|Home | About | Journals | Submit | Contact Us | Français|
Following trauma, the central nervous system (CNS) of the medicinal leech, unlike the mammalian CNS, has a strong capacity to regenerate neurites and synaptic connections that restore normal function. Here, we show that this regenerative process is enhanced by a controlled bacterial infection, suggesting that induction of regeneration of normal CNS function may depend critically upon the co-initiation of an immune response. We explore the interaction between the activation of a neuroimmune response and the process of regeneration by assaying the potential roles of two newly characterized antimicrobial peptides. Our data provide evidence that microbial components differentially induce the transcription, by microglial cells, of both antimicrobial peptide genes, the products of which accumulate rapidly at sites in the CNS undergoing regeneration following axotomy. Using a preparation of leech CNS depleted of microglial cells, we also demonstrate the production of antimicrobial peptides by neurons. Interestingly, in addition to exerting antibacterial properties, both peptides act as promoters of the regenerative process of axotomized leech CNS. These data are the first to report the neuronal synthesis of antimicrobial peptides and their participation in the immune response and the regeneration of the CNS. Thus, the leech CNS appears as an excellent model for studying the implication of immune molecules in neural repair.
Because of its multiple vital functions, it is critical that the central nervous system (CNS) be successfully defended against pathogens. Until recently, this organ was considered to be immunologically inert and isolated from the peripheral immune system. However, it is now well established that immune surveillance and inflammatory responses do occur within this compartment (1). Indeed, in response to either cerebral injury or systemic bacterial infection, the CNS launches a well organized immunological reaction that encompasses both neural components and peripheral immune system cells. Within the mammalian CNS, resident glial cells, including astrocytes and microglia, have been shown to initiate a characteristic innate immune response by producing and releasing antimicrobial peptides, cytokines and chemokines (2). These circulating molecules promote the destruction of the invading bacteria, the permeabilization of the blood brain barrier (BBB), and the recruitment of peripheral leukocytes to the CNS and the activation of their effector functions, including further production of cytokines as well as phagocytosis by peripheral macrophages. The specific outcome of this neuroinflammatory response, which has both beneficial and detrimental aspects, depends on the context of the insult and on the duration of the inflammation. On the positive side, increased immune activity rapidly initiates the killing of bacteria and the removal of apoptotic cells and cellular debris, while also playing an important role in neuroprotection and repair by inducing the production of neurotrophic factors. In fact, several recent observations suggest that induction of regeneration of normal CNS function may depend critically upon the co-initiation of an immune response (3). On the negative side, excessive and/or chronic glial reactivity, in conjunction with the presence of adaptive immune cells within the CNS, can damage the CNS by inducing neuronal death and by blocking axonal myelination. An important question that remains unanswered is how the vertebrate immune system can be both friend and foe to the damaged tissue. Some of the difficulty in obtaining an answer probably lies in the conflation of the innate and adaptive immune responses, which makes the vertebrate immune response quite complex and difficult to dissect.
In this context, we have begun a detailed study of the immune response that can be evoked in the injured CNS of an invertebrate, namely the medicinal leech (Hirudo medicinalis), in which the immune response appears to lack a deleterious component. Invertebrates, being devoid of adaptive immunity, are excellent model systems for exploring the molecular basis of innate immunity. For example, the initial evidence for the pivotal role of the Toll receptor family in immunity was discovered in Drosophila, and only later in mammals (4). Another example is the discovery of the first antimicrobial peptides by Hans Boman in the insect Hyalophora cecropia (5). Antimicrobial peptides are now considered as important effectors of the innate immune systems of both invertebrates and vertebrates (6).
Most reports on immune effectors in invertebrates have tended to focus on their involvement in the systemic antiinfectious response. However, some studies have described the presence of immune molecules in the nervous systems of insects and nematodes, both members of the ecdysozoan group. Indeed, several Toll-like receptors (TLR) and some molecules of the TLR signalling pathway have been detected in glial and neuronal cells of Drosophila, and appear to have a role in neural development in the larvae (7). In Caenorhabditis elegans, an ortholog of the Drosophila toll gene was shown to be expressed in pharyngeal neurons, where it participates in defensive behaviour by discouraging the worm from ingesting pathogenic bacteria (8). In the medicinal leech, a member of the lophotrochozoan group, Emes et al. reported that several central neurons express HmCRIP, an ortholog of the mammalian Cysteine Rich Intestinal Protein (CRIP) that is known to regulate the inflammatory response through control of Th1/Th2 differentiation in rodents (9).
Several features make the CNS of the medicinal leech particularly attractive as a model system for the exploration of interactions between the nervous and immune systems. These features include simplicity, a fixed number of neurons, and consistency from animal to animal, which allow the recognition, characterization and repeated study of identified neurons, at all developmental stages and following specific perturbations, such as mechanical or septic trauma (10). The leech CNS is comprised of a fixed number of midbody segmental ganglia linked to each other by longitudinal nerves known as connectives. Most segmental ganglia have a complement of ~400 neurons and 8 giant glial cells, along with a large population of microglial cells.
In the context of the observations we report here, however, the most important feature is the capacity of the medicinal leech CNS to regenerate and restore normal function in response to injury. If the nerve cord of this annelid is crushed or partially cut, axons grow across the lesion and conduction of signals through the damaged region is restored within a few days, even when the nerve cord is removed from the animal and maintained in culture. By contrast, when the mammalian spinal cord is injured, regeneration of normal connections generally fails. In the leech, the process of regeneration begins with a rapid activation of microglial cells leading to their accumulation at the lesion site. Like their counterparts in the mammalian brain, leech microglial cells are involved in phagocytosis of damaged tissue (11).
