Cloning and expression of porcine Colony Stimulating Factor-1 (CSF-1) and Colony Stimulating Factor-1 Receptor (CSF-1R) and analysis of the species specificity of stimulation by CSF-1 and Interleukin 34

Abstract

Macrophage Colony Stimulating Factor (CSF-1) controls the survival, differentiation and proliferation of cells of the mononuclear phagocyte system. A second ligand for the CSF-1R, Interleukin 34 (IL-34), has been described, but its physiological role is not yet known. The domestic pig provides an alternative to traditional rodent models for evaluating potential therapeutic applications of CSF-1R agonists and antagonists. To enable such studies, we cloned and expressed active pig CSF-1. To provide a bioassay, pig CSF-1R was expressed in the factor-dependent Ba/F3 cell line. On this transfected cell line, recombinant porcine CSF-1 and human CSF-1 had identical activity. Mouse CSF-1 does not interact with the human CSF-1 receptor but was active on pig. By contrast, porcine CSF-1 was active on mouse, human, cat and dog cells. IL-34 was previously shown to be species-specific, with mouse and human proteins demonstrating limited cross-species activity. The pig CSF-1R was equally responsive to both mouse and human IL-34. Based upon the published crystal structures of CSF-1/CSF-1R and IL34/CSF-1R complexes, we discuss the molecular basis for the species specificity.

Fig. 1

Alignment of cloned porcine CSF-1 with human, mouse, canine and feline CSF-1. Alignment was performed using Clustal W (http://www.ebi.ac.uk/Tools/msa/clustalw2/) and demonstrates the high level of homology that exists between porcine and human CSF-1. The red colour represents small hydrophobic amino acids, blue colour represents acidic amino acids, magenta denotes basic amino acids, and the green colour corresponds to hydrophilic or polar amino acids. Identical amino acids are represented by “∗”, conserved substitutions are represented by “:” and semi-conserved substitutions are represented by “.”. (Note: signal peptide amino acid sequence is missing from feline CSF-1).

Fig. 2

Alignment of the cloned porcine CSF-1R extracellular domain with human, mouse, canine and feline CSF-1R. Alignment was performed using Clustal W (http://www.ebi.ac.uk/Tools/msa/clustalw2/) and demonstrates the high level of homology that exists between porcine and human CSF-1R. The red colour represents small hydrophobic amino acids, blue colour represents acidic amino acids, magenta denotes basic amino acids, and the green colour corresponds to hydrophilic or polar amino acids. Identical amino acids are represented by “∗”, conserved substitutions are represented by “:” and semi-conserved substitutions are represented by “.”.

Fig. 3

Expression of cloned porcine CSF-1 (A) and CSF-1R (B). (A) Western blot of secreted cloned porcine CSF-1 by HEK293T cells transfected with porcine CSF-1_pEF6 expression construct or empty pEF6 construct. Under non-reducing conditions, two bands of approximately 37 KDa and 50 KDa were detected, whereas two smaller bands (20 and 25 KDa) were detected in the presence of dithiothreitol (DTT). Porcine CSF-1 is secreted as a disulphide linked dimer that is variably glycosylated. (B) Western blot of cloned expressed porcine CSF-1R transfected into Ba/F3 cells. Cells were cultured in either rh-CSF-1, IL-3 or both factors combined prior to collection of cell lysate. Differential levels of receptor expression were noted when the cells were cultured in these different conditions. Receptor expression is reduced when Ba/F3 cells expressing CSF-1R are cultured in rh-CSF-1 alone.

Fig. 4

Demonstration of biological activity of cloned secreted porcine CSF-1 transfected into HEK293T cells and porcine CSF-1 expressed in E.Coli. (A) An MTT cell viability assay was performed using Ba/F3 cells expressing porcine CSF-1R and supernatant collected from transfected HEK293T cells with porcine CSF-1_pEF6 expression construct. Using 80% of the HEK293T transfected cell supernatant produced viable cells. (B) Porcine bone marrow cells cultured with 20% supernatant collected from HEK293T cells transfected with porcine CSF-1_pEF6 expression construct differentiated into BMDMs, adhered to the tissue culture plate and proliferated compared to cells cultured with supernatant from HEK293T cells transfected with empty pEF6 construct which did not adhere, proliferate or survive after 7 days in culture (C). (D) Bacterially expressed Porcine CSF-1 has also shown to be biologically active on the porcine CSF-1R expressed in Ba/F3 cells in an MTT cell viability assay. (E) An MTT cell viability assay was performed using mouse BMMs cultured with supernatant collected from transfected HEK293T cells with porcine CSF-1_pEF6 expression construct. Using either 50% and 20% supernatant produced viable cells. All of the assays shown are representative of three separate experiments with either triplicate or quadruplicate determinations.

