Interaction of HmC1q with leech microglial cells: involvement of C1qBP-related molecule in the induction of cell chemotaxis

Background In invertebrates, the medicinal leech is considered to be an interesting and appropriate model to study neuroimmune mechanisms. Indeed, this non-vertebrate animal can restore normal function of its central nervous system (CNS) after injury. Microglia accumulation at the damage site has been shown to be required for axon sprouting and for efficient regeneration. We characterized HmC1q as a novel chemotactic factor for leech microglial cell recruitment. In mammals, a C1q-binding protein (C1qBP alias gC1qR), which interacts with the globular head of C1q, has been reported to participate in C1q-mediated chemotaxis of blood immune cells. In this study, we evaluated the chemotactic activities of a recombinant form of HmC1q and its interaction with a newly characterized leech C1qBP that acts as its potential ligand. Methods Recombinant HmC1q (rHmC1q) was produced in the yeast Pichia pastoris. Chemotaxis assays were performed to investigate rHmC1q-dependent microglia migration. The involvement of a C1qBP-related molecule in this chemotaxis mechanism was assessed by flow cytometry and with affinity purification experiments. The cellular localization of C1qBP mRNA and protein in leech was investigated using immunohistochemistry and in situ hybridization techniques. Results rHmC1q-stimulated microglia migrate in a dose-dependent manner. This rHmC1q-induced chemotaxis was reduced when cells were preincubated with either anti-HmC1q or anti-human C1qBP antibodies. A C1qBP-related molecule was characterized in leech microglia. Conclusions A previous study showed that recruitment of microglia is observed after HmC1q release at the cut end of axons. Here, we demonstrate that rHmC1q-dependent chemotaxis might be driven via a HmC1q-binding protein located on the microglial cell surface. Taken together, these results highlight the importance of the interaction between C1q and C1qBP in microglial activation leading to nerve repair in the medicinal leech.


Background
In the mammalian central nervous system (CNS), microglial cells constitute the resident immune cells, maintaining the integrity of the nervous system and able to respond to any kind of brain damage [1]. In healthy brain, resting microglial cells show a ramified morphology [2]. Under pathophysiological conditions, they rapidly change their morphology and change to amoeboid activated microglia. This activation is controlled by 'on' or 'off' signals [3]. Complement proteins are potential candidates to exert such 'on' signals on microglia and can induce neuronal cell death [4]. Indeed, the complement system can be activated by three different pathways, the classical, the lectin-dependent and the alternative pathways [4]. Moreover, following human brain infection or injury, production of complement by resident cells has been clearly demonstrated to be highly increased upon activation [5]. Interestingly, C1q, the first component of the classical complement pathway, may serve as a reliable marker of microglial activation, ranging from undetectable levels of C1q biosynthesis in resident microglia to high C1q expression in activated, non-ramified microglia. C1q synthesized and released by activated microglia has been shown to maintain and regulate microglial activation in diseased CNS tissue [4,6,7]. Thus, C1q plays an important role in microglia regulation after nerve injury.
Unlike mammals, the medicinal leech Hirudo medicinalis can fully regenerate its CNS after injury and restore function of individual neurons [8,9]. For this reason, the leech CNS, which combines simplicity and well known organization [10], has been an attractive model in neurobiology for decades. After injury, leech microglia immediately move toward the lesion site. This phenomenon has been shown to be essential to promote axon sprouting and successful nervous system repair [11][12][13][14]. Leech microglial cells exhibit morphological changes similar to vertebrate ones in the course of migration in response to tissue damage [15,16]. In our laboratory, we were interested to assess the role of C1q in microglial cell accumulation after leech CNS injury.
We previously characterized, for the first time in an invertebrate nervous system, a C1q domain-containing (C1qDC) factor named HmC1q [17]. Of interest, its involvement in leech microglia recruitment following experimental injury has been clearly demonstrated. In order to study its interaction with CNS cells and elucidate its role in microglial cell chemotaxis, the recombinant form of HmC1q (rHmC1q) was produced in the yeast Pichia pastoris. In the present report, we demonstrate the chemotactic activity of the recombinant protein on leech microglial cells and we used rHmC1q to tightly explore its functions in the leech nervous system following trauma. In vertebrates, C1q has been demonstrated to exert its chemotactic activity through C1q receptors expressed on immune cells [18]. Finally, the interaction between rHmC1q and leech CNS cells was investigated, allowing the identification of a C1qBPrelated molecule, which was named HmC1qBP, homologous to the mammalian C1q receptor (alias gC1qR, p32, p33, C1qBP, HABP1; SF2p32, TAP) [19]. Therefore the involvement of a C1q domain-containing factor in microglial activation is demonstrated for the first time in a CNS.

Recombinant HmC1q production and purification Expression vector construction
The cDNA encoding the HmC1q (Genbank accession number EU581715) [17] was amplified by PCR from total leech CNS cDNA as template. Amplification was performed using specific forward (5'gcgccctacgtaatgaaag-tatttctggaaatcctcgc3') and reverse (5'taattgcggccgct-cactttctgcttgcaatt3') primers containing SnaBI (forward, bold) and NotI (reverse, bold) restriction sites, respectively, together with the predicted natural signal peptide sequence (forward, underlined). PCR amplifications were carried out on a Thermal Cycler (Eppendorf, Hamburg, Germany) with 150 ng of cDNA in a solution containing 1.25 U of Hot Start Proofreading DNA polymerase (Accu Prime™ Pfx, Life Technologies, Grand Island, NY, USA), 0.3 μM of each PCR primer, 1 × DNA polymerase manufacturer's buffer containing deoxyribonucleotide triphosphate (dNTP) in a final volume of 50 μl. The reaction cycles were performed as follows: 95°C for 2 minutes, followed by 35 cycles of 15 s at 95°C, 30 s at 60°C and 1 minute at 68°C. A single PCR product was obtained and ligated into the SnaBI and NotI digested pPIC3.5 K vector with T4 DNA Ligase (Life Technologies, Grand Island, NY, USA) according to the instructions of the Multi-Copy P. pastoris Expression Kit manual (Life Technologies, Grand Island, NY, USA). The plasmid DNA HmC1q/pPIC3.5 K was amplified into Escherichia coli Top10F' chemically competent cells (Life Technologies, Grand Island, NY, USA). Cloning steps were verified by both strands DNA sequencing (Eurogentec S.A., Liege, Belgium).

