A recombinant immunotoxin against the tumor-associated antigen mesothelin reengineered for high activity, low off-target toxicity, and reduced antigenicity

Abstract

SS1P is a recombinant immunotoxin (RIT) engineered for the targeted elimination of malignant cells that express the tumor-associated antigen mesothelin. It is composed of an antimesothelin antibody variable fragment (Fv) linked to a cytotoxic fragment of Pseudomonas exotoxin A (PE) that includes domains II and III of native PE. The clinical use of SS1P is limited by its propensity to induce neutralizing antibodies and to cause a dose-limiting capillary leak syndrome (CLS) in patients. In this article, we describe a reengineered SS1P with improved properties that overcome these deficits. The redesign of SS1P consists of (i) removing the bulk of PE domain II (residues 251-273 and 284-394 of native PE), leaving only an 11-residue furin cleavage site, (ii) adding a Gly-Gly-Ser peptide linker after the furin cleavage site, and (iii) replacing eight highly solvent-exposed residues in the catalytic domain of PE. The new molecule, SS1-LR/GGS/8M, has cytotoxic activity comparable with SS1P on several mesothelin-expressing cell lines and remarkably improved activity on primary cells from patients with mesothelioma. In a mouse xenograft tumor model, high doses of SS1-LR/GGS/8M elicit antitumor activity superior to the activity of SS1P at its maximum-tolerated dose. In addition, SS1-LR/GGS/8M has greatly decreased ability to cause CLS in a rat model and reduced antigenicity or reactivity with antibodies to the sera of patients previously treated with SS1P.

Figure 1

Recombinant immunotoxins. The RIT SS1P consists of the ds V_H and V_L polypeptide chains of the Fv from the anti-mesothelin monoclonal antibody SS1 coupled to a 38 kDa fragment of PE38. Both the V_H and V_L fragments contain an internal disulfide bond in addition to the interchain disulfide bond engineered between the two fragments. A. A structural model of SS1P. The Fv is recombinantly joined to PE38, which is divided into domain II, domain III, and part of domain Ib from native PE38. Disulfide bonds are indicated as space-filling structures. The model contains a gap in the structure that corresponds to the deletion of residues 365-380 in domain Ib. B. A model of SS1-LR. SS1-LR lacks the majority of domains Ib and II of PE, but includes an 11-residue stretch of domain II that contains the furin cleavage site. Both models are hypothetical arrangements based on the structures of native PE and IgG; they do not represent actual structure determinations. C. Various constructs were created with mutations around the furin cleavage site of SS1-LR. The furin site from native PE is shaded. Black arrows indicate the position of cleavage between Arg-279 and Gly-280. SS1-LR/GGS R279G contains a point mutation (underlined) that prevents furin cleavage by replacing Arg-279 with Gly.

Figure 2

Relative cytotoxicity in vitro. A panel of eight cell lines that express mesothelin was used to compare the cytotoxicities of SS1P, SS1-LR, SS1-LR/GGS, and SS1-LR/GGS/8M. For each cell line, the average EC₅₀ of each RIT relative to the average EC₅₀ of SS1P is presented.

Figure 3

Rat model of capillary leak syndrome. A. Rats were grouped and treated intravenously as indicated, and evaluated for health prior to sacrifice after 24 hours. Hydrothorax fluid was removed and measured. B. The lungs of rats from each group were fixed, sectioned, and stained with hematoxylin and eosin. Representative microscopy fields at 200X magnification are shown.

Figure 4

Antitumor activity of SS1-LR/GGS/8M. Nude mice with A431/H9 xenograft tumors were grouped and treated intravenously with SS1-LR/GGS/8M. Initial treatment groups consisted of (A) vehicle (n=5), (B) 0.4 mg/kg SS1P (n=4), (C) 2.5 mg/kg SS1-LR/GGS/8M (n=6), and (D) 5.0 mg/kg SS1-LR/GGS/8M (n=6). Tumor size was measured over the course of 22 days. Experiments were subsequently repeated to evaluate the effect of using a greater number of doses and higher dose levels. These treatment groups consisted of (E) 5.0 mg/kg SS1-LR/GGS/8M (n=10) and (F) 10.0 mg/kg SS1-LR/GGS/8M (n=8). Tumor size was measured over the course of 31 days. All graphs indicate the tumor sizes of individual mice. Arrowheads indicate days when treatment was administered.

Figure 5

Human antigenicity of SS1-LR/GGS/8M. The reactivity of SS1P and SS1-LR/GGS/8M with preexisting antibodies in human sera was compared using a displacement assay to determine the concentration at which the two RITs reduced the signal of an ELISA to detect serum antibodies by 50% (IC₅₀). The IC₅₀ values of SS1-LR/GGS/8M relative to SS1P are plotted. Example curves for individual patient sera are provided in Supplementary Fig. S4.

Figure 6

Cytotoxicity of SS1-LR/GGS/8M on patient cells. Cells cultured from the pleural fluid or ascites of patients with mesothelioma were treated with increasing concentrations of SS1P (white bar) or SS1-LR/GGS/8M (grey bar). After 4 days, cells were evaluated for viability using a crystal violet assay. The mean values and standard errors from three replicates are plotted. Asterisks indicate significant differences of p < 0.01 (**), or p < 0.001 (***).