A CAR-T Study for Solid Tumors: In Vivo Pharmacology

Evaluating the efficacy and safety profile of CAR T in vivo

T cell therapies are novel and, in some cases, curative, but they also come with formidable limitations. Follow our three-part series, Driving the CAR-T, and see how a CAR-T cell therapy is engineered and produced, and the preclinical pharmacology studies that are typically required for CAR-T products. Yesterday, we talked about HER2 targeting CAR-T cells developed at Charles River’s lab in Leiden. Today we are looking at what pharmacological studies of those CAR-T cells revealed.

Before new drugs are approved for use in humans they go through extensive pre-clinical evaluation. The purpose of this testing is to provide evidence of activity that would benefit patients and demonstrate an acceptable safety profile. Most small molecule and antibody drugs follow a well-established pathway of pre-clinical development. The development paradigm for cellular therapies is still evolving, however, the basic principles of pharmacological and toxicological testing also apply to this new class of drugs, as our Discovery Oncology labs in Morrisville and Freiburg showed recently when they evaluated HER2 targeting CAR-T cells developed by our Early Discovery site Leiden.

Many of us know HER-2 as an oncogene that is over-expressed, and gene amplified in breast and other types of cancers. Although most breast cancer patients are treated these days with a curative intention, there are still many patients progressing to metastatic disease. The advent of anti-HER2 therapy significantly delayed disease progression and prolongs survival for those patients with HER2+ tumors. Beside the classical approaches which include small molecule drugs, such as Lapatinib, monoclonal antibodies such as trastuzumab or antibody-drug conjugates against HER2, the use of CAR-T cell therapy in solid cancers is gaining traction with an extensive pipeline of engineered products undergoing development.

We know that alterations in the ERBB family of genes, including HER2, can cause tumor progression in multiple solid cancers, and abnormally activate HER2 signaling, resulting in normal cells developing malignant properties, tumor growth and progression. So, we set out to perform CAR-T studies to demonstrate the efficacy and safety of the HER2 targeting CAR-T cells, in 2D and 3D models and in animal studies.

What we learned from in vitro studies of HER2 targeting CAR-T cells

As is the case in any well-established drug development project, our CAR-T studies first aimed to prove efficacy of the HER2 CARs in 2D as well as 3D in vitro experiments. A panel of human cancer cell lines (SK-OV3, Hs578T and JIMT-1) with different HER-2+ expression levels was treated with the HER2 CARs using normal T cells and the monoclonal antibody drug trastuzumab as control treatment arms.

The HER2 CAR´s eradicated the 3D spheroids of the HER2+ cell lines SK-OV3 and JIMT-1 in a dose-dependent manner. Trastuzumab was active in the SK-OV3 model and inactive in the JIMT-1 line, which is in line with clinical data, as JIMT-1, although HER2+, is derived from a trastuzumab refractory patient. Of note, the control T cells in our CAR-T studies did not influence the tumor growth at all. The HER2- cell line Hs578T served as a negative control and proved to be resistant to any treatment in this CAR-T study.

What we learned from in vivo studies of HER-2 targeting CAR-T cells

To underpin those results of the CAR-T study in the JIMT- cell lines, the HER2-targeting CAR-T cells were applied in vivo in two different doses to JIMT-1 tumor-bearing NSG mice, with tumor volume measured over time. The HER2 CAR-T cells were able to induce a complete remission three weeks after the start of treatment. In contrast, the tumors displayed progressive disease under therapy with the normal T cells or Trastuzumab. Interestingly, the CAR-T cells induced a tumor swelling between 8-12 days post injection, which is typical in patients, too. Four weeks following injection, the human T cells could be detected in spleen, peripheral blood, liver, and bone marrow of the JIMT- 1 bearing mice. The implication of this finding for possible side effects of the HER2 CAR´s remains to be elucidated.

We also performed CAR T efficacy and safety testing in NCG mice bearing subcutaneous OE19 tumors—an eosophageal adenocarcinoma (EAC) expressing high levels of HER2. NCG mice are highly immunodeficient making them excellent hosts for engraftment of human cells. OE19 is a robust model with uniform tumor progression and a high take rate.

We designed a 4-arm CAR-T study, testing CAR-T cells at a single dose, which was predicted to show efficacy based on the potent tumor killing activity observed in vitro studies. Ten animals per group were used in the CAR-T study to provide robust statistical power and additional cohorts served as controls. Mice were randomized into groups when tumors were established, and treatments were started intravenously on the same day.

The number of CAR T cells detected in blood increased significantly over time, suggesting that CAR T cells expanded, and developed effector function. Cytokines that play roles in chemoattraction, and enhance T cell function, were elevated. The favorable pharmacodynamic profile suggested that CAR T cells should elicit anti-tumor responses.

To our surprise, however, the anti-tumor activity of CAR T cells was relatively modest. The obvious question was why. Clues to understanding these results emerged when we performed a tissue analysis of tumor samples at the end of the CAR-T study and observed notable differences in the morphology of tumor cells between treatment groups. In the control animals a small proportion of the neoplastic cells were arranged in an epithelial conformation while the remaining tumor cells appeared to have to a mesenchymal structure. This epithelial morphology was not present in tumors from CAR T treated animals.

Additionally, HER2 staining was only present on tumor cells having an epithelial-like morphology and thus limited to tumors from animals in the control groups. IHC and flow cytometry staining confirmed HER2 was absent on tumors in the CAR-T treatment group. This finding seemed to be consistent with the CAR-T treatment eliminating HER2-positive neoplastic cells in that group. The removal of the HER2- positive cells did not however lead to a marked reduction in subcutaneous tumor size since most of the tumors were not HER2-positive.

In CAR-T studies loss of target a significant issue

In a separate CAR-T study we evaluated how quickly changes in tissue morphology and HER2 expression could be detected. As early as day 4 a significant fraction of tumor cells were HER2 negative and by day 12, when we would typically begin treatment, approximately half of all tumor cells lacked expression of HER2. Similar findings have been observed in breast cancer models.

So while CAR-T cells appeared to be highly effective at removing HER2-expressing tumor cells with epithelial morphology, loss of the CAR target masked their overall efficacy. The findings of our CAR-T studies provide insights into the complex tumor biology at play and highlight the need for cautious interpretation of CAR-T study data. In fact, loss of surface CAR target is now recognized as a significant clinical problem for the development of CAR T cells, leading in many cases to patient relapse.

In an attempt to provide long-lasting durable responses, researchers are developing many new strategies including the development of CAR-Ts reactive to dual tumor targets. There are unique features of the OE19/NCG preclinical model we developed which mirror some of the tumor heterogeneity seen in clinical settings and could provide an attractive translational system for evaluating some of these new approaches to treatment.

Series logo designed by Chelsea Bates, Charles River Laboratories