The authors have declared that no competing interests exist.
Conceived and designed the experiments: LB CS ID. Performed the experiments: LB RW XC FJ YLD ID. Analyzed the data: LB RW MD YLD IR CS ID. Wrote the paper: LB CS ID.
Oligodendrocyte differentiation is temporally regulated during development by multiple factors. Here, we investigated whether the timing of oligodendrocyte differentiation might be controlled by neuronal differentiation in cerebellar organotypic cultures. In these cultures, the slices taken from newborn mice show very few oligodendrocytes during the first week of culture (immature slices) whereas their number increases importantly during the second week (mature slices). First, we showed that mature cerebellar slices or their conditioned media stimulated oligodendrocyte differentiation in immature slices thus demonstrating the existence of diffusible factors controlling oligodendrocyte differentiation. Using conditioned media from different models of slice culture in which the number of Purkinje cells varies drastically, we showed that the effects of these differentiating factors were proportional to the number of Purkinje cells. To identify these diffusible factors, we first performed a transcriptome analysis with an Affymetrix array for cerebellar cortex and then real-time quantitative PCR on mRNAs extracted from fluorescent flow cytometry sorted (FACS) Purkinje cells of L7-GFP transgenic mice at different ages. These analyses revealed that during postnatal maturation, Purkinje cells down-regulate Sonic Hedgehog and up-regulate vitronectin. Then, we showed that Sonic Hedgehog stimulates the proliferation of oligodendrocyte precursor cells and inhibits their differentiation. In contrast, vitronectin stimulates oligodendrocyte differentiation, whereas its inhibition with blocking antibodies abolishes the conditioned media effects. Altogether, these results suggest that Purkinje cells participate in controlling the timing of oligodendrocyte differentiation in the cerebellum through the developmentally regulated expression of diffusible molecules such as Sonic Hedgehog and vitronectin.
Oligodendrocytes are central nervous system macroglial cells that synthesize myelin, a multilayered membrane ensheathing axons which facilitates rapid nerve conduction
Neuron maturation affects oligodendrocyte survival and the timing of myelin formation, OPCs nonetheless differentiate into mature oligodendrocytes and generate a myelin sheath in the absence of axons in vitro
The role of neurons in the switch between OPC proliferation and differentiation into oligodendrocytes remains unclear. The timing of this switch depends on both the intracellular timer and extrinsic factors
In this study, we investigated the existence of neuronal soluble factors controlling oligodendrocyte differentiation in an integrated system. For that purpose, we used cerebellar organotypic cultures, in which neuron-glial interactions mimic those occurring in vivo and in which only one type of neuron, the Purkinje cell, is myelinated
To analyze the timing of the oligodendrocyte differentiation process in cerebellar slice cultures, we focused on the expression of MBP because this protein is expressed in mature oligodendrocytes (both pre- and myelinating oligodendrocytes
P0 cerebellar slices grown for 7 DIV had very few MBP+ oligodendrocytes (
At 10 DIV, MBP+ oligodendrocytes were present throughout the slices (
After 14 DIV, the number of MBP+ internodes increased (
Our results therefore showed that most OPCs differentiated into MBP+ oligodendrocytes between 7 and 10 DIV and most of the myelination process occurred between 10 and 14 DIV. P0 slices were thus considered to be immature for oligodendrocyte differentiation during the first 7 DIV and to be mature during the following week, in which rapid oligodendrocyte differentiation occurred.
