Introduction

Antibody-drug conjugates (ADCs) are a class of immunotherapeutic agents that enable the delivery of cytotoxic drugs to target malignant cells. ADCs passively accumulate in solid tumour tissue through the enhanced permeability and retention effect1 and actively accumulate in target malignant cells because of the specific antibody-antigen binding2. The selection of a suitable monoclonal antibody (mAb), an anticancer agent (ACA), and a linker have supported the clinical success of ADCs3. The ADC strategy should be confined to highly toxic ACAs and not applied to ordinary ACAs, such as taxane, adriamycin, and others, because fewer than four ACA molecules should be conjugated to the mAb to prevent a decrease in the affinity of the mAb when too many ACA molecules are attached4. Monomethyl auristatin E (MMAE) is one of the most useful and potent ACAs for the clinical development of novel ADCs5,6. MMAE inhibits cell division by blocking the polymerization of tubulin.

In our laboratory, we have created an anti-human tissue factor (TF) mAb attached to valine-citrulline (Val-Cit)-MMAE (human TF ADC) and have reported its antitumour effect against xenografts of a human pancreatic cancer cell line, BxPC-37. The Val-Cit-MMAE has been designed for maximum serum stability and efficient release into the tumour environment8. Once human TF ADC binds to the target malignant cells, it is internalized by endocytosis, and MMAE is theoretically released into the tumour environment through the action of the lysosomal enzymes on the linker. TF is a transmembrane glycoprotein involved in the initiation of the extrinsic pathway of blood coagulation9, is expressed in various types of cancer, and plays a role in cancer progression, angiogenesis, tumour growth, and metastasis10. Because pancreatic cancer tissue expresses high levels of TF, it is a useful target antigen for this condition11,12,13.

To optimize the efficacy of an ADC against TF-positive solid tumours, a preclinical pharmacological evaluation of the ADC should be performed to determine whether the human TF ADC has been optimally designed. Regarding antitumour effects, ACAs must penetrate the tumour tissue efficiently and be retained there at a high and biologically active concentration14,15. For such analyses, high-performance liquid chromatography (HPLC) or liquid chromatography mass spectrometry (LC-MS) is generally used. However, these techniques do not provide information about the drug distribution in a specific target area, although they allow optimization of the drug design to a certain extent, enabling more efficiently targeted delivery. Autoradiography can be used to examine the tissue distribution of radiolabelled small molecules16. However, this method cannot distinguish between a radiolabelled drug conjugated to an ADC and free radiolabelled drug released from the ADCs23.

LC-MS/MS

The samples were analysed using a liquid chromatograph tandem mass spectrometer API3200 LC-MS/MS system (AB SCIEX, Framingham, MA). The analytical column, a reversed-phase LC column (4.0 μm polar, 80 Å; 50 × 3.0 mm, Synergi; Shimazu), was heated to 40 °C. The injection volume was 10 μm, and the flow rate was 0.4 mL/min. The autosampler was equipped with a cooling stack set at 4 °C. Acetonitrile and a 0.1% (w/v) formic acid solution were used as the mobile phases. For the gradient elution, the mobile phase composition was as follows: 5% acetonitrile for 1.0 min, increased to 40% for 2.0 min, increased to 100% for 2.0 min, maintained at 100% for 2.0 min, and then decreased back to 5.0% for 2.0 min. The mobile phase was introduced into the spectrometer via electrospray ionization in positive-ion mode under multiple reaction monitoring (MRM) conditions. The MRM transitions were used for MMAE (m/z 718.4/152.2) and MMAF as an internal standard (m/z 732.4/170.3). The standard curve had a linear range from 0.10 nM to 100 nM. Samples from the non-treated BxPC-3 tumour tissues were used as a negative control.

Immunohistochemistry

The 10-μm frozen sections were fixed with 4% paraformaldehyde (Wako) for 10 min. To block endogenous peroxidase, 3% hydrogen peroxide for 20 min was used. After being blocked with 3% skim milk in PBS for 30 min, the sections were incubated with an anti-rat IgG goat antibody (Histostar; MBL, Nagoya, Japan) for mAb/ADC according to the manufacturer’s instructions. The continuous sections were incubated with an anti-mouse CD31 goat antibody for vascular endothelial cells (10 μg/mL; R&D Systems, Minneapolis, MN) for 1 h at room temperature. After being washed with PBS, the sections were incubated with a horseradish peroxidase (HRP)-conjugated anti-goat IgG antibody (Jackson IR, West Grove, PA) according to the manufacturer’s instructions. Counterstaining was performed using haematoxylin. Negative controls included the replacement of the primary antibody by PBS and isotype antibodies.

Statistics

Measured differences were considered significant at P < 0.05. For the quantification of MMAE using LC-MS/MS, statistical significance was determined using the Tukey-Kramer multiple comparison test. To distinguish between MMAE alone and MMAE conjugated to a mAb, statistical significance was determined using the Tukey-Kramer multiple comparison test. For the high-resolution MALDI-IMS of MMAE in tumour tissues, statistical significance was determined using the two-sided Student’s t-test. The bar graph data were expressed as the mean ± standard deviation (SD). The error bars indicated the SD. The statistical analyses were performed using Statcal QC (The publisher OMS, Saitama, Japan).

Additional Information

How to cite this article: Fujiwara, Y. et al. Imaging mass spectrometry for the precise design of antibody-drug conjugates. Sci. Rep. 6, 24954; doi: 10.1038/srep24954 (2016).