Transformation of P. pastoris strain and screening for protein expression
The recombinant plasmid HmC1q/pPIC3.5 K (see above) was linearized with SacI and used to transform GS115 P. pastoris strain by electroporation according to the method described in the manufacturer's manual (Life Technologies, Grand Island, NY, USA). Selection of His+/Mut+ transformants was achieved as previously described [20]. The recombinant clones were screened on yeast/peptone/dextrose (YPD) agar plates containing growing doses of G418 (Geneticin; Life Technologies, Grand Island, NY, USA) for the presence of multiple inserts. A total of 20 clones were inoculated in 10 ml of buffered glycerol-complex (BMGY) medium (1% w/v yeast extract, 2% w/v peptone, 1.34% w/v yeast nitrogen base, 4 μg/ml D-biotin, 100 mM potassium phosphate, pH 6.0, 1% v/v glycerol) and incubated at 29°C and 225 rpm. After 48 h, cells were pelleted by centrifugation for 2 minutes at 1,000 g at room temperature (RT). The pellets were gently resuspended in 2 ml of basal minimum medium (BMM) (1.34% w/v yeast nitrogen base, 4 μg/ml D-biotin, 100 mM potassium phosphate, pH 6.0, 0.5% v/v MeOH) and incubated at 29°C and 225 rpm. After 48 h, the cultures were centrifuged (10,000 g, 10 minutes, 4°C), and the supernatants were dried under vacuum to be checked for protein expression by western blot. The clone corresponding to the highest production of protein was stored in glycerol at -80°C.

Purification of recombinant protein
Aliquots of supernatants obtained from 32 cultures of 2 ml were centrifuged at 12,000 g for 10 minutes at 4°C, filtered through a 0.8 μm filter and concentrated until a 1 ml volume (Centricon YM-10, Millipore, Billerica, MA, USA). Purification was achieved in one step by reverse-phase high performance liquid chromatography (RP-HPLC) with a C 8 column (250 × 4.1 mm, Grace-Vydac, Columbia, MD, USA) with a linear gradient of acetonitrile (ACN) in acidified water (0,1% trifluoroacetic acid) from 2% to 32% ACN for 60 minutes at a flow rate of 1 ml/min. The presence of rHmC1q in the eluted fractions was checked by western blotting. The RP-HPLC fraction containing the recombinant protein was dried under vacuum and stored at -20°C.
For further analyses, the transformed P. pastoris culture supernatant, the RP-HPLC-purified recombinant protein and the non-transformed P. pastoris culture supernatant will be respectively referred to as 'rHmC1q supernatant', 'purified rHmC1q' and 'control supernatant'.

Western blotting
Samples (either rHmC1q supernatant, control supernatant or purified rHmC1q) were reconstituted in Laemmli buffer before loading onto a 12% acrylamide running gel and a 4% acrylamide stacking gel as previously described [21]. Briefly, migration was carried out using a cathode buffer (0.6% Tris base, 2.5% taurine, and 0.1% SDS) and an anode buffer (0.6% Tris base, 2.8% glycine, and 0.1% SDS). Gels ran at 70 V for 15 minutes and at 120 V for 45 minutes. Separated proteins were transferred to Nitrocellulose Transfer Membrane Protran BA 83 (Schleicher & Schuell Bioscience, Dassel, Germany) by electroblotting. After preincubation in blocking solution (BS) (phosphate-buffered saline (PBS) containing 0.05% Tween 20 and 2% ovalbumin fraction V) membranes were incubated overnight at 4°C with either rabbit polyclonal anti-HmC1q antibody or preimmune serum (dilution 1:1,000 in BS). Specific rabbit polyclonal anti-HmC1q antibodies were raised using a synthetic peptide corresponding to predicted His 197 -Thr 212 region of HmC1q protein (Agro-bio, La Ferté Saint Aubin, France) [17]. After three PBS washes, goat anti-rabbit or anti-mouse IgG antibodies conjugated with horseradish peroxidase (dilution 1:20 000 in BS) (Jackson Immunoresearch, West Grove, PA, USA) were added for 1 h at RT. The final washes were performed in PBS and immunolabelled proteins were revealed with the ECL Kit SuperSignal West Pico Chemoluminescent Substrate (Thermo Fisher Scientific, Rockford, IL, USA) and Kodak X-Omat LS film (Sigma-Aldrich, St. Louis, MO, USA).

