Advertisement
Preclinical Development

Single-Chain Antibody-Based Immunotoxins Targeting Her2/neu: Design Optimization and Impact of Affinity on Antitumor Efficacy and Off-Target Toxicity

Yu Cao, James D. Marks, Qian Huang, Stephen I. Rudnick, Chiyi Xiong, Walter N. Hittelman, Xiaoxia Wen, John W. Marks, Lawrence H. Cheung, Kim Boland, Chun Li, Gregory P. Adams and Michael G. Rosenblum
DOI: 10.1158/1535-7163.MCT-11-0519 Published January 2012

Abstract

Recombinant immunotoxins, consisting of single-chain variable fragments (scFv) genetically fused to polypeptide toxins, represent potentially effective candidates for cancer therapeutics. We evaluated the affinity of various anti-Her2/neu scFv fused to recombinant gelonin (rGel) and its effect on antitumor efficacy and off-target toxicity. A series of rGel-based immunotoxins were created from the human anti-Her2/neu scFv C6.5 and various affinity mutants (designated ML3-9, MH3-B1, and B1D3) with affinities ranging from 10−8 to 10−11 mol/L. Against Her2/neu-overexpressing tumor cells, immunotoxins with increasing affinity displayed improved internalization and enhanced autophagic cytotoxicity. Targeting indices were highest for the highest affinity B1D3/rGel construct. However, the addition of free Her2/neu extracellular domain (ECD) significantly reduced the cytotoxicity of B1D3/rGel because of immune complex formation. In contrast, ECD addition had little impact on the lower affinity constructs in vitro. In vivo studies against established BT474 M1 xenografts showed growth suppression by all immunotoxins. Surprisingly, therapy with the B1D3-rGel induced significant liver toxicity because of immune complex formation with shed Her2/neu antigen in circulation. The MH3-B1/rGel construct with intermediate affinity showed effective tumor growth inhibition without inducing hepatotoxicity or complex formation. These findings show that while high-affinity constructs can be potent antitumor agents, they may also be associated with mistargeting through the facile formation of complexes with soluble antigen leading to significant off-target toxicity. Constructs composed of intermediate-affinity antibodies are also potent agents that are more resistant to immune complex formation. Therefore, affinity is an exceptionally important consideration when evaluating the design and efficacy of targeted therapeutics. Mol Cancer Ther; 11(1); 143–53. ©2011 AACR.

Introduction

Immunotherapeutic approaches using antibodies have been widely explored against a variety of tumors, but an effective treatment of solid tumors remains a potential problem because therapeutic antibodies must diffuse into tumors through a disordered vasculature and against a hydrostatic pressure gradient (1, 2). Because low–molecular weight antibody fragments have been shown to have better tumor diffusion properties (3), single-chain variable fragments (scFv) were favored to deliver protein-based toxins to cancer cells (4, 5).

A variety of scFv-based immunotoxins have been engineered that are suitable for diverse therapeutic applications. An anti-CD174 scFv designated SGN-10 fused with Pseudomonas exotoxin (PE) was developed for optimal tumor penetration but clinical studies were limited by renal toxicity and gastritis (6, 7). Chaudhary and colleagues (8) and Powell and colleagues (9) generated LMB-2, anti-CD25 scFv-PE immunotoxin and described promising preclinical efficacy on malignant cells from patients with adult T-cell leukemia. However, common toxicities included transaminase evaluation. The therapeutic window for this class of constructs may be optimized by various design changes to lower the efficacious dose, improve specificity by reducing off-target effects, thereby allowing an increase in the maximal tolerated dose (10–12).

Tumor–antigen affinity and specificity of scFvs are important variables that may impact off-target tissue distribution and toxicity in vivo. These attributes have led to the commonly held concept that scFv must have high affinity to be therapeutically relevant. However, studies by Adams and colleagues (13) and Rudnick and Adams (14) have suggested that high-affinity scFv may be suboptimal vehicles and that lower affinity scFv appear to diffuse more uniformly throughout the tumor interior. In addition, because the presence of shed tumor antigen has the potential to misdirect the targeted constructs through immune complex formation (10, 15), higher affinity scFv could potentially be at risk compared with lower affinity constructs.

Although previous studies primarily focused on the in vivo behavior of scFv, few companion studies have been conducted to determine whether scFv-based immunotoxins display the same behavior with regard to the relationship between affinity, tumor penetration, tumor residence, and efficacy. Our present knowledge of the affinity/function relationship of scFv-based immunotoxins is insufficient to afford accurate predictions as to whether a given scFv is appropriate for toxin delivery. A comprehensive head-to-head comparison of recombinant immunotoxins with different affinities targeting the same epitope on an antigen would be useful to guide the developmental strategy for future immunotoxins.

We previously reported the construction and characterization of anti-Her2/neu immunotoxins constructed by fusing scFv C6.5 with the recombinant gelonin (rGel). These constructs showed highly efficient activity against Her2/neu-positive tumor cells (16). In the current study, we generated a series of rGel-containing fusion constructs composed of C6.5 and its mutants with varying affinities to Her2/neu and examined the impact of affinity on in vitro cytotoxicity, pharmacodynamics, and antitumor efficacy. In addition, we investigated the effect of antibody affinity on behavior in the presence of soluble antigen, formation of immune complexes, and the coincident development of off-target toxicity.

Materials and Methods

Plasmid construction

The gene encoding human anti-Her2/neu scFv (C6.5 and its affinity mutants, ML3-9, MH3-B1, and B1D3 created by site-directed amino acid substitutions in the CDR3s; ref. 17) were supplied by Dr. James D Marks (University of California, San Francisco, San Francisco, CA; Fig. 1A). Illustrations of the immunotoxin constructs are shown in Fig. 1B. Recombinant immunotoxins containing each scFv and rGel were constructed by overlapping PCR and were designated C6.5/rGel, ML3-9/rGel, MH3-B1/rGel, and B1D3/rGel, respectively.

Figure 1.
Figure 1.

