A novel reduced immunogenicity bispecific targeted toxin simultaneously recognizing human epidermal growth factor and interleukin-4 receptors in a mouse model of metastatic breast carcinoma

Abstract

Purpose: To develop a targeted biological drug that when systemically injected can penetrate to metastatic breast cancer tumors, one needs a drug of high potency and reduced immunogenicity. Thus, we bioengineered a novel bispecific ligand-directed toxin (BLT) targeted by dual high-affinity cytokines with a PE(38)KDEL COOH terminus. Our purpose was to reduce toxin immunogenicity using mutagenesis, measure the ability of mutated drug to elicit B-cell antitoxin antibody responses, and show that mutated drug was effective against systemic breast cancer in vivo.

Experimental design: A new BLT was created in which both human epidermal growth factor (EGF) and interleukin 4 cytokines were cloned onto the same single-chain molecule with truncated Pseudomonas exotoxin (PE(38)). Site-specific mutagenesis was used to mutate amino acids in seven key epitopic toxin regions that dictate B-cell generation of neutralizing antitoxin antibodies. Bioassays were used to determine whether mutation reduced potency, and ELISA studies were done to determine whether antitoxin antibodies were reduced. Finally, a genetically altered luciferase xenograft model was used; this model could be imaged in real time to determine the effect on the systemic malignant human breast cancer MDA-MB-231.

Results: EGF4KDEL 7mut was significantly effective against established systemic human breast cancer and prevented metastatic spread. Mutagenesis reduced immunogenicity by approximately 90% with no apparent loss in in vitro or in vivo activity.

Conclusions: Because EGF4KDEL 7mut was highly effective even when we waited 26 days to begin therapy and because immunogenicity was significantly reduced, we can now give multiple drug treatments for chemotherapy-refractory breast cancer in clinical trials.

Figure 1

A) In vitro activity of BLT EGF4KDEL and monospecific EGFKDEL and cpIL4KDEL was determined by measuring ³H-leucine incorporation of MDA-MB-231 breast cancer cells following 72 hours incubation with varying concentrations of cytotoxins. Data is mean of triplicate samples ±SD and expressed as percentage of ³H-leucine relative to control cells incubated in media alone (% Control Response). IC₅₀ (inhibitory concentration 50%) is the concentration of BLT that inhibits 50% of the percent control activity. B) In vitro activity of bispecific EGF4KDEL was compared to an equimolar mixture of monospecific EGFKDEL and IL4KDEL against MDA-MB-231 cells in a cell proliferation assay. Irrelevant control BLT 2219 KDEL, DT2219ARL and Bic3 were also tested. Control counts for untreated MDA-MB-231 cells = 17,644 ± 2,362 cpm/20 ×10⁴ cells. C) To study selectivity, EGF4KDEL was tested on EGF⁻IL-4⁻ HPBMLT cells in the proliferation assay. Control counts for HPB-MLT cells = 104,705 ± 10,522/20 ×10⁴ cells D) To test the ability of anti-EGF and anti-IL-4 to block, individual antibodies were added EGF4KDEL and then tested against MDA-MD-231 cells in proliferation assays. Control counts for untreated MDA-MB-231 (15,714 ± 1,789/20 ×10⁴ cells. Center) EGF4KDEL consisted of the genes for human EGF, human IL-4, and 362 amino acid of the PE38 molecule with c-terminal REDLK replaced with KDEL. The final target gene was spliced into the pET21d bacteria expression vector.

