Research Article: Chemical Screens Identify Drugs that Enhance or Mitigate Cellular Responses to Antibody-Toxin Fusion Proteins

Date Published: August 24, 2016

Publisher: Public Library of Science

Author(s): Antonella Antignani, Lesley Mathews Griner, Rajarshi Guha, Nathan Simon, Matteo Pasetto, Jonathan Keller, Manjie Huang, Evan Angelus, Ira Pastan, Marc Ferrer, David J. FitzGerald, Craig J. Thomas, Ferenc Gallyas.


The intersection of small molecular weight drugs and antibody-based therapeutics is rarely studied in large scale. Both types of agents are currently part of the cancer armamentarium. However, very little is known about how to combine them in optimal ways. Immunotoxins are antibody-toxin gene fusion proteins engineered to target cancer cells via antibody binding to surface antigens. For fusion proteins derived from Pseudomonas exotoxin (PE), potency relies on the enzymatic domain of the toxin which catalyzes the ADP-ribosylation of EF2 causing inhibition of protein synthesis leading to cell death. Candidate immunotoxins have demonstrated clear value in clinical trials but generally have not been curative as single agents. Therefore we undertook three screens to discover effective combinations that could act synergistically. From the MIPE-3 library of compounds we identified various enhancers of immunotoxin action and at least one major class of inhibitor. Follow-up experiments confirmed the screening data and suggested that immunotoxins when administered with everolimus or nilotinib exhibit favorable combinatory activity and would be candidates for preclinical development. Mechanistic studies revealed that everolimus-immunotoxin combinations acted synergistically on elements of the protein synthetic machinery, including S61 kinase and 4E-BP1 of the mTORC1 pathway. Conversely, PARP inhibitors antagonized immunotoxins and also blocked the toxicity due to native ADP-ribosylating toxins. Thus, our goal of investigating a chemical library was justified based on the identification of several approved compounds that could be developed preclinically as ‘enhancers’ and at least one class of mitigator to be avoided.

Partial Text

Antibody-based therapeutics show great promise for the treatment of patients with cancer [1]. Ideally, the chosen antibody binds to surface antigens on malignant cells and not to healthy tissues. One successful strategy for producing therapeutic antibodies is the construction of antibody-toxin fusion proteins, also known as recombinant immunotoxins [2, 3]. Immunotoxins produced from truncated versions of Pseudomonas exotoxin (PE) kill cells via the ADP-ribosylation of elongation factor 2 leading to inhibition of protein synthesis [2, 3]. Several immunotoxins have been evaluated already in clinical trials with some striking results including a high percentage of complete remissions in patients diagnosed with hairy cell leukemia when treated with immunotoxins targeting surface-expressed CD22 [4–6]. However, the same immunotoxins produced fewer responses in other CD22-positive B-cell malignancies such as CLL or NHL [4]. Similarly, an immunotoxin targeting mesothelin (SS1P) produced few objective responses when evaluated as a single agent in patients diagnosed with mesothelioma [7, 8]. To achieve maximum benefit, it is likely that immunotoxins will need to be administered in combination with small molecular weight drugs or other types of therapies. Ideally, suitable combinations can be identified that are synergistic for killing cancer cells while avoiding increased systemic toxicity. To identify effective immunotoxin-drug combinations we devised screens using both epithelial (KB3-1) and hematological (Nalm-6) cell lines. KB3-1 cells were incubated with the SS1P immunotoxin while Nalm-6 cells were treated with HA22, the immunotoxin targeting CD22. The goal was to find compounds that enhanced immunotoxin activity, with a preference for those drugs that were approved already for human use. Further, the screen should also identify potential mitigators, i.e. combinations to be avoided. Here we report on the outcomes of screens using the MIPE-3 small molecule library of ‘cancer focused’ compounds containing both approved and investigational drugs [9]. Immunotoxins at fixed concentrations were added to cells that were treated with eleven concentrations of each drug spanning a 4.5 log10 range. To avoid trivial differences related to immunotoxin design, both immunotoxins, SS1P and HA22, were constructed as disulfide-stabilized antibody Fvs joined with truncated PE38 from Pseudomonas exotoxin [3]. Purified immunotoxins of clinical grade were used in both screens. As outlined below, the screening effort was successful in identifying a number of approved compounds that enhanced the action of both immunotoxins. Likewise, mitigators were also identified, including, prominently, PARP inhibitors. And more generally our results confirmed the utility of screening drugs and antibody-based therapeutics using cell-based viability assays in a multiwell format.

Immunotoxins were championed as anti-cancer agents because of the reputed potency associated with the enzymatic domain of the bacterial toxin joined to the targeting antibody [2, 3]. However, clinical evaluation of several immunotoxins has demonstrated variable results ranging from a high rate of complete remissions in Hairy Cell Leukemia to more modest response rates for other B-cell malignancies including NHL, CLL and ALL [4–6, 28]. The reason for these disparate responses is currently poorly understood. While a detailed study of patient tumor cell biology might uncover why HCL cells are very sensitive to immunotoxin treatments while other B-cell malignancies are less so, a more pragmatic approach would seek out drugs that can convert resistant cells to sensitive ones. When mesothelin was targeted with the immunotoxin, SS1P, there were no objective responses from two early trials, one involving the continuous administration of agent and the other involving bolus infusions [7, 8]. Only when SS1P was administered in the presence of pemetrexed and cisplatin, was there evidence of objective responses [29]. Thus, there is clinical precedence for the use of drug combinations to enhance immunotoxin outcomes. While this is not a great surprise, as most cancer treatments use combination therapies to achieve superior outcomes, it was far from obvious which drugs would be best suited to accompany immunotoxin treatments.