Leech CNS and microglial cell preparation
All protocols regarding the use of leeches were carried out in strict accordance with the French legislation and European Treaty, and were in compliance with the Helsinki Declaration. H. medicinalis adult leeches were obtained from Ricarimpex (Eysines, France). The leech nerve cord (CNS) is constituted of 23 metameric ganglia joined by structures, called connectives, containing the axonal processes and glial cells [10]. After anesthesia in 10% ethanol at 4°C for 15 minutes, animal CNSs were dissected out in a sterile Ringer solution (115 mM NaCl, 1.8 mM CaCl 2 , 4 mM KCl, 10 mM Tris maleate, pH 7.4) under a laminar flow hood After isolation, samples were placed in three successive baths of antibiotics (100 UI/ml penicillin, 100 μg/ml streptomycin and 100 μg/ml gentamycin) for 15 minutes and further incubated in Leibovitz L-15 medium (Life Technologies, Grand Island, NY, USA) containing 2 mM L-glutamine, 0.6% glucose and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (complete medium). The experimental injury was performed by crushing the connectives between the third and fourth ganglia. Nerve cords were used for ex vivo recruitment assays, whole mount immunohistochemistry, fluorescence in situ hybridization or nerve cell preparation.
For total nerve or microglial cell isolations, nerve cords treated as indicated above were placed in 35 mm Petri dishes with 200 μl of complete L-15 medium. Each ganglion was carefully decapsulated by removing the collagen layer enveloping the nerve cord with microscissors. Nerve cells, neurons and microglial cells were mechanically resuspended by gentle scraping (total nerve cells). After a filtration through 7 μm nylon mesh as described [17,22], the enriched microglial cell population was then collected and centrifuged at 1,000 g for 10 minutes at RT. The cell pellet was resuspended in L-15 medium (100 μl per nerve cord) for migration assays.

Chemotaxis assays
In vitro chemotaxis assays were performed by using the double-P assay as described by Köhidai and colleagues with minor modifications [23]. Petri dishes (35 mm) were filled with 1 ml of a 0.5% agar and 1% gelatin solution. After drying, two 6 mm diameter wells were made, each one presenting a parallel individual channel. One well was filled with 50 μl of purified microglial cells (see above) and the other one with chemotactic factors or negative controls reagents. A channel was further created perpendicularly to others using a coverslip. By 1 h later, cells in the chemoattractant containing well were collected. Either rHmC1q supernatant (0.1, 1, 3, 8 and 15 μl) or control supernatant (0.1, 1, 3, 8 and 15 μl) were used as chemotactic factors. For inhibitory chemotactic experiments, cells were preincubated for 1 h at RT either with rabbit polyclonal anti-HmC1q antibody or with preimmune serum as negative control (both 1:250); and either with rabbit polyclonal anti-human C1qBP antibody (1:1,000) or with rabbit IgG isotype as negative control (1:1,000). The number of migrating cells was counted on a hemocytometer (three different counts) under Axioskop microscope (Zeiss, Oberkochen, Germany). Additional experiments were also performed with RP-HPLC-purified rHmC1q as chemoattractant in similar conditions. Experiments were performed in triplicate. The results were expressed as the mean cell number ± SD. Comparisons between means were made using the Student's t test. Statistical differences were considered to be significant if p was < 0.01.
Ex vivo microglial cell recruitment assays Ganglia 2, 3, 4 and 5 were dissected from the animal and pinned in separate plastic 35 mm Petri dishes (Falcon 3005, Becton Dickinson, Franklin Lakes, NJ, USA) coated with silicone rubber (Sylgard 184, Dow Corning Corp., Midland, MI, USA) and placed in L-15 complete medium. The following products were respectively injected (8 μl) inside the connectives separating the ganglia 3 and 4: PBS; rHmC1q supernatant; rHmC1q supernatant + anti-HmC1q antibody (dilution 1:5,000); rHmC1q supernatant + preimmune serum (dilution 1:5,000) or the control yeast supernatant. For injections, patch pipettes were pulled from borosilicate glass capillaries (outer diameter 1.5 mm, Clark GC 150 F-10) using a two-stage horizontal micropipette puller (model P-97, Sutter Instrument Co., Novato, CA, USA) (pipette resistance 3 to 5 MΩ). The connectives were crushed immediately after injection with fine forceps on both side of the injection site and the tissues were fixed in buffered 4% paraformaldehyde, pH 7.4 4 h after the injection. The Hoechst 33342 (Life Technologies, Grand Island, NY, USA) fluorescent dye (dilution 1:1,000 in L-15 medium) was then applied to injured nerve cords for 30 minutes to counterstain the nuclei of microglial cells. Microglial cells movement in response to these different injections was then observed with an inverted microscope (DMIRE2, Leica Microsystems, Wetzlar, Germany).

Immunohistochemistry
In experiments with anti-human C1qBP antibody, analyses were performed on nerve cords dissected out as described above and incubated 6 h in complete L-15 medium. They were fixed for 1 h at 4°C, immediately after dissection (T0) or 6 h (T6h) and 24 h (T24h) after incubation in complete L-15 medium, in 4% paraformaldehyde, washed in PBS, permeabilized by a 24 h-incubation at RT in 1% Triton X100 in PBS and preincubated for 8 h at RT in 1% Triton, 3% normal donkey serum (NDS) and 1% ovalbumin in PBS. Samples were then incubated overnight at 4°C with specific rabbit polyclonal anti-human C1qBP antibody (1:250) diluted in a PBS solution containing 1% bovine serum albumin (BSA), 0.05% Triton, 1% NDS and 1% ovalbumin (AB solution). After three washes with PBS, samples were incubated 1 h at room temperature with anti-rabbit donkey antibody (Life Technologies, Grand Island, NY, USA) conjugated to Alexa Fluor 488 (1:2,000 in the AB solution), rinsed with PBS and finally mounted with Glycergel (Sigma-Aldrich, St. Louis, MO, USA). Prior to mounting, the cell nuclei were counterstained by Hoechst dye as previously described. Samples without the addition of primary antibody were used as negative control. Slides were kept at 4°C in the dark until observation, realized with a Zeiss LSM780 confocal microscope (Zeiss, Oberkochen, Germany). We opted to present microglia nuclei in white for a better display of the results.