Fig. 5

Activity of porcine and human recombinant CSF-1 on canine and feline bone marrow progenitor cells. Canine and feline bone marrow cells were cultured with either 10⁴ U/ml rh-CSF-1, 300ng/ml porcine E.Coli expressed CSF-1 or no growth factors. By day 5 of differentiation, canine and feline BMDMs with no growth factors were dead (canine A & feline D). By day 5 of differentiation for canine and day 12 for feline, BMDMS were attaching to the culture dish when cultured with either rh-CSF-1 (canine B & feline E) or porcine CSF-1 (canine C & feline F).

Fig. 6

Activity of recombinant human and mouse IL-34 on porcine CSF-1R expressed in Ba/F3 cells. An MTT cell viability was used to assess the biological activity of human and mouse IL-34 on expressed porcine CSF-1R. (A) Both human and mouse IL-34 are biologically active on the porcine receptor. (B) Human IL-34 has demonstrates similar activity to rh-CSF-1 on the porcine CSF-1R in an MTT assay.

Fig. 7

3D models of non-conserved contact amino acids between human and mouse CSF-1R and IL-34. 3D models demonstrating the charged amino acid changes of human and mouse CSF-1R and IL-34 were generated using the PDB file 4DKD (human) and 4EXP (mouse). Published contact amino acids for both human and mouse CSF-1R and IL-34 were analysed and non-conserved contact amino acids highlighted using FirstGlance (http://firstglance.jmol.org). (A) 3D model of human CSF-1R non-conserved amino acids for IL-34 binding and (B) mouse CSF-1R non-conserved amino acids for IL-34 binding. (C) 3D model of human IL-34 non-conserved contact amino acids and (D) 3D model of mouse IL-34 non-conserved contact amino acids. Positively charged atoms are represented by blue colour and negatively charged atoms by red colour. Medium blue coloured atoms denote partially charged atoms.

Fig. 8

3D models of non-conserved contact amino acids between human and mouse and porcine CSF-1 and CSF-1R. 3D models demonstrating the charged amino acid changes of human, porcine and mouse CSF-1 were generated using the PDB file for mouse CSF-1 (3EJJ) was used as a template for both human and porcine CSF-1 models. Published contact amino acids of mouse CSF-1 were analysed and non-conserved contact amino acids highlighted using FirstGlance (http://firstglance.jmol.org). (A) 3D model of mouse CSF-1 non-conserved contact amino acids, (B) 3D model of porcine CSF-1 non-conserved contact amino acids and (C) human CSF-1 non-conserved contact amino acids. Positively charged atoms are represented by blue colour and negatively charged atoms by red colour. Medium blue coloured atoms denotes partially charged atoms.

Fig. 9

3D models of non-conserved contact amino acids between mouse, human and porcine CSF-1 binding sites of CSF-1R. 3D models highlighting the non-conserved CSF-1 contact binding sites of mouse, human and porcine CSF-1R. Human and mouse CSF-1R models were generated using the PDB file 4DKD (human) and 4EXP (mouse). Porcine CSF-1R was generated using human CSF-1R (4DKD) as template. The 6 non-conserved amino acids involved in CSF-1 binding to CSF-1R (Glu78, Asn13, Asn85, His6, Phe55, Arg79) were identified and the coresponding binding site of CSF-1R identified (A) Mouse CSF-1R highlighting the position of Lys151 and Gly232. (B) Human CSF-1R and (C) Porcine CSF-1R highlighting these amino acids. Whilst the change from Lys to His on the human CSF-1R does not alter the amino acid properties, there is an increase in both the molecular and residue weight which could potentially constrain the binding of the mouse protein to the human receptor. The substitution of Gly232 in the mouse receptor with the larger Asn232 in human might produce a steric hindrance that is not seen when histidine is substituted as in the pig. Positively charged atoms are represented by blue colour and negatively charged atoms by red colour. Medium blue coloured atoms denotes partially charged atoms.