Dual-fluorescence isogenic high-content screening for MUC16/CA125 selective agents

Structural basis for antibody recognition of the proximal MUC16 ectodomain

BACKGROUND

Mucin 16 (MUC16) overexpression is linked with cancer progression, metastasis, and therapy resistance in high grade serous ovarian cancer and other malignancies. The cleavage of MUC16 forms independent bimodular fragments, the shed tandem repeat sequence which circulates as a protein bearing the ovarian cancer biomarker (CA125) and a proximal membrane-bound component which is critical in MUC16 oncogenic behavior. A humanized, high affinity antibody targeting the proximal ectodomain represents a potential therapeutic agent against MUC16 with lower antigenic potential and restricted human tissue expression.

RESULTS

Here, we demonstrate the potential therapeutic versatility of the humanized antibody as a monoclonal antibody, antibody drug conjugate, and chimeric antigen receptor. We report the crystal structures of 4H11-scFv, derived from an antibody specifically targeting the MUC16 C-terminal region, alone and in complex with a 26-amino acid MUC16 segment resolved at 2.36 Å and 2.47 Å resolution, respectively. The scFv forms a robust interaction with an epitope consisting of two consecutive β-turns and a β-hairpin stabilized by 2 hydrogen bonds. The VH-VL interface within the 4H11-scFv is stabilized through an intricate network of 11 hydrogen bonds and a cation-π interaction.

CONCLUSIONS

Together, our studies offer insight into antibody-MUC16 ectodomain interaction and advance our ability to design agents with potentially improved therapeutic properties over anti-CA125 moiety antibodies.

Dual Fluorescence Isogenic Synthetic Lethal Kinase Screen and High-Content Secondary Screening for MUC16/CA125 Selective Agents

Significant strides have been made in the development of precision therapeutics for cancer. Aberrantly expressed glycoproteins represent a potential avenue for therapeutic development. The MUC16/CA125 glycoprotein serves as a biomarker of disease and a driver of malignant transformation in epithelial ovarian cancer. Previously, we demonstrated a proof-of-principle approach to selectively targeting MUC16+ cells. In this report, we performed a synthetic lethal kinase screen using a human kinome RNAi library and identified key pathways preferentially targetable in MUC16+ cells using isogenic dual florescence ovarian cancer cell lines. Utilizing a separate approach, we performed high-content small-molecule screening of 6 different libraries of 356,982 compounds for MUC16/CA125 selective agents and identified lead candidates that showed preferential cytotoxicity in MUC16+ cells. Compounds with differential activity were selected and tested in various other ovarian cell lines or isogenic pairs to identify lead compounds for Structural Activity Relationship (SAR) selection. Lead siRNA and small molecule inhibitor candidates preferentially inhibited invasion of MUC16+ cells in vitro and in vivo and we show that this is due to decreased activation of MAP kinase, and non-receptor tyrosine kinases. Taken together, we present a comprehensive screening approach to the development of a novel class of MUC16-selective targeted therapeutics and identify candidates suitable for further clinical development.

Tox_(R)CNN: Deep learning-based nuclei profiling tool for drug toxicity screening

Toxicity is an important factor in failed drug development, and its efficient identification and prediction is a major challenge in drug discovery. We have explored the potential of microscopy images of fluorescently labeled nuclei for the prediction of toxicity based on nucleus pattern recognition. Deep learning algorithms obtain abstract representations of images through an automated process, allowing them to efficiently classify complex patterns, and have become the state-of-the art in machine learning for computer vision. Here, deep convolutional neural networks (CNN) were trained to predict toxicity from images of DAPI-stained cells pre-treated with a set of drugs with differing toxicity mechanisms. Different cropping strategies were used for training CNN models, the nuclei-cropping-based Tox_CNN model outperformed other models classifying cells according to health status. Tox_CNN allowed automated extraction of feature maps that clustered compounds according to mechanism of action. Moreover, fully automated region-based CNNs (RCNN) were implemented to detect and classify nuclei, providing per-cell toxicity prediction from raw screening images. We validated both Tox_(R)CNN models for detection of pre-lethal toxicity from nuclei images, which proved to be more sensitive and have broader specificity than established toxicity readouts. These models predicted toxicity of drugs with mechanisms of action other than those they had been trained for and were successfully transferred to other cell assays. The Tox_(R)CNN models thus provide robust, sensitive, and cost-effective tools for in vitro screening of drug-induced toxicity. These models can be adopted for compound prioritization in drug screening campaigns, and could thereby increase the efficiency of drug discovery.

Expression of the Carboxy-Terminal Portion of MUC16/CA125 Induces Transformation and Tumor Invasion

The CA125 antigen is found in the serum of many patients with serous ovarian cancer and has been widely used as a disease marker. CA125 has been shown to be an independent factor for clinical outcome in this disease. In The Cancer Genome Atlas ovarian cancer project, MUC16 expression levels are frequently increased, and the highest levels of MUC16 expression are linked to a significantly worse survival. To examine the biologic effect of the proximal portion of MUC16/CA125, NIH/3T3 (3T3) fibroblast cell lines were stably transfected with the carboxy elements of MUC16. As few as 114 amino acids from the carboxy-terminal portion of MUC16 were sufficient to increase soft agar growth, promote matrigel invasion, and increase the rate of tumor growth in athymic nude mice. Transformation with carboxy elements of MUC16 was associated with activation of the AKT and ERK pathways. MUC16 transformation was associated with up-regulation of a number of metastases and invasion gene transcripts, including IL-1β, MMP2, and MMP9. All observed oncogenic changes were exclusively dependent on the extracellular “ectodomain” of MUC16. The biologic impact of MUC16 was also explored through the creation of a transgenic mouse model expressing 354 amino acids of the carboxy-terminal portion of MUC16 (MUC16^c354). Under a CMV, early enhancer plus chicken β actin promoter (CAG) MUC16^c354 was well expressed in many organs, including the brain, colon, heart, kidney, liver, lung, ovary, and spleen. MUC16^c354 transgenic animals appear to be viable, fertile, and have a normal lifespan. However, when crossed with p53-deficient mice, the MUC16^c354:p53^+/- progeny displayed a higher frequency of spontaneous tumor development compared to p53^+/- mice alone. We conclude that the carboxy-terminal portion of the MUC16/CA125 protein is oncogenic in NIH/3T3 cells, increases invasive tumor properties, activates the AKT and ERK pathways, and contributes to the biologic properties of ovarian cancer.

Increasing the Content of High-Content Screening: An Overview

Target-based high-throughput screening (HTS) has recently been critiqued for its relatively poor yield compared to phenotypic screening approaches. One type of phenotypic screening, image-based high-content screening (HCS), has been seen as particularly promising.

In this article, we assess whether HCS is as high content as it can be. We analyze HCS publications and find that although the number of HCS experiments published each year continues to grow steadily, the information content lags behind. We find that a majority of high-content screens published so far (60−80%) made use of only one or two image-based features measured from each sample and disregarded the distribution of those features among each cell population. We discuss several potential explanations, focusing on the hypothesis that data analysis traditions are to blame. This includes practical problems related to managing large and multidimensional HCS data sets as well as the adoption of assay quality statistics from HTS to HCS. Both may have led to the simplification or systematic rejection of assays carrying complex and valuable phenotypic information.

We predict that advanced data analysis methods that enable full multiparametric data to be harvested for entire cell populations will enable HCS to finally reach its potential.