Construction and preparation of scFv/rGel immunotoxins. A, amino acid mutations and affinity parameters of the C6.5 and its mutants, ML3-9, MH3-B1, and B1D3. The listed amino acids for each scFv indicate mutations to the sequence and the substituting amino acids. Dashes indicate no changes from the original sequence. B, schematic diagram of immunotoxin constructs containing scFv (C6.5, ML3-9, MH3-B1, or B1D3) and rGel. C, SDS-PAGE analysis of purified immunotoxins.

Protein expression and purification

The immunotoxins were expressed in Escherichia coli strain AD494 (DE3) pLysS and purified by immobilized metal affinity chromatography (IMAC) essentially as previously described (16).

Binding affinity of immunotoxins

The binding affinity and specificity of the immunotoxins were tested by ELISA on Her2/neu-positive (SKBR3, BT474 M1) and -negative (MCF7) cells. Rabbit anti-rGel antibody and horseradish peroxidase–conjugated goat anti-rabbit IgG were used as a tracer in this assay as described previously (16).

Internalization and competitive inhibition analysis

Immunofluorescence-based internalization studies were conducted on Her2/neu-positive (SKBR3, BT474 M1) and -negative (MCF7) cells. Immunofluorescence staining and competitive inhibition were analyzed as described in Supplementary Methods.

Cytotoxicity of scFv/rGel and competitive cytotoxicity assays

The cytotoxicity of immunotoxins on log-phase Her2/neu-positive and -negative cell lines were tested with the crystal violet staining method, and competitive assays were conducted as described in Supplementary Methods (18).

Western blot analysis of apoptosis and autophagy

The detection of apoptosis and autophagy on BT474 M1 cells treated with immunotoxins was analyzed as described in Supplementary Methods.

In vivo efficacy studies

BALB/c nude mice bearing subcutaneous BT474 M1 tumors were established and treated (intravenously, tail vein) with immunotoxins, as described in Supplementary Methods.

Tissue distribution study

The MH3-B1/rGel and B1D3/rGel was labeled with IRDye800CW according to the manufacturer's protocol. The tissue distribution assays and the imaging analysis are further described in Supplementary Methods.

Coimmunoprecipitation assay

Liver samples from mice after treatment with MH3-B1/rGel or B1D3/rGel were collected. Samples were examined for the presence of antigen:immunotoxin complexes as described in Supplementary Methods.

In situ immunofluorescent detection

Samples of liver tissues from mice were further prepared for immunofluorescence staining tracing Her2/neu antigen and scFv/rGel immunotoxins as described in Supplementary Methods.

Liver toxicity study

Hepatotoxicity was investigated by measuring activities of alanine transaminase (ALT), aspartate transaminase (AST), and lactate dehydrogenase (LDH) in collected serum from treated mice according to an assay kit (Roche). The histologic examination for hepatotoxicity was assessed by hematoxylin and eosin staining. Further details are presented in Supplementary Methods.

Statistical analysis

Statistical analyses were conducted with SPSS version 17.0.2 software (SPSS Inc.). Data were presented as mean ± SD, and significance was determined using a 2-sided Student t test, unless otherwise noted. A value of P < 0.05 was considered statistically significant.

Results

Preparation of scFv/rGel fusion constructs

The scFv/rGel constructs were created from human anti-Her2/neu scFv C6.5 and various affinity mutants (designated ML3-9, MH3-B1, and B1D3, in increasing affinity order). The affinities of the scFv ranged from 10−8 to 10−11 mol/L (Fig. 1A; refs. 17, 19). The immunotoxin genes were cloned into vector pET-32a(+) separately (Fig. 1B). Sequenced DNA clones were subsequently transformed into E. coli AD494 (DE3) pLysS for protein expression. As shown in Fig. 1C, after purification, all the immunotoxins migrated on SDS-PAGE at the expected molecular weight of 55 kDa under both reducing and nonreducing conditions.

Binding and cellular internalization of the fusion constructs

To ensure that immunotoxins retained antigen-binding ability, the fusion proteins were compared in an ELISA-based–binding assay using Her2/neu-positive (SKBR3, BT474 M1) and -negative (MCF7) cells. All the scFv/rGel constructs showed specific and significant ELISA binding to Her2/neu-positive cells with negligible binding to negative cells (Fig. 2A). The equilibrium dissociation constants (Kd) were calculated (GraphPad Prism), and the affinities of immunotoxins for BT474 M1 cells were found to be 53.13 nmol/L (C6.5/rGel), 1.45 nmol/L (ML3-9/rGel), 0.18 nmol/L (MH3-B1/rGel), and 27 pmol/L (B1D3/rGel). The correlation between the Kd values of the scFvs and fusion constructs was found to be significant with a correlation coefficient of 0.939 (P < 0.01), indicating that introduction of the rGel component did not affect the binding affinity of the scFv.

Figure 2.
Figure 2.

Characterization and comparison of scFv/rGel immunotoxins. A, evaluation binding activity of the scFv/rGel to Her2/neu-positive (SKBR3, BT474 M1) and -negative (MCF7) cells by whole-cell ELISA. B, quantification of internalization rate of the immunotoxins on Her2/neu-positive and -negative cells. Cells were subjected to immunofluorescent staining with anti-rGel antibody (fluorescein isothiocyanate–conjugated secondary antibody). The bar graphs were calculated from relative fluorescence estimation, and the values are expressed as mean ± SD (n > 50).

We next examined whether the various affinity scFv/rGel fusions could specifically internalize into target cells. Immunofluorescence staining was conducted on Her2/neu-positive and -negative cells. As quantified by relative fluorescence (Fig. 2B), the internalization efficiency exhibited a moderate increase with increasing binding affinity in Her2/neu-positive cells. For BT474 M1 cells, the relative fluorescence intensities were 56.30 (C6.5/rGel), 73.69 (ML3-9/rGel), 86.29 (MH3-B1/rGel), and 90.41 (B1D3/rGel). There was a good correlation of between increases in apparent affinity and internalization efficiency (r2 = 0.8289; P < 0.01) indicating that efficient binding to the cell surface appears to be primarily responsible for rapid internalization after cell exposure.

