(T1330-10-58) Design and Development of Protease-Activated Chemically Self-Assembled Nanorings (Pro-CSANs) for Cancer Immunotherapy

Purpose: Our lab has previously developed a non-genetic platform to redirect T cells to combat solid tumors. This approach is based on fusing a tumor-targeting protein or a T cell-targeting anti-CD3 (αCD3) single-chain variable fragment to a dimer of Escherichia coli (E. coli) dihydrofolate reductase (1DD). Incubating a mixture of tumor-targeting and T cell-targeting 1DD fusion proteins with a chemical dimerizer, bis-methotrexate (bisMTX), led to the spontaneous assembly of bispecific, multivalent nanorings, or CSANs. These bispecific CSANs bound to CD3⁺ T cells and redirected them to target breast cancer. However, tumor-associated antigen targeting relies on differential antigen expression between cancer and healthy cells. The high affinity and avidity of the multivalent bispecific CSANs may potentially target healthy tissue with low antigen expression, thereby exhibiting “on-target, off-tumor” toxicity. Matrix metalloproteinases are overexpressed in a variety of solid tumors. Therefore, to improve tumor specificity, we hypothesized that masking the αCD3 moiety in our CSANs with a bulky protein, through a matrix metalloproteinase-2 (MMP-2) sensitive linker will prevent engagement and activation of T cells outside the tumor microenvironment (TME). Once the Pro-CSANs enter the TME, cleavage by MMP-2 will lead to unmasking of the αCD3 moiety, followed by engagement and activation of T cells, causing targeted cancer cell lysis (Figure 1).

Methods: An αCD3-1DD fusion protein containing an αCD3 nanobody was designed. To sterically mask the anti-CD3 nanobody’s binding site, monomeric human histidine triad nucleotide binding protein 1 (hHINT1) was genetically fused to the nanobody’s N-terminus via a flexible glycine-serine linker embedding an MMP-2 substrate sequence. The genes encoding the αCD3-1DD and hHINT1-αCD3-1DD fusion proteins were cloned into pET28a (+) vector. The recombinant plasmids were then used to transform SHuffle T7 Competent E. coli. Recombinant proteins were expressed and purified via immobilized metal-affinity chromatography. Proteins were characterized via analytical size exclusion chromatography (SEC) to evaluate nanoring formation. To demonstrate unmasking of the hHINT1-anti-CD3-DHFR fusion protein, a time-dependent in vitro cleavage assay was performed. hHINT1-αCD3-1DD was incubated with recombinant human MMP-2 at room temperature. The results of the cleavage assay were analyzed using SDS-PAGE. Flow cytometry was utilized to compare the binding of hHINT1-αCD3-1DD and αCD3-1DD fusion proteins to the surface of T cells. Both hHINT1-αCD3-1DD and αCD3-1DD were mixed in a 1:1 ratio with monomeric superfolder green fluorescent protein-2-1DD (msGFP2-1DD) fusion protein in the presence of bisMTX, thus forming fluorescent bispecific CSANs. These bispecific CSANs were then incubated with human T cells and binding was analyzed using the fluorescence of GFP as the readout. msGFP2-1DD monospecific CSANs were used as the negative/fluorescence control.

Results: SEC analysis confirmed that incubation of monomeric proteins with bisMTX (1:3 equivalents) for 1 hour at room temperature led to complete oligomerization to form nanorings (Figure 2). Further characterization of nanorings using dynamic light scattering and cryo-transmission electron microscopy is currently underway. In vitro cleavage assay demonstrated that ~100% of hHINT1-αCD3-1DD was cleaved by MMP-2 resulting in a band of ~53 kDa that corresponds to αCD3-1DD within 24 hours (Figure 2). Flow cytometry histograms of αCD3-1DD/msGFP2-1DD bispecific CSANs showed an increase in median GFP value with an increase in CSAN concentration, demonstrating a concentration-dependent binding to T cells (Figure 3). In contrast, histograms of all the tested concentrations of hHINT1-αCD3-1DD/msGFP2-1DD bispecific CSANs did not show an increase in median GFP value, thereby showcasing the masking of the αCD3 nanobody. Furthermore, when hHINT1-αCD3-1DD/msGFP2-1DD bispecific CSANs were incubated with MMP-2, the CSANs bound to T cells in a concentration-dependent manner, thereby demonstrating the unmasking of the αCD3 nanobody (Figure 3).

Conclusion: Nanobody-based αCD3-1DD and hHINT1-αCD3-1DD fusion proteins were successfully designed, expressed, and purified from E. coli. These fusion proteins oligomerized in the presence of bisMTX to form CSANs. Preliminary studies demonstrated that the hHINT1 steric mask was completely removed in vitro by MMP-2. Flow cytometry analysis established a proof-of-concept that the masked hHINT1-αCD3-1DD fusion protein showed reduced binding to T cells but regained binding upon unmasking by MMP-2. As the in vitro 2D and 3D cytotoxicity analysis are currently on-going, the data generated thus far demonstrate that the Pro-CSANs would significantly mitigate “on-target, off tumor” toxicity, and serve as a valuable immunotherapeutic in the quest to combat solid tumors.

References: 1. Jacob R. Petersburg, Jingjing Shen, Clifford M. Csizmar, Katherine A. Murphy, Justin Spanier, Kari Gabrielse, Thomas S. Griffith, Brian Fife, and Carston R. Wagner ACS Nano 2018 12, 6563-6576.
2. Ozgun Kilic, Marcos R. Matos de Souza, Abdulaziz A. Almotlak, Yiao Wang, Jill M. Siegfried, Mark D. Distefano, and Carston R. Wagner
Journal of Medicinal Chemistry 2020 63, 10235-10245.
3. Shima Moradi-Kalbolandi, Azadeh Sharifi-K, Behrad Darvishi, Keivan Majidzadeh-A, Neda jalili, Solmaz Sadeghi, Marjan Mosayebzadeh, Hassan Sanati, Malihe Salehi, Leila Farahmand, Evaluation the potential of recombinant anti-CD3 nanobody on immunomodulatory function, Molecular Immunology 2020 118, 174-181.
4. Valbuena, FM, Fitzgerald, I, Strack, RL, Andruska, N, Smith, L, Glick, BS. A photostable monomeric superfolder green fluorescent protein. Traffic. 2020 21, 534– 544.

Acknowledgements: NIH R01CA247681


Figure 1. Mechanism of action of Pro-CSANs. The Pro-CSANs bind to tumor cells first, followed by unmasking of the anti-CD3 nanobody via MMP-2, leading to engagement and activation of T cells specifically in the TME.


Figure 2. Characterization of αCD3-1DD and hHINT1-αCD3-1DD fusion proteins. A: SEC analysis for αCD3-1DD as monomer and upon incubation with bisMTX resulting in the formation of nanorings; B: SEC analysis for hHINT1-αCD3-1DD as monomer and upon incubation with bisMTX resulting in the formation of nanorings; C: SDS-PAGE analysis demonstrating the in vitro cleavage of hHINT1-αCD3-1DD to form αCD3-1DD within 24 hours.


Figure 3. Flow cytometry histograms depicting binding of proteins to T cells. A: Qualitative binding analysis of αCD3-1DD; B: Qualitative binding analysis of hHINT1-αCD3-1DD; C: Qualitative binding analysis of hHINT1-αCD3-1DD after cleavage with recombinant MMP-2 for 3 hours at 37 °C followed by incubation with T cells. Tables indicate median fluorescence intensity values. msGFP2 = monomeric superfolder green fluorescent protein.