Recombinant REIC-Dkk-3 proteins were expressed using a previously developed supergene expression (SGE) system (27,28). The expression plasmid DNA [pIDT-SMART (C-TSC)-REIC] for expression of the full-length REIC/Dkk-3 (FL-REIC) protein was described previously (27). The cDNA fragment encoding an N-terminal truncated form of C-REIC [Arg142-Ile350] was amplified with PCR primers containing the EcoRI and BamHI restriction sites. The PCR products were first cloned into the p3xFLAG-CMV-9 expression vector (Sigma-Aldrich, St. Louis, MO, USA) to express FLAG-tag fused C-REIC protein. To obtain efficient recombinant protein expression with the SGE system, the open reading frame was cloned into the pIDT-SMART (C-TSC) vector.

Both FL-REIC and C-REIC were transiently expressed in FreeStyle™ 293-F cells (Life Technologies, Carlsbad, CA, USA) using Freestyle 293 Expression Medium and the 293 Fectin transfection reagent (Life Technologies), according to the manufacturer’s instructions. Briefly, exponentially growing cells (1×10⁶ cells/ml) with 180 ml media were prepared in a 500-ml flask. After transfection with 180 μg each of expression plasmid DNAs and 293 Fectin complex, the cells were cultivated using an orbital shaker (125 rpm) at 37°C in the presence of 8% CO₂ for 4 days. Secreted proteins in the culture media were concentrated by Amicon Ultra centrifugal filter units (Millipore, Billerica, MA, USA), and the buffer was then replaced with 20 mM HEPES buffer (pH 7.2) using Sephadex G25M column chromatography (GE Healthcare, Piscataway, NJ, USA). Subsequently, the proteins were purified by anion exchange column chromatography (DEAE-Toyopearl 650M; Tosoh, Tokyo, Japan) and eluted with a linear NaCl gradient (0 to 0.7 M). The solution of the recombinant proteins was changed to PBS using a Sephadex G25M column, and then sterilized with 0.22 μm Millex-GV syringe filters (Millipore) and stored at −70°C until use for the biological experiments.

During the optimization of purification procedures for REIC/Dkk-3 proteins, degraded products were often detected on SDS-PAGE. This degradation converged to a 17-kDa band on SDS-PAGE, which was no longer degraded with long incubation times. This limited degradation product (C17-REIC) was analyzed for its amino-terminal sequence with a protein sequencer (Applied Biosystems 491), and carboxyl terminal amino acids were determined by amino acid analyzer (L-8500; Hitachi, Japan) after hydrazinolysis of the protein.

Human peripheral blood monocytes (PBMCs) were prepared from the blood of healthy donors by a standard method involving Ficoll-Paque centrifugation. The cell collection rate was determined by the trypan blue exclusion method. The survival rate was confirmed to be 99% or greater. For preparation of the monocytes, PBMCs were resuspended in LGM-3 (serum-free lymphocyte growth medium-3; Lonza, Walkersville, MD, USA). The cells adhering to a plastic dish (subjected to incubation in a 10-cm dish at 37°C for 2 h) were used as monocytes. In some experiments, CD14⁺ monocytes were separated using CD14⁺ magnetic-activated cell sorting microbeads (MACS; Miltenyi Biotec, Bergisch Gladbach, Germany). Purified CD14⁺ monocytes were resuspended in LGM-3 medium.

CD14⁺ monocytes were cultured in LGM-3 medium with or without DC differentiation factors. As a positive control, 2 ng/ml each of GM-CSF and IL-4 (both from R&D Systems, Minneapolis, MN, USA) were added to the medium. As for REIC/Dkk-3 proteins, 10 μg/ml of purified recombinant proteins was added. After cultivation for 7 days, the solution was stirred manually, and after 3 min, the number of DC-like cells per randomly selected visual field was counted with magnification of the slightly expanded photographs. The data were converted into a graph (n=5 visual fields). The cells were observed with a phase contrast microscope.

Purified CD14⁺ monocytes were incubated for 6 h in LGM-3 medium with 2 ng/ml GM-CSF or 10 μg/ml REIC protein. Total cellular proteins were prepared from the treated cells, and western blot analysis was performed as previously described (21). Proteins were identified using the following antibodies: anti-phospho-Akt (Ser473), anti-phospho-glycogen synthase kinase 3β (GSK-3β) (Ser9), anti-GSK-3β, anti-phosphorylated signal transducers and activators of transcriptions (STAT)3 (Tyr705) and anti-phospho-STAT5 (Tyr694) (Cell Signaling Technology, Beverly, MA, USA).

Murine renal carcinoma (RENCa) cells (1×10⁶) were subcutaneously injected into mice (BALB/c, female, n=5). On days 3, 5, 7, 10, 12 and 14 after injection (provided that day 3 after injection was designated as the day of the start of administration of REIC proteins), 100 μg each of FL-REIC or C17-REIC, both proteins dissolved in 100 μl of PBS, or PBS as a control was intraperitoneally injected into mice. On day 17, the therapeutic effects were evaluated as tumor volume, and anticancer immune activity was measured before mice were euthanized. All experiments were conducted in accordance with the guidelines for animal experiments of our institution.

EDTA (0.2% solution, 30 μl) was added to 750 μl of mouse blood collected from the inferior vena cava as an anticoagulant. Antibodies (1 μl each) with different fluorescent labels (purchased from eBioscience) were added to 30 μl of blood, stirred and incubated at 4°C for 60 min to stain immunocompetent cells as follows: DCs (anti-CD11c antibody and anti-CD80 antibody) or cytotoxic T cells (anti-CD8 antibody and anti-CD69 antibody).