The leech nerve cord consists of a chain of interconnected segmental ganglia that resides within the ventral blood sinus. It is encapsulated by a tough fibrous sheath that may, like the mammalian blood-brain barrier, limit the exchange of macromolecules and cells with the blood which continuously bathes it. The virtually intact CNS (except for the cut lateral nerves) can be easily removed from the animal and maintained in culture for weeks in the absence of peripheral immune system components and blood cells that might infiltrate the CNS after injury. Consequently, studies can be focused exclusively on the immune response of the CNS itself.
In the study reported here, we sought to determine whether the injured leech CNS is able to regenerate connections ex vivo in the presence of bacteria isolated from the natural environment of this annelid, and if so, whether this involves an antiinfectious process which relies on antimicrobial peptides. Two novel antibacterial peptides, neuromacin and Hm-lumbricin, were isolated from the medicinal leech nervous system and molecularly characterized. We obtained evidence that microbial components differentially induce the transcription of both genes in the leech CNS. We also found that neuromacin and Hm-lumbricin, which are expressed by microglial cells and by neurons themselves, accumulated at the cut ends of nerves, a location where they might maximally exert their antimicrobial activity. Of particular interest, neuromacin and Hm lumbricin are shown to enhance the capacity of the leech CNS to regenerate. To the best of our knowledge, this is the first report describing the role of antimicrobial peptides in neural immunity and neural repair of an invertebrate.
Adult leeches, Hirudo medicinalis species, weighting 2-3g were purchased from a commercial supplier (Ricarampex, France). Animals were kept in artificial pond water and starved up to 3 weeks before use. Leeches were anaesthetised in 10% ethanol-spring water for 20 min and the nerve cords were removed according to the protocol of Nicholls et al (12). After three successive baths in antibiotics, cords were placed in Leibowitz-15 (L-15, Gibco) culture medium supplemented with 2% fetal calf serum, 0.6% glucose and gentamicin (10mg/ml) (13). Connectives between ganglia were injured in a standard manner using a pair of sterilized fine iridectomy scissors. Axotomized nerve cords were separately incubated in L-15 medium containing different microbial components: 3 × 107 CFU/ml of heat killed or live bacteria, 10 μg/ml of polyI:C (Invivogen), 100 μg/ml of zymosan (Invivogen,) and 100 μg/ml of mannane (Invivogen) for different time (t=0 to t=8 days) at room temperature. Incubations without microbial components or bacteria were performed in the same conditions as controls. All the steps were performed under sterile conditions.
Nerve cords depleted of microglial cells were obtained 6 hours after having opened the capsule surrounding the ganglia with fine scissors.
The Gram-positive and Gram-negative bacteria Micrococcus nishinomiyaensis and Aeromonas hydrophila respectively were isolated from the natural environment of Hirudo medicinalis. These bacterial colonies, which are present in the water, were selected from agar plate under aerobic conditions at room temperature using a random isolation grid. The Gram-positive bacteria Bacillus megaterium was used for the SYTOX Green assay.
Dissected nerve cords (three per conditions) were pinned in a dish coated with silicone rubber (Sylgard 184) containing supplemented L-15 medium under sterile conditions and one connective of each nerve cord was cut. To measure the effect of bacteria on the regeneration process, a mix of either live or heat-killed A. hydrophila and M. nishinomiyaensis (3×107 CFU/ml) was added to the culture medium. Controls were performed in the absence of bacteria. To determine the effect of the antimicrobial peptides on the capacity of the leech nerve cord to regenerate, anti Hm-lumbricin and/or anti neuromacin antibody(ies) at a 1:100 dilution was(ere) added to the culture medium containing killed bacteria. Native neuromacin was also added to the medium at the final concentration of 4 μM in absence of bacteria.
Pictures were taken every 24 hours for 1 week (objective ×5) using a LEICA inverted microscope DMIRE2. Images were taken using the Bioposition V3.0 software developed on the Matrox MIL7.5 Base Library by Gilles Courtand from the CCMIC (Centre Commun de Mesures Imagerie Cellulaire) of the University of Lille 1, France.
After each purification step, antibacterial activity was monitored by a solid plate assay as described in previous studies (14). The minimal inhibitory concentration (MIC) and the minimal bactericidal concentration (MBC) were determined according to the method of Hancock (http://cmdr.ubc.ca/bobh/showmethod.php?methodid=79). Permeabilization of bacterial membranes and pore-forming activity were assayed as previously described (15, 16)
100 nerve cords challenged for 6 hours with a mix of killed bacteria were homogenized in phosphate buffered saline pH 7.5. Liquid was immediately centrifuged at 10,000g at 4°C for 20 min and the supernatant was acidified by adjusting the pH to 3.9 with 1 M HCl. Centrifugation (10,000g at 4°C for 20 min) was then used to clarify the supernatants, which were loaded onto Sep-Pak C18 Vac cartridges (Waters). Elution steps were performed with 2% and 60% acetonitril (ACN) in 0.01% trifluoroacetic acid water. The pre-purified fractions were then lyophilized, reconstituted in pure water and tested for antimicrobial activity as described below. Only the 60% ACN eluted fractions were active and submitted to purification by reversed-phase high pressure liquid chromatography (RP-HPLC). All the purification steps were carried out on a Beckman™ Gold HPLC system.