Fluorescent in situ hybridization (FISH)
Nerve cords were fixed for 1 h at 4°C in 4% paraformaldehyde just after dissection. The 5' biotin-labeled specific antisense probe and sense probe (negative control) were generated from a specific sequence (corresponding to the nucleotide sequence 154 to 859 of C1qBP molecule; Genbank accession number JN207836). After PCR amplification and the insertion of the product in pGEM-T easy vector system (Promega, Madison, WI, USA), the RNA sequence of interest was obtained by in vitro transcription using DIG/Biotin RNA-labeling kit according to the manufacturer's instructions (Roche Diagnostics, Rotkreuz, Switzerland). The hybridization protocol was performed as previously described [24]. Nerve cords were incubated with a secondary anti-biotin antibody conjugated to Alexa Fluor 488 (dilution 1:5,000 in PBS) (Life Technologies, Grand Island, NY, USA). Final rinsing and mounting steps for confocal microscopy observation were performed as described above.

Flow cytometry analyses
Total nerve cells isolated as described above from eight nerve cords were equally distributed (about 10 6 cells per tube) and kept for 24 h at RT in L-15 medium either alone or with purified rHmC1q. Then, they were incubated for 30 minutes with the fluorescein-labeled mouse monoclonal anti-human C1qBP (60.11) antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) (dilution 1:250 in L-15 medium). Nerve cells were then washed with L-15 medium, centrifuged for 8 minutes at 1,000 g at 4°C. Cell pellet was resuspended in 500 μl L-15 medium and finally, examined by fluorescence-activated cell sorter (FACS) (EPICS XL4-MCL, Beckman Coulter, Diagnostics division, Brea, CA, USA) equipped with an argon ion laser with an excitation power of 15 mW at 488 nm. Forward scatter (FSC) and side scatter (SSC) were analyzed on linear scales, while green (FL1) was analyzed on logarithmic scales. Data acquisition and analysis were performed using Expo32 software.

Microglial cell protein extraction
Protein extraction was performed by trichloroacetic acid/acetone precipitation and was resuspended in lysis buffer ( Protein extracts were desalted by Zeba Desalt Spin Columns, 2 ml (Pierce) following the manufacturer's guidelines. Protein concentration was determined using the Bradford method (Bio-Rad, Hercules, CA, USA) and protein extracts were stored at -20°C.

Human C1q biotinylation and streptavidin affinity purification
The biotinylation of the recombinant human C1q (Prospecbio, Rehovot, Israel) was carried out by using the sulfo-NHS-SS-biotin kit (Thermo Fisher Scientific, Rockford, IL, USA) according to the manufacturer's instructions. Briefly, proteins were biotinylated in PBS with a 20-fold molar excess for 30 minutes at RT. Unreacted sulfo-NHS-SS-biotin was removed using the Zeba Desalt Spin Columns (Thermo Fisher Scientific, Rockford, IL, USA). Biotinylated human C1q was immediately fixed onto a streptavidin column (Thermo Fisher Scientific, Rockford, IL, USA), previously equilibrated with five volumes of PBS 0.1 M. The interaction between biotin and streptavidin occurred at RT for 10 minutes. Microglia protein extract (800 μg) was added in the column, incubated overnight at 4°C and rinsed ten times with PBS 0.1 M. Captured microglial cell proteins were eluted from the streptavidin-agarose with 5% 2-mercaptoethanol/PBS 0.1 M at 30°C for 30 minutes. Proteins were precipitated in 10% trichloroacetic acid/ acetone at -20°C for 45 minutes, and centrifuged at 13,000 g for 15 minutes. The protein pellet was washed in cold acetone, air dried and dissolved in Laemmli buffer. Two other columns were used for the negative controls: the first one containing the biotinylated human C1q with no microglia protein extract and the second one containing only the microglia protein extract to evaluate the unspecific reaction between streptavidin and microglial cell components. Samples were loaded on a 12% SDS-PAGE gel and analyzed by western blotting using a rabbit polyclonal anti-human C1qBP antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), diluted at 1:5,000 in BS (see above).