Figure 2

The selective efficacy of EGF4KDEL against MDA-MB-231-luc in vivo. (A) To determine the metastatic ability of the MBA-MB-231-luc tumor, a representative tumor injected mouse was organ imaged on day 60 following intrasplenic (IS) tumor injection on day 0. Two images are shown for each organ. The image on the left is without bioluminescent imaging and on the right, the same organ with bioluminescent imaging. Bioluminescence intensity is shown as a function of photons/sec/sr/cm² and is listed for each image. B) Bioluminescent imaging for Experiment 1 in which nude mice were given an IS injection of 10⁶ MBA-MB-231 cells on day 0. Mice M1-M5 were untreated, while mice M6-M11 were treated IP with 10 courses (40 injections) of EGF4KDEL beginning on day 5. Mice M12 and M13 received treatment with the identical dose and schedule, but with anti-B cell 2219ARLKDEL instead. Also shown on the figure is a graph of average total photon tumor activity from the mice in this experiment. Averaged data are compared by Student t test. (p< 0.01).

Figure 3

Anti-tumor effect of EGF4KDEL given 26 days post tumor. A) Bioluminescent imaging from Experiment 2 in which nude mice were given an IS injection of 10⁶ MBA-MB-231 cells on day 0. Mice M14-M18 were treated IP with 10 courses (40 injections) of EGF4KDEL beginning on day 5. In contrast, mice M19-M22 were treated with 10 courses EGF4KDEL beginning on day 26. M23-M27 were injected equimolar mixture of monospecific EGFKDEL + IL4KDEL with same schedule with EGF4KDEL. Bioluminescence intensity is shown as a function of photons/sec/sr/cm² B) Mice M28-M31 were treated with EGF4 devoid of toxin as part of experiment 1. C) Mice 32-M36 were part of experiment 3 whereby mice given IS injection of 2 × 10⁶ MBA-MB-231 cells on day 0, were treated with mutated EGF4KDEL 7mut. Mice were treated IP with 10 courses (40 injections) of EGF4KDEL 7mut beginning on day 5.

Figure 4

Total photon tumor activity of the individual mice. The data is pooled from Experiment 1 and 2. Mice were either A) untreated or B) treated with 10 courses of EGF4KDEL beginning on day 5. Data are expressed as total photon activity graphed over time for each individual animal. Bioluminescence intensity is shown as a function of photons/sec/sr/cm². A and B differ significantly (p<0.01). C) Representative mice were treated with only 3 courses of EGF4KDEL. Data are expressed as total photon activity graphed over time for each individual animal. Bioluminescence intensity is shown as a function of photons/sec/sr/cm².

Figure 5

EGF4KDEL 7mut has reduced immunogenicity, but not reduced activity A) To detect anti-toxin antibodies, immuncompetent BALB/c mice were immunized with either non-mutated parental EGF4KDEL or mutated EGF4KDEL. Serums from individual mice were analyzed in a modified ELISA measuring ug/ml anti-toxin IgG. Data were represented as the average ug IgG/ml. The two groups significantly differed (p=0.0317) B) To determine the effect of EGF4KDEL 7mut on MDA-MB-231 cells, increasing concentrations of EGF4KDEL 7mut was compared to non-mutated EGF4KDEL in a proliferation assay. C) To determine whether EGF4KDEL 7mut was efficacious, animals with MBA-MB-231 tumors were treated with 10 IP courses beginning on day 5. Total photon tumor activity was plotted for the individual mice. Bioluminescence intensity is shown as a function of photons/sec/sr/cm². D) The mean total photon activity is graphed for the 4 remission mice in C that were treated with EGF4KDEL 7 mut as well as two additional treatment groups given 2219KDEL using an identical schedule or a no treatment control group. The EGF4KDEL 7mut group statistically differed from other 2 groups (p<0.01). The one relapsed mouse M36 in the EGF4KDEL 7 mut group was not included in this graph.

Figure 6

The toxicity of mutated and non-mutated EGF4KDEL. A) Average body weight is plotted for the EGF4KDEL 7mut mice shown in Figure 5C and 5D. No significant differences were observed with the EGF4KDEL 7mut-treated mice compared to the 2219ARLKDEL 7mut and non treatment controls. B) Kidney section prepared by H & E staining and magnified 200x taken from 202 day survivor receiving multiple injections of BLT.