In vitro cytotoxicity of scFv/rGel fusion constructs

All the scFv/rGel constructs and rGel were tested against a number of different tumor cell lines (Table 1). As expected, there appeared to be a good correlation (r2 = 0.7812; P < 0.01) between apparent affinity and IC50 values. Targeting indices were found to be highest for the highest affinity B1D3/rGel construct. This study showed that for the scFv/rGel immunotoxins, binding affinity appears to mediate internalization efficiency and this appeared to directly impact the overall cytotoxic effects observed. Furthermore, against Her2/neu-negative cells, there was little or no specific cytotoxicity of the constructs compared with rGel itself.

View this table:
Table 1.

Comparative IC50 values of fusion constructs against various types of tumor cell lines

Effects of various fusion constructs on cytotoxic mechanisms

The cytotoxic effects mediated by scFv/rGel immunotoxins were analyzed in BT474 M1 cells. As shown in Fig. 3A, scFv/rGel fusions did not activate caspase-dependent apoptosis in target cells, showing no cleavage of the caspase substrate PARP. We next assessed LDH release and found that exposure of BT474 M1 cells to immunotoxins did not induce necrotic cell death (data not shown).

Figure 3.
Figure 3.

Cell-killing mechanism analysis of the immunotoxins on BT474 M1 cells. A, Western blot analysis of PARP cleavage after 24 and 48 hours of scFv/rGel fusion treatment. B, analysis of LC3 after treatment with the scFv/rGel fusions for 24 and 48 hours. C, analysis of cell extract and medium for HMGB1 protein after immunotoxin treatment for 48 hours.

Then, we examined whether the cytotoxic effects of these immunotoxins activate autophagic signaling. As shown in Fig. 3B, the ratio of LC3-II formation to the β-actin control was shown to be increased after treatment with the fusion constructs, showing that autophagic flux was induced by rGel-based immunotoxins in BT474 M1 cells. In addition, autophagic induction by fusion constructs was further validated by the observed selective release of cellular HMGB1 (Fig. 3C; ref. 20). These data indicated that the observed cytotoxic effects of scFv/rGel fusions in BT474 M1 cells appeared to be mediated not through an apoptotic or necrotic mechanisms but by the efficient induction of autophagic cell death.

Influence of soluble Her2/neu extracellular domain on immunotoxin activity

Shedding of target antigen from the surface of tumor cells into circulation may present obstacles for antibodies to effectively target tumor cells in vivo (21). To investigate the impact of soluble antigen on scFv/rGel immunotoxins, we evaluated the internalization of the immunotoxins in the presence of free Her2/neu extracellular domain (ECD) in BT474 M1 cells (Fig. 4A and Supplementary Fig. S1). The addition of ECD reduced the internalization for all the constructs. The highest affinity B1D3/rGel construct showed a significant reduction (P < 0.01), whereas the lower affinity fusions exhibited the lesser impact of ECD on internalization.

Figure 4.
Figure 4.

Competitive analysis of the scFv/rGel in the presence of free Her2/neu ECD. A, quantification of competitive internalization rate of the fusion proteins on BT474 M1 cells. Cells were treated with the mixture of 20 nmol/L immunotoxins and different concentration of ECD. Values are expressed as mean ± SD (n > 50). B, coimmunoprecipitation of scFv/rGel and ECD complex. The mixture supernatants were subjected to Her2/neu immunoprecipitation (IP), followed by immunoblotting (IB) for rGel. C, competitive cytotoxicity of the immunotoxins in the presence of 20 nmol/L ECD. Values are presented as IC50.

We next applied coimmunoprecipitation to examine whether the decreased internalization observed was because of the immune complexes of the immunotoxins with ECD. As shown in Fig. 4B, the highest affinity construct (B1D3/rGel) and the lowest affinity construct (C6.5/rGel) formed the highest and lowest amount of immune complex with ECD, with intermediate-affinity molecules (ML3-9/rGel and MH3-B1/rGel) generating intermediate levels of immune complexes. Therefore, the significant reduction in cell internalization observed with B1D3/rGel fusion was the result of immune complexes formation with soluble ECD, further preventing binding of the immunotoxins via cell-associated antigen.

A competitive cytotoxicity assay was conducted on SKBR3 and BT474 M1 cells by adding 20 nmol/L ECD to various concentrations of each fusion construct (Fig. 4C). All the fusion constructs showed an increase in IC50 value in the presence of ECD. Constructs with low and medium affinity showed the least impact of ECD on cytotoxic effects whereas B1D3/rGel showed the greatest influence. Furthermore, the addition of various concentration of ECD to a fixed (20 nmol/L) dose of immunotoxins showed similar effects (Supplementary Fig. S2). B1D3/rGel was impacted to the greatest extent in the presence of ECD, whereas the constructs with low or medium affinity showed less impact on cytotoxicity.

Her2/neu antigen shed from tumor cells

Because the Her2/neu antigen can be shed from target cells and may impact the cytotoxic effects observed with scFv/rGel immunotoxins, we measured the endogenous Her2/neu antigen levels in cell media and serum from mice bearing tumor xenografts by quantitative ELISA. The medium of Her2/neu-positive cells (SKBR3, BT474 M1, NCI-N87, and Calu3) were collected daily for 7 days. As shown in Supplementary Fig. S3A, the Her2/neu antigen was present in the medium of all Her2/neu-positive cells, and the concentrations increased relative to cell number. Levels of antigen in culture media were ≤0.4 nmol/L from all lines tested during a 72-hour cytotoxicity assay and appeared to be well below levels that would impact immunotoxin efficacy.

