Cloning, Expression and Characterization of a Peptibody To Deplete Myeloid Derived Suppressor Cells in a Murine Mammary Carcinoma Model

Myeloid derived suppressor cells (MDSCs) are an immature heterogeneous population of myeloid lineage that attenuate the anti-tumor immune responses. Depletion of MDSCs has been shown to improve ecacy of cancer immunotherapeutic approaches. Here, we produced and characterized a recombinant peptibody capable of recognizing and depleting murine MDSCs. Using SOE-PCR, the coding sequence of the MDSC binding peptide and linker were synthesized and then ligated into a home-made expression plasmid containing mouse IgG2a Fc. The peptibody construct was transfected into CHO-K1 cells by lipofectamine 3000 reagent and the resulting fusion protein was puried with protein G column and subsequently characterized by ELISA, SDS-PAGE and immunoblotting. The binding prole of the peptibody to splenic MDSCs and its MDSC depletion ability were then tested by ow cytometry. The puried peptibody appeared as a 70 kDa band in Western blot. It could bind to 98.8% of splenic CD11b + /Gr-1 + MDSCs. In addition, the intratumoral MDSCs were signicantly depleted after peptibody treatment compared to their PBS-treated negative control counterparts (P <0.05). In this study, a peptibody capable of depleting intratumoral MDSCs, was produced. Our results imply that it could be considered as a potential drug effective for immunotherapy of cancers.


Introduction
Myeloid derived suppressor cells (MDSCs) are known as an immunosuppressive population of myeloid lineage that are enriched in cancer and chronic infections [1,2]. MDSCs are primarily generated from myeloid precursors of bone marrow in response to tumor-derived factors such as VEGF, TGF-β, IL-6, IL-10, CSF-1 and GM-CSF, and are subsequently recruited to blood, spleen and tumor bed [3]. MDSCs are functionally distinguished from the normal myeloid cells based on their ability to inhibit the immune responses mediated by T cells, B cells and natural killer (NK) cells [4][5][6]. These immune inhibitory effects of MDSCs mostly contribute to expression of the regulatory mediators including arginase1, S100A8/A9, NO, reactive oxygen species, immunosuppressive cytokines and surface immune checkpoints [7]. Moreover, MDSCs facilitate metastasis of primary tumors to distant organs through formation of premetastatic niches [8]. MDSCs are involved in matrix remodeling by induction of MMP8 and MMP9, angiogenesis process and support the metastatic potency of circulating cancer cells [9,10]. In a recent study it was reported that epigenetic therapy could disrupt the pulmonary premetastatic niches in murine tumor models by down-regulation of CXCR2 and CCR2 chemokine receptors that mediate the MDSC migration to peripheral organs [11]. Overall, MDSCs take part in the immune escape, invasion and progression of tumors to other organs [12,13].
MDSC-targeting agents are involved in diverse mechanisms to overcome immunosuppression mediated by MDSCs by interfering in MDSC differentiation, hampering their recruitment to tumor site, depleting MDSCs and by reprogramming MDSCs in tumor microenvironment [14]. In this context, early studies demonstrated that all-trans retinoic acid (ATRA) induce differentiation of MDSCs into mature macrophages and DCs and augment the e cacy of anti-VEGFR2 therapy in a breast cancer animal model [15,16]. Besides, cytotoxic chemotherapeutic agents including 5-uorouracil, carboplatin, paclitaxel or gemcitabine have been shown to diminish MDSCs in blood circulation and improve anti-tumor immune responses [17][18][19][20]. However, these drugs reduce other rapidly proliferating cell populations such as T cells and highlight the need for development of more speci c agents without off-targeting effects. Francis Mussai et al. discovered depleting effects of an immunotoxin, anti-CD33 monoclonal antibody (gemtuzumab) conjugated to ozogamicin, on human MDSCs that improved the e cacy of CAR-T cell immunotherapy [21]. This immunotoxin was investigated in a phase II clinical trial and could signi cantly deplete CD33+ MDSCs with promising results [22,23]. Moreover, anti-GR-1 monoclonal antibody is also able to signi cantly reduce the MDSC frequency in tumor bearing mice [24,25]. However it mainly reacted with PMN-MDSCs and thus, M-MDSCs could escape from the elimination by Gr-speci c antibodies. Recently, Hong Qin et al. generated a novel therapeutic peptibody that e ciently depleted both PMN-MDSCs and M-MDSCs in tumor-bearing mice [26]. Here, a peptibody containing the MDSC-binding peptide and the Fc domain of mouse IgG2a was produced in a eukaryotic system and was characterized in terms of recognition and depletion of MDSCs in 4T1 mouse tumor model.