We investigated whether, during the phase of oligodendrocyte differentiation, cells in the slice culture (between 7 and 14 DIV) were able to increase oligodendrocyte differentiation on immature slices (7 DIV). To this aim, P0 cerebellar slices were grown in culture until 7 DIV, when the differentiation of oligodendrocytes was beginning and could be called “mature”. At this time point, fresh P0 slices (immature) were added to the Millicell (
Then to confirm that this increase of MBP immunostaining was indeed due to an increase of the number of oligodendrocytes, we evaluated the density of APC+/OLIG2+ cells on the slice cultures. APC is expressed by oligodendrocytes and some astrocytes
These results showed that mature slices release factors that increase oligodendrocyte differentiation in immature ones. They also suggest that the switch from OPC proliferation to oligodendrocyte differentiation in cerebellar organotypic cultures is controlled by factors released from cerebellar cells at critical stages of maturity. Then we investigated whether Purkinje cells might influence the timing of oligodendrocyte differentiation in cerebellar slices. Indeed the number of Purkinje cells can be manipulated during the first week in vitro, i.e. before the addition of immature slices.
To investigate whether Purkinje cells might control the secretion of these differentiation factors, we used two different slice-culture models in which the numbers of Purkinje cells were markedly different. To increase the number of Purkinje cells, we applied a BrdU treatment that kills dividing cells during the first 3 DIV of the culture
Photomicrographs of P0–14 DIV cerebellar slices double-labeled with antibodies against MBP (A1, B1, and C1) and Calbindin (Calb1 = CaBP; A2, B2, and C2) and double-labeled with antibodies against Calb1 (green) and Parvalbumin (PV, red) for A3, B3, and C3. A. P0–14 DIV control slice: note the presence of numerous MBP+ internodes (A1) and Calb1/CaBP positive Purkinje cells (A2). B. P0 cerebellar slice treated during the first 3 days with a high dose of BrdU. Note the absence of MBP staining (B1) and the high density of Calb1+ Purkinje cells (B2). C. The P0 slices were axotomized after 2 DIV (dotted line represent the lesion). The resulting density of Purkinje cells (C2) is much lower than that under the conditions for A and B. In all 3 conditions (A3, B3, and C3), Purkinje cell were differentiated since some of them presented elaborated dendritic tree with spines and expressed parvalbumin (yellow Purkinje cells). Scale bar is 390 µm for A1, B1, C1, A2, B2, and C2 and 30 µm in A3, B3, and C3.
To decrease the number of Purkinje cells, we took advantage of the fact that newborn Purkinje cells axotomized after 2 DIV, die in large numbers in organotypic culture. Indeed, axotomized slices contained small numbers of surviving Purkinje cells in the second week of culture (
We then evaluated whether the Purkinje cells have reached a differentiated stage at 14 DIV in all 3 different conditions (Ctrl, BrdU-treated slices or axotomy-treated slices).
We previously showed that at the time of the culture, newborn Purkinje cells were bipolar, whereas after 7 DIV, most Purkinje cells have retracted their primitive processes and developed numerous perisomatic protrusions
Thus, the use of conditioned medium taken from BrdU treated slices (BrdU-CM) allowed us to study the effect of slice cultures containing high numbers of differentiated Purkinje cells and no MBP+ oligodendrocytes. Whereas taking the CM from the “axotomy” slices (Axt-CM) allowed us to study of the effect of slice cultures containing very few differentiated Purkinje cells.
