Recombinant Immunocytokines Targeting the ... - ACS Publications

mice, respectively. These results indicate that biologically effective IL2 and GM-CSF immunocytokines combine the targeting ability of an antitransfer...
1 downloads 0 Views 178KB Size
482

Bioconjugate Chem. 1998, 9, 482−489

Recombinant Immunocytokines Targeting the Mouse Transferrin Receptor: Construction and Biological Activities Torsten Dreier,† Holger N. Lode,† Rong Xiang,† Carrie S. Dolman,† Ralph A. Reisfeld,*,† and Angray S. Kang‡ Department of Immunology and Department of Molecular Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037. Received February 16, 1998; Revised Manuscript Received May 7, 1998

Localized cytokine therapies with recombinant monoclonal antibody-cytokine fusion proteins, designated immunocytokines, have become of increasing interest for tumor immunotherapy, since they direct immunomodulatory cytokines into the tumor microenvironment. To investigate their mechanisms of action in a variety of syngeneic tumor models, recombinant mouse cytokines IL2 and GM-CSF were engineered as fusion proteins to the carboxyl terminus of a chimeric rat/mouse antitransferrin receptor antibody, ch17217 and expressed in stable-transfected Chinese hamster ovary cells. The recombinant immunocytokines were readily purified by affinity chromatography and their binding characteristics were identical to those shown for the ch17217 antibody. The IL2 immunocytokine had an activity similar to recombinant mouse IL2, whereas the GM-CSF immunocytokine had enhanced cytokine activity relative to recombinant mouse GM-CSF. The clearance rates of ch17217 and the GM-CSF and IL2 immunocytokines were relatively similar with elimination phases (t1/2R) of 1.8 h and distribution phases (t1/2β) of 83, 88, and 91 h, respectively. Both immunocytokines demonstrated effective antitumor activity by suppressing the growth of hepatic metastases of mouse neuroblastoma and pulmonary metastases of mouse colon carcinoma in syngeneic A/J and BALB/c mice, respectively. These results indicate that biologically effective IL2 and GM-CSF immunocytokines combine the targeting ability of an antitransferrin receptor monoclonal antibody with the immunomodulatory functions of each cytokine. Because of the universal expression of the transferrin receptor on mouse tumor cell lines, these constructs should prove useful to determine their efficacy in a wide variety of syngeneic mouse tumor models and to perform detailed studies of their modes of action.

INTRODUCTION

The increased availability of recombinant cytokines in pharmacologic quantities has stimulated a number of research efforts to assess the effect of these immunomodulatory molecules on the host immune response to cancer. Among a variety of cytokines, IL2 has been most widely applied, since it is one of the most potent antitumor cytokines with biological activities that include the stimulation of proliferation of such antitumor effectors as cytotoxic T cells, natural killer (NK) cells, and macrophages (1, 2). This cytokine has been applied with some success for the treatment of patients with melanoma and renal carcinoma (3). However, the systemic administration of IL2 is limited by serious side effects, even at nanomolar concentrations (4). Therefore, investigators started to focus on local cytokine treatments, e.g., intratumoral cytokine injection or cytokine gene therapy which, in contrast to systemic application, comply with the paracrine mode of action of these immunomodulators being optimally effective within a few cell diameters from their cell of origin (5). Gene therapy, i.e., the ex vivo transfection of cytokine genes into autologous tumor cells that are subsequently injected into a host, has been widely applied in clinical trials and preclinical animal models. When such cyto* To whom correspondence and reprints should be addressed. Phone: 619-784-8105. Fax: 619-784-2708. E-mail: reisfeld@ scripps.edu. † Department of Immunology. ‡ Department of Molecular Biology.

