Antibody-Based Delivery of Cytokine Payloads to Carbonic Anhydrase IX Leads to Cancer Cures in Immunocompetent Tumor-Bearing Mice

Abstract

Antibody-cytokine fusion proteins can have the potential to increase the density and activity of subsets of leukocytes within the tumor mass. Here, we describe the design, production, and characterization of four novel antibody-cytokine fusion proteins directed against human carbonic anhydrase IX, a highly validated marker of hypoxia that is overexpressed in clear cell renal cell carcinoma and other malignancies. As immunomodulatory payloads we used TNF, IL2, IFNα2 (corresponding to products that are in clinical use), and IL12 (as this cytokine potently activates T cells and NK cells). Therapy experiments were performed in BALB/c mice, bearing CT26 tumors transfected with human carbonic anhydrase IX, in order to assess the performance of the fusion proteins in an immunocompetent setting. The biopharmaceuticals featuring TNF, IL2, or IL12 as payloads cured all mice in their therapy groups, whereas only a subset of mice was cured by the antibody-based delivery of IFNα2. Although the antibody fusion with TNF mediated a rapid hemorrhagic necrosis of the tumor mass, a slower regression of the neoplastic lesions (which continued after the last injection) was observed with the other fusion proteins, and treated mice acquired protective anticancer immunity. A high proportion of tumor-infiltrating CD8⁺ T cells was specific to the retroviral antigen AH1; however, the LGPGREYRAL peptide derived from human carbonic anhydrase IX was also present on tumor cells. The results described herein provide a rationale for the clinical use of fully human antibody-cytokine fusions specific to carbonic anhydrase IX.

Figure 1. Ex vivo immunofluorescence analysis of tumor and organ sections of BALB/c mice bearing CT26 tumors transfected with human CAIX.

Mice received one injection of 150 μg FITC-labelled IgG(XE114) and were sacrificed 24h later. Binding of IgG(XE114) was detected in green using an anti-FITC antibody and an AF488 labelled secondary antibody while blood vessels were stained in red using an anti-CD31 antibody and an AF594 labelled secondary antibody. Tu=tumor, St=stomach, In=intestine, Ki=kidney, Li=liver, Lu=lung, He=heart, Sp=spleen, Mu=muscle. Scale bars represent 100 μm.

Figure 2. In vitro characterization of newly designed fusion proteins.

a) XE-mTNF. b) XE-hIL2. c) XE-mIFNα2. i) cloning scheme. ii) expected protein structure, yield of TGE in CHO-S cells and expected molecular weight. iii) SDS-PAGE analysis (kDa: marker in kDa, nr: non-reduced, r: reduced). iv) Size exclusion profile of the respective immunocytokine and standard proteins indicated by black arrows (1: IgG, 145 kDa, 2: BSA, 67 kDa, 3: Diabody, 25 kDa). v) MS profile. vi) BIAcore profile. vii) Activity assay for the respective payload.

Figure 3. Activity of XE-mTNF, XE-hIL2, XE-mIFNa2 in BALB/c mice bearing CT26 tumors transfected with human CAIX.

a) Tumor bearing mice received three injections of either 3 μg XE-mTNF, 50 μg XE-hIL2 or 150 μg XE-mIFNα2 or saline (PBS) as control, starting when tumors reached a size of 90 mm³. Data are represented as mean (n=5) ± SEM. Statistical significance towards PBS control group is shown on day 13, *p<0.05, **p<0.005. b) Body weight change during the treatment phase. c) Pictures of representative tumor growth taken before treatment (day 7) and 24 h after each injection (day 8, 10, 12) as well as follow up observations (day 15, 25).

Figure 4. Tumor re-challenge and tumor infiltrate studies.

a) Cured and naïve mice received a subcutaneous tumor cell injection of CT26 cells expressing human CAIX and tumor growth was monitored for 20 days. b-e) Mice bearing human CAIX expressing CT26 tumors received one or two injections of either 150 μL of PBS as control, 3 μg of XE-mTNF, 50 μg of XE-hIL2 or 150 μg of XE-mIFNα2. (b) Time line of i.v. injections (indicated by black arrows) and tumor harvesting (indicated by a black cross) for immunohistological experiments: 24 hours after the last injection, tumors were excised and sections were stained in green for activated Caspase 3 (c), CD4+ T cells (d), or CD8⁺ T cells (e), in red for CD31 positive cells and in blue for nuclei. Scale bars represent 100 μm.

Figure 5. FACS analysis of AH1 positive CD8⁺ T cells.

The number of AH1 positive CD8⁺ T cells in human CAIX expressing CT26 tumors and draining lymph nodes (DLN) was measured 24 hours after BALB/c mice received two injections of either one injection of PBS or XE-mTNF (3 μg) or two injections of XE-hIL2 (50 μg) or of XE-mIFNα2 (150 μg). Representative FACS plots of AH1 positive CD8⁺ T cells in tumor bed and draining lymph nodes are shown. Injections and tumor and lymph node harvesting were performed as indicated in the time line included in Figure 4b.

Figure 6. Mass spectrometric identification of the LGPGREYRAL peptide eluted from MHC class I preparations of CT26 cells transfected with human CAIX.

MHC class I-peptide complexes were purified from CT26 cell lysates with the M1/42 monoclonal antibody. MHC class I bound peptides were eluted in acid, further purified on a C18 resin and analyzed by LC/MS, yielding a spectrum corresponding to human CAIX-derived peptide LGPGREYRAL, with a Xcorr of 2.03 (upper panel). The peptide differs from is murine version by a single R to Q substitution (i.e. LGPGQEYRAL). The sequence of the endogenous peptide was confirmed with a synthetic peptide which was recorded in accordance with MHC peptidome analyses (lower panel). Fragment ions (b-, y and immonium ions) typically observed in HCD fragmentation were labelled in the spectrum of the synthetic peptide.