Bram Herpers1, Kuan Yan1, Thomas Dijkmans1, Leo Price1

1OcellO B.V. Leiden, The Netherlands


Antibody-drug conjugates (ADCs) are the result of clever combinations of a non-specific toxin and a specific antibody targeting the tumor. Only upon ADC binding the toxin can be activated, increasing specificity and pharmacological safety. ADC binding, internalization, toxin activation, tumor penetration, tumor killing and resistance formation are (some of) the biological processes involved, but are difficult to study in vivo. To visualize and quantify ADC activity on human tumors, we generated in vitro 3D cultures from Patient-Derived Xenograft (PDX) models (in vitro PDX).


Cells derived from PDX models expressing the HER2 receptor were embedded in a laminin/collagen-rich matrix and an optimized culture medium was added to sustain the outgrowth of tumor cell clusters (tumoroids). ADC exposure was performed to study the effect on the outgrowth of developing tumoroids or to investigate the penetration rate of ADCs in established tumoroids. Also, long-term low dose T-DM1 (Kadcyla) exposure was performed to generate and study ADC resistance mechanisms. These tumoroids, grown and tested in 384-wells plates, are visualized using automated microscopy and analyzed in OcellO’s advanced 3D image analysis methods.


Differential sensitivity to (HER2) ADCs is not only observed between different patients or tissue origins, but also the rate at which ADCs penetrate and kill the tumor depends on toxin conjugation methods. Furthermore, the growth factor environment determines ADC sensitivity. An in vitro PDX model generated from a T-DM1 resistant patient showed high T-DM1 sensitivity in naïve conditions, but became T-DM1 resistant upon long-term exposure.


Our 3D culture, imaging and analysis methods are powerful tools to study ADC sensitivity, cell surface binding, tissue penetration, and resistance formation. The 384-wells plate format allows quick testing of variations in antibody-conjugation-toxin combinations in a visual and quantifiable manner.