Production of recombinant HmC1q in P. pastoris
HmC1q was expressed by the yeast P. pastoris under control of the methanol-inducible alcohol oxidase promoter and secreted using its natural signal peptide. A positive clone was selected following the detection of a signal using the polyclonal anti-HmC1q antibodies by western blot because of its relevance to the native HmC1q detection as previously described ( Figure 1A, control) [17]. No significant difference was detected between the recognition of rHmC1q in the transformed yeast supernatant and native HmC1q released in the conditioned medium ( Figure 1A, positive clone vs control). The negative control performed with a non-transformed yeast clone did not show any immunoreactive bands. The presence of three bands on SDS-PAGE shows slight differences in the recombinant protein, suggesting different levels of glycosylation. The shared immunoreactivity suggests that rHmC1q has the same size and migration profile as the native HmC1q. The expression of rHmC1q was observed up to 4 days without any significant proteolytic degradation (data not shown). The protein was purified by RP-HPLC and eluted as the major peak at 10% of acetonitrile (arrow in Figure 1B). The corresponding molecule (purified rHmC1q) was then analyzed by western blot by using the same antiserum and displayed specific immunoreactivity ( Figure 1B).
In vitro chemotactic activity of rHmC1q on leech microglial cells Native HmC1q has been shown to recruit leech microglial cells in a dose-dependent manner [17]. In the present study, the biological activity of the rHmC1q supernatant was compared to the control supernatant one using similar volumes in respective chemotaxis assays (0.1, 1, 3, 8, 15 μl). Freshly prepared leech microglial cells were demonstrated to migrate towards rHmC1q supernatant also in a dose-dependent manner. The optimal effect was observed by using 8 μl of rHmC1q supernatant while a higher amount (15 μl) caused a decreased migration of the microglia. This effect is a characteristic of cytokines having chemotactic activity and has been described for other chemoattractant factors of leech CNS conditioned medium [22]. Of interest, similar dose-dependent results were obtained by using purified rHmC1q (data not shown). The control supernatant used as negative control did not exert any significant effect on microglial cell recruitment (Figure 2A).
In order to maintain a relevant comparison between rHmC1q and a pertinent negative control, we used the rHmC1q supernatant (8 μl) as positive control and the control supernatant (8 μl) as negative control. Chemotaxis assays were performed using the rHmC1q supernatant and two negative control media (L-15 and control P. pastoris supernatant) in the presence of blocking antibodies directed against the leech protein (anti-HmC1q) and the human C1q receptor molecule (anti-C1qBP), respectively ( Figure 2B). Compared to rHmC1q supernatant, negative controls did not exert any significant chemotactic effect ( Figure 2B).
Interestingly, the rHmC1q supernatant-mediated microglial recruitment was inhibited by using rabbit polyclonal anti-HmC1q antibodies ( Figure 2B, black hatched bars) whereas no significant inhibitory effect was detected with the preimmune serum ( Figure 2B, gray hatched bars). In addition, no effect was observed by preincubating cells with polyclonal anti-HmC1q antibodies in negative controls (L-15 medium or the control P. pastoris supernatant). Finally, the involvement of an HmC1q binding protein as receptor in mediating microglia chemotaxis in response to rHmC1q was investigated. Pretreatment of microglial cells with rabbit polyclonal anti-human C1qBP antibodies significantly abrogated the chemotactic response to rHmC1q ( Figure  2B, black dotted bars) whereas no significant inhibitory effect was detected using rabbit IgG isotype as negative control ( Figure 2B, gray dotted bars). No neutralizing effect was observed in negative control assays (L-15 or control supernatant) ( Figure 2B).
Taken together, these results clearly indicate that leech microglial cells specifically respond to rHmC1q in a dose-dependent manner and strongly suggest that HmC1q-mediated recruitment is exerted through a C1qBP-related molecule.

Ex vivo chemotactic effect of rHmC1q on resident microglial cells in injured nerve cords
The in vitro chemotaxis assays were corroborated by the ex vivo experiments performed on whole collected and injured nerve cords. The cell movement was analyzed by Hoechst dye counterstaining because microglia are the only circulating resident cells present in connectives ( Figure 2C). The microglial cell migration at the site of injury was evaluated 4 h after injection of different molecules in nerve cords, immediately followed by experimental lesion. In the positive control experiment, the microglial cell accumulation at the lesion site strongly increased after nerve injury ( Figure 2C, a vs b). Interestingly, the injection of rHmC1q supernatant enhanced the microglia recruitment compared to those obtained in the presence of PBS only ( Figure 2C, b vs c) or in the presence of the control supernatant ( Figure  2C, c vs d). Of interest, while preincubation with rabbit preimmune serum did not have any significant neutralizing effect on rHmC1q-mediated recruitment ( Figure  2C, c vs e), the injection of rabbit polyclonal anti-HmC1q antibodies reduced the rHmC1q supernatantmediated chemotactic activity ( Figure 2C, c vs f). These results specifically confirmed the capacity of rHmC1q to increase ex vivo the microglial cell recruitment following injury, compared to a normal cell accumulation in the presence of PBS.

Characterization of a leech C1qBP related molecule
Because of the neutralizing effect observed using antihuman C1qBP antibodies in chemotaxis assays (see above), the presence of a potential C1q-binding protein (receptor-like) was investigated in an expressed sequence tag (EST) library constructed from total RNA of nerve cords isolated from adult leeches. A full-length mRNA sequence for a C1q binding proteins (C1qBP) related molecule was identified (Genbank JN207836) encoding a 289-amino-acid sequence ( Figure 3A). The theoretical molecular weight was predicted to be 32,973 Da according to the complete protein sequence. Similar to mammalian C1qBP, this sequence presents a MAM33 domain ( Figure 3A) and contains the mature form as described for the human C1qBP [19]. Blast-P analyses of the leech sequence revealed homologies exclusively with many known C1qBP sequences ( Figure 3B) with high conservation for the MAM33 domain [25,26].
Considering that only C1qBP molecules were matched from databases, whatever the rate of homology, we restricted the presentation of the Blast-P analysis to the highest homology percentages (ranging from 47% to 55%) in the table ( Figure 3B). Multiple alignments were realized to show the similarities in primary structure of human, mouse and leech forms ( Figure 3C). Interestingly, seven of the nine residues essential for conformation and ligand binding properties in the human C1qBP (Glu-89, Arg-122, Lys-123, Leu-231, Asp-232, Arg-246, Gly-247, Glu-264 and Tyr-268) are conserved in the leech sequence (Glu-104, Arg-136, Asp-240, Arg-254, Gly-255, Glu-272 and Tyr-276) [27]. These residues, which suggest comparable physicochemical features, are indicated in black boxes ( Figure 3A) and are underlined in Figure 3C. The residues Lys-123 and Leu-231 in the human sequence are replaced by Thr-137 and Met-239 in the leech one, respectively (asterisks, Figure 3C). From all these elements, the leech molecule was named HmC1qBP for H. medicinalis C1qBP.