Measurements of shed Her2/neu antigen in the blood of mice bearing BT474 M1 tumor showed that Her2/neu levels increase in parallel with tumor size (Supplementary Fig. S3B). Levels of Her2/neu in serum increased from 2 to 12 nmol/L for mice with 200 mm3 tumors up to 1,800 mm3, respectively. The correlation between shed Her2/neu levels and tumor volume was found to be significant with a correlation coefficient of 0.797 (P < 0.01). At sufficiently high Her2/neu levels, the efficacy of high-affinity–targeted therapeutics could be impacted.

Antitumor activity of scFv/rGel fusions in xenograft models

We next evaluated the ability of various scFv/rGel immunotoxins to inhibit the growth of established BT474 M1 tumor xenografts in BALB/c nude mice after systemic administration. BT474 M1 cells were implanted into mice and tumors were allowed to grow to 200 mm3 in volume. Mice were then treated with each fusion protein and rGel as control at a total dose of 24 mg/kg. As shown in Fig. 5A, treatment with the scFv/rGel fusions all showed great antitumor effects, with the intermediate-affinity MH3-B1/rGel showing more enhanced and long-lasting tumor inhibition effects than lower affinity C6.5/rGel and ML3-9/rGel. There was little obvious toxicity observed with the administration of the immunotoxins with the exception of the highest affinity B1D3/rGel. As shown in Fig. 5B, mice treated with this agent showed considerable body weight loss (∼27%), and all the mice in this group died after fourth injection. Further studies were initiated to examine the reason of the toxicity of B1D3/rGel compared with other constructs.

Figure 5.
Figure 5.

In vivo study of the immunotoxins against BT474 M1 tumor xenografts in nude mice. A, treatment of BT474 M1 tumors with immunotoxins at the dose of 24 mg/kg. Mean tumor volume was calculated by W × L × H as measured by digital calipers. B, average body weight of the mice during the immunotoxin treatment. C and D, whole body imaging results of the mice intravenously injected with either IR-MH3-B1/rGel or IR-B1D3/rGel at 4, 24, 48, and 72 hours. Arrows point at tumor (T) and liver (L). E and F, the comparison of tissue-to-muscle ratio of IR-MH3-B1/rGel and IR-B1D3/rGel at 24 and 72 hours after injection into xenograft mice (n = 5).

In vivo optical imaging

To examine the in vivo toxicity observed in mice treated with the highest affinity B1D3/rGel, we used a fluorescent molecular imaging probe (IRDye 800CW) to label MH3-B1/rGel and B1D3/rGel for in vivo biodistribution studies (22). Nude mice bearing BT474 M1 tumors were injected (intravenously) with 1.5 nmol/L IRDye800-MH3-B1/rGel (IR-MH3-B1/rGel) or IRDye800-B1D3/rGel (IR-B1D3/rGel). The mice were then imaged at different times (4, 24, 48, and 72 hours) with the IVIS Optical Imaging System (Fig. 5C and D). Both IR-MH3-B1/rGel and IR-B1D3/rGel were shown to accumulate in BT474 M1 tumors with the first 4 hours of postinjection (tumor to contralateral background ratio of 1.85 ± 0.12 and 1.85 ± 0.09 for each) and reached maximal concentrations at 48 hours (tumor to contralateral background ratio of 2.70 ± 0.31 and 2.58 ± 0.15, respectively). In addition, accumulation of IR-B1D3/rGel was observed in the liver postinjection from 48 to 72 hours, compared with IR-MH3-B1/rGel.

To avoid any measuring errors caused by limited tissue penetration of fluorophores, animals were sacrificed; tumor and major organs were collected at the 24-hour and 72-hour time points and were subjected immediately to NIRF imaging (Supplementary Fig. S4). At 24 hours postinjection, there were no significant differences between the tissue distribution of IR-MH3-B1/rGel and IR-B1D3/rGel (Fig. 5E and Supplementary Fig. S5A). However, a better biodistribution of IR-MH3-B1/rGel over IR-B1D3/rGel could be identified 72 hours after injection, with the tissue-to-muscle ratio of tumor being 1.5 times higher than the later one (e.g. 2.84 ± 0.23 vs. 1.89 ± 0.23, P < 0.002; Fig. 5F and Supplementary Fig. S5B). Correspondingly, a 1.6-fold higher liver retention of IR-B1D3/rGel over IR-MH3-B1/rGel was observed (4.65 ± 0.61 vs. 2.89 ± 0.56, respectively; P < 0.01). Both IR-MH3-B1/rGel and IR-B1D3/rGel were found to accumulate in the kidneys likely due to renal clearance of the low–molecular weight agents. In the other major tissues we collected, there were no significant differences in the distribution of these immunotoxins to tissues such as heart, lung, spleen, muscle.

Accumulation of Her2/neu antigen and B1D3/rGel in the liver

To identify the cause of the liver distribution with the highest affinity B1D3/rGel, we administered B1D3/rGel and MH3-B1/rGel to nude mice with or without tumors at a dose of 1.5 nmol/L. Mice were sacrificed 72 hours after injection. After homogenization, the liver samples were subjected to immunoprecipitation using anti-Her2/neu antibody and immunoblotted to assess the fusions. As shown in Fig. 6A, we found immune complexes of Her2/neu antigen and the B1D3/rGel in tumor-bearing mice but not in tumor-free mice. In contrast, we found lower amounts of immune complexes in the livers of mice treated with MH3-B1/rGel.

Figure 6.
Figure 6.

Analysis of the accumulation of Her2/neu antigen and scFv/rGel (MH3-B1/rGel or B1D3/rGel) complex–driven liver toxicity after 72 hours of injection. A, coimmunoprecipitation of Her2/neu antigen and scFv/rGel in the liver from the mice with or without BT474 M1 tumors. Liver homogenates were subjected to Her2/neu immunoprecipitation (IP), followed by Western blotting for rGel. IB, immunoblotting. B, immunofluorescence colocalization of scFv/rGel (green) and Her2/neu antigen (red) in the liver of the BT474 M1 burden mice. C, enzymatic activities of ALT in the serum of mice with or without tumors, treated with rGel, MH3-B1/rGel, or B1D3/rGel. Values are expressed as mean ± SD (n = 5). ***, statistically different from own control mice at P < 0.002. D, histologic findings of the livers from tumor burden mice after immunotoxin application. All the sections were stained with hematoxylin and eosin (primary magnification, ×100). Inset shows multiple necrosis of hepatocytes (primary magnification, ×400).