Subsequently, erythrocytes were lysed in a red blood cell lysis buffer. Cells were washed twice with PBS and resuspended in 200 μl of PBS to generate a solution for analysis. A total of 3×10⁴ cells were collected using a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA) and analyzed using CellQuest software (Becton Dickinson). An appropriate gate was set on the basis of the forward scatter pattern characteristic of these cells, and only cells within the gate were analyzed.

Data are expressed as the means ± standard error. Differences between two groups were analyzed using the unpaired Student’s t-test, and p<0.05 was considered statistically significant.

To elucidate the molecular mechanism underlying the induction of anticancer immune responses by REIC/Dkk-3, the FL-REIC/Dkk-3 protein and the C terminal domain of REIC/Dkk-3 (C-REIC) containing two cysteine (Cys)-rich domains were produced in Freestyle 293-F cells (Life Technologies) (Fig. 1A). Secreted REIC/Dkk-3 protein was recovered from the culture medium of transfected 293-F cells on day 4. Approximately 100 mg of purified FL-REIC was obtained from a 1-liter culture using this system. Since the expression of C-REIC protein with the original signal peptide showed a low yield, the signal peptide was replaced by the Met-preprotrypsin leader sequence (PPT LS) preceding the FLAG coding sequence of the p3xFLAG-CMV-9 vector.

The stability of the FL-REIC protein was tested by incubation at 37°C, which resulted in the detection of degraded products on SDS-PAGE. Although the degradation mechanism was unclear, this proteolytic degradation was enhanced with unpurified FL-REIC protein in a low-salt buffer. However, the degraded protein products converged in a band of ~17 kDa on SDS-PAGE, and longer incubation periods did not result in additional degradation products. Amino acid sequencing of this product resulted in the identification of Ser135 as the amino terminal residue and Phe288 as the carboxyl terminal residue (Fig. 1A). The purity of the REIC/Dkk-3 protein was determined as greater than 90% by SDS-PAGE (Fig. 1B).

Previous studies from our group showed that FL-REIC is a DC-like cell differentiation factor for monocytes when used at a range of 1–10 μg/ml (21). To investigate whether the C-REIC and C17-REIC proteins induce DC differentiation, CD14⁺ monocytes were incubated with 10 μg/ml of purified REIC proteins. After a 7-day culture, imaging with a phase-contrast microscope showed DC-like differentiation of cells treated with GM-CSF/IL-4 and each of the three REIC proteins (FL-REIC, C-REIC and C17-REIC) (Fig. 2A). The number of DC-like cells per randomly-selected visual field (n=5) was counted, and the three types of REIC proteins had a comparable effect (Fig. 2B). These results indicated that C17-REIC, composed of two Cys-rich domains, is essential for the induction of DC-like cell differentiation.

Recently, intracellular activation of both STAT3 and STAT5 was found to play a role in the development of DCs (29–31). In our previous study, we showed that REIC/Dkk-3 phosphorylates STAT1 and STAT3 in monocytes (21). Furthermore, phosphorylation of GSK-3β on Ser9 by GM-CSF is thought to be involved in DC maturation (32). To evaluate the phosphorylation of STAT3, STAT5 and GSK-3β induced by REIC/Dkk-3, monocytes were treated for 6 h with REIC/Dkk-3 or GM-CSF. Both REIC/Dkk-3 and GM-CSF induced the phosphorylation of STAT3, STAT5 and GSK-3β in the monocytes, although the effective dose for REIC/Dkk-3 was much higher than that of GM-CSF (Fig. 3A and B). By contrast, heat treatment of REIC/Dkk-3 protein abrogated the effect on the phosphorylation of STAT3 in lymphocytes (Fig. 3A). Consequently, activation of STAT signaling and GSK-3β inactivation depending on Ser9 phosphorylation is a biological activity unique to the REIC/Dkk-3 protein, and it is not caused by possible contaminants, such as lipopolysaccharides. Since phosphorylated Akt also induces the phosphorylation of GSK-3β (33,34), we analyzed the activation of the PI3K/Akt pathway. Our results showed that Akt was not activated in response to REIC/Dkk-3 treatment (Fig. 3B). Taken together, these results revealed that REIC/Dkk-3 induces the phosphorylation of GSK-3β in monocytes independently from the PI3K/Akt pathway.

We previously demonstrated that intratumoral administration of FL-REIC inhibited tumor growth in vivo through the induction of cancer immunity (21). To investigate the antitumor potential of REIC/Dkk-3, FL-REIC and C17-REIC proteins were intraperitoneally injected into tumor-bearing mice (Fig. 4A). Significant tumor growth suppression was observed 17 days after the injection of the REIC/Dkk-3 proteins. Tumor volumes were statistically significantly smaller in the group treated with both FL-REIC and C17-REIC than in the group treated with PBS (Fig. 4B). These antitumor effects of the REIC/Dkk-3 proteins were accompanied by in vivo induction of CTL (CD69⁺/CD8⁺) and DCs (CD11c⁺/CD80⁺) in the peripheral blood (Fig. 4C). Taken together, these results revealed that intraperitoneally injected REIC/Dkk-3 proteins exhibited antitumor effects mediated by the activation of systemic immunity, and the cysteine-rich core domain was essential for these biological responses.