The highest-producing of HFI fusion clone was harvested after the HFI expression vector had been transfected into adherent Chinese Hamster Ovary (CHO-K1) cells (ATCC No. CRL-9096), incubated with the selection medium containing 300 μg/ml G418 (also known as geneticin) (31,32). The primary cell bank (PCB) of the HFI engineering cells was built when the adherent CHO cells was domesticated into suspension cells in tissue culture with HyClone SFM4CHO (HyClone) or EX-CELL™325 (Sigma-Aldrich) cell culture medium, which is a protein-free formulation that contained no components of bovine origin. The master cell bank (MCB) and the working cell bank (WCB) with the stable expressing HFI were obtained from the PCB in tissue culture roller bottles with the cell culture medium, as described above. The supernatant containing fusion protein of HFI was harvested as the cell culture process of the Cell Bioreactor (BIOSTAT^® Bplus), which was determined through the batch culture or the continuous flow culture. The fusion protein of HFI was purified over the following sophisticated purification processes, including microfiltration, ion-exchange, affinity and molecular sieve chromatography.

Quantification of the HFI was assayed by ELISA. Goat anti-human (100 μl) IgG (1:200 dilution) per well was coated onto 96-well ELISA plates overnight at 4°C. After washing the plates, serial dilution of the samples, which were maintained in PBS with 3% bovine serum albumin (BSA) were added to the plates and incubated at 37°C for 2 h. Then after washing the plates, 100 μl per well of horse-radish peroxidase-conjugated goat anti-human IgG (1:1,000 dilution) was bound and incubated at 37°C for 1 h. Color was generated by adding peroxidase substrate and absorbance was measured on a plate reader at 492 nm. Dilutions of the HFI, for which the concentration was calibrated by the Kjeldahl method and the Lowry method, were used to generate a standard curve to estimate the concentration of unknown samples.

The cytokine activity of the HFI was determined by IL-2 dependent T cell proliferation assay. Recombinant human IL-2 (rhIL-2) standard was obtained from National Institute for the control of Pharmaceutical and Biological products. The mouse CTLL-2 cell line expressing high-affinity IL-2 receptor was used to determine the IL-2 bioactivity by the T cell proliferation assay. Serially diluted samples and IL-2 standards were incubated with 4×10⁴ cells per well in triplicate in 96-well flat bottom plates for 24 h at 37°C in a humidified atmosphere of 5% CO₂. MTT (5 mg/ml) reagent was added 4 h before the end of culture, and then cells were lysed with 10% SDS and 50% N,N-dimethyl formamide, pH was maintained at 7.2. The OD values were read at 570 nm. The mouse CTLL-2 cell was provided by Professor N. Guo (Institute of Basic Medical Sciences, China).

About 1×10⁶ cells per well were incubated with varying concentrations of the HFI or the same molar concentrations of the different proteins in a final volume of 0.2 ml. After incubation on ice for 30 min, cells were washed twice with PBS-BSA, and a fluorescein isothiocyanate (FITC)-conjugated goat anti-human IgG antibody (Beijing Dingguo Biotechnology Co. Ltd.) was added. Incubation was continued for an additional 30 min on ice. After two washes with PBS-BSA, the cells were analyzed by flow cytometry (FACSCalibur, Becton Dickinson).

The three types of mice including female FVB/neu mice, BALB/c mice and female Balb/C athymic nude mice of 7 to 10-weeks-age, were obtained from Shanghai Laboratory Animal Center of the Chinese Academy of Sciences. All mice were housed under specific pathogen-free conditions. Experiments were carried out according to the National Institute of Healthy Guide for Care and Use of Laboratory Animals, and were approved by the Institutional Animal Care and Use Committee at the Shanghai Institute of Materia Medica, Chinese Academy of Sciences.

BALB/c mice (48 animals per group) were injected with 25 or 100 μg of the purified HFI sample in a volume of 0.2 ml in the tail vein using a slow push, respectively. At various time points, small blood samples (6 animals per time point) were taken by retro-orbital bleeding and collected in tubes coated with heparin to prevent clotting. After centrifugation, the plasma was assayed by ELISA with goat anti-human IgG and goat anti-human IL-2 antibody. The results were normalized to the initial concentration in the serum of each mouse taken immediately after injection.