Immunofluorescence staining confirmed colocalization and accumulation of B1D3/rGel with Her2/neu antigen in the liver of tumor-bearing mice, but less staining of antigen was observed from the livers in the MH3-B1/rGel treatment group (Fig. 6B).

In vivo toxicity of immunotoxins in mice

We then examined the influence of tumor-derived shed antigen on the hepatotoxicity of the B1D3 and MH3-B1 fusion toxins. Serum samples were collected 72 hours after the administration of the fusions or rGel to mice with or without tumors, and enzymatic activities of liver enzymes were determined (Fig. 6C and Supplementary Fig. S6). We found slight increases in the serum ALT, AST, and LDH levels in tumor-bearing versus tumor-free mice treated with rGel or MH3-B1/rGel. The greatest increases in all 3 markers were found with tumor-bearing mice treated with the high-affinity B1D3/rGel construct. This increase was not observed in tumor-free mice confirming that the observed hepatotoxicity was due to immune complexes of B1D3/rGel and Her2/neu antigen localizing in the liver.

To verify the hepatotoxicity in animals, mouse livers were harvested 72 hours after immunotoxin injection and examined. Compared with animals treated with rGel or MH3-B1/rGel, the livers of mice treated with B1D3/rGel showed severe liver damage characterized by marked necrosis and vacuolar degeneration of hepatocytes and extensive hemorrhage (Fig. 6D). This suggests that the in vivo efficacy of high-affinity immunotoxins can be significantly impaired by the presence of shed antigen. Furthermore, the immune complexes formed by the immunotoxins and the shed antigen contribute to significant hepatotoxicity.

Discussion

Using a panel of scFv/rGel fusions specific for the same Her2/neu epitope (17) and nearly identical in their sequences and structure, we were able to assess the influence of affinity. This appears to be one of the first comprehensive examinations of the impact of affinity on the in vitro and in vivo behavior of immunotoxins. More importantly, the current study clearly described the increased potential for high-affinity immunotoxins to form immune complexes in vivo resulting in off-target toxicity in liver responsible for clearing these complexes. These data have potential relevance to a number of anti-Her2/neu approaches (23, 24), as well as other targets for which there is a level of circulating antigen present (25, 26).

Previous studies suggested that the binding affinity for antigen plays a pivotal role in the total concentration and penetration of scFv into tumors (13, 14). On the basis of scFv/rGel immunotoxins, we showed that increasing affinity could improve cell-binding ability, internalization efficiency, and autophagic cytotoxicity on Her2/neu-positive cells. However, fusion toxins with a 10-fold increase in affinity did not show either a corresponding improvement in internalization or a concomitant improvement in cytotoxic effects. This suggests that the internalization rate of the construct may primarily be associated with Her2/neu receptor recycling and the rate of antigen endocytosis and appears to be largely unaffected by the affinity of immunotoxin binding (27, 28). Moreover, intracellular trafficking and distribution of the toxin component to the ribosomal compartment may be critical factors that can define immunotoxin sensitivity.

In addition to the characteristics of the targeting moiety, shed antigen levels are clearly a factor in the potential therapeutic application of targeted agents. Pharmacokinetic studies of anti-Her2/neu antibodies in patients showed an inverse association between serum concentrations of antibody and the levels of shed Her2/neu antigen (29, 30). This observation may be explained, in part, by formation of soluble antigen:antibody complexes leading to a more rapid clearance by the reticuloendothelial system (RES). Clinical studies of Herceptin showed that Her2/neu plasma concentrations greater than 500 ng/mL (∼10 nmol/L) were associated with shorter serum half-life and subtherapeutic trough levels of the antibody (31, 32). We showed the presence of shed Her2/neu antigen in cell culture medium and in blood from tumor-bearing mice in the nanomolar range. In vitro studies showed that the activity of the highest affinity B1D3/rGel construct was the most vulnerable to soluble Her2/neu antigen compared with the immunotoxins with lower affinity. As a result, soluble Her2/neu antigen formed immune complexes with B1D3/rGel leading to a reduction in in vitro cytotoxicity and a significant increase in in vivo toxicity. Biodistribution studies in BT474 M1 tumor–bearing mice indicated that there was a statistically significant decrease in tumor localization for B1D3/rGel compared with MH3-B1/rGel 72 hours after injection and a corresponding increase in liver accumulation. These data are consistent with the finding that B1D3/rGel high-affinity construct formed immune complexes with soluble, tumor-derived antigen leading to clearance by the hepatic RES and a reduction in ability to distribute to the tumor.

The presence of hepatotoxicity has limited the clinical dose-escalation of some immunotoxins. Anti-Her2/neu immunotoxins containing PE have shown potent antitumor activity in animal models (23, 33) but resulted in unexpected hepatotoxicity in all patients likely due to the normal presence of Her2/neu on hepatocytes (34, 35). Studies by Onda and colleagues found that overall positive charge on anti-Tac(Fv) PE38 immunotoxin contributed to nonspecific binding to liver cells and resulted in dose-limiting liver toxicity (36, 37). In this study, analyzing the behavior of different affinity immunotoxins offered a comprehensive insight into the relationships between antibody affinity, scFv/rGel tumor uptake, in vivo efficacy, toxicity, and hepatic accumulation of immune complexes. The importance of complex formation leading to toxicity appears to be underscored by lethal toxicity observed in tumor-bearing but not in tumor-free mice after administration of the high-affinity immunotoxin.