Aliquots of the 60% Sep-Pak fractions were subjected to RP-HPLC on a Sephasyl C18 column (250 × 4.1 mm, 218TP54 Vydac™). Elution was performed with a linear gradient of 2-62 % ACN in acidified water over 90 min at a flow rate of 1 ml/min. Fractions corresponding to absorbance peaks were collected in polypropylene tubes, lyophilized, reconstituted in water and tested for antimicrobial activity.
Active fractions were further loaded onto a C18 column (250 × 2.1 mm, 218TP52 Vydac™) with a gradient consisting in 2-25% ACN in acidified water for 10 min and 25-35% ACN for 40 min at a flow rate of 0.2 ml/min. Fractions were collected and treated as above.
One additional step was performed on a narrow bore C18 reversed phase column (150 × 2 mm, Waters) at a flow rate of 0.2 ml/min using the ACN gradient described in the step 2.
The purity assessment and the molecular mass determination of the peptides were carried out by matrix-assisted laser desorption/ionisation-time of flight (MALDI-TOF) instrument (DE STR PRO, Applied Biosystem). N terminal sequencing of the purified peptides was performed by automated Edman degradation on a pulse liquid automatic peptide sequenator (Beckman).
The culture medium of 20 nerve cords challenged with zymosan (see collection and treatment of the nerve cords) was acidified and prepurified as described above. The entire 60 % Sep Pak fraction was loaded onto a C18 column (250 × 2.1 mm, 218TP52 Vydac™) with a gradient consisting in 2-62% ACN in acidified water for 90 min. Fractions were collected, lyophilized, reconstituted in 10μl of pure water and tested by Dot Immunobinding Assay (DIA) according to Salzet et al. (17). Briefly, one of the 10μl samples was spotted onto a nitrocellulose membrane and was incubated with the Hm-lumbricin antiserum (1/1000). Bound antibodies were detected with a goat antirabbit IgG conjugated to horseradish peroxidase by using a chemoluminescence kit (ECL; Amersham).
Complete cDNA for Hm-lumbricin was obtained from an expressed sequence tag (EST) library created from the nervous system of Hirudo medicinalis (http://www.cns.fr/externe/English/Projets/Projet_PE/PE.html). Hm-lumbricin cDNA was amplified from the cDNA prepared for neuromacin by polymerase chain reaction (PCR) using the sense oligonucleotide 5′-ATGGAGGAGGAAATTGAAGAACTCC-3′and the antisense oligodT primer.
cDNA for neuromacin was cloned using two steps PCR amplification:
Reverse-transcriptase PCR (RT-PCR): Total RNA from leech nervous system was extracted using Trizol (Life Technologies). RNA (3 μg) was transcribed into single stranded cDNA using oligo(dT)18-adaptator primer, 5′-CGAGTCGACATCGATCG(T)18 -3′ (Kit superscript Tm, GIBCO/BRL, protocol of the manufacturer). One fourth of the reaction was amplified by PCR using the oligodT primer and degenerate sense oligonucleotide pool whose sequence is deduced from Asp1-Trp6 with a designed 5′ flanking sequence 5′-GA(C/T)TG(C/T)TA(C/T)GA(A/G)GA(C/T)TGG -3′.PCR was performed for 25 cycles using one unit of Taq polymerase (Appligene quantum) in 1.5 mM of MgCl2. The cycling parameters were: 94°C for 1 min., 50°C for 1 min. and 72°C for 1 min.
Rapid Amplification of 5′cDNA End. Reverse-transcription was performed using antisense oligonucleotides 5′- CAGTCAGGAATGACGTTCCAGGCG -3′deduced from the neuromacin cDNA sequence previously obtained. After first strand cDNA synthesis and addition of a poly-d(C) tail at its 3′ end using a terminal transferase (GIBCO, protocol of the manufacturer), PCR was performed with an oligodG anchor primer and internal antisense primers deduced from the cDNA obtained in step 1: 5′- TCCCCTGTGACCGAGTTC -3′. PCR parameters were identical to those described in step 1.
All PCR products were ligated into the PGEMT easy vector (according to the protocol provided by the manufacturer) and transformed into competent Escherichia coli JM 109 cells (Promega).
Plasmids DNA were sequenced with a FM13/RM13 sequencing kit (Pharmacia Biotech) according to the manufacturer's instructions.
Twenty leeches were used per conditions. Nerve cords incubated in the presence or absence of killed bacteria, poly I:C, mannane or zymosan (see above for details) were crushed, in 2ml tubes prefilled with Qiazol reagent (Qiagen, France) and 1.4 mm ceramic beads, by shaking the mixture twice for 45 s at 6500 rpm in a Precellys®24 homogenizer (Bertin distributed by Ozyme, France). RNA extraction was performed according to the manufacturer and extracted total RNA was treated with RQ1Dnase1 (Promega, France) to prevent contaminations with genomic DNA. First strand cDNA was generated from 2 μg of total RNA using random primers (Promega, France) and Superscript III reverse transcriptase (RT) kit (Invitrogen, France) in a final volume of 60 μl. Omitting RT or RNA from the reaction mixture resulted in non amplification control and non template control respectively. cDNA were treated with RNaseH (Promega, France) to optimize the amplification.