Materials And Methods
Design and cloning of the peptibody coding sequence The MDSC binding peptibody contains a short speci c peptide that is linked to the Fc domain of mouse IgG2a antibody (Gene ID: 380793) by a glycine-serine linker (GS linker) together with an IL-2 signal peptide that is located upstream of the sequence (Fig. 1a) [26]. SmaI and BstbI restriction sites were inserted upstream of the peptide and downstream of the linker sequences, respectively. Two slow codon pairs in GS linker were exchanged with a fast counterpart by only a single nucleotide change of A to T (GGAGGC to GGTGGC). In order to produce codon optimized MDSC-speci c peptide and linker sequences (84 bp), three overlapping primers were designed and applied in Splicing by Overlap Extension (SOE) PCR (Bioneer, Daejeon, Korea) ( Table 1). The product of each PCR was used as a template for the next PCR. The nal PCR product was digested with SmaI and BstbI enzymes (Fermentas, Tokyo, Japan) and subsequently cloned into a home-made expression plasmid containing mouse IgG2a Fc and glutamine synthetase (GS) gene as a selection marker. The recombinant construct was ligated using T4 DNA ligase (Fermentas) and transformed into chemically competent E. coli DH5a host cells through heat shock approach. Positively transformed colonies were screened by colony PCR using primers designed for CMV promoter (forward primer: 5-CAGACATAATAGCTGACAGACTAAC-3) and SV40 poly-A tail (reverse primer: 5-ATACCTACCAGTTCTGCGC-3) (Bioneer). The PCR reaction was carried out with an initial melting step at 94°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 Sec, annealing at 60°C for 30 Sec, and extension at 72°C for 1 min with a nal extension step at 72°C for 6 min. Moreover, the selected colonies were con rmed using the enzymatic digestion and DNA sequencing (Applied Biosystems 3500, CA, USA). Table 1. Overlapping primers applied in SOE PCR to extend MDSC-speci c peptide and linker sequences. To produce MDSC speci c peptide and linker GS, the two overlapping sequences were designed and fused using the complementary part of sequences (underlined in table) in a PCR cycle. Then, the SmaI and BstbI restriction sites were joined in 5 and 3 ends of the product, respectively by overhanging primer pairs 1 and 2 carrying special annealing sequence of the template molecule ends. SmaI restriction site: CCCGGG, BstbI

Stable expression of the peptibody by transfected CHO-K1 cells
The culture medium was replaced with glutamine free DMEM medium (Sigma) supplemented with 10% dialyzed FBS (HyClone™, GE Healthcare, Marlborough, MA, USA) and L-Methionine sulfoximine (MSX, 25µM) (Sigma) 48 hours post-transfection. After two weeks, the production level of the peptibody by each colony in the supernatant was measured by ELISA. Colonies with the highest expression level were selected and subcloned in 96-well plates by limiting dilution method. The nal clone was cultured in serum free First CHOice medium (UGA Biopharma, Hennigsdorf, Germany) containing Feed alpha and Feed beta supplements (UGA Biopharma) and maintained for two weeks at 31°C in 5% CO 2 .
Puri cation of the peptibody from the culture supernatant by protein G column The collected supernatant of stable CHO-K1 clone producing the peptibody was centrifuged (4000 ×g for 5 min at 4°C) and then the clear supernatant was concentrated using a stirred ultra-membrane cell with a 10-kDa cut-off (Amicon 8400, Minden, Germany) under N2 gas pressure. The concentrated sample was then passed through a HiTrap Protein G HP column (GE Healthcare). The peptibody was eluted from the protein G column by a glycine-HCl buffer (0.1M, pH 2.7). The concentration of the peptibody was determined using BCA protein assay kit (Thermo scienti c, Rockford, IL, USA).

Characterization of the peptibody by SDS-PAGE and Western blot
The puri ed peptibody was mixed with 6X sample buffer with and without β-mercaptoethanol and boiled.

Determination of binding speci city of peptibody to MDSCs
To prepare FITC conjugate, the puri ed peptibody was dialyzed against bicarbonate buffer (0.1M, pH 8.3) overnight at 4°C. One milligram FITC was dissolved in 1 ml DMSO and subsequently 20 µl FITC was mixed with the peptibody (1 mg) in a nal volume of 1 ml at room temperature for one hour. Next, the mixture was dialyzed against PBS overnight at 4°C [27].

