Embargoed until 2027-04-25
Author
Date
2024Type
- Doctoral Thesis
ETH Bibliography
yes
Altmetrics
Abstract
Colorectal cancer (CRC) ranks as the second leading cause of cancer-related deaths globally, resulting in about one million people dying from it each year. At the time of diagnosis, most CRC cases are in an advanced stage, often with the cancer already spread to other organs. The current standard of care for these patients provides limited benefit, which is reflected in a 5-year survival rate of only 14%. This highlights the urgent medical need for new therapeutic modalities to effectively treat CRC. In this thesis, we explored immunotherapy as a strategy to enhance the body’s own immune system to fight cancer. In particular, the use of T cell engaging bispecific antibodies (TCBs) presents a promising approach. TCBs are engineered to simultaneously bind to both T cells and tumor cells, redirecting the cytotoxic power of T cells to target and eliminate tumor cells. The design of TCBs described in this work is inspired by the natural mechanism of T cell priming, which requires two signals for the full activation of T cells. ‘Signal one’ is initiated through the engagement of the CD3/T cell receptor (TCR) complex, and ‘signal two’ is a costimulatory signal mediated by the interaction of the CD28 receptor.
The first chapter of this thesis was dedicated to optimizing the design and geometry of TCBs targeting carcinoembryonic antigen (CEA), which is a well-validated overexpressed antigen in more than 98% of CRC patients. Four different TCB formats were developed, employing the antibodies F4 (targeting human CEA) and 2C11 (targeting mouse CD3). The 2C11 antibody, serving as a surrogate for anti human CD3, was used to facilitate in vivo testing in immunocompetent mouse models. TCB formats included both IgG-based and antibody fragmentbased constructs, differing in size, valency, and geometry. The 2+1 configuration, comprising an anti-CEA single-chain diabody (scDbCEA) fused to an anti-CD3 single-chain variable fragment (scFvCD3), emerged as the most potent design. It showed effective in vitro tumor cell killing at subnanomolar concentrations in different CEA+ cell lines. The scDbCEA x scFvCD3 format was further evaluated in two different syngenic mouse models of colorectal cancer, where it demonstrated tumor growth retardation. Tumor sections from the TCB treated group showed an increase in CD8+ T cell infiltration, underscoring the potential of TCBs to convert immunologically ‘cold’ tumors into ‘hot’ ones. This study highlighted essential characteristics for the design of TCBs targeting CEA: bivalent tumor targeting, monovalent T cell targeting, and short spatial separation between the binding arms.
The second chapter focused on the development of novel anti-human CD28 antibodies with specific desired properties. While CD3-targeting TCBs have shown great clinical outcomes, the lack of a costimulatory signal via CD28 typically results in early T cell exhaustion and inadequate T cell activation. Therefore, combining CD28-targeting therapeutics with CD3-targeting TCBs presents an attractive strategy to boost T cell activity in CRC. However, this approach faced skepticism following TeGenero’s phase I clinical trial in 2006 using the superagonistic anti-CD28 antibody TGN1412, which led to severe and life-threatening side effects, including multiple-organ failures due to cytokine release syndrome. This incident significantly halted progress in CD28-targeting therapies for several years. Superagonistic anti-CD28 antibodies can induce polyclonal activation of T cells without the need for ‘signal one.’ In this report, we employed antibody phage display technology to develop a new conventional agonistic anti-CD28 antibody termed ‘E1P2.’ E1P2 binds near the epitope of the natural CD28 ligands, CD80/CD86, which is different from the lateral and membrane-proximal epitope targeted by TGN1412. E1P2 was subjected to head-tohead comparisons with TGN1412 in multiple in vitro and in vivo experiments. E1P2, unlike TGN1412, showed no signs of in vitro superagonistic activities in human T cells from various healthy donors. In an in vivo safety study using humanized NSG mice, E1P2 did not induce cytokine release, in contrast to the response observed with TGN1412. In an in vitro activity assay utilizing human T cells, the combination of E1P2 in an IgG format with a CD3-targeting TCB enhanced tumor cell killing and increased T cell proliferation. Collectively, these findings underscore the therapeutic potential of E1P2 to boost the activity of T cells in a desired safety manner only in the presence of ‘signal one’ provided by TCR/CD3 activating constructs.
The third chapter of this thesis assessed the therapeutic potential of combining CD3- and CD28- targeting TCBs, both directed against CEA, for the treatment of CRC. This section aimed to integrate the knowledge and tools developed in the previously described chapters one and two. Fully human reagents were used in this part, and all constructs were designed to retain the Fc region, ensuring an IgG-like structure with an extended serum half-life. CD3xCEA TCBs were generated using a bivalent tumor targeting arm (F4 antibody targeting human CEA), a monovalent T cell targeting arm (SP34 antibody targeting human CD3), and a short interdomain distance between the respective binding moieties. CD28xCEA TCBs were developed based on a 2+2 configuration, comprising the E1P2 antibody against human CD28 and either the F4 antibody for CEA or the Sm3E antibody for a different epitope on CEA. An in vitro killing assay including human T cells and LS174T (human CRC cells naturally expressing CEA) was conducted to evaluate the therapeutic potential of the combination strategy. The combined use of CD3xCEA and CD28xCEA TCBs, each directed against a different CEA epitope, demonstrated substantial anti-tumor activity. In summary, this study emphasized the potential benefits of combining CD3- and CD28-targeting TCBs in the treatment of challenging forms of cancers such as CRC. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000670140Publication status
publishedExternal links
Search print copy at ETH Library
Publisher
ETH ZurichOrganisational unit
03816 - Halin Winter, Cornelia / Halin Winter, Cornelia
More
Show all metadata
ETH Bibliography
yes
Altmetrics