Real time PCR were performed with the Quantitect SYBRgreen PCR kit (Qiagen, France) by combining 1μl of cDNA, 0.8 μM of each primer and 1× of SYBRgreen reagent in a final volume of 25μl. The primers were designed with the Primer3 Input software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) according to the conditions described in the PCR kit. Hm-lumbricin: 5′-CCAATCGGCTCTCCTTACTC-3′, 5′-ACAGGCCCTCAGTTCATTTC-3′; Neuromacin: 5′-GTGGACGCCTGGAACGTCAT-3′, 5′-CTTGAGGACTCCAGGGCAGT-3′ 18S: 5′-TGCGGTTATTTCGATTGTCA-3′, 5′-AGACAAATCGCTCCACCAAC-3′ Real time PCR was conducted on an Applied Biosystem 5700 using hot start with cycle conditions, 40 cycles: 94°C, 15 s; 56°C, 30 s and 72°C, 30 s, followed by a final 72°C extension for 3 min. Analysis of relative gene expression data was performed using the ΔΔCt method. A plot of the log cDNA dilution versus ΔCt was generated for each target gene tested to validate the qPCR experiments. The slopes of regression lines for neuromacin and Hm-lumbricin were 0.0698 and 0.0864 respectively suggesting equivalent efficiencies of amplification (data not shown) (18)
Microglia deficient nerve cords were fixed in a solution containing 4% paraformaldehyde at 4°C overnight. Plasmids containing the coding region of neuromacin and Hm-lumbricin probes were used as templates for the synthesis of the probes. Digoxigenin (DIG)-UTP-labeled antisense and sense riboprobes were generated from linearized cDNA plasmids by in vitro transcription using RNA-labeling-kit (Roche™). DIG–labeled riboprobes (40-100 ng per slide) were hybridized as previously described (19). Slides were observed under a Zeiss™ Axioskop microscope. As a control, antisense riboprobes were replaced by sense riboprobes.
Single cell RT-PCRs were carried out on fresh dissociated neurons and microglial cells. Cells were individually collected using a patch-clamp micromanipulator (20). Patch pipettes were filled with the internal pipette solution (10 μl of 140 mM KCl, 2 mM MgCl2, 1.1 mM EGTA and 10 mM Hepes, pH 7.25). Neurons and microglial cells were aspired under visual control. The pipette content (~8 μl) was ejected into a 0.2 ml PCR microtube without touching the wall, immediately frozen with nitrogen azote and kept at -80°C until use. A new patch pipette was used for each new collection. The RT reaction was performed with one single cell using oligodT(18) primer according to Roudbaraki et al (20) The following sense primers were designed to amplify with an oligodT(34) primer cgagtcgacatcgatcg(t)18 a PCR size product under 500bp : Neuromacin: 5′-CTCTTCTCAACAAGTTGCTCTGC-3′ and Hm-lumbricin: 5′-CAGCAAGTACGAGAGGCAAAAGGACA-3′ PCR was performed by using 4 μl RT template according the protocol of the AmpliTaq Gold (Applied Biosystems). The conditions of DNA amplification included an initial denaturation step of 10 min at 95°C, 45 cycles of 45 s at 95°C, 45 s at 60 C, and 45 s at 72° C; and finally 5 min at 72° C. The PCR products were loaded onto a 1% agarose gel.
Because of the high conservation between neuromacin and theromacin epitopes, the theromacin antiserum was used for our analysis. This rabbit antibody recognizes a synthetic peptide corresponding to theromacin (Gln49-Arg63) as previously described (19). The Hm-lumbricin antiserum was produced in the laboratory. The chemically synthesized region of Hm-lumbricin (Gln5-Pro30) was coupled to ovalbumin and used for the immunization procedure of two New Zealand White rabbits (Saprophyte Pathogen free) according to the protocol previously described (21)
Nerve cords depleted or not of microglial cells were fixed overnight at 4°C in 4% paraformaldehyde at different times after bacterial exposure. Membranes were permeabilized by incubating the samples in PBS containing 1% Triton X-100 for 24 h at room temperature. Non-specific background staining was blocked with a PBS solution containing 1% Triton, 3% normal goat serum (NGS) and 1% ovalbumin for 8 h at room temperature. Samples were then incubated with whether rabbit neuromacin antibodies (1:800) or rabbit Hm-lumbricin antibodies (1:200) diluted in the AB solution (PBS containing 1% BSA 0,05% Triton 1% NGS 1% NDS 1% Ovalbumin) overnight at 4°C. Primary antibodies were removed and samples were incubated with Alexa Fluor 546 conjugated to goat anti-rabbit secondary antibody (Invitrogen) diluted 1:4000 in the AB solution. As a control, the immunolabeling procedure was carried out with the preimmune serum. Slides were mounted and examined using a confocal microscope (Zeiss LSM 510) or an inverted microscope (Leica DMIRE2).
Neurons dissociated from nerve cords incubated for 6hours with bacteria were fixed for 30min. at 4°C by addition of ice-cold 4%paraformaldehyde solution. Cells were then centrifuged on slides (12, 000 neurons/slide) and immersed for 10 minutes in Tris-buffered saline (TBS; 0.1mol/L Tris, pH 7.5 or less, 0.9% NaCl). After 1 hour of incubation in TBS containing 3% normal goat serum (NGS), cells were incubated overnight at 20°C with rabbit anti-lumbricin antibodies (1/100) or anti-neuromacin antibodies (1/400) in TBS containing 2%NSG and 0.01% Triton X-100. Next, cells were rinsed 3 times with TBS and incubated for 2 hours at room temperature with FITC–conjugated goat antirabbit IgG diluted 1/100 (Jackson Immunoresearch).Cells were examined using a confocal microscope (Zeiss LSM 510).