For P0 slices grown for seven days in the presence of conditioned medium from P0 7DIV–14DIV cultures (Ctrl-CM,
We then tested the CMs from the two different culture conditions described above to manipulate the number of Purkinje cells. In all experimental conditions, we did not detect slices in Group II (i.e. presenting a significant number of internodes). Thus, the density of MBP might be used as an index of oligodendrocyte differentiation. P0 slices grown with BrdU-CM (
To understand how Purkinje cells could be involved in the timing of oligodendrocyte differentiation, we looked for genes which are: regulated during postnatal development, expressed by Purkinje cells and are coding for extracellular proteins. We used Affymetrix arrays for transcriptome analysis on cerebellar cortical regions, at time points P0, P3, P5, P7, and P10. Statistical and gene ontology analyses identified 4 extracellular proteins encoded by genes displaying an expression that decreased, and 6 displaying an expression that increased by more than twofold during the first 10 postnatal days (see
Affymetrix ID | Gene Symbol | P0/P0 | P3/P0 | P5/P0 | P7/P0 | P10/P0 | Expression Pattern |
1450716_at | Adamts1 | 1.000 | 0.825 | 0.679 | 0.599 | 0.364 | decreased |
1426670_at | Agrn | 1.000 | 0.783 | 0.772 | 0.616 | 0.409 | decreased |
1451119_a_at | Fbln1 | 1.000 | 1.172 | 0.806 | 0.720 | 0.412 | decreased |
1436869_at | Shh | 1.000 | 1.158 | 0.843 | 0.673 | 0.421 | decreased |
1424186_at | Ccdc80 | 1.000 | 1.644 | 3.097 | 4.046 | 4.829 | increased |
1449581_at | Emid1 | 1.000 | 1.816 | 1.848 | 1.993 | 2.525 | increased |
1424010_at | Mfap4 | 1.000 | 1.344 | 1.982 | 2.699 | 2.063 | increased |
1417678_at | Mmp24 | 1.000 | 0.980 | 1.367 | 1.374 | 2.041 | increased |
1451342_at | Spon1 | 1.000 | 2.148 | 2.245 | 2.652 | 3.499 | increased |
1420484_a_at | Vtn | 1.000 | 1.640 | 2.284 | 2.717 | 3.334 | increased |
Expressions of the 10 transcription factors belonging to GO Cellular Component term ‘extracellular region/space’ (GO:0005576) and presenting a significant increase or decrease of expression measured by Affymetrix from cerebellar cortices during the first 10 postnatal days. Values for P0, P3, P5, P7, and P10 are expressed as a ratio of P0 values.
Then to check the expression patterns of Shh and vitronectin in developing Purkinje cells, we performed real-time quantitative PCR using mRNAs extracted from fluorescent flow cytometry sorted (FACS) Purkinje cells of L7-GFP transgenic mice at different ages (P0, P3, P7 and P10). In these experiments, the results obtained from P0 and P3 Purkinje cells were pooled together and compared to those of pooled P7 and P10. From this comparison, it clearly appeared that over the ten first postnatal days Shh presented a decreased expression, whereas vitronectin displayed an increased expression in Purkinje cells (
As Shh and vitronectin were good candidate modulators of the timing of cerebellar oligodendrocyte differentiation, we analyzed their roles in the control of the OPC proliferation/differentiation balance in organotypic cultures.
We investigated the effects of Shh treatment on OPC proliferation in P0–7DIV cultures by counting the OPCs (OLIG2+) incorporating BrdU following a short pulse. Treatment with Shh increased the number of OLIG2+/BrdU+ cells around Purkinje cell axons close to deep nuclear neuron regions (see
We then investigated whether Shh inhibited OPC differentiation. We added Shh to the culture medium between the 7th and 10th days of culture (
We then investigated whether Shh inhibition promoted oligodendrocyte differentiation. We added cyclopamine —an alkaloid known to block the Shh signaling pathway
Finally, we investigated whether Shh blocked oligodendrocyte differentiation in immature slices induced by the factors released by mature slices. We used BrdU-treated mature slices (depleted of OPCs;
These results suggest that Shh prevents OPCs from exiting the cell cycle and inhibits the effects of oligodendrocyte differentiation factors —released by mature slices— on immature slices.
We investigated the effects of vitronectin on immature slices. Addition of vitronectin (5 µM) during the first 7 DIV increased the density of MBP staining on P0–7DIV slices (
To affect oligodendrocyte differentiation, the extracellular matrix protein vitronectin must be released from cerebellar slices. Experiments with serum-free medium were also performed (see
We also compared the amounts of vitronectin in control, axotomized and BrdU treated P0–10DIV slices. As expected, we observed more vitronectin in BrdU treated slices than in control and axotomized slices (
Finally, we used blocking antibodies to determine whether vitronectin present in the CM induced the increase in OPC differentiation. In the presence of these antibodies, the CM did not promote OPC differentiation in immature slices (
Our findings suggest that Sonic Hedgehog and vitronectin play important antagonistic roles in controlling the timing of OPC differentiation during cerebellar development.