Statistical analysis
Statistical analyses were conducted by Graph Pad Prism 8 software (Graph Pad Software Inc., San Diego, CA, USA) using two-tailed t-test to compare MDSC percentages in treated group of mice with the control group that received PBS. P value < 0.05 was de ned statistically signi cant and represented as *.

MDSC binding peptide production in eukaryotic CHO-K1 cells
The sequence of the MDSC binding peptide and the GS linker were assembled by SOE-PCR which resulted in production of a 111 bp fragment (Fig. 1b). The ampli ed sequence was then inserted into the plasmid that carried the coding sequences of mouse IgG2a Fc protein. Transformed E. coli colonies harboring the correct construct had a 1317 bp amplicon in colony PCR using speci c primers (Fig. 1c). Two selected colonies (C12 and C13) were con rmed using enzymatic digestion (Fig. 1d). Furthermore, DNA sequencing analysis of these two colonies indicated that the MDSC binding peptide was successfully cloned into the expression vector in a correct frame shift without any mutation.
It is interesting to mention that before codon pair optimization of GS linker sequence, the maximum transient expression level of the transfected cells was 12 ng/ml despite several attempts to improve the production level. Codon pair optimization of the MDSC-peptide coding and GS linker sequences improved peptibody production level by 20.8 folds (250 ng/ml) after transient transfection. Finally, after multiple rounds of subcloning and selection of stable clones, a stable high producer clone which was producing 8 µg/ml of the peptibody was selected and expanded.
Characterization of MDSC binding peptide using SDS-PAGE and immunoblotting The size and structure of the puri ed peptibody were investigated by SDS-PAGE and the peptibody migrated as a single band. The molecular weight of the peptibody was around 70 kDa under non-reducing conditions and its reduced form showed a smaller 35 kDa band (Fig. 2a). The recombinant peptibody was also detected by immunoblotting with HRP conjugated sheep anti-mouse IgG as a 70 kDa protein band (Fig. 2b).
We next determined the binding characteristics of the peptibody to MDSCs. When the 4T1 tumor size reached 760 mm 3 , the splenocytes were harvested and stained with the FITC-conjugated peptibody and APC-anti-CD11b and PE-anti-Gr1 antibodies as MDSC markers. The peptibody bound to 98.8% of CD11b/Gr1 positive MDSCs, however, the control peptibody failed to recognize this cell population (Fig. 3). The CD45/CD11b/Gr-1 positive MDSCs substituted 66.2% of the splenocytes obtained from the 4T1 tumor bearing mice.

MDSCs depletion by the peptibody in 4T1 tumor model
The peptibody was intraperitoneally administrated at 50 µg/mouse/day; 17, 18 and 19 days after induction of 4T1 tumors (tumor volume: 1500 mm 3 ). Flow cytometric analysis showed that intratumoral MDSCs were signi cantly depleted after peptibody treatment compared with the control group that received PBS alone (39.6% compared to 57.3%) (P <0.05) (Fig. 4). However, no signi cant reduction of MDSCs was observed in splenocytes of tumor bearing mice (Fig. 4).