Localization of HmC1qBP mRNA and protein in leech microglia
In order to specify the cell expressing the HmC1qBP transcripts in the leech nervous system, specific fluorescence in situ hybridization (FISH) was carried out on injured nerve cords. The transcripts were mainly  1, 1, 3, 8, 15 μl). The diagram illustrates the chemotactic activity of rHmC1q in a dose-dependent manner. Asterisks denote cell migration of the indicated sample that is significantly different (p <0.01) than the corresponding control value. (B) Inhibitory effects on rHmC1q-mediated leech microglia chemotaxis. Leech microglial migration was evaluated with either L15 medium, control Pichia pastoris supernatant or rHmC1q-containing supernatant as chemoattractant. An inhibitory effect was demonstrated by preincubation of cells either with rabbit polyclonal anti-HmC1q antibodies (black hatching in each condition) or with a rabbit polyclonal anti-human C1qBP (alias gC1qR) antibody (black dots in each condition). Negative controls consisted of cells preincubated with rabbit preimmune serum or rabbit IgG isotype respectively. Asterisks denote cell migration of the indicated sample that is significantly different (p <0.01) than the corresponding controls. (C) Chemotactic effect of rHmC1q on leech microglia in ex vivo assays. Microglial cells, which are the only circulating resident cells present in the connectives, were counterstained with Hoechst dye. The Hoechst dying was observed immediately after crush injury (see arrows) with injection of phosphate-buffered saline (PBS) (a) or 4 h after injury with injection of PBS (b) illustrating microglial accumulation at the lesion site. Injection of rHmC1q increased microglial cells recruitment (c) compared either to the presence of PBS only (b) or to control yeast supernatant (d). When tissues were injected with polyclonal anti-HmC1q antibody (f), the chemotactic activity of rHmC1q was not observed whereas the corresponding preimmune serum did not exert any neutralizing effect (e).  Figure 3D) while no specific signals were detected with sense riboprobes used as negative control (Figure 3D').
Cells containing HmC1qBP protein were investigated by immunohistochemistry with anti-human C1qBP antibodies on whole mounted injured nerve cords immediately after dissection, 6 h, or 24 h after lesion (Figure 4). The microglia recruitment was simultaneously studied by using a Hoechst dye (white nuclei). Immediately following the crush (T0), no immunostaining and no accumulation of microglial cells was observed ( Figure 4A). At 6 h after injury C1qBP staining was detected at the lesion site of connectives and some microglial cells accumulated at this site ( Figure 4B). Of interest, magnification of the injury site 6 h after injury showed that the C1qBP staining was exclusively located in some microglial cells among the whole recruited microglial population ( Figure 4C). Therefore, the timecourse evolution of C1qBP staining is directly associated with the progressive migration of microglial cells. Interestingly, since some microglial cells are still C1qBP negative it suggests the existence of different states of reactivity in microglia ( Figure 4C). The C1qBP signal is stronger 24 h after injury at the lesion site and is still correlated with the increase of accumulated microglia ( Figure  4D). Negative controls performed using the secondary antibodies alone did not show any signal ( Figure 4A',B', D'). This timecourse accumulation of C1qBP positive microglial cells highly suggests the involvement of C1qBP in the microglia recruitment.

Identification of a C1qBP molecule on leech microglia
Flow cytometry analyses were then carried out in the presence of a monoclonal anti-human C1qBP antibody ( Figure 5A-D). A dot plot of leech nerve cells incubated overnight in L-15 medium was deduced ( Figure 5A, on the right). Among the selected gate that fits on cells having parameters corresponding to those of microglial cells, 2.86% of the nerve cells (shown in green on the dot plot) were autofluorescent ( Figure 5A, on the left). When cells were exposed to the fluorescein-labeled mouse monoclonal anti-human C1qBP antibody, 46.26% were positive ( Figure 5B, shown in red on the dot plot and FL1 scale). This antibody recognized a leech protein localized on the cell surface and did not affect the cell dot plot overview. However, when the nerve cells were preincubated overnight with the purified rHmC1q, the number of immunopositive cells strongly decreased to 5.90% ( Figure 5C, on the right for the dot plot and left for FL1 scale). Indeed, flow cytometry histogram overlays clearly showed the fluorescence shift observed when A flow cytometry histogram overlay shows a shift in fluorescence when the nerve cells are incubated with anti-human C1qBP antibody (red) compared to either nerve cells that were previously incubated with rHmC1q supernatant (blue) or nerve cells alone (gray). These data are representative of three independent experiments. (E) Western blot analysis using polyclonal anti-human C1qBP antibodies from affinity enrichment with biotinylated human C1q. A specific signal corresponding to a 33-kDa molecule, according to the protein ladder (L), was detected when biotinylated human C1q was incubated with leech microglia protein extracts and then eluted on a streptavidin column (1). No immunopositive signal was observed when leech microglial protein extracts were eluted alone on streptavidin column (2) or when biotinylated human C1q was eluted alone with the streptavidin column (3). No signal was detected using the secondary antibody alone as a negative control (data not shown).
Tahtouh et al. Journal of Neuroinflammation 2012, 9:37 http://www.jneuroinflammation.com/content/9/1/37 the leech cells were preincubated with rHmC1q and exposed to anti-human C1qBP compared to the cell incubation with the anti-human C1qBP alone ( Figure  5D). These results show that rHmC1q strongly competed for the target protein specifically recognized by the anti-human C1qBP antibody and suggest that C1qBP might be the HmC1q binding protein.
Because the above results strongly suggested an interaction between HmC1q and a C1qBP, copurification experiments were performed. Preliminary attempts with rHmC1q were undertaken. However, because rHmC1q autoaggregated within some steps of the biotinylation protocol, human C1q was preferred. Indeed, human C1q does not present any massive aggregation and was shown to exhibit a chemotactic effect on leech microglia [17], similarly to rHmC1q. Once biotinylated, human C1q was incubated with leech microglia protein extracts. Following elution on an activated streptavidin column, the interactants of C1q were analyzed by western blotting using polyclonal anti-human C1qBP antibodies ( Figure  5E). The polyclonal anti-human C1qBP antibodies specifically recognized a unique 33-kDa molecule ( Figure 5E, lane 1), which corresponds to the predicted molecular weight of HmC1qBP. In the first negative control, when microglia protein extracts were incubated on a streptavidin column alone ( Figure 5E, lane 2) no signal was obtained, showing that leech microglial proteins cannot recognize streptavidin. In the second negative control, when the sulfo-NHS-SS-biotin-labeled human C1q was loaded alone on streptavidin column ( Figure 5E, lane 3), no immunoreactivity was detected, indicating that antihuman C1qBP antibodies do not react with the human C1q. Therefore, these results gave evidence of a specific interaction between human C1q and a C1q-binding protein present in leech microglia protein extract.