When the signals of HER2 were blocked on the surface of the HER2 overexpression cells, the growth of cells was inhibited or the cells entered apoptosis. The three types of cells including SKBR3 cells (HER2⁺⁺⁺), SKOV3 cells (HER2⁺⁺) and MCF7 cells (HER2^±) were provided by Professor N. Guo (Institute of Basic Medical Sciences, China), and were chosen to evaluate HFI antitumor activity in vitro(33). Firstly, the cells were placed on 96-well cell plates, then, serial dilution of the samples were added to the plates and incubated at 37°C in a humidified atmosphere of 5% CO₂. Apoptosis of the cells were assayed for different time points by MTT cell proliferation assay methods.

i) Six to seven-week-old female Balb/C athymic nude mice were subcutaneously injected 1×10⁷ Calu-3 cells (human non-small cell lung cancer) into the right flank. The mice were treated with PBS (iv: 0.2 ml per mice) (control) or HFI (iv: 1.0, 0.5 and 0.25 mg/kg mice weight, respectively). Seven or eight mice were used in each group. ii) The antitumor activity of HFI was evaluated in FVB/neu transgenic mouse model, which expressed the ErbB2/neu proto-oncogene and closely recapitulated the ontogeny and progression of human breast cancer. When tumors reached a size of more than 300 mm³, the mice were grouped (n=5) and PBS or HFI (1.0 mg/kg mice weight) or Herceptin (3.0 mg/kg mice weight) in a volume of 0.2 ml, was injected i.v. into the lateral tail vein. Dosage regimen: the first week for five continuous days, then once every other day. The mice weight were monitored daily and tumor volume was measured with a caliper once per two days, using the formula volume = length × width²/2.

Splenocytes were harvested on day 20 after HFI administration and were cultured in triplicate (5×10⁵/well) in 96-well culture plates (Corning) for 24 h with 0.5 μg/ml of concanavalin A (Con A) or 1 μg/ml of lipopolysaccharide (LPS). The cells were pulsed with 0.5 μCi/well of [³H]thymidine for 8 h and harvested onto glass fiber filters. The incorporated radioactivity was then counted using a Beta Scintillation Counter (MicroBeta Trilux, PerkinElmer Life Sciences, Boston, MA).

Splenocytes were cultured in 24-wells culture plates (Corning) for 24 and 48 h with 0.5 μg/ml of Con A or 1 μg/ml of LPS. Culture supernatants were harvested and submitted to assays for the determination of cytokine concentrations. Cytokines in culture supernatant was assayed using mouse IFN-γ ELISA kit (BD Pharmingen), according to the manufacturer’s instructions.

Murine YAC-1 cells were used as targets in the NK assay. Target tumor cells were labeled with 250 kBq [³H]thymidine by incubating overnight in a culture dish (Corning) containing 10 ml of culture medium. Targeted cells (2×10⁶) were cultured in triplicate with different numbers of effector cells (effector: Target ratio = 100:1, 50:1 or 12.5:1) for 4, 8 and 12 h in 96-wells flat-bottomed plates. Cells were harvested onto glass fiber filters. The incorporated radioactivity was then counted using a Beta Scintillation Counter (MicroBeta Trilux, PerkinElmer Life Sciences, Boston, MA). Cytotoxic activity was calculated and expressed as the mean of percent cytotoxicity. Percent cytotoxicity = (cpm in effector-free wells − cpm in effector positive wells)/(cpm in effector-free wells) ×100.

A single-cell suspension of spleen cells harvested on day 20 after HFI administration were blocked with anti-mCD16/CD32 (2.4G2, purified inhouse), stained with fluorescein isothiocyanate (FITC)-conjugated, phycoerythrin (PE)-conjugated, (PerCPCy5.5)-conjugated, (APC)-conjugated, and analyzed using a FACSCalibur system with Cell Quest software (Becton Dickinson). (FITC), (PE), (PerCPCy5.5) or (APC)-conjugated mAb against CD3 (145-2c11), CD4 (Gk1.5), CD8 (Ly-2) (53-6.7), CD1 (1d3), TCR-γ/δ (GL3) were purchased from BD Pharmingen. PerCPCy5.5-conjugated anti-mouse/rat FoxP3 (FJK-16s), PerCPCy5.5-conjugated rat IgG2a and the FoxP3 staining buffer set were purchased from eBioscience.

The statistical significance of the time course data was determined by two-way analysis of variance testing and Student’s t-test was used to compare two treatment groups. Statistical significance was set at p<0.05.