kines, including IL2 and GM-CSF, are produced by the transfected tumor cells, they induce a local inflammatory response, resulting in some cases, in a systemic immune response effective against challenge with wild-type tumor cells (6, 7). Some encouraging preclinical and early clinical data were obtained with this strategy (8); however, the feasibility of its broad clinical application as a cancer therapy is limited by its patient-specific nature. We have developed an alternative approach for tumor immunotherapy by using immunocytokines with the unique ability to direct multifunctional cytokines to the tumor microenvironment (9). This approach conforms with the paracrine function of most cytokines and is not limited by systemic toxicities or patient specificity. We previously demonstrated in syngeneic mouse tumor models of neuroblastoma (10), colon carcinoma (11), and melanoma (12) that recombinant immunocytokines comprised of different tumor-specific antibodies and IL2 can eradicate established disseminated tumor metastases, followed in some cases by a long-lived tumor-protective immune response (11). However, the antibody specificity for one distinct target antigen limited the use of such constructs to only a small number of syngeneic mouse tumor models. Here, we describe the construction, characterization and biological activity of two recombinant immunocytokines comprised of a monoclonal antibody against mouse transferrin receptor and mouse IL2 and GM-CSF, respectively. In both cases, the antigen-binding capacity and cytokine functions were maintained, and effective suppression of metastasis was demonstrated in two

S1043-1802(98)00020-2 CCC: $15.00 © 1998 American Chemical Society Published on Web 06/03/1998

Recombinant Immunocytokines

syngeneic models of mouse neuroblastoma and colon carcinoma. The universal expression of the transferrin receptor on mouse tumor cell lines should make these immunocytokines applicable to ascertain their antitumor activity in all syngeneic mouse tumor models and to perform in-depth studies of their underlying effector mechanisms. EXPERIMENTAL PROCEDURES

Antibodies and Cell Lines. The rat hybridoma R17217, secreting an antibody of IgG2b isotype, directed against the mouse transferrin receptor (Tf-R), was kindly made available by Dr. Jayne F. Lesley (Salk Institute, La Jolla, CA). This antibody was extensively characterized and demonstrated not to affect functions of Tf-R involved in iron transport into cells or to inhibit the growth of tumor cells (13, 14). The GM-CSF-dependent acute mouse hematopoietic stem cell line 32D(G) was kindly provided by Dr. D. Santoli (Wistar Institute, Philadelphia, PA). The IL2dependent mouse leukemia CTLL-2 and CHO cell lines were obtained from the American Type Culture Collection (Rockville, MD). Mice. Syngeneic female A/J and BALB/c mice were obtained at 6-8 weeks of age from Jackson Laboratory (Bar Harbor, ME). The mice were maintained under specific pathogen-free conditions. All experiments were performed according to NIH Guidelines for Care and Use of Laboratory Animals. Construction of Immunocytokines. The variable region of the rat immunoglobulin chains were amplified by RT-PCR using mRNA isolated from R17217 hybridoma as a template. To access the VL chain, a 5′-primer was designed according to sequence information available for these regions (15). We selected GAYGTNCARATGACNCARTCNCC as the 5′-primer for amplification of the variable region of the rat VL. The 3′-primer was designed to anneal to the rat κ light chain region (GGATGATGTCTTATGAACAA). In a second PCR reaction, the OKT3-leader sequence was added upstream and a NarI site downstream of the primary VL PCR product. The construction of the chimeric rat-mouse light chain was completed by joining this VL via the NarI site with a mouse Cκ fragment (Figure 1A). The variable heavy chain (VH) was amplified in a similar manner. Briefly, we used the 5′-primer GAGGTGCARCTNGTNGAATCNGGAGGAGG and selected the 3′-primer from the region of the rat J-minigene (GGTGACCATGGTTCCTGGGCCCCAGAAGTC). In a second PCR, the N418 hamster immunoglobulin-derived leader sequence was added upstream and a Csp45I site downstream of the original VH-PCR-product, respectively (Figure 1A). This VH construct was ligated to a mouse IgG2a constant heavy-chain backbone via the Csp45I site, which had been engineered at the 5′ end of the IgG2a. The mouse IgG2a backbone was constructed with an SmaI site located seven amino acids upstream of the original 3′ end of mouse IgG2a. In the case of the recombinant antibody, we introduced a synthetic oligonucleotide in this SmaI site, providing a stop codon followed by an EcoRI site. For the construction of the immunocytokines, the mature sequences of the mouse cytokines IL-2 and GM-CSF were amplified from the RNA of Concanavalin-A stimulated mouse splenocytes by RT-PCR with the following primers: GGACCCCGGGAGCACCCACTTCAAGCTCC (IL-2 5′); CTGAATTCTTATTGAGGGCTTGTTGAGATGATGC (IL-2 3′); GGACCCCGGGAGCACCCACCCGCTCACCC (GM-CSF 5′); and