Discussion
In mammals, microglial cells are regulators of tissue homeostasis and are involved in pathological processes orchestrating tissue remodeling. They are currently considered to function as sensors in the brain [28]. Among the mediators expressed by microglial cells and neurons, the subunit C1q belonging to C1 complement factor seems to be a key molecule in neuroinflammatory diseases [29][30][31][32][33]. This complement protein is involved in a large array of vital functions such as the modulation of various immune cells, clearance of apoptotic cells and unwanted synapses, phagocytosis and chemotaxis [34][35][36]. In several neurodegenerative diseases, tight interactions between C1q and microglial cells may be crucial in the regulation of neuroinflammation [37]. C1q biosynthesis rapidly increases when microglial cells are activated. C1q might be thus considered as a reliable marker for microglia activation.
Of interest in vertebrates, soluble C1q appears to be a potent chemoattractant factor. Indeed, it recruits human immature dendritic cells (DC), neutrophils, eosinophils and mast cells [18,38] through a C1q receptor-dependent mechanism [18,38]. The globular C1q-binding proteins (also called C1qBP, gC1qR, p32, p33 or TAP) interact with the globular heads of C1q and participate in C1qmediated chemotaxis of human neutrophils [39], human eosinophils [40] and murine mast cells [19]. However, a specific chemotactic activity of C1q through a C1qBP has not previously been reported for mammalian nerve cells.
In the medicinal leech, previous reports have demonstrated that microglial cell recruitment is essential for efficient repair of injured CNS tissue [14]. In this original report we elucidate the role of HmC1q as a chemotactic factor by using a recombinant protein and by identifying a C1q-binding protein (HmC1qBP) as a potential receptor present on microglial cells. This study in the medicinal leech is the first evidence of the interaction between C1q domain-containing protein and C1qBP enabling microglia recruitment in injured CNS.
In the first part of the present study, recombinant HmC1q was produced in the yeast P. pastoris, an organism that can be easily manipulated at the molecular genetic level and may express proteins at high levels, intracellularly or extracellularly. In addition, P. pastoris performs 'higher eukaryotic' protein modifications, such as glycosylation, disulfide bond formation, and proteolytic processing, compared to bacteria such as Saccharomyces cerevisiae or baculovirus [41]. The selected clone studied in this report was immunopositive using anti-HmC1q antibodies, with three distinct immunoreactive bands located in the region corresponding to that of native HmC1q (32.6 kDa). As usually observed in recombinant protein production, there was a slight shift in the final recombinant protein size that can be easily caused by differences in the number and type of added sugar units. Indeed, P. pastoris is able to add both Olinked and N-linked carbohydrate moieties to secreted proteins. Of interest, the recombinant HmC1q exhibited chemotactic activity toward leech microglial cells similar to that of HmC1q-containing medium, previously shown to act as human C1q [17]. A similar dose-dependent chemotactic effect was observed with rHmC1q supernatant as well as with purified rHmC1q (data not shown) and, in these two cases, microglial cell recruitment was specifically neutralized when cells were preincubated with anti-HmC1q antibodies. It must be underlined that (i) microglial recruitment was normal when the cells were preincubated with rabbit preimmune serum, and (ii) control yeast supernatant, without rHmC1q, did not exhibit any chemotactic effect. Therefore, this study shows that the chemotactic effect is dependent on the presence of rHmC1q.
Importantly, HmC1q exhibits an in vitro chemotactic effect only on a fraction of microglial cells, suggesting the existence of a subpopulation that is HmC1q-dependent. We have recently shown that crushed nerve cordconditioned medium contains another chemoattractant factor, homologous to the mature form of interleukin (IL)-16 and named HmIL-16, which also promotes microglial cell migration [22]. As observed for HmC1q, HmIL-16-dependent recruitment is limited to some microglial cells. Therefore, in leech, the involvement of several activation and migration signals acting on different subsets of microglial cells at the lesion site could be taken into account as suggested for mammals [3,42].
Additional ex vivo experiments in injured nerve cords have clearly shown that the recombinant HmC1q conserves its functional properties in the whole nerve cord. Indeed preinjection of rHmC1q into injured nerve cords enhances microglial cell migration at the lesion site (rHmC1q vs PBS) and this accumulation is specifically inhibited by anti-HmC1q antibodies but not by preimmune serum. Taken together, these results emphasize the in vitro and ex vivo chemotactic activity of rHmC1q in the recruitment of resident microglial cells present in leech CNS. In order to specify the in vitro chemotactic mechanisms of rHmC1q on microglia, we first used antibodies against a potential C1q receptor as described in the literature [43]. Under our experimental conditions, HmC1q-dependent cell accumulation was abrogated after cell preincubation with the anti-human C1qBP antibody, but not with the isotype control. Therefore it must be assumed that this HmC1q-dependent chemotaxis involves a homolog of human C1qbinding protein (C1qBP, alias gC1qR).