Shedding of cell surface antigens is an important biologic process that is used by cells to modulate responses to signals in the extracellular environment (38, 39). Our biodistribution study indicated that there was little difference in the liver and tumor uptake of MH3-B1/rGel and B1D3/rGel at 24 hours, but there was a significant difference between the two at 72 hours. These results suggested that complex formation between the Her2/neu antigen and B1D3/rGel immunotoxin may play a role in later events in the biodistribution patterns of high-affinity constructs. The observations were consistent with our findings that clearance of antigen:immunotoxin complexes are an important event in the observed off-target toxicity found in liver and that high-affinity antibodies are most susceptible to formation of these complexes.

The emergence of clinically relevant, nonspecific toxicity has limited the therapeutic potential of numerous immunotoxins (40–42) and is a significant limitation to the clinical development of this class of molecules in general. Therefore, it is of paramount importance to design immunotoxins with minimal toxicity to normal tissues. Previous concepts focused on using the highest affinity antibodies available for use in immunotoxin constructs. The current study suggests that high-affinity constructs are the most vulnerable to the presence of shed antigen. Intermediate-affinity antibodies appear to be preferable for this application because their targeting ability is high enough to show reliable specificity and targeted toxicity but low enough to avoid interference by shed antigen.

The toxicity of Her2/neu constructs to liver as a result of normal tissue expression is a common problem that has been addressed in clinical trials. Because the scFv antibodies we used for this study do not cross-react with murine Her2/neu (17, 19), antigen expression leading to toxicity was therefore not a factor contributing to hepatic toxicity. Our study clearly showed that when high-affinity immunotoxins form complexes in vivo with tumor-derived Her2/neu antigen, this, in addition, results in significantly increased hepatic toxicity.

It is important to note that murine studies of species-specific anti-human Her2/neu antibody constructs may underrepresent the eventual clinical toxicity of these agents against normal tissues expressing low levels of Her2/neu antigen. Such dose-limiting toxicities include renal and liver toxicity as well as vascular leak syndrome (43–45). The use of rGel-based constructs may be particularly important to be considered in cases of low-level expression of antigen on normal tissues because previous studies have shown that there appears to be a relatively high minimal threshold of Her2/neu antigen sites (∼150,000) present on a target cell before specific toxicity is enabled. This appears to be unique compared with studies with other toxins.

The present study suggests that the efficacy of high-affinity immunotoxins appears to be most easily impaired by the presence of circulating Her2/neu antigen and the toxicity of these agents is more easily increased because of the facile formation of immune complexes leading to toxicity in RES organs (i.e., liver). The use of fusion constructs with intermediate affinity appears to be most appropriate because their antitumor efficacy and toxicity do not appear to be significantly impacted by the presence of soluble antigen. These factors should be taken into consideration when designing antitumor immunotoxins as cancer therapeutics.

Disclosure of Potential Conflicts of Interest

M.G. Rosenblum has a patent with the Clayton Foundation for Research. No potential conflicts of interest were disclosed by the other authors.

Grant Support

This work was conducted by the Clayton Foundation for Research and supported by NCI grant R01 CA118159 (G.P. Adams).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Acknowledgments

The authors thank Heidi Simmons of the FCCC Developmental Therapeutics Program for her excellent technical assistance as well as Dr. Margaret Einarson and the FCCC High Throughput Screening and Translational Research Facility.

Footnotes

  • Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).

  • Received July 15, 2011.
  • Revision received November 8, 2011.
  • Accepted November 8, 2011.