Discussion
It is well known that tumor microenvironment is in ltrated with various types of immune suppressive cells including the immature myeloid cells, MDSCs [28]. MDSCs certainly have critical roles in tumor development and are documented as a major immunosuppressive population against e cient immunotherapy [29,30]. In this context, MDSC-targeting strategies lead to MDSC depletion or inhibition of their immunosuppressive function in malignancies. ATRA, an isomer of retinoic acid, can mediate differentiation of MDSCs to mature cells such as macrophages, dendritic cells and granulocytes in human and mouse tumors. By induction of MDSC differentiation, ATRA enhances anti-tumor immune responses of CD4 + and CD8 + T cells and improves the e cacy of cancer vaccines [31]. In a lymphoma model, 5-uorouracil and gemcitabine dramatically decreased MDSC numbers in spleen and tumor tissues. This inhibition of MDSC population was associated with induction of the INFγ production in tumor-speci c CD8 + T cells and development of anti-tumor T cell dependent immune responses [32,33]. In addition, anti-Gr-1 monoclonal antibody eliminates Gr-1 + granulocytic MDSCs and can restore the antitumor T cell responses.
However, it causes elimination of mature granulocytes and consequently induces a severe systemic immunosuppression [34]. Thus, the nding of a speci c MDSC-depleting agent with poor off-targeting activity is valuable to overcome the robust tumor-induced immunosuppression and leading to enhancement of cancer immunotherapies in clinical practice. In order to discover an e cient MDSC-targeting agent, Hong  Table 1). We exchanged the slow codon pairs with a fast counterpart in GS linker by only a single nucleotide change of A to T (GGAGGC to GGTGGC). In agreement with Trinh study [35], we observed that a fast codon pair improved peptibody production levels from 12 ng/ml to 250 ng/ml in the CHO-K1 transiently transfected host. Our efforts to produce stable cell line with original sequences also led to increased production to 8 ug/ml. Thus, in both transient and stable transfection plans, codon pair optimization resulted in a signi cant improvement in peptibody production level.
It was suggested that overrepresented codon pairs may reduce e ciency of transition process and may thus act as ribosomal pause sites. These slow pairs give the intact polypeptides an opportunity to fold properly by chaperones. So, it seems that elimination of the slow codon pairs to accelerate the translation step may result in increasing the yield of recombinant protein production [36,37].
When 4T1 murine mammary carcinoma cells were implanted orthotopically into mammary fat pad of female Balb/c mice, they could spontaneously metastasize to bone, liver, lungs and spleen organs which was also affected in late-stage cancer patients [38,39]. The invasive 4T1 tumor growth causes a leukemoid reaction and splenomegaly accompanied by massive myeloid cell in ltration that was partly mediated via colony-stimulating factor and chemokines secreted by 4T1 tumor cells [40]. The predominant population of these leukocytes has been reported to be CD11b + /Gr-1 + immature myeloid MDSCs that accumulate in primary tumors, spleen, lung and peripheral blood [30,41]. Transcripts of G-CSF and GM-CSF and myeloid cell chemokines MCP-1, KC, RANTES, MIP-1a and MIP-1b have been detected in cultured 4T1 cells in vitro which suggests that 4T1 tumor cells individually cause a remarkable increase in myeloid in ltrations of primary tumors and metastatic foci. Our ndings consistently showed that 4T1 tumor promoted the MDSCs expansion in spleen and tumor beds [40]. Interestingly, in our experiments 4T1 induced MDSCs to constitute 89% and 57% of CD45 + population in spleen and primary tumor, respectively, four weeks after tumor inoculation (Fig. 4). This suppressive population probably participates in establishment and survival of 4T1 tumors within 3-4 weeks using blocking of innate and adaptive anti-tumor immune responses directed against 4T1 cell antigens and as a result reduction of immune surveillance.
The puri ed peptibody bound to 98.8% of splenic MDSCs isolated from 4T1 tumor implanted mice and could deplete MDSC population in vivo. Intraperitoneal injection of three doses of peptibody signi cantly reduced intratumoral MDSC numbers but this reduction was not signi cant for splenic MDSCs. One explanation for this nding is that four weeks after tumor inoculation the MDSCs constitute 89% of the leukocytes in spleen. Thus, administration of 3 doses of peptibody may not be su cient to induce signi cant reduction of MDSCs. However, the frequency of MDSCs in tumor microenvironment is not so high to counteract the peptibody depletion effects. The dose and stability of the peptibody administered into the peritoneum might also affect its functional failure in the spleen. Assessment of the effect of peptibody on MDSCs at early stages of tumor development, during the rst 2 weeks of implantation, may help to elucidate the underlying reasons.

Conclusion
In this study, we successfully constructed and produced a fusion protein containing MDSC-binding peptide in CHO-K1 cells and improved the expression level of the peptibody by codon pair optimization in GS linker.
The peptibody was able to deplete MDSC population resident in 4T1 tumor microenvironment. This peptibody could be applied for combination immunotherapy to overcome the MDSC immune suppressive function.

Declarations
Ethics approval and consent to participate: Not applicable Consent for publication: Not applicable Availability of data and materials: All data generated or analyzed during this study are included in the manuscript.
Competing interests: The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.
Funding:    Flow cytometric analysis of MDSCs in the spleen and primary tumor of 4T1 bearing mice after treatment with peptibody. Peptibody was intraperitoneally injected at 50 µg dose per day for three days, and then spleens and tumors were harvested to analyze by ow cytometry. Control group was given PBS and in each group was included 4 female Balb/c mouse. Peptibody treatment could decrease the splenic MDSCs to 70.5% proportion compared to PBS received group and signi cantly depleted the intratumoral MDSCs to 39.6% (a, b). PEP: peptibody

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryTable.1.docx