To reinforce this observation, in the second part of the study, attempts were undertaken to characterize the potential receptor of HmC1q on the surface of leech microglial cells. This goal was achieved through several experimental approaches: the identification of a C1qBPrelated sequence in leech CNS EST databases, the localization of the leech form of C1qBP, binding competition of HmC1q for the protein specifically recognized by an anti-C1qBP antibody, and the purification of this binding protein to a biotinylated human C1q.
A 33 kDa C1qBP-related molecule was characterized from leech CNS EST databases. This sequence contains a MAM33 domain, which is attributed to an acidic protein of the mitochondrial matrix involved in oxidative phosphorylation and specifically related to the human complement receptor C1qBP [44]. The leech molecule named HmC1qBP exhibits strong similarities with vertebrate and invertebrate known C1qBPs. The comparison between the leech form of C1qBP (HmC1qBP) and the human and mouse ones shows the presence of key residues described as essential for receptor folding and ligand binding properties [27] Most of them (seven out of nine) are perfectly conserved. The residues Thr-137 and Met-239 in the leech sequence are respectively related to residues Lys-123 and Leu-231 in the human one, showing comparable physicochemical features.
Interestingly, the fluorescence in situ hybridization led to us specifically locating HmC1qBP mRNA in microglial cells. Following lesions in the leech nerve cord, a timecourse analysis of C1qBP immunostaining was performed using anti-human C1qBP antibodies in order to localize the HmC1qBP protein. The accumulation of microglia was simultaneously observed using Hoechst dye because only microglial cells are able to circulate inside the connectives.
Immediately after a lesion, no specific C1qBP staining was observed in the damaged connectives while no microglial cells accumulated at the crush site. Of interest, over the timecourse analysis (0, 6 and 24 h following the crush) the C1qBP molecule was progressively detected at the lesion site. This increase in cell HmC1qBP staining was exclusively observed where Hoechst-dyed microglia accumulated. Analysis of the lesion site with high magnification revealed, at 6 h following the lesion, that HmC1qBP protein is only present in microglia. Importantly, only a part of the numerous Hoechst-dyed recruited microglial cells are HmC1qBP positive. Therefore, these data suggest that HmC1qBP could be involved in specific recruitment of a well defined microglial cell subpopulation whose chemotaxis is mediated by recognition between HmC1q and HmC1qBP.
Subsequent competition binding assays analyzed using flow cytometry revealed that recombinant HmC1q may share the same protein target recognized by the antihuman C1qBP antibody on the surface of leech microglial cells. Indeed, whereas the anti-human C1qBP antibody alone was found to bind to at least 46% of microglial cells, preincubation with rHmC1q markedly decreased the percentage of cells with specific fluorescence to 5%. This observation shows that rHmC1q acts as a competitor for anti-C1qBP antibodies, indicating that both molecules bind to the same ligand.
To definitively demonstrate that a C1qBP is implicated in leech microglial cell recruitment, a copurification strategy was undertaken using biotinylated human C1q. From the specific complexes eluted, a 33-kDa molecule was specifically detected by western blotting. Interestingly, this product presented the same molecular weight as the predicted HmC1qBP protein. This result, in conjunction with the in vitro competitive effect between HmC1q and anti-human C1qBP antibodies, confirms that HmC1q is able to bind to a C1qBP-related molecule in leech microglia. Therefore the present report demonstrates that HmC1qBP has structural and functional analogies with its human counterpart. Though consistent data are provided in this report that clearly demonstrate the involvement of a C1q-binding protein acting as a receptor for HmC1q, the involvement of other potential receptors in microglial cell recruitment during nerve injury cannot be excluded and requires further experiments.
Indeed, in mammals, the C1qBP molecule interacts with the globular heads of C1q [19,45] whereas the other type of C1q-binding protein, the cC1qR (calreticulin) binds to the collagenous portion of C1q [46]. Human C1q has been shown to be a chemotactic factor for human immature dendritic cells. This migration is mediated through ligation of both C1qBP (gC1qR) and cC1qR [18,47]. As the collagen-like sequence is present in HmC1q [17], additional experiments will investigate the existence of a cC1qR-related molecule that might be involved in HmC1q-dependent microglia recruitment.

Conclusions
Recombinant HmC1q was demonstrated to exert efficient chemoattractant activity in vitro and in injured nerve cords. In addition, HmC1q was shown to act on a microglial cell subpopulation through HmC1qBP. In mammals, such an interaction was identified in dendritic cells, but has never been shown in nerve cells. In summary, the production of rHmC1q and the evidence of the involvement of HmC1qBP contribute to a better understanding of microglial activation leading to leech nerve repair. These data from the leech CNS highlight C1q domain-containing factor functions in the integrity of the CNS, as recently suggested in mammals [36,48].