The stable HFI expression was first assayed in the differential course of cell passage by using a pair of goat anti-human IgG antibodies in a ‘sandwich ELISA’, SDS-PAGE and IL-2 dependent T cell proliferation. The fusion proteins of HFI were stably expressed in different cell lines, such as the primary cell line (PCL), the master cell line (MCL), the working cell line (WCL) and the production of end-stage cell line (ECL) in 1,000 ml tissue culture roller bottles with a culture volume of 200–300 ml and at (2–4)×10⁶ cells/ml density at 30°C for 7 days. The culture supernatants were harvested following 1,500 rpm for 5 min to 8,000 rpm for 30 min. The molecular weight of HFI was found in a bivalent form of 140 kDa (Fig. 1A) and a monovalent form of 70 kDa (Fig. 1B) on 8% SDS-PAGE stained with Coomassie Brilliant Blue. The HFI expression of the four kinds of cells cultured for 7 days was examined at 100–150 mg/l by ELISA (Fig. 1C). The IL-2 bioactivity of HFI was assessed at (2–2.5) ×10⁶ IU/mg by the standard IL-2 dependent T cell proliferation assay and ELISA calibration concentration (data not shown). Fig. 1 shows that the HFI gene copy number and HFI expression were relatively stable in the differential course of cell passage.

Since the expression of HFI in the cell lines was stable, the purification and stability of HFI was crucial to the ability of this fusion protein. The supernatant of HFI reaction liquid was sampled at 30°C and assayed for one month by the Cell Bioreactor under a continuous flow mode. In the HFI purification process, insoluble cell debris and other irrelevant substances were first removed from the HFI supernatant by 0.2 or 0.45 μm microfiltration membrane bag (Millipore), then the microfiltration sample was concentrated with the 30–50 kD nitrocellulose membrane bag (Millipore) to a proper volume. This concentration step was critical for HFI purification, as it could not only remove a large number of cellular components, but also deplete certain host proteins and small molecules in culture medium. In order to protect the affinity chromatography media from erosion, the ultra-filtration sample was purified by DEAE ion exchange in 0.2–0.4 M NaCl, 40 mM phosphate-citrate buffer and pH 7.2 was maintained, this process was mainly to remove endotoxin, nucleic acids and other substances. Purity of antibody protein reached 97% or higher after affinity chromatography purification. Therefore, the HFI was purified over Mabselect Sure (GE Healthcare) affinity chromatography, eluting with acidic conditions. The purified solution of the HFI was obtained by Superdex 200 (GE Healthcare) chromatography media to remove HFI monomer and polymer (Fig. 2). After the protein purification process, the HFI yield was measured by a microplate spectrophotometer at 490 nm, and the final HFI sample was more than 40% yield, the HFI purity was examined by reducing and non-reducing SDS-PAGE (Fig. 2A), the quantity of HFI monomers and polymers were less than 5% of HFI dimers (Fig. 2A; lanes 7 and 8), more than 98% purity of the HFI sample was determined by HPLC (Fig. 2B), fluorescent PCR, and host protein ELISA. Endotoxin was monitored by LAL agglutination. The HFI bioactivity was consistent with IL-2 activity, measured by a standard IL-2-dependent T cell proliferation assay (Fig. 2C). Meanwhile, the stability of the purified solution of the HFI was also analyzed at 4°C for 2 months. These results suggested that the fusion protein of the HFI was relatively stable, and could be suitable for further drug development.

The HFI built targeting was based on the binding properties of the ant-ErbB2 antibody and interleukin-2 bioactivity. The biological activity of IL-2 of the HFI was based on the ability of the HFI to support IL-2 dependent T cell proliferation (Fig. 2C). Human CD3⁺ and CD16⁺ cells were detected from the peripheral blood of mice at several doses of the HFI treated groups (33). The binding characteristics of the HFI was identified, and then three HER2 expressing cells, SKBR3 (HER2⁺⁺⁺), SKOV3 (HER2⁺⁺) and MCF7 cells (HER2^±) were used to test HFI binding activity in different tumor cell lines. In the same molar concentration, HFI, Herceptin (Trastuzumab Injection, Roche Pharma Ltd.) and HF (antibody part of HFI, self-made sample) showed the same binding ability (Fig. 3). The HFI binding capacity was not affected by IL-2 fusion with HFI as compared to HF (Fig. 3A–D). This result demonstrated that the HFI binding ErbB2 protein characteristics depended on the expression of the ErbB2 protein (Fig. 3A–D). HFI showed a similar characteristic of binding HER2 with the Herceptin (Fig. 3A–D).