Sectioning of one side of the paired connective nerve linking adjacent segmental ganglia was performed on excised nerve cords maintained in culture (shown schematically in Fig. 1A). To monitor the progress of nerve repair, micrographs of the damaged nerve cords were taken every 24 h, in the presence or absence of bacteria. Under sterile conditions, as documented in Figure 1B, restoration of the connective nerve across the cut begins at ~4 days post-axotomy (panel J4) and is finished 4 days later. This is in line with the observations reported by Müller et al (22), who demonstrated that even the synaptic connections and normal functions of axotomized leech neurons were restored after 8 days. In comparison, nerve repair is evident sooner in the presence of bacteria, reconnection starting after 2 days or 3 days with live or heat-killed bacteria (Fig. 1C and 1D). In the case of incubation with live bacteria, however, signs of degeneration are soon present (Fig. 1C, panel J4), probably as a consequence of the uncontrollable expansion of the bacterial population in the rich culture medium, which becomes deleterious for the CNS. By contrast, incubation with a defined amount of heat-killed bacteria appears to promote the regeneration process relative to sterile conditions.
Neuromacin and Hm-lumbricin were isolated from a preparation of 100 nerve cords of Hirudo medicinalis under acidic conditions (Fig. 2). Identification and characterization of both molecules were performed by combining N-terminal sequencing by Edman degradation, MALDI-TOF mass spectrometry and cDNA cloning (Fig. 3). Amino acid sequences deduced from the cDNAs allowed the conclusion that neuromacin and Hm-lumbricin are both cationic peptides with calculated pI of 8.49 and 9.16, respectively. Data bank analysis (BLAST program in Swiss-Prot) revealed that the primary structure of neuromacin was 67% identical to that of theromacin, an antibacterial peptide previously isolated by our group from the body fluid of the leech Theromyzon tessulatum (19). Neuromacin was named according to its apparent homology to theromacin and its production being restricted to the nervous system. Searches in databases from Genbank and EST libraries revealed the presence of putative neuromacin-like gene products, often named theromacin-like in various invertebrate species such as the mollusks Aplysia californica and Biomphalaria glabrata and the annelid Lumbricus rubellus (Fig. 4A).
Blast analysis results (Fig. 4B) show that Hm-lumbricin shares 78% sequence identity with lumbricin-1, an antimicrobial peptide originally isolated from the earthworm Lumbricus rubellus, and 81% sequence identity with PP-1, a lumbricin-1-like gene product cloned from another earthworm, Pheretima tschiliensis (23, 24).
In liquid growth inhibition assay, the purified neuromacin was active against Micrococcus nishinomiyaensis (MIC 1.95-3.8 μM; MBC 7.8-15.6 μM). No activity was found towards the Gram negative A. hydrophila at the same concentration. To investigate the mode of action of neuromacin against Gram positive bacteria, we used B. megaterium and the fluorescent dye SYTOX Green. Results of this assay demonstrate that neuromacin rapidly permeabilizes bacterial membranes and is thus potently active against these Gram-positives (Fig 5A). Since neuromacin contains four histidine residues in its primary structure (in contrast to none in theromacin), we evaluated the effect of a more acidic pH, at which the histidine residues are protonated, on the membrane-permeabilizing activity (Fig. 5B). The efficacy of neuromacin to induce membrane lesions in viable bacteria was found to be weaker at pH 5.2 than at pH 7.4. Within the measurement period, the activity of neuromacin against the target bacterium tested appeared to be over one order of magnitude lower than those for the well-known antimicrobial peptides cecropin P1 and magainin II. Similar activity against the Gram-negative representative E. coli was not detected up to a neuromacin concentration of 25 μM (data not shown). We also measured the pore-forming activity of neuromacin to further characterize its mode of action by using a minimalistic membrane system. More precisely, we monitored the dissipation of a membrane potential induced in liposomes composed of azolectin, a crude phospholipid mixture from soy bean. Pore-forming activity was not detected at final concentrations up to 2 μM, whereas the positive control alamethicin gave a very strong signal at 50 nM (data not shown).
Based on the amino acid homologies with lumbricin-1 (Fig 4B), we speculate that Hm-lumbricin and lumbricin-1 must have similar antimicrobial properties. Notably, the N-terminal sequence, which is the most conserved part between lumbricin 1 and Hm-lumbricin, has been shown by Cho et al to exhibit a stronger activity than the entire lumbricin-1 against fungi, Gram positive and Gram negative bacteria, without exerting any hemolytic activities (23).
The capacity of both peptides to promote the regeneration of the leech nerve cord was also tested ex vivo by adding the neuromacin and/or Hm-lumbricin antibody (ies) to axotomized nerve cords in presence of killed bacteria (Fig. 6). Due to the presence of bacteria, the reconnection process should have started 3 days post-axotomy (Fig. 1D). It appeared that the presence of antibodies in the culture medium blocked the regeneration process since no reconnection was observed even 7 days post-axotomy. That suggests that both peptides are involved in the regenerative process of the leech CNS. These observations were corroborated by the data obtained by adding native neuromacin to axotomized nerve cords under aseptic conditions (Fig. 6). Nerve repair was evident sooner in the presence of neuromacin, reconnection starting in less than 24 hours instead of 4 days without neuromacin. So, in addition to exerting antimicrobial properties, Hm-lumbricin and neuromacin have the capacity to enhance the regenerative responses of the leech CNS.
Gene expression sites of neuromacin and Hm-lumbricin were investigated by single cell RT-PCR and by ISH (Fig. 7). Interestingly, both genes are expressed in neurons after septic exposure of the CNS. This is the first reported observation of the transcription of antimicrobial peptides by neurons in an invertebrate. Neuromacin and at a lower level Hm-lumbricin transcripts were also detected in microglial cells.
Gene expression patterns of neuromacin and Hm-lumbricin during an induced immune response were investigated by real time PCR. As illustrated in figures 8A and 8B, neuromacin and Hm-lumbricin transcript levels were rapidly enhanced by bacterial challenge, but no significant changes were measured following axotomy under sterile conditions, indicating that the induction was not due to the axotomy itself, but rather to the incubation of the nerve cords with bacteria.
We have further attempted to analyze whether the leech nervous system was able to discriminate among different microorganisms. The data presented in Figure 8B indicate that treatment of the nerve cord with the gram positive bacterium M. nishinomiyaensis induces an increase in both Hm-lumbricin and neuromacin mRNA levels, the first more pronounced than the second. In contrast, no significant change in neuromacin expression but a large increase in the level of Hm-lumbricin transcripts was observed following challenge with zymosan, a component of yeast membrane. This distinct pattern of induction appears to correlate with the activity of the peptides encoded by these genes. Indeed, the neuromacin gene, which encodes a product with anti-Gram positive activity, is induced by Gram positive bacteria, and the Hm-lumbricin gene, which encodes an anti-Gram positive and anti-fungal peptide, is induced by both Gram positives and zymosan. These results suggest that the leech CNS responds to infection with a pathogen-specific pattern of gene expression.
To get further information on possible roles of both antimicrobial peptides in the immune response of the leech CNS, double staining was performed on challenged nerve cords at t=0 h and t=6 h post-axotomy by using Hoechst 33258 and either the anti-Hm-lumbricin polyclonal antibody or the anti-neuromacin polyclonal antibody (Fig. 9). Challenge with M. nishinomiyaensis was chosen because of its efficacy to induce the expression of both Hm-lumbricin and neuromacin genes. By contrast with the CNS fixed immediately after dissection (Fig. 9A, B), nerve cords fixed 6 h after septic challenge present a strong immunoreactivity with both antibodies at the lesion (Fig. 9E, F), along with an accumulation of microglial cells as revealed by nuclear staining with Hoechst 33258 (Fig. 9 G,H). Microglial cells are resident cells evenly distributed in leech ganglia and in the bundle of axons that connect them. After damage to the CNS, these cells have been shown by Müller and co-workers to migrate to the site of the lesion, where they accumulate (25). There, microglial cells phagocytose damaged tissue and produced laminin, an extracellular matrix molecule known to promote neurite outgrowth in the leech and in vertebrates (11). Immunolocalization of both peptides at the axotomized site where microglial cells accumulate suggests a production and a release of Hm-lumbricin and neuromacin by these cells. Six hours post-challenge, neuromacin but not Hm-lumbricin was also strongly immunolocalized in microglial cells covering the segmental ganglia and the connectives (insets Fig. 9E). It has been shown that nearly all microglial cells are able to move, but that after the cord has been crushed, only 15-40% of them migrate to the injured site (26). We hypothesize that the induction of neuromacin synthesis is relevant to the global population of microglial cells, whereas the induction of Hm-lumbricin synthesis only pertains to the cells recruited to the lesion site.
Interestingly, ISH combined with single cell RT-PCR demonstrated the presence of neuromacin and Hm-lumbricin mRNAs in neurons (Fig.7). In order to determine whether the detection of both peptides at the lesion site could also reflect a neuronal production, we have developed an ex vivo model of leech CNS almost completely devoid of microglial cells. Each segmental ganglion of the medicinal leech nerve cord is enclosed in a capsule. As revealed by nuclear staining with Hoechst 33258 (Fig. 10G and 10H), mechanical destruction of the capsule resulted in a marked reduction, within 6 hours, in the numbers of microglial cells surrounding the nerve cord. With these conditions, the accumulation of microglial cells which is normally observable (Fig. 9G, ,9H)9H) at the axotomized site does not occur anymore (Fig. 10G and 10H). As presented in figures 10E and 10F, it appeared that this depletion does not affect the accumulation of both peptides at the site of axotomy. Confocal microscopy analysis (insets Fig. 10E and 10F) more precisely evidenced the presence of Hm-lumbricin and neuromacin in the cell body of neurons and in the axons of the injured connectives, suggesting a neuronal production presumably followed by the axonal transport of the antimicrobial peptides to the lesion site. Thus, the presence of both Hm-lumbricin and neuromacin at the axotomized site implicates peptide production by neurons and by the microglial cells recruited at the lesion site.
Precursor sequences were also deduced from the cDNA sequences (Fig. 3). Signal PVI software analysis revealed that the N-terminal Asp of neuromacin is preceded by a putative signal peptide comprising Met1-Pro23 (Fig. 3A). This leads to the notion that neuromacin may be generated through conventional processing mechanisms and can be secreted to the extracellular medium. By contrast, Hm-lumbricin precursor lacks a typical signal sequence making its extracellular presence unlikely. In order to determine whether Hm-lumbricin could be released upon microbial stimulation, peptide purification was investigated from the incubation medium of nerve cords challenged with zymosan (Fig 11). Zymosan was chosen because of its specificity and efficiency to induce the expression of the gene encoding Hm-lumbricin (Fig. 8B). RP-HPLC analysis coupled to DIA evidenced the presence of Hm-lumbricin in the culture medium demonstrating that despite lacking a signal peptide, this antimicrobial peptide can be secreted and so can exert its biological effects. It is noteworthy that several molecules devoid of a signal peptide have been shown to be secreted. For example, PP1, a lumbricin-like peptide found in P. tschiliensis, has been detected in the mucus it secretes (23). In vertebrates, the rat PEBP was shown to be secreted into epididymal fluid despite lacking a signal peptide (27)
Altogether these results provide direct evidence of the existence of infection-inducible antimicrobial peptides in the CNS of an invertebrate. Two newly characterized antimicrobial peptides, Hm-lumbricin and neuromacin have been shown to be produced by microglial cells and by neurons themselves in response to CNS injury. Surprisingly, these peptides, in addition to exerting antimicrobial properties, appeared as promoters of the neural repair of the leech CNS.
Neuromacin is a relative of theromacin, a cysteine-rich antimicrobial peptide previously identified by our group from the body fluid of the leech Theromyzon tessulatum (19) Theromacin, which has also been characterized in the medicinal leech (Hm-theromacin), possesses a longer C-terminal domain than neuromacin. That probably results in two different conformations, resulting in different biological activities for the two peptides. Further investigation of this relationship between structure and function is warranted. Hm-theromacin and neuromacin present a differential tissue expression. Hm-theromacin is expressed by a peripheral tissue functionally resembling the insect fat body, whereas neuromacin expression is restricted to the nervous system. Curiously, only neuromacin-like molecules have been found in other invertebrates, such as mollusks and other annelids. Thus far, the larger theromacin appears to be unique to leeches. Neither neuromacin-like nor lumbricin-like molecules have been found in the genomes of ecdysozoan invertebrates such as Caenorhabditis elegans and Drosophila melanogaster, underlining the importance of enlarging the number of invertebrate models dedicated to study innate immunity.
Neuromacin, like theromacin, displayed potent activity against Gram-positive bacteria (19). For antimicrobial peptides in general, a two-step mechanism is proposed to explain their killing effect on microorganisms. Most cationic peptides bind to the target cell surface due to their content of negatively charged phospholipids or other moieties. Many of them subsequently disturb the membrane integrity (28) (29). Likewise, killing of bacteria by neuromacin was accompanied by permeabilization of the bacterial membranes comparable to the action of cecropin P1 and magainin II, although apparently not as rapidly as these other antimicrobial peptides. The outer membrane of gram-negative bacteria may constitute a barrier that hinders the peptide from reaching the target, i.e. the cytoplasmic membrane.
The importance of neuromacin and Hm-lumbricin in the antiinfectious immunity of the leech CNS is emphasized by the presence of their transcripts in neuronal cells and by the fact that their gene expression is upregulated by some microbial components. Admittedly, a neuromacin-like peptide reported as theromacin-like has recently been detected by sequencing cDNA libraries from the CNS of the mollusk A. californica, but the roles and the production sites of the peptide were not detailed (30). In C. elegans, several genes encoding neuropeptide-like proteins named NLP-29, NLP-31 and NLP-33, the sequences of which were deduced from an in silico analysis of an EST library, have been shown to be induced in the hypodermis by fungal infection (31). Interestingly, the chemically synthesized NLP-31 exhibited antifungal activity (31) but in contrast to most of the nlp family members, NLP-29, NLP-31 and NLP-33 were not detected in neurons using GFP reporter genes (32).
In humans, antimicrobial peptides, such as defensins, have been shown to be produced in the central nervous system. HBD1 mRNA appeared to be constitutively expressed by astrocytes, microglia and meningeal fibroblasts but not by neurons, whereas HBD2 mRNA expression is inducible in astrocytes upon lipopolysaccharide (LPS), IL-1-β or TNF challenge (33). The bovine defensin gene lap is expressed in the cerebellum, in the choroid plexus, which produces cerebrospinal fluid, and other areas of the brain (34). By immunohistochemical analysis, the rat homologue of the human cathelicidin LL-37 was detected in specific CNS regions, such as some neurons in the olfactory bulb and the grey matter (35). The enhancement of Hm-lumbricin and neuromacin transcript levels after bacterial challenge suggests a regulation of these antimicrobial peptide genes similar to that of the human HBD2 in the CNS (33). Moreover, multiple examples of neuropeptides, such as the α-Melanocyte stimulating hormone in human and proenkephalin A-derived peptides in both mammals and/or leeches, have also been described to possess antimicrobial properties in vitro (36) (14). Thus, as in humans, antimicrobial peptides may be involved in the innate immune system of the leech CNS.