References

  1. 1.↵
    1. Jain RK
    . Delivery of molecular and cellular medicine to solid tumors. Adv Drug Deliv Rev 2001;46:149–68.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Markman B,
    2. Tabernero J
    . Monoclonal antibodies in solid tumours. Curr Clin Pharmacol 2010;5:160–5.
    OpenUrlCrossRefPubMed
  3. 3.↵
    1. Kreitman RJ,
    2. Pastan I
    . Immunotoxins in the treatment of hematologic malignancies. Curr Drug Targets 2006;7:1301–11.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Beckman RA,
    2. Weiner LM,
    3. Davis HM
    . Antibody constructs in cancer therapy: protein engineering strategies to improve exposure in solid tumors. Cancer 2007;109:170–9.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Schaedel O,
    2. Reiter Y
    . Antibodies and their fragments as anti-cancer agents. Curr Pharm Des 2006;12:363–78.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Posey JA,
    2. Khazaeli MB,
    3. Bookman MA,
    4. Nowrouzi A,
    5. Grizzle WE,
    6. Thornton J,
    7. et al.
    A phase I trial of the single-chain immunotoxin SGN-10 (BR96 sFv-PE40) in patients with advanced solid tumors. Clin Cancer Res 2002;8:3092–9.
    OpenUrlAbstract/FREE Full Text
  7. 7.↵
    1. Smith SV
    . SGN-10. Seattle genetics. Curr Opin Investig Drugs 2001;2:1314–9.
    OpenUrlPubMed
  8. 8.↵
    1. Chaudhary VK,
    2. Queen C,
    3. Junghans RP,
    4. Waldmann TA,
    5. FitzGerald DJ,
    6. Pastan I
    . A recombinant immunotoxin consisting of two antibody variable domains fused to Pseudomonas exotoxin. Nature 1989;339:394–7.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Powell DJ Jr.,
    2. Felipe-Silva A,
    3. Merino MJ,
    4. Ahmadzadeh M,
    5. Allen T,
    6. Levy C,
    7. et al.
    Administration of a CD25-directed immunotoxin, LMB-2, to patients with metastatic melanoma induces a selective partial reduction in regulatory T cells in vivo . J Immunol 2007;179:4919–28.
    OpenUrlAbstract/FREE Full Text
  10. 10.↵
    1. Pennell CA,
    2. Erickson HA
    . Designing immunotoxins for cancer therapy. Immunol Res 2002;25:177–91.
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Kreitman RJ
    . Immunotoxins for targeted cancer therapy. AAPS J 2006;8:E532–51.
    OpenUrlCrossRefPubMed
  12. 12.↵
    1. Pastan I,
    2. Hassan R,
    3. Fitzgerald DJ,
    4. Kreitman RJ
    . Immunotoxin therapy of cancer. Nat Rev Cancer 2006;6:559–65.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Adams GP,
    2. Schier R,
    3. McCall AM,
    4. Simmons HH,
    5. Horak EM,
    6. Alpaugh RK,
    7. et al.
    High affinity restricts the localization and tumor penetration of single-chain fv antibody molecules. Cancer Res 2001;61:4750–5.
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    1. Rudnick SI,
    2. Adams GP
    . Affinity and avidity in antibody-based tumor targeting. Cancer Biother Radiopharm 2009;24:155–61.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Zhang Y,
    2. Pastan I
    . High shed antigen levels within tumors: an additional barrier to immunoconjugate therapy. Clin Cancer Res 2008;14:7981–6.
    OpenUrlAbstract/FREE Full Text
  16. 16.↵
    1. Cao Y,
    2. Marks JD,
    3. Marks JW,
    4. Cheung LH,
    5. Kim S,
    6. Rosenblum MG
    . Construction and characterization of novel, recombinant immunotoxins targeting the Her2/neu oncogene product: in vitro and in vivo studies. Cancer Res 2009;69:8987–95.
    OpenUrlAbstract/FREE Full Text
  17. 17.↵
    1. Schier R,
    2. McCall A,
    3. Adams GP,
    4. Marshall KW,
    5. Merritt H,
    6. Yim M,
    7. et al.
    Isolation of picomolar affinity anti-c-erbB-2 single-chain Fv by molecular evolution of the complementarity determining regions in the center of the antibody binding site. J Mol Biol 1996;263:551–67.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Rosenblum MG,
    2. Cheung LH,
    3. Liu Y,
    4. Marks JW III.
    . Design, expression, purification, and characterization, in vitro and in vivo, of an antimelanoma single-chain Fv antibody fused to the toxin gelonin. Cancer Res 2003;63:3995–4002.
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    1. Schier R,
    2. Marks JD,
    3. Wolf EJ,
    4. Apell G,
    5. Wong C,
    6. McCartney JE,
    7. et al.
    In vitro and in vivo characterization of a human anti-c-erbB-2 single-chain Fv isolated from a filamentous phage antibody library. Immunotechnology 1995;1:73–81.
    OpenUrlCrossRefPubMed
  20. 20.↵
    1. Thorburn J,
    2. Horita H,
    3. Redzic J,
    4. Hansen K,
    5. Frankel AE,
    6. Thorburn A
    . Autophagy regulates selective HMGB1 release in tumor cells that are destined to die. Cell Death Differ 2009;16:175–83.
    OpenUrlCrossRefPubMed
  21. 21.↵
    1. Carter PJ,
    2. Senter PD
    . Antibody-drug conjugates for cancer therapy. Cancer J 2008;14:154–69.
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Adams KE,
    2. Ke S,
    3. Kwon S,
    4. Liang F,
    5. Fan Z,
    6. Lu Y,
    7. et al.
    Comparison of visible and near-infrared wavelength-excitable fluorescent dyes for molecular imaging of cancer. J Biomed Opt 2007;12:024017.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. King CR,
    2. Fischer PH,
    3. Rando RF,
    4. Pastan I
    . The performance of e23(Fv)PEs, recombinant toxins targeting the erbB-2 protein. Semin Cancer Biol 1996;7:79–86.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Mazor Y,
    2. Noy R,
    3. Wels WS,
    4. Benhar I
    . chFRP5-ZZ-PE38, a large IgG-toxin immunoconjugate outperforms the corresponding smaller FRP5(Fv)-ETA immunotoxin in eradicating ErbB2-expressing tumor xenografts. Cancer Lett 2007;257:124–35.
    OpenUrlCrossRefPubMed
  25. 25.↵
    1. Chan SY,
    2. Gordon AN,
    3. Coleman RE,
    4. Hall JB,
    5. Berger MS,
    6. Sherman ML,
    7. et al.
    A phase 2 study of the cytotoxic immunoconjugate CMB-401 (hCTM01-calicheamicin) in patients with platinum-sensitive recurrent epithelial ovarian carcinoma. Cancer Immunol Immunother 2003;52:243–8.
    OpenUrlPubMed
  26. 26.↵
    1. Zhang Y,
    2. Xiang L,