The pharmacokinetic properties of immunocytokines were reported to be dependent on their uptake by the FcR-bearing cells and the intracellular proteolysis (28). Half-life of drugs was related to their molecular size. The structure of HFI was determined by the in vivo stability, and properties of IL-2 and anti-ErbB2 protein. The half-life of HFI was detected in the mouse serum at different time points by ELISA. The results showed that the concentration of HFI decreased quickly during the first 4 h, and showed a relatively stable plateau following 64 h (Fig. 4). At 100 μg HFI maintained relatively higher levels than 25 μg HFI, demonstrating that the higher levels were maintained with higher concentrations (Fig. 4). The results indicated that the HFI structural design was conducive to recruitment around the cells, and also extended the half-life of IL-2.

The HFI has a combination of activities, i.e., binding properties of HER2 receptor and interleukin-2 bioactivity. The ability of the HFI binding HER2 receptor depends on the HER2 overexpression on the cell surface. In this study, whether the HFI inhibited the cell proliferation when the HFI was bound to HER2 receptor was evaluated. The three cell types, breast cancer (SKBR-3, HER2⁺⁺⁺), breast cancer (MCF-7, HER2^±) and ovarian cancer cells (SKOV-3, HER2⁺⁺) were incubated with the different concentrations of HFI for 72 h. The results showed that the HFI did not inhibit cells proliferation, and proved that the HFI did not have chemical toxicity. In vitro, the role of HFI was to inhibit tumor cell growth by activating the effector cells.

The in vivo antitumor activity was investigated using a Calu-3 xenograft model. Calu-3 is known to be a HER2-overpressing tumor. A total of 1×10⁷ Calu-3 cells suspended in 0.15 ml PBS were injected subcutaneously into the right flank of the mice. Normally, 9–12 days after tumor cell implantation, when tumors were clearly palpable and reached a size of 200 to 350 mm³, the mice were grouped (n=7) and PBS or the HFI was administered intravenously into the lateral tail vein, respectively (Fig. 5). To determine the antitumor activity of HFI at the different doses and the effect of treatment, therapy was started in groups of mice treated with PBS or the HFI (1.0, 0.5 or 0.25 mg/kg mice weight), respectively. The HFI was administered intravenously once daily for the first 5 days (0–4 day). At day 20 of the treatment, the therapeutic effect observed in the mice treated with the 1.0 mg/kg dose was considerably better than that observed with the 0.5 mg/kg dose (P<0.01), and the 0.25 mg/kg dose (P<0.01). Fig. 6 demonstrates that the HFI had a significant antitumor activity on models expressing high levels of HER2 in vivo.

The fusion proteins (HFI) retained the activities of both HER2 binding and IL-2. The role of IL-2 was mainly to activate lymphocytes to release cytokines and other cells which indirectly killed tumor cells. Thus, nude mice could not fully reflect the HFI antitumor activity in vivo. In order to verify the integrity of antitumor activity of HFI, the high expression of HER2 transgenic mouse model (immunocompetent) of spontaneous tumor formation was chosen, and the mice were treated with HFI (Fig. 6). After 4-day treatment, HFI (1.0 mg/kg mouse weight) caused a dramatic suppression of FVB/neu in the transgenic mouse model. Moreover, after treatment for 15 days, the tumor volume which was initially <400 mm³, was undetectable.

The ex vivo immunoenhancement of HFI was evaluated on splenocyte proliferation induced by Con A and LPS. The results showed that the splenocytes of FVB/neu tumor bearing mice treated with 1 mg/kg HFI for 3 weeks had significant proliferation activity compared to those treated with PBS induced by LPS (Fig. 7A) or Con A (Fig. 7B). The production of IFN-γ was induced by FVB/neu in mice treated with HFI (Fig. 7C) or Herceptin (Fig. 7D). HFI also increased the NK cell-mediated cytotoxicity (Fig. 7E–G).

The therapeutic mechanisms of HFI were investigated ex vivo to determine whether its effects were attributable to influenced FoxP3⁺ Treg cells, and T cells. Intracellular staining of transcription factors showed that HFI treatment significantly reduced the expression of FoxP3 (Fig. 8A and C). HFI also increased the T cell population (Fig. 8B and D).