Bioconjugate Chem., Vol. 9, No. 4, 1998 483

CTGAATTCTGGGCTTCCTCATTTTTGGCTTGG (GMCSF 3′). The resulting PCR products were inserted as SmaI-EcoRI fragments at the 3′ end of the heavy chain construct. Construction of the Expression Vectors. For expression purposes, the light-chain construct was cloned into the multiple cloning site of the pcDNA3.1/zeo vector (Invitrogen, Carlsbad, CA) using the NheI and XhoI restriction sites. The heavy-chain constructs were cloned as NheI to EcoRI fragments into the pBK-CMV vector (Stratagene, La Jolla, CA). The entire heavy-chain expression unit was isolated after digestion with NsiI followed by blunt ending with mung bean nuclease and a MluI digest. A double expression vector was constructed by cloning the heavy-chain expression unit into the pcDNA3.1/zeo vector containing the light-chain expression unit using ScaI and MluI restriction sites (Figure 1B), thus sacrificing the ampicillin gene used for selection in prokaryotic cells. However, the selectable zeocin marker is retained. Cell Culture and Transfection. Chinese hamster ovary (CHO) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum and 1% nonessential amino acid solution (Biowhittaker, Walkersville, MD). Transfection with the double expression vectors was done at cell densities of 70-80% using SuperFect (Qiagen, Chatsworth, CA) according to the manufacturer’s protocol. Twenty-four h after transfection, the cells were transferred to 96-well plates and drug selection was initiated by supplementing the media with 200 µg/mL zeocin (Invitrogen, Carlsbad, CA). Zeocin-resistant colonies developed within 2 weeks after transfection. The supernatants of these colonies were assayed for mouse IgG2a production by ELISA and dot blot analyses. The highest producers (10-20 µg/mL) were expanded to produce the desired immunocytokines. Production and Purification of Immunocytokines. For ease of purification, the medium of 90% confluent cells cultured in roller bottles was changed to serum-free IS-CHO medium (Irvine Scientific, Santa Ana, CA) and culture continued in this medium for another 10 days. The immunocytokines produced were isolated from the spent culture media by protein A Sepharose chromatography and concentrated by ultrafiltration to a final concentration of 1-2 mg/mL. The molecular size and purity of the proteins was analyzed by SDS-PAGE. Immunoassays. Identification of the chimeric antibodies or immunocytokines was performed by ELISA and dot blot analyses. The capture antibody used in the ELISA was an Fc-specific goat anti-mouse antibody, and the detection antibody was an alkaline phosphataselabeled κ-chain-specific goat anti-mouse antibody (Sigma, St. Louis, MO). An alkaline phosphatase-labeled goat anti-mouse antibody was used for dot blots. The mouse IgG2a antibody UPC10 (Sigma, St. Louis, MO) served as a positive control and the supernatant of nontransfected CHO cells as a negative control. Cytokine Activity Assays. IL-2 activity was measured by a mouse T-cell proliferation assay (16). Briefly, the IL-2 dependent murine leukemia cell line CTLL-2 (ATCC) was cultured for 48 h without IL2. Then, 3 × 104 cells were incubated for 24 h with serial dilutions of either recombinant murine (rm) IL2 or the ch17217-IL2 immunocytokine, respectively, in a total volume of 200 µL. GM-CSF activity was detected in a proliferation assay with the GM-CSF-dependent cell line, 32D(G). The cells were added to 96-well plates (5000 cells/well) together with serial dilutions of either rmGM-CSF or the

484 Bioconjugate Chem., Vol. 9, No. 4, 1998

Dreier et al.