    3. Hassan R,
    4. Pastan I
    . Immunotoxin and Taxol synergy results from a decrease in shed mesothelin levels in the extracellular space of tumors. Proc Natl Acad Sci U S A 2007;104:17099–104.
    OpenUrlAbstract/FREE Full Text
  27. 27.↵
    1. Ackerman ME,
    2. Pawlowski D,
    3. Wittrup KD
    . Effect of antigen turnover rate and expression level on antibody penetration into tumor spheroids. Mol Cancer Ther 2008;7:2233–40.
    OpenUrlAbstract/FREE Full Text
  28. 28.↵
    1. Austin CD,
    2. De Maziere AM,
    3. Pisacane PI,
    4. van Dijk SM,
    5. Eigenbrot C,
    6. Sliwkowski MX,
    7. et al.
    Endocytosis and sorting of ErbB2 and the site of action of cancer therapeutics trastuzumab and geldanamycin. Mol Biol Cell 2004;15:5268–82.
    OpenUrlAbstract/FREE Full Text
  29. 29.↵
    1. Pegram MD,
    2. Lipton A,
    3. Hayes DF,
    4. Weber BL,
    5. Baselga JM,
    6. Tripathy D,
    7. et al.
    Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients with HER2/neu-overexpressing metastatic breast cancer refractory to chemotherapy treatment. J Clin Oncol 1998;16:2659–71.
    OpenUrlAbstract
  30. 30.↵
    1. Pegram M,
    2. Hsu S,
    3. Lewis G,
    4. Pietras R,
    5. Beryt M,
    6. Sliwkowski M,
    7. et al.
    Inhibitory effects of combinations of HER-2/neu antibody and chemotherapeutic agents used for treatment of human breast cancers. Oncogene 1999;18:2241–51.
    OpenUrlCrossRefPubMed
  31. 31.↵
    1. Baselga J,
    2. Tripathy D,
    3. Mendelsohn J,
    4. Baughman S,
    5. Benz CC,
    6. Dantis L,
    7. et al.
    Phase II study of weekly intravenous recombinant humanized anti-p185HER2 monoclonal antibody in patients with HER2/neu-overexpressing metastatic breast cancer. J Clin Oncol 1996;14:737–44.
    OpenUrlAbstract/FREE Full Text
  32. 32.↵
    1. Leary AF,
    2. Hanna WM,
    3. van de Vijver MJ,
    4. Penault-Llorca F,
    5. Ruschoff J,
    6. Osamura RY,
    7. et al.
    Value and limitations of measuring HER-2 extracellular domain in the serum of breast cancer patients. J Clin Oncol 2009;27:1694–705.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    1. Wels W,
    2. Harwerth IM,
    3. Mueller M,
    4. Groner B,
    5. Hynes NE
    . Selective inhibition of tumor cell growth by a recombinant single-chain antibody-toxin specific for the erbB-2 receptor. Cancer Res 1992;52:6310–7.
    OpenUrlAbstract/FREE Full Text
  34. 34.↵
    1. Pai-Scherf LH,
    2. Villa J,
    3. Pearson D,
    4. Watson T,
    5. Liu E,
    6. Willingham MC,
    7. et al.
    Hepatotoxicity in cancer patients receiving erb-38, a recombinant immunotoxin that targets the erbB2 receptor. Clin Cancer Res 1999;5:2311–5.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Azemar M,
    2. Djahansouzi S,
    3. Jager E,
    4. Solbach C,
    5. Schmidt M,
    6. Maurer AB,
    7. et al.
    Regression of cutaneous tumor lesions in patients intratumorally injected with a recombinant single-chain antibody-toxin targeted to ErbB2/HER2. Breast Cancer Res Treat 2003;82:155–64.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Onda M,
    2. Willingham M,
    3. Wang QC,
    4. Kreitman RJ,
    5. Tsutsumi Y,
    6. Nagata S,
    7. et al.
    Inhibition of TNF-alpha produced by Kupffer cells protects against the nonspecific liver toxicity of immunotoxin anti-Tac(Fv)-PE38, LMB-2. J Immunol 2000;165:7150–6.
    OpenUrlAbstract/FREE Full Text
  37. 37.↵
    1. Onda M,
    2. Kreitman RJ,
    3. Vasmatzis G,
    4. Lee B,
    5. Pastan I
    . Reduction of the nonspecific animal toxicity of anti-Tac(Fv)-PE38 by mutations in the framework regions of the Fv which lower the isoelectric point. J Immunol 1999;163:6072–7.
    OpenUrlAbstract/FREE Full Text
  38. 38.↵
    1. Arribas J,
    2. Borroto A
    . Protein ectodomain shedding. Chem Rev 2002;102:4627–38.
    OpenUrlCrossRefPubMed
  39. 39.↵
    1. Dello SP,
    2. Rovida E
    . Transmodulation of cell surface regulatory molecules via ectodomain shedding. Biol Chem 2002;383:69–83.
    OpenUrlCrossRefPubMed
  40. 40.↵
    1. Ghetie MA,
    2. Ghetie V,
    3. Vitetta ES
    . Immunotoxins for the treatment of B-cell lymphomas. Mol Med 1997;3:420–7.
    OpenUrlPubMed
  41. 41.↵
    1. Bang S,
    2. Nagata S,
    3. Onda M,
    4. Kreitman RJ,
    5. Pastan I
    . HA22 (R490A) is a recombinant immunotoxin with increased antitumor activity without an increase in animal toxicity. Clin Cancer Res 2005;11:1545–50.
    OpenUrlAbstract/FREE Full Text
  42. 42.↵
    1. Mathew M,
    2. Verma RS
    . Humanized immunotoxins: a new generation of immunotoxins for targeted cancer therapy. Cancer Sci 2009;100:1359–65.
    OpenUrlCrossRefPubMed
  43. 43.↵
    1. Baluna R,
    2. Vitetta ES
    . Vascular leak syndrome: a side effect of immunotherapy. Immunopharmacology 1997;37:117–32.
    OpenUrlCrossRefPubMed
  44. 44.↵
    1. Vallera DA,
    2. Panoskaltsis-Mortari A,
    3. Blazar BR
    . Renal dysfunction accounts for the dose limiting toxicity of DT390anti-CD3sFv, a potential new recombinant anti-GVHD immunotoxin. Protein Eng 1997;10:1071–6.
    OpenUrlAbstract/FREE Full Text
  45. 45.↵
    1. Brown J,
    2. Rasamoelisolo M,
    3. Spearman M,
    4. Bosc D,
    5. Cizeau J,
    6. Entwistle J,
    7. et al.
    Preclinical assessment of an anti-EpCAM immunotoxin: locoregional delivery provides a safer alternative to systemic administration. Cancer Biother Radiopharm 2009;24:477–87.
    OpenUrlCrossRefPubMed
Back to top
Molecular Cancer Therapeutics: 11 (1)
January 2012
Volume 11, Issue 1

Sign up for alerts

View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
Single-Chain Antibody-Based Immunotoxins Targeting Her2/neu: Design Optimization and Impact of Affinity on Antitumor Efficacy and Off-Target Toxicity
Yu Cao, James D. Marks, Qian Huang, Stephen I. Rudnick, Chiyi Xiong, Walter N. Hittelman, Xiaoxia Wen, John W. Marks, Lawrence H. Cheung, Kim Boland, Chun Li, Gregory P. Adams and Michael G. Rosenblum
Mol Cancer Ther January 1 2012 (11) (1) 143-153; DOI: 10.1158/1535-7163.MCT-11-0519
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement