| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Medical Physiology and
University Medical Imaging Center, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands
2Correspondence: Department of Medical Physiology, University Medical Center Groningen, University of Groningen, Ant. Deusinglaan 1, 9713AV Groningen, The Netherlands. E-mail: k.p.h.biber{at}med.umcg.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: chemokines • neuroimmunology • vesicles • microglia • neuronal injury • neuron-glia communication
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, 2)
. Neurohormones, neuropeptides, and neurotrophins are localized in larger vesicles that can be sorted either to the constitutive release pathway or to secretory granules [large dense-core vesicles (DCVs)] of the regulated release pathway (3
4
5
6)
.
As recognized recently, cytokines and chemokines (signaling molecules of the immune system) are also expressed in neurons (7
8
9
10)
. Although the brain can be considered an immune-privileged organ (for recent reviews, see refs. 11
, 12
), it is clear today that inflammatory factors such as cytokines and chemokines are functional in the brain (10
11
12
13
14
15
16)
. Especially, their expression in neurons has lately gained much interest (9
, 10)
. It is discussed at the moment whether neuronal chemokines participate in the communication between neurons and other cells, particularly in directed neuron-microglia signaling that often occurs after neuronal damage (17)
.
We have recently shown that under stress conditions, neurons specifically express the microglia-activating CCL21 (7
, 18
19
20)
. Moreover, neuronal CCL21 was found in vesicle-like structures that were distributed throughout neurons (8)
. Accordingly, CCL21 was suggested to be a neuronal signal to induce microglia activation at distant sites from a primary lesion (8)
. A prerequisite for this, however, would be the sorting of neuronal CCL21 into an appropriate secretory system that would allow its directed release.
We have therefore addressed the question of which neuronal vesicle type contains CCL21. Coexpression studies of fluorescent-labeled CCL21 with markers of neuronal vesicles, as well as detailed live-imaging analysis of CCL21-loaded vesicles, indicate that CCL21 in primary neurons and in differentiated neuroblastoma cells is sorted to secretory granules of the regulated release pathway. Vesicles loaded with CCL21 are preferentially sorted into neuronal axons. Taken together, our data show that neuronal CCL21 is sorted and transported in large DCVs, a system that allows neurons the targeted release of CCL21, thus corroborating the assumption that neuronal CCL21 is a signal to activate microglia at distant sites from a primary lesion.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Cell culture
Cell cultures were maintained at 37°C in a humidified atmosphere containing 5% CO2.
Neuroblastoma-glioma 108 (NG-108) cells
NG-108 neuroblastoma-glioma cells were cultured and transfected in culture medium [Dulbecco modified Eagle medium (DMEM) with 10% fetal calf serum (FCS), 0.01% penicillin/streptomycin, and 1% sodium pyruvate]. Neuronal-like differentiation of NG-108 cells was done by transferring to differentiation medium (DMEM with 0.5% FCS, 0.01% penicillin/streptomycin, 1% sodium pyruvate, 5-N ethylcarboxamide adenosine 10–5 M, and 3-isobutyl-1-methylxantine 10–6 M).
Primary cortical neurons
Cultures of mixed glia-cortical neurons were prepared from P2 57BL/6 mouse brains. Briefly, cortices were dissected in ice-cold HBSS containing 10 mM HEPES, pH 7.6. After removal of the meninges, cortices were sliced and placed in a 12 U/ml papain solution (Worthington Biochemical Corp., Lakewood, NJ, USA) at 35°C for 40 min. Afterward, cells were carefully dissociated by trituration and centrifuged through a BSA column at 1000 g for 7 min to remove debris. Cells were resuspended in complete neurobasal medium containing 2% B27, 5% FCS, and 0.5mg/ml Primocin (Amaxa Inc., Gaithersburg, MD, USA) and 500µM GlutaMAX (Invitrogen, Carlsbad, CA, USA) and seeded on poly-D-lysine (10 µg/ml) -coated glass coverslips at a density of 105/cm2. Cells were treated with Ara-C (10 mM) on day 5 to stop glial proliferation, and medium was refreshed afterward twice per week by replacing half of the volume with fresh complete neurobasal medium (without FCS).
For live-cell imaging, cultures of purified cortical neurons were established, as described elsewhere (7)
. Briefly, pregnant mice (NMRI) were anesthetized with isoflurane and sacrificed by cervical dislocation, and embryonic day (ED) 16 embryos were removed. Cortices were dissected in ice-cold HBSS supplemented with 30% glucose. After meninges were removed, cortices were placed in a 0.25% trypsin solution at 37°C for 20 min. Subsequently, tissue was gently dissociated by trituration and then filtered through a cell strainer (70 µm; BD Biosciences, San Jose, CA, USA). After one washing step (100 g for 10 min), neurons were seeded on poly-D-lysine (10 µg/ml) -coated glass at a density of 5 x 104/cm2 in complete neurobasal medium containing 2% B27, 0.01% penicillin/streptomysin, sodium pyruvate, and 2mM GlutaMAX). Medium was refreshed by replacing half of the volume with fresh complete neurobasal medium.
Plasmids
pEGFP-N2 vector was purchased from Clontech (Leusden, The Netherlands). Dr. R. Toonen (VUMC, Amsterdam, The Netherlands), kindly supplied pmRFP-N1, pNPY-mRFP, and pVAMP2-mRFP. The expression vector for pCCL21-EGFP fusion protein was cloned as described elsewhere (8)
. The following primers have been used to amplify the full (pCCL21-mRFP) or N-terminal deleted CCL21 (pCCL21delN-mRFP) for subcloning into XhoI–BamH I sites of pmRFP-N1: CCL21F 5'-ATA CTC GAG ATG GCT CAG ATG ACT CTG AGC CTC; CCL21del1F 5'-ATA CTC GAG ATG AGT GAT GGA GGG GGA CAG; and CCL21B 5'-GGT GGA TCC GCT CCT CTT GAG GGC TGT GTC TGT TC. The signal peptide plasmid was prepared using the following oligonucleotides (Isogen BV, Maarssen, The Netherlands): sense 5'-TCG AGA TGG CTC AGA TGA TGA CTC TGA GCC TCC TTA GCC TGG TCC TGG CTC TCT GCA TCC CCT GGA CCC AAG GCA G and antisense GAT CCT GCC TTG GGT CCA GGG GAT GCA GAG AGC CAG GAC CAG GCT AAG GAG GCT CAG AGT CAT CAT CTG AGC CAT C. The construct was subcloned into XhoI–BamH I sites of pEGFP-N2.
Transfection
Plasmids were transfected into differentiated NG-108 cells using transfection lipid nanofectamin (PAA Laboratories, Colbe, Germany) according to the manufacturer’s instructions. NG-108 cells were transfected in culture medium for 5 h and transferred back to differentiation medium. Primary cortical neurons were transfected using Ca2+-PO4 transfection method. Primary glia-cortical neurons cultures were transfected on days in vitro (DIV) 9–10 whereas purified primary cortical neurons were transfected on DIV 4–6. Briefly, conditioned medium was collected and kept warm at 37°C. Cells were rinsed once with neurobasal medium (without additives). After rinsing, Ca2+-PO4 transfection was performed in neurobasal medium (without additives) for 5 h. After transfection, cells were rinsed once with neurobasal after which cells were cultured in the original conditioned medium.
Sucrose gradient subcellular fractionation
Differentiated NG-108 cells were transfected with either pCCL21-EGFP, pNPY-mRFP, or pEFGP. The next day, cells were homogenized in homogenization buffer (HEPES 4 mM, EDTA 1 mM, sucrose 0.21M, pH7.4, containing 1x complete protease inhibitor cocktail (Roche Diagnostics Corp., Indianapolis, IN, USA)). The homogenate was centrifuged 10 min at 1000 g to pelletize nuclei and unlysed debris. The supernatant was added to a linear sucrose gradient, consisting of 9 layers of sucrose from 0.6 to 3.2 M, and centrifuged for 18 h at 65,000 g in an ultracentrifuge (Sorvall Discovery 90SE, Kendro Laboratories Products, Newtown, CT, USA; rotor SW41Ti, Beckman Coulter, Fullerton, CA, USA). All sucrose solutions were prepared in 4 mM Hepes, 1 mM EDTA, pH 7.4. After centrifugation, 11 fractions of 1 ml each were collected and proteins were precipitated with trichloroacetic acid and processed for Western blotting.
Western blot analysis
A denaturing 10% SDS-PAGE gel was used to resolve NG-108 protein extracts. Gels were blotted overnight in 25 mM ethanolamine/glycine, pH 9.5, to Hybond-ECL nitrocellulose membranes (Amersham Biosciences Corp., Piscataway, NJ, USA). Blocking in Tris-buffered saline Tween 20 (TBST) (50 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween), containing 5% low-fat milk powder (Nutricia, Zoetermeer, The Netherlands) was followed by overnight incubation with rabbit anti-CCL21 1:500 (PeproTech Inc., Rocky Hill, NJ, USA) or monoclonal green fluorescent protein (GFP) 1:500 (Chemicon International Inc., Temecula, CA, USA). After washing, the membrane was incubated at room temperature for 2.5 h in TBST containing 1% milk powder and secondary horseradish peroxidase-conjugated antibody (Amersham Biosciences). Chemiluminescence was developed by ECLplus Western blotting detection system (Amersham Biosciences) and visualized by exposing X-ray films (Eastman Kodak, Rochester, NY, USA). Recombinant CCL21 was used as a positive control (PeproTech).
For protein extracts derived from sucrose fractions, a denaturing 12.5% SDS-PAGE gel was used for migration. Gels were blotted semidry on nitrocellulose membranes. Membranes were blocked with Odyssey blocking buffer (OBB; Li-COR, Lincoln, NE, USA) followed by overnight incubation with the following antibodies diluted in OBB and TBST: rabbit anti-CCL21, 1:4000 (PeproTech); goat anti-neuropeptide Y (NPY), 1:2000 (Sanver Tech, Boechout, Belgium); mouse anti-GFP, 1:3000 (Chemicon). The following day, membranes were washed several times in PBS Tween 20 (PBST); incubated with IRDye 800CW donkey anti-rabbit (1:8000), IRDye 680 donkey anti-mouse (1:8000), or IRDye 800CW donkey anti-goat (1:6000) secondary antibodies (Li-COR) for 1 h; and washed again in PBST. After a quick wash in PBS, membranes were scanned with the Odyssey infrared imaging system (Li-COR). Protein quantification was done with ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA).
Immunocytochemistry
Cells were fixed in 4% paraformaldehyde for 10 min, washed in PBS, blocked in PBS containing 10% FCS, and incubated overnight with the following primary antibodies: mouse antisynaptophysin, 1:50 (Chemicon); mouse anti-tau, 1:1000 (Cell Signaling Technology, Danvers, MA, USA); mouse anti-β-3 tubulin, 1:500 (Chemicon); and rabbit anti-chromogranin A+B, 1:100 (Novus Biologicals, Littleton, CO, USA). Antibody binding was visualized using secondary antibodies coupled to Alexa 488 (Molecular Probes Inc., Eugene, OR, USA) or to cyanine 3 (Cy3; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA). Control experiments for all immunocytochemical staining were done by incubating cells in the absence of primary antibodies.
Fluorescence imaging
Imaging of the immunofluorescent staining and the expression of the various fusion proteins was performed on a Leica AOBS_TCS SP2 confocal laser scanning microscope using an x63 NA 1.4 oil-immersion objective (Leica Microsystems, Rijswijk, The Netherlands) and either an AR/Kr laser for GFP and Alexa 488 visualization, or an He laser for RFP and Cy3 visualization.
Time-lapse imaging
Vesicular CCL21 transport was visualized in transfected NG-108 cells and purified cortical neurons. Imaging was performed using a live-cell imaging setup (Solamere Technology Group, Salt Lake City, UT, USA). The system consists of an AR/Kr laser (Dynamic Laser, Salt Lake City, UT, USA) with controlled power output. Selection of the different laser lines was done using an acousto-optic tunable filter (AOTF), which was connected to a CSU10 spinning Nipkow disk (Yokogawa, Tokyo, Japan). Lab-Tek chambers were placed in a motorized xyz stage of a DM IR2 Leica microscope, which was placed in a custom made incubation chamber with an 37°C humidified atmosphere containing 5% CO2. A Stanford Photonics XR MEGA-10 Gen III iCCD camera (Stanford Photonics, Palo Alto, CA, USA) was used for acquisition. The setup was controlled by InVivo software (Media Cybernetics, Bethesda, MD, USA). Images were acquired at 3 frames/s (3 Hz) with 260 ms integration time per image using an x63 NA 1.4 oil-immersion objective + 1.5 Optovar (Leica Microsystems).
Data analysis
Confocal images and acquired time series were corrected for background noise and were deconvoluted using Huygens Pro Software (Scientific Volume Imaging, Hilversum, The Netherlands). Quantification of vesicle colocalization was performed using single channels of red-green confocal images saved separately. Images were then superimposed, and yellow and green puncta were counted. We also used the method described in de Wit et al. (21)
, in which circular regions were placed around individual puncta in one image and transferred to the other image. Both methods gave similar results. Vesicles were tracked using the MtrackJ plug-in (E. Meijering, Erasmus MC, Rotterdam, The Netherlands) for ImageJ (22
23)
. A vesicle was regarded as moving when it moved for at least 3 frames in the same direction. Vesicles were divided into 3 classes based on their type of movement (anterograde, retrograde, and bidirectional) using the distance to origin function of MtrackJ in ImageJ. For every vesicle, the average speed was calculated by dividing the total traveled distance by the total time. Speed correction for stationary periods was applied.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
for CCL21-EGFP and Fig. 2A
for CCL21-mRFP). In contrast, the expression of the fluorescent label alone resulted in ubiquitous cytoplasmic distribution not only in NG-108 cells (Fig. 1A
for EGFP and Fig. 2B
for mRFP) but also in primary neurons (data not shown).
|
|
Because vesicles may contain multiple proteases, it was investigated whether the CCL21-fusion protein remained intact by Western blot analysis in transfected NG-108 cells, employing anti-CCL21 and anti-GFP antibodies. As expected, anti-GFP antibody recognized EGFP (27 kDa) in pEGFP-transfected NG108 cells as well as the 42 kDa CCL21-EGFP fusion protein in pCCL21-EGFP-transfected cells (Fig. 1B
). When anti-CCL21 antibody was used, it recognized recombinant CCL21 (15 kDa) as positive control and the 42 kDa CCL21-EGFP fusion protein from pCCL21-EGFP-transfected cells (Fig. 1C
). Further control experiments in NG-108 cells and primary cortical neurons revealed a complete overlay of the immunocytochemical staining for anti-CCL21 and anti-GFP antibodies in pCCL21-EGFP-transfected cells (data not shown). These experiments show two things: 1) the sorting of CCL21 is not due to the attached label (both labels show the same expression), and 2) the fusion protein remains intact after sorting.
Vesicle sorting is often due to specific motifs (signal peptides) in proteins. Analysis of the protein sequence of murine CCL21 using the signal peptide finder SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP/) (24)
revealed the possible presence of a 23-amino acid signal peptide at the N-terminal region of CCL21 with the following sequence: MAQMMTLSLLSLVLALCIPWTQG.
To investigate whether the N-terminal region of CCL21 was responsible for the vesicle sorting of CCL21, an N-terminal deletion was introduced directly after the signaling peptide of the CCL21-mRFP fusion protein (pCCL21delN-mRFP). Expression of pCCL21delN-mRFP resulted in ubiquitous cytoplasmic distribution of the pCCL21delN-mRFP fusion protein (Fig. 2C
), not distinguishable from the expression of mRFP alone (Fig. 2B
), suggesting that CCL21 lacking its N-terminal signaling peptide lost its ability to get sorted into vesicles. To demonstrate that this signal peptide is necessary for vesicle sorting, we generated a construct wherein the signal peptide of CCL21 was attached to EGFP. NG-108 cells transfected with this fusion protein (pCCL21sp-EGFP) displayed a punctate/vesicle-like EGFP staining (Fig. 2D
), different from the normal ubiquitous cytoplasmic EGFP expression (Fig. 1A
). These data confirmed that the signal peptide of CCL21 is indeed responsible for its vesicle sorting.
Expression of CCL21-EGFP vesicles in primary neurons
We have previously shown the vesicle-like expression of CCL21 in damaged neurons (8)
. Since CCL21-EGFP transient transfection in NG-108 cells resulted in the same vesicle-like expression, we investigated the expression and localization of CCL21-EGFP containing vesicles in primary cortical neurons. Cultures were transfected with CCL21-EGFP (green fluorescent signal) and immunostained with various neuronal markers (red fluorescent signal) (Fig. 3
). Apart from the soma, CCL21-EGFP was also found in neuronal tau-stained axons (Fig. 3A
, arrowheads) and β-3-tubulin-stained axons and dendrites (Fig. 3B
, arrowheads). Moreover, CCL21-EGFP was frequently observed in direct vicinity of synaptophysin-positive sites (Fig. 3C
, arrows) suggesting that CCL21-vesicles are directed to synapses. This staining pattern, taken together with our earlier findings, clearly shows that CCL21-EGFP containing vesicles are localized throughout neurons.
|
CCL21-EGFP is sorted into DCVs
To further characterize the nature of the CCL21 containing vesicles, NPY-mRFP and VAMP2-mRFP expression plasmids were cotransfected together with CCL21-EGFP. Because NPY is specifically stored in DCVs (25
26)
, we used it as a specific DCV marker. Cotransfection experiments using pCCL21-EGFP (Fig. 4A
) and pNPY-mRFP (Fig. 4B
) revealed a partial overlap of both fluorescent markers (37.2±7.5%, mean±SEM; n=1599 puncta in 10 cells) in NG-108 cells (Fig. 4C
). Similar results have been obtained in primary cortical neurons (Fig. 4D
). VAMP2 is a vSNARE that facilitates synaptic vesicle fusion with the plasma membrane and has been considered a synaptic vesicle marker (25)
. However, recent data have demonstrated that VAMP2 is also localized in the membrane of DCVs and is thus considered one of the various markers necessary for DCV identification (27)
. Coexpression of pCCL21-EGFP (Fig. 4E
) and pVAMP-2-mRFP (Fig. 4F
) also showed a partial colocalization (22.3±3.8%; n=1923 puncta in 14 cells) of both markers (Fig. 4G
). Again, similar results were observed in primary cortical neurons (Fig. 4H
). Chromogranin A and B, which are implicated in the sorting of cargo into DCVs, were also used as markers to further characterize the DCV nature of CCL21-containing vesicles. We transfected differentiated NG-108 cells with pCCL21-EGFP (Fig. 4I
) and performed an immunostaining against chromogranin A and B after fixation (Fig. 4J
). The result showed that 16.8 ± 3.3% (n=910 puncta in 7 cells) of the vesicles were positive for chromogranin A + B (Fig. 4K
), thus supporting the data obtained with NPY and VAMP2 to identify CCL21-containing vesicles as DCVs. Finally, because DCVs originate from the trans-Golgi apparatus (28)
, we immunostained NG-108 cells transfected with CCL21-EGFP (Fig. 4L
) against TGN38, a protein present in rodent trans-Golgi apparatus (Fig. 4M
), where a complete overlap was observed (Fig. 4N
).
|
To further confirm the presence of CCL21 in DCVs, we performed a sucrose gradient fractionation of differentiated NG-108 cells transfected with either pCCL21-EGFP, pNPY-mRFP, or pEGFP. CCL21-EGFP was observed in the same fractions where NPY was detected. (fractions 5, 6, and 7; Fig. 4O
; Supplemental Fig. 1). On the contrary, EGFP was detected in the first three fractions, where cytoplasmic proteins are normally found (Fig. 4O
; Supplemental Fig. 1).
Live-cell imaging of CCL21-EGFP-containing vesicles in NG-108 cells and primary neurons
Live-cell imaging of NG-108 cells (data not shown) and primary cortical neurons enabled us to visualize and follow CCL21-EGFP-containing vesicles over time. Vesicles were visualized as punctuated green signals that showed complex transport behavior from the Golgi/trans-Golgi network (in neurons typically in the axonal hillock) into dendrites and axons (see also Supplemental Movie 1). Most vesicles in axons moved continuously either in the anterograde (Fig. 5
, red arrows) or retrograde directions (Fig. 5
, blue arrow). Few vesicles showed bidirectional movements (Fig. 5
, green arrow) or were not moving (Fig. 5
, yellow arrow). Similar transport behavior was observed in dendrites (see Supplemental Movie 1).
|
CCL21-EGFP-containing vesicles were analyzed for transport direction (anterograde vs. retrograde), transport speed, and localization (axon vs. dendrites). Of the vesicles analyzed, 77% were located in the axons or were transported toward it, compared to 23% into dendrites. These data suggest the existence of preferential transport of CCL21-filled vesicles toward the axon. In axons, 89% of the vesicles moved in an anterograde direction, whereas 11% of the vesicles were transported retrogradely (Fig. 6
and Table 1
). In dendrites, 53% of the vesicles showed anterograde transport, whereas 47% of the vesicles were transported retrogradely (Fig. 6
and Table 1
). Average speed for vesicles in axons (anterograde and retrograde) was 
0.6 µm/s, and slightly slower in the dendrites (see also Table 1
for details).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Furthermore, it has been reported that neuronal injury may lead to microglia activation at distant sites from the primary injury, indicating neuron-microglial communication (19
, 30
31)
. Although it has been recognized that neurons express inflammatory signals to control microglial activity (17
, 32
33)
, it has not yet been elucidated whether neurons are capable of targeting these signals in order to trigger microglial activity at specific locations.
Recent evidence showed that endangered neurons specifically express the chemokine CCL21, which activates microglia via chemokine receptor CXCR3. Furthermore, it has been reported that the lack of CCL21-CXCR3 signaling may inhibit the microglial response after neuronal injury (7
8
, 18
19
20)
. Because neuronal CCL21 was moreover localized in vesicle-like structures, we hypothesized that neurons express and sort CCL21 in vesicles that enable directed communication.
Neuronal CCL21 is targeted to secretory granules
Because there are various neuronal systems to sort and transport signaling molecules, we have determined the type of vesicle that contains CCL21 in primary cortical neurons. We have shown that expression of 2 different CCL21-fluorescent fusion proteins yields a punctate localization indicative of vesicles clearly distinguishable from the cytoplasmic expression of the fluorescent proteins alone. A 23-amino acid N-terminal signal peptide was identified by bioinformatics (SignalP 3.0). Removal of this N-terminal signal peptide of CCL21 abrogated its sorting into vesicles, whereas insertion of this signal peptide upstream of EGFP (a normally nonsorted protein) resulted in punctuate/vesicle-like expression of the EGFP protein. These data indicate that the N-terminal 23-amino acid signal peptide of CCL21 is necessary and sufficient for its sorting into vesicles.
We have also demonstrated that CCL21-EGFP-containing vesicles are present throughout both transfected differentiated NG-108 cells and primary cortical neurons, and are colocalized with NPY-positive vesicles, a marker for large DCVs (25
26
, 34)
. In addition, CCL21-EGFP-containing vesicles costained with other markers for DCVs, such as the vSNARE VAMP2 (27)
chromogranins A and B, which are believed to be major players in the regulation of the sorting of proteins inside DCVs (27)
, and TGN38, a trans-Golgi marker, where DCVs originate from (28)
. However, in contrast to de Wit et al. (21)
, who used the same markers, only a partial colocalization was observed. In the present study, we observed only 37% colocalization with CCL21 and NPY compared to over 80% for NPY and semaphorin 3A observed by de Wit et al. (21)
. It could be assumed that, because CCL21 and NPY belong to two different classes of signaling molecules (neuroimmune factors vs. neuropeptides) and their function in the CNS is completely different, the sorting of these proteins might accordingly be differentially regulated. We also observed a small overlap with overexpressed-VAMP2, similar to that observed for semaphorin 3A by de Wit et al. (21)
, which can be explained by possible competition between the endogenous and overexpressed forms of VAMP2. Against our expectations, low levels of colocalization of CCL21 with chromogranin A and B were observed. Several new lines of evidence suggest that the DCV sorting machinery, with respect to chromogranins, is highly complex and not fully understood (35)
, suggesting the existence and implication of other major players in the sorting machinery. Further experiments will therefore be needed to fully understand the molecules implicated in the sorting of CCL21.
It is known that the cargo of large DCVs can be targeted toward pre- or postsynaptic sites as well as somatodendritic sites (4
, 5
, 36)
. Here, we show that CCL21-EGFP-filled vesicles are transported into axons and dendrites, resembling the transport behavior of large DCVs containing other cargo (4
, 5
, 25)
. More important, CCL21-EGFP-containing vesicles are preferentially transported into axons. Several studies have shown the role of dynein, dynactin, and kinesins in DCV microtubule retrograde and anterograde transport (37
38
39)
. Furthermore, it has been shown that the presence or absence of various kinesins may be responsible for the targeted transport of DCVs toward axons and/or dendrites (38
, 40
41)
. Further studies will be necessary to determine which motor proteins are implicated in CCL21-vesicle transport toward dendrites or axons. Based on our findings, it is concluded that neurons sort CCL21 into large DCVs, the secretory granules of the regulated release pathway (5
, 42)
. Thus, neurons sort CCL21 into the same system that is used for neurotrophins, neurohormones, and neuropeptides, indicating that neurons are able to regulate and target the release of inflammatory signals.
Implications for neuroimmunology
Neurons have long been implicated as solely targets of immunological activity in the brain, innocent victims of activated immune cells. Currently, neurons are more and more seen as active contributors to the local immune environment (12
, 32
33
, 43)
. Therefore, it seems reasonable to assume that such a contribution is facilitated by a cellular mechanism that allows specific targeting of inflammatory signaling. Our data corroborate this assumption, showing for the first time a neuronal chemokine preferentially transported into axons. Moreover, the cargo of large DCVs has so far mainly been discussed in the context of neuron-neuron communication (4
, 5
, 44
45)
; the finding of CCL21 in secretory granules also adds a new aspect to large DCV function, namely their involvement in neuron-glia communication.
The chemokine CCL21 is not the only inflammatory mediator present in neurons. Various reports clearly show that neurons may also express several other cytokines and chemokines, such as IL-1β, IL-6, TGF-β, CXCL10, CXCL12, CCL2, and CX3CL1 (44
45
46
47)
. Little is yet known, however, about the subcellular localization of these signaling molecules in neurons. Findings that suggest vesicle-based expression of IL-6 and TGF-β in neuronal cell lines (44
45)
and the description of neuronal CCL2 and CXCL12 in colocalization with neurotransmitters and neuropeptides in synaptic regions (9)
may indicate that CCL21 is not the only targeted inflammatory signal in neurons. While this article was under revision, it was shown that GABAergic interneurons use CXCL12 in their communication with neural progenitors in the dentate gyrus (48)
.
Thus, the data presented here strongly indicate that neurons not only sort and target signaling molecules for neuron-neuron interaction, they also target inflammatory mediators that control microglia activity in a directed manner. It is therefore proposed herein that neurons can modulate their immune-environment using inflammatory signaling molecules.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication January 31, 2008. Accepted for publication July 17, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
