The NKG2D/NKG2D ligand pathway mediates cancer immunosurveillance by cytotoxic lymphocytes. Natural killer group 2 member D (NKG2D) is a homodimeric C-type lectin like receptor expressed on almost all human cytotoxic lymphocytes (NK, CD8+ and γδ T cells) and serves as a primary activating receptor on NK cells. In humans, there are eight NKG2D ligands (MICA/MICB and ULBP1-6), among which MICA/B (MHC class I chain-related sequence A/B) have been best characterized. MICA and MICB are ´sister´ molecules that share around 91 % identity in their coding sequences. The domain structure of MICA/B consists of three extracellular domains, the α1α2 MHC class I-like superdomain and the Ig-like α3 domain. Under normal physiological conditions, MICA/B glycoproteins usually are not expressed in healthy tissues, but MICA/B cell surface expression can be induced by various stress conditions (e.g., virus infection, malignant transformation). MICA/B-NKG2D interaction promotes immune recognition and cytolysis of MICA/B-expressing stressed cells. However, tumor cells often escape from NKG2D-mediated recognition, apparently also by shedding of MICA molecules from the cell surface by various metalloproteases (e.g., ADAM10 and ADAM17). There is a negative correlation of elevated soluble MICA/B (sMICA/B) levels in serum of cancer patients (as a consequence of MICA/B shedding) with overall survival. Moreover, sMICA/B was suggested as a cancer marker. Hence, these properties mark the shedding of MICA/B as an attractive target for cancer studies. Numerous clinical studies attempted to target metalloproteases in order to block their action to treat malignant diseases in patients. Unfortunately, these attempts were unsuccessful due to important biological roles of these proteases in processes. As a result, there is an urgent need to develop other means of cancer treatment that lead to inhibition of MICA shedding from tumors.

The aim of this thesis was to characterize the MICA/B-monoclonal antibody (mAb) BAMO3, which inhibits MICA/B shedding, and test the therapeutic efficacy of BAMO3-mediated inhibition of MICA shedding in aggressive syngeneic mouse tumor models. In addition, particular amino acids in the transmembrane and α3 domain in the MICA molecule and their role in MICA shedding and eventually their role in affecting BAMO3-mediated MICA shedding inhibition were studied. Lastly, therapeutic efficacy of another MICA antibody conjugated to toxin was tested. Previously, the Steinle laboratory generated the mAb BAMO3 specific for MICA and MICB. This thesis demonstrates BAMO3 proteolytic shedding inhibition of MICA/MICB from different tumor cells in a dose-dependent and cell-specific manner in vitro. The shedding inhibition was very efficient as BAMO3 was able to prevent MICA/B shedding at very low concentrations, as low as 0.1 µg/ml, and at 10 µg/ml the MICA/B shedding was almost completely blocked. BAMO3 is widely applicable as it prevented shedding of different MICA and MICB allelic variants expressed by various cancer cell lines. To elucidate the molecular action of BAMO3, the epitope of BAMO3 was mapped to the membrane proximal site of the MICA α3 domain suggesting that BAMO3 acts by steric hindrance of metalloprotease cleavage.

MICA is highly polymorphic and there are over 100 different alleles recognized in the current literature. Several alleles possess two, four or even nine GCT repeats, in the transmembrane region of MICA, respectively. The results of this study suggest that the length of alanine repeats in the transmembrane domain of MICA does not affect MICA shedding inhibition by BAMO3.

Next, a cysteine in position 250 in MICA (positioned in the α3 domain), that does not seem to form an intramolecular disulfide bond with any cysteine partner, was studied. Mutation of this cysteine (Cys250Ala) led to mitigation of shedding intensity of Cys250Ala MICA from B16F10 cells. Moreover, the surface expression of Cys250Ala MICA was increased while ADAM10 metalloprotease surface expression was decreased. Further, co-immunoprecipitation experiments indicated a lower intensity interaction between Cys250Ala MICA and ADAM10.

Finally, anti-tumor effects of anti-MICA antibodies were studied in vivo. Administration of BAMO3 delayed growth of RMA-MICA lymphoma in a syngeneic mouse model in MICA transgenic (MICAgen) mice and thereby improved survival of tumor-bearing mice. Tumor-infiltrating NK cells expressed higher NKG2D levels upon BAMO3 treatment as compared to control-treated mice. Overall, these data demonstrated therapeutic efficacy of BAMO3 treatment in the mouse RMA lymphoma model, and, together with other data from our laboratory, endorse BAMO3 treatment as a potential novel immunotherapeutic approach for treatment of human malignant diseases.

In addition, therapeutic efficacy of a drug-conjugated antibody against MICA (19E9-PBD) was tested in MICAgen mice bearing various aggressive mouse tumors. 19E9-PBD demonstrated a tremendous efficiency in delaying the tumor growth and improving survival of MICAgen mice. A decreased MICA expression on immunosuppressive myeloid-derived suppressor cells (MDSCs) was detected in the tumors. Moreover, the combination therapy of 19E9-PBD together with an anti-PD-1 antibody showed therapeutic efficacy in delaying the growth of MICA-expressing tumors.

In conclusion, these results presented powerful antibody-based agents for tumor elimination of MICA-positive tumors. They may be used to complement and enhance the antineoplastic effects of other anti-cancer therapeutics or immune checkpoint inhibitors that are currently used for patients with various malignancies.