous or intraperitoneal injection with GrB(C210A)-Fc-IT4 (100 mg/kg) following the same scheduled (QODX5) showed significant tumor growth delay independent of the route of administration (online supplemental figure 1S). There were no changes in body weight in control versus treated mice for the duration of both studies indicating that 100 mg/kg total dose was well tolerated in mice.

In order to determine whether the antitumor effect observed was mediated by the cytotoxic payload (GrB) and not because of TWEAK/Fn14 axis inhibition by the scFvIT4 targeting moiety, we treated nude mice with subcutaneous A549 NSCLC tumors every day for 10 days with either GrB(C210A)-Fc-IT4 or the unconjugated anti-Fn14 antibody, enavatuzumab at equimolar drug concentrations.27 28 Administration of enavatuzumab resulted in transient tumor growth inhibition followed by rapid regrowth of the tumors back to and exceeding control tumor growth (figure 5B). In contrast, treatment with our GrB-based construct was able to induce a sustained prolonged growth delayed when compared with controls.

Finally, we examined the efficacy of GrB(C210A)-Fc-IT4 in a human lung PDX model. Tumor cells were subcutaneously (s.c.) implanted in the flank of mice and once tumors reached about 200 mm³, animals were grouped randomly (six mice per group) and treated with PBS or GrB(C210A)-Fc-IT4 (100 mg/kg). We observed a 50% reduction in tumor burden in GrB(C210A)-Fc-IT4 treated mice compared with the vehicle group (figure 5C). Overall, these data indicate that GrB(C210A)-Fc-IT4 is effective against breast and lung tumor xenograft models causing a dramatic and prolonged tumor growth inhibition, and also reduces tumor growth rate in a lung PDX tumor model.