Deep neural networks identify signaling mechanisms of ErbB-family drug resistance from a continuous cell morphology space.

It is well known that the development of drug resistance in cancer cells can lead to changes in cell morphology. Here, we describe the use of deep neural networks to analyze this relationship, demonstrating that complex cell morphologies can encode states of signaling networks and unravel cellular mechanisms hidden to conventional approaches. We perform high-content screening of 17 cancer cell lines, generating more than 500 billion data points from ∼850 million cells. We analyze these data using a deep learning model, resulting in the identification of a continuous 27-dimension space describing all of the observed cell morphologies. From its morphology alone, we could thus predict whether a cell was resistant to ErbB-family drugs, with an accuracy of 74%, and predict the potential mechanism of resistance, subsequently validating the role of MET and insulin-like growth factor 1 receptor (IGF1R) as drivers of cetuximab resistance in in vitro models of lung and head/neck cancer.

Cetuximab Resistance in Squamous Carcinomas of the Upper Aerodigestive Tract Is Driven by Receptor Tyrosine Kinase Plasticity: Potential for mAb Mixtures.

Squamous cell carcinomas (SCC) arising in upper parts of the aerodigestive tract are among the leading causes of death worldwide. EGFR has been found to play an essential role in driving the malignancy of SCC of the upper aerodigestive tract (SCCUAT), but, despite this, clinical results using a range of different EGFR-targeted agents have been disappointing. Cetuximab is currently the only EGFR-targeted agent approved by the FDA for treatment of SCCUAT. However, intrinsic and acquired cetuximab resistance is a major problem for effective therapy. Thus, a better understanding of the mechanisms responsible for cetuximab resistance is valuable for development of the next generation of antibody therapeutics. In order to better understand the underlying mechanisms of cetuximab resistance in SCCUAT, we established from cetuximab-sensitive models cell lines with acquired resistance to cetuximab by continuous selective pressure in vitro and in vivo Our results show that resistant clones maintain partial dependency on EGFR and that receptor tyrosine kinase plasticity mediated by HER3 and IGF1R plays an essential role. A multitarget mAb mixture against EGFR, HER3, and IGF1R was able to overcome cetuximab resistance in vitro To our surprise, these findings could be extended to include SCCUAT cell lines with intrinsic resistance to cetuximab, suggesting that the triad consisting of EGFR, HER3, and IGF1R plays a key role in SCCUAT. Our results thus provide a rationale for simultaneous targeting of EGFR, HER3, and IGF1R in SCCUAT. Mol Cancer Ther; 15(7); 1614-26. ©2016 AACR.

Pan-HER, an Antibody Mixture Simultaneously Targeting EGFR, HER2, and HER3, Effectively Overcomes Tumor Heterogeneity and Plasticity.

PURPOSE: Accumulating evidence indicates a high degree of plasticity and compensatory signaling within the human epidermal growth factor receptor (HER) family, leading to resistance upon therapeutic intervention with HER family members. EXPERIMENTAL DESIGN/RESULTS: We have generated Pan-HER, a mixture of six antibodies targeting each of the HER family members EGFR, HER2, and HER3 with synergistic pairs of antibodies, which simultaneously remove all three targets, thereby preventing compensatory tumor promoting mechanisms within the HER family. Pan-HER induces potent growth inhibition in a range of cancer cell lines and xenograft models, including cell lines with acquired resistance to therapeutic antibodies. Pan-HER is also highly efficacious in the presence of HER family ligands, indicating that it is capable of overcoming acquired resistance due to increased ligand production. All three target specificities contribute to the enhanced efficacy, demonstrating a distinct benefit of combined HER family targeting when compared with single-receptor targeting. CONCLUSIONS: Our data show that simultaneous targeting of three receptors provides broader efficacy than targeting a single receptor or any combination of two receptors in the HER family, especially in the presence of HER family ligands. Pan-HER represents a novel strategy to deal with primary and acquired resistance due to tumor heterogeneity and plasticity in terms of HER family dependency and as such may be a viable alternative in the clinic.

Targeting Three Distinct HER2 Domains with a Recombinant Antibody Mixture Overcomes Trastuzumab Resistance.

HER2 plays an important role in the development and maintenance of the malignant phenotype of several human cancers. As such, it is a frequently pursued therapeutic target and two antibodies targeting HER2 have been clinically approved, trastuzumab and pertuzumab. It has been suggested that optimal inhibition of HER2 is achieved when utilizing two or more antibodies targeting nonoverlapping epitopes. Superior clinical activity of the trastuzumab plus pertuzumab combination in metastatic breast cancer supports this hypothesis. Because trastuzumab and pertuzumab were not codeveloped, there may be potential for further optimizing HER2 targeting. The study herein evaluated functional activity of anti-HER2 antibody combinations identifying optimal epitope combinations that provide efficacious HER2 inhibition. High-affinity antibodies to all four extracellular domains on HER2 were identified and tested for ability to inhibit growth of different HER2-dependent tumor cell lines. An antibody mixture targeting three HER2 subdomains proved to be superior to trastuzumab, pertuzumab, or a combination in vitro and to trastuzumab in two in vivo models. Specifically, the tripartite antibody mixture induced efficient HER2 internalization and degradation demonstrating increased sensitivity in cell lines with HER2 amplification and high EGFR levels. When compared with individual and clinically approved mAbs, the synergistic tripartite antibody targeting HER2 subdomains I, II, and IV demonstrates superior anticancer activity.

Rational identification of an optimal antibody mixture for targeting the epidermal growth factor receptor.

The epidermal growth factor receptor (EGFR) is frequently dysregulated in human malignancies and a validated target for cancer therapy. Two monoclonal anti-EGFR antibodies (cetuximab and panitumumab) are approved for clinical use. However, the percentage of patients responding to treatment is low and many patients experiencing an initial response eventually relapse. Thus, the need for more efficacious treatments remains. Previous studies have reported that mixtures of antibodies targeting multiple distinct epitopes are more effective than single mAbs at inhibiting growth of human cancer cells in vitro and in vivo. The current work describes the rational approach that led to discovery and selection of a novel anti-EGFR antibody mixture Sym004, which is currently in Phase 2 clinical testing. Twenty-four selected anti-EGFR antibodies were systematically tested in dual and triple mixtures for their ability to inhibit cancer cells in vitro and tumor growth in vivo. The results show that targeting EGFR dependent cancer cells with mixtures of antibodies is superior at inhibiting their growth both in vitro and in vivo. In particular, antibody mixtures targeting non-overlapping epitopes on domain III are efficient and indeed Sym004 is composed of two monoclonal antibodies targeting this domain. The superior growth inhibitory activity of mixtures correlated with their ability to induce efficient EGFR degradation.