Figure 1. Schematics depicting DNA constructs encoding for the ch17217-IL2 and GM-CSF immunocytokines and their expression vector. (A) Schematic of heavy and light chains; the variable region of the rat R17217 light chain (VL) was constructed downstream of an OKT3 leader sequence (OL) and subsequently cloned upstream of the constant region of a mouse κ light chain (Cκ) using the NarI restriction site. The variable region of the rat R17217 heavy chain (VH) was combined with a hamster-derived N418 leader sequence (NL) and introduced in the heavy-chain backbone of mouse Ig2a by using the Csp 45I restriction site. (B) Double expression vector with heavy and light chains. The ch17217 light-chain expression unit was cloned into the multiple cloning site of the pcDNA 3.1/zeo vector as an NheI XhoI fragment. Heavy-chain expression units for each fusion protein were introduced as ScaI MluI fragments, generated as described in the Experimental Procedures. The zeocin resistance gene remained intact as a functional selection marker in both eukaryotic and prokaryotic cells.

ch17217-GM-CSF immunocytokine and incubated for 24 h. [methyl-3H]Thymidine uptake was measured 16 h after addition of 0.5 µCi/well. All samples were tested in triplicate. Radiolabeling. Radiolabeling with 125I was performed, as previously described (17). Briefly, 1 mg of the respective protein was incubated for 30 min on ice with 1 mCi of 125I (100 mCi, 3.75 Gbq)/mL (Amersham, Arlington Heights, IL) in polystyrene tubes coated with 100 µg of Iodo-Gen reagent (Pierce, Rockford, IL). Nonincorporated 125I was removed by gel filtration on PD10 columns (Pharmacia, Piscataway, NJ).

Saturation Binding Assays. Binding assays were performed with ch17217 and both immunocytokines on several mouse tumor cell lines (RENCA, NXS2, CT26). Although all of these cell lines expressed Tf-R, we selected RENCA mouse renal carcinoma cells for our analyses, since these were adherent and provided the most reproducible saturation binding data. For that purpose, 9 × 104 cells were added to 24-well microtiter plates and cultured for 24 h. Saturation binding was then determined by adding the respective 125I-radiolabled proteins at various concentrations to the cells followed by incubation on ice for 2 h. Cells were washed three

Recombinant Immunocytokines

times with ice cold PBS and 1% BSA, and cell-bound radiolabel was determined in a γ-scintillation counter. Nonspecific binding was determined for all proteins in the presence of a 200-fold excess of unlabeled mAb R17217. The number of binding sites and the binding constants for each construct were determined by Scatchard plot analysis (18). Pharmacokinetics. Six week old BALB/c mice were used to determine the blood clearance of mAb ch17217 and immunocytokines, ch17217-IL2 and ch17217-GMCSF. Groups of four mice each were injected into the lateral tail vein with 25 µg of each of these radiolabeled proteins (specific activity of 0.2 µCi/µg). Blood samples were collected from the orbital sinus at 0.5, 1, 2, 4, 8, 24, 48, and 72 h after injection of the radiolabeled proteins and assayed for 125I activity in a γ-scintillation counter. Elimination (t1/2R) and distribution (t1/2β) phase half-lives were calculated assuming exponential clearance according to the formula A ) Ao e-kt using the two earliest and latest time points, respectively (19). Syngeneic Tumor Models. Pulmonary colon carcinoma and hepatic neuroblastoma metastases were induced in syngeneic BALB/c and A/J mice, respectively, as described previously (10, 11). Briefly, 1 × 105 or 1 × 104 NXS2 mouse neuroblastoma cells were injected into the lateral tail vein of A/J mice to induce hepatic metastases. CT26 mouse colon carcinoma cells (1 × 104) were injected intravenously (i.v.) for induction of pulmonary metastases. Mice were treated 24 h after tumor cell inoculation with daily i.v. injections (×7) of either PBS, 5 and 30 µg ch17217-IL2, or 30 µg of ch17217-GMCSF. Macroscopic metastases to livers and lungs were determined 24-28 days after tumor cell inoculation. Since the metastatic foci appeared fused, metastases were scored according to the percentage of lung and liver surface involvement. Statistics. The statistical significance of differential findings between experimental groups of mice was determined by the nonparametric Mann-Whitney rank sum test. Results were regarded significant if two-tailed P values were