Christine Ambrose1, Lihe Su1, Lan Wu1, Fay Dufort1, Roy R. Lobb1, Andreas Hombach2, Hinrich Abken2, Paul D. Rennert1.
1: Aleta Biotherapeutics, Natick, MA USA; 2: University of Cologne, Cologne, Germany
1 – Introduction
Remarkable progress has been made in the treatment of relapsed/refractory Acute Lymphocytic Leukemia and Non-Hodgkin Lymphoma with CAR-CD19 T cells. In contrast, progress against CD19-negative hematological cancers and solid tumors has been limited. Intensive efforts to optimize cellular therapeutics for better efficacy include provision of cytokine support and countering immuno-suppression. However, lack of sufficient antigen is a significant additional hurdle that CAR-T therapeutics for solid tumors must overcome. We present a novel strategy to utilize CD19 for sustained antigen presentation in order to promote cellular therapeutic expansion, efficacy and persistence. The strategy, called IMPACTTM (Integrated Modules oPtimize Adoptive Cell Therapy), employs a methodology that is modular in design and can be applied to diverse antigens and tumor types, yet retains the well-established advantages of CAR T cells directed to CD19.
2 – Technology Overview
IMPACTTM fusion proteins are created by cloning the extracellular domain (ECD) of a CAR T cell target protein (e.g. CD19) to an scFv that recognizes a second target protein. The system is modular: diverse ECD-scFv fusion proteins have been designed and expressed. In one iteration, this fusion protein is purified and utlized in conjunction with an exising CAR T cell, e.g. a CD19-anti-Her2 fusion protein is purified and added to a culture of CAR19 T cells and Her2+ tumor cells, creating a bridge that triggers CAR19 T cell cytotoxic activity. In another iteration, the CAR19 T cells expresses the CD19- anti-Her2 fusion protein, and this expression is sufficient to create the “cytotoxic bridge” as shown in Figure 1. The normal B cell pools ensures expansion and persistance of the CAR19 T cells in vivo.
Figure 1. Bridging CAR19 T cells to Her2+ tumor cells
3 – Purified IMPACTTM fusion proteins
We cloned the human CD19 ECD (two IgC domains) and the scFv derived from the anti-Her2 mAb trastuzumab into a lentiviral expression system for transfection then purification by affinity chromatography and SEC (Lake Pharma). Monomeric protein was used in our experiments; this monomer is stable in solution. The fusion protein carries a C-terminal His tag. Control proteins (CD19 ECD-His and CD22 ECD 1-3 trastuzumab scFv) were also created.
Table 1. Affinities of fusion proteins for target molecules, determined by ELISA analyses
Figure 2. FACs binding of CD19-anti-Her2 scFv and CD22- anti-Her2 scFv fusion proteins to SKOV-3 ovarian cancer cells
4 – Construction and characterization of CAR19 T cells
The scFv from anti-CD19 mAb FMC63 was cloned in frame with a FLAG-tagged linker, the CD28 transmembrane sequence and the cytoplasmic domains of CD28, 4-1BB and CD3ζ. The FLAG tag was encoded to facilitate detection of the CAR by FACs analysis following transduction into primary human T cells (Figure 3). The CAR19 T cells were tested for cytotoxic activity in a cell killing assay using B cell lymphoma lines (ATCC). One example is shown in Figure 3.
Figure 3. Expression of the CD19 CAR and cytotoxicity against a CD19+ B cell lymphoma line
5 – Fusion protein mediated cytotoxicity
The ability of the fusion proteins to bridge to CAR19 cells and mediate killing of Her2+ SKOV-3 tumor cells was evaluated. First we performed a titration.
Figure 4. Cytotoxicity induced by titration of the IMPACTTM CD19-anti-Her2 scFv fusion protein with CAR19 T cells and SKOV3 tumor cells
Next, 1ug/ml of protein or was added either to the CAR19 T cells, or to the SKOV-3 tumor cells, then incubated for 30′ on ice, before mixing with the cognate cell partner. In the control experiment both cell types and the protein were added simultaneously.
Figure 5. Cytotoxicity following different orders of addition of fusion protein to cells
6 – CAR19 with the IMPACTTM fusion protein encoded as an integrated gene
The development candidates are being constructed as integrated genes (i-genes) using lentiviral vectors and packaging systems. A prototype schematic is shown here:
SvFv and bispecific scFv CD19 fusion protein constructs are being optimized. Our current programs use scFvs to CD20, BCMA, CLL-1, ROR-1 and Her2.
Transduction of primary human T cells with the viral particles yielded cell surface CAR expression and detectible fusion protein (< 5ng/ml). These cells were placed into culture with SKOV-3 tumor cells in order to assess cytotoxicity.
Figure 6. CAR19 T cells secreting an IMPACTTM fusion protein (CAR130) demonstrate redirected cytotoxicity
We conclude that IMPACTTM fusion proteins mediate redirected tumor cell killing at very low concentrations, are not ‘shut-down’ by the soluble protein, and can be successfully secreted from CAR T cells. The first in vivo study using purified fusion proteins, CAR19 cells and SKOV-3 tumors is in progress.