In addition to manifesting antibacterial properties, neuromacin and Hm-lumbricin exert impressive regenerative effects on the leech CNS. In vertebrates, one study provides evidence for the positive effects of an antimicrobial peptide on the restoration of the functions of a lesioned peripheric nerve. Indeed, the addition of neutrophil defensin NP-1 on the lesioned sciatic nerve in rats leads to increase the rate of growth of regenerative nerve fibers by 30% (37). Thus, the data presented here are the first evidencing the participation of an antimicrobial substance produced by the nervous system itself in the regeneration process of the CNS.
A difference in neuromacin and Hm-lumbricin gene expression was observed after infection with different microbial components, suggesting that the antibacterial response of the medicinal leech CNS is specific to the antigens presented. This was not expected, given that in the leech T. tessulatum our previous work demonstrated the non-specificity of the humoral antimicrobial response to infection (19). This ability to discriminate pathogens might be relevant to the use, in the experiments presented here, of bacteria naturally living in the environment of the leech. A significant and antigen-specific increase of the level of neuromacin and Hm-lumbricin transcripts was also noticed in the nerve cord of intact leeches exposed to alive bacteria (unpublished data) confirming that the induction is due to the presence of bacteria and not to the injury. The perception mode of infection by leeches remains unknown. In many species, including flies and mammals, molecular components of a particular group of pathogens can be recognized by TIR-domain containing receptors that act as sentinels. Their activation triggers an intracellular signaling pathway followed by the up-regulation of defense genes (38). We have recently detected the presence of a Toll-like receptor in the leech nervous system, and preliminary observations on the effects of silencing this gene suggest that this receptor is involved in the induction of as gene coding for a chemo-attractant protein, but not in the induction of either the Hm-lumbricin or the neuromacin genes (unpublished data).
Having determined the gene expression sites and demonstrated their up-regulation, we then sought to characterize further the regulation of the peptides synthesis by immunohistochemistry. Our results indicated an increase in the amounts of both peptides in the challenged CNS. This enhancement occurred in the ganglia and at the axotomized sites, but only in the presence of bacterial components. No change was observed under sterile conditions (data not shown). Hm-lumbricin and neuromacin peptides accumulated in the microglial cells recruited to the lesion site. Unlike what happens in the mammalian brain, this accumulation is known to be the result of cell migration and does not reflect microglial cell division in response to leech CNS injury (26). However, the synthesis of antimicrobial peptides by the leech microglial cells confirms once again the functional similarity of these invertebrate cells with human microglial cells. Moreover, by developing a preparation of leech CNS depleted in microglial cells, we have demonstrated that the accumulation at the axotomized site of both peptides is also due to a neuronal production.
In conclusion, our data show for the first time that an invertebrate CNS is able to mount a specific antibacterial response by discriminating some pathogen components. We have also visually demonstrated that, as in mammals, the leech nervous system uses a common panel of proteins to initiate an antimicrobial response and regrowth programs. Indeed, we demonstrate that microbial challenge promotes the regenerative process of the injured central nervous system of the medicinal leech by inducing the synthesis of antimicrobial peptides in neurons and glia. The relative simplicity of the leech CNS in combination with its having complex mechanisms to react to infection suggests that the study of the neural immunity in H. medicinalis will contribute to a better understanding of the implication of immune molecules in the neural repair of the CNS in mammals.
The technical assistance of Annie Desmons, Christelle Van-Camp and Mathilde Verstraete is gratefully acknowledged. We thank Heidrun Lieβegang for her technical assistance in measuring permeabilization of bacterial membranes and pore-forming activity and Henning Fedders for evaluating the data of these assays. We also thank Dr Anna Paula Teixera (INRA UR IASP site 311, France) for N-terminal sequencing. The authors are indebted to Gilles Courtand (CCMIC, USTL, Lille1, France) for access to the Cellular Imaging Center and for his help in acquiring pictures. We also thank the Genoscope for the EST sequencing facilities.
This work was supported by the Centre National de la Recherche scientifique (CNRS), the Ministère de l'Enseignement, de la Recherche et des Technologies (MERT), a Minority and Health Disparities International Research Training grant from the National Center for Minorities and Health Disparities and the Fogarty International Center of the National Institutes of Health (MD001429), the Genoscope, and the Deutsche Forschungsgemeinschaft (SFB 617, TP 18). The nucleotide sequences reported in this paper have been submitted to the GenBankTM/EBI Data Bank with accession numbers EU156754, EU156755, EU156756 and EU164975
David Schikorski, CNRS, FRE2933, Laboratoire de Neuroimmunologie des Annélides, Université de Lille1, Villeneuve d'Ascq, F-59655 France.
Virginie Cuvillier-Hot, CNRS, FRE2933, Laboratoire de Neuroimmunologie des Annélides, Université de Lille1, Villeneuve d'Ascq, F-59655 France.
Matthias Leippe, Department of Zoophysiology, Zoological Institute, University of Kiel, Olshausenstr. 40, 24098 Kiel, Germany.
Céline Boidin-Wichlacz, CNRS, FRE2933, Laboratoire de Neuroimmunologie des Annélides, Université de Lille1, Villeneuve d'Ascq, F-59655 France.
Christian Slomianny, INSERM, U800, Laboratoire de Physiologie Cellulaire, Université de Lille1, Villeneuve d'Ascq, F-59655 France.
Eduardo Macagno, Section of Cell and Developmental Biology, Division of Biological Sciences, University of California, San Diego, La Jolla, California 92093-0380, USA.