Different cell types differentiate in parallel during development. The synchronization of cell development is particularly important for cell types that have strong interactions, such as oligodendrocytes and neurons. OPCs start to differentiate if a mitogenic stimulus is removed or a differentiation stimulus is added. This cell differentiation is inhibited in the presence of a mitogenic stimulus and the absence of a differentiation stimulus
Mature slices promoted oligodendrocyte differentiation in immature slices. We modified the number of Purkinje cells present in the slices to investigate whether this cell type is involved in the stimulation of oligodendrocyte differentiation during the second week of culture. We have previously reported, that high doses of BrdU treatment during the first three days of culture generates cerebellar slices with large numbers of Purkinje cells and reactive astrocytes, but with markedly decreased numbers of oligodendrocytes and microglial cells
In the present study, CM from mature slices treated with BrdU had a stronger effect on oligodendrocyte differentiation than CM from control mature slices. The almost complete absence of oligodendrocytes and microglial cells from mature slices treated with BrdU rules out the self-stimulation of oligodendrocyte differentiation
Altogether our findings indicate that among the cells in mature slices, Purkinje cells and astrocytes might be responsible for directly or indirectly releasing factors promoting oligodendrocyte differentiation in immature cultures. We verified through the analysis of two parameters (dendritic morphology and expression of parvalbumin) that Purkinje cell differentiation occurred after the two types of treatment (BrdU and axotomy).
We then hypothesized that during their maturation Purkinje cells might express differentially genes to synchronize OPC differentiation with their own differentiation. Transcriptome analysis performed on cerebellar cortex during the first postnatal week showed a decreasing expression profile of 4 genes and an increasing expression profile of 6 genes. Interestingly, among them, Shh and vitronectin were of special interest. Shh is well known to be produced by Purkinje cells
The transcriptome analysis has been performed on cerebellar cortex. Nevertheless, the same profiles were observed for Shh and vitronectin in GFP expressing PCs sorted by FACS, demonstrating that during their maturation Purkinje cells down regulate the expression of Shh and up regulate their vitronectin expression. Thus, Purkinje cells might indeed control the differentiation of OPC by regulating the expression of Shh and vitronectin during their differentiation.
Then, we checked whether Shh and vitronectin were able to participate to the control of OPC differentiation process in organotypic culture.
We showed that Shh stimulated proliferation of OPCs and inhibited oligodendrocyte differentiation in our cultures. Shh affects OPC in different ways. First, Shh plays an important role in OPC specification in the spinal cord
The addition of vitronectin to immature slices enhanced oligodendrocyte differentiation and vitronectin-blocking antibodies inhibited the effects of mature-slice induced differentiation on immature slices. Furthermore, when the slices cultures contained more Purkinje cells (following BrdU treatment), levels of vitronectin were elevated. Only a slight difference between control and axotomized slices was detected likely because we are closed to the threshold of detection. Thus, our results clearly demonstrate that vitronectin is required for mature slices to promote oligodendrocyte differentiation in immature slices. Previous studies reported no effect of vitronectin on OPC differentiation
Organotypic culture is an integrated system, in which cell interactions mimic those occuring in vivo, and is easier to manipulate than in vivo models. Furthermore only one type of neuron is myelinated in this system: the Purkinje cell. Using this system, we showed that the maturation of the Purkinje cell is involved in controlling the timing of oligodendrocyte differentiation. Indeed, our results suggest that Purkinje cells release different factors during their maturation, which have opposing effects on oligodendrocyte differentiation. This temporal regulation probably synchronizes the differentiation of oligodendrocytes and Purkinje cells. However, as discussed above, oligodendrocyte differentiation occurs even when the number of Purkinje cells is reduced (axotomy experiment). This suggests the presence of other differentiating factors in cultured slices or medium. Many of the factors known to affect oligodendrocyte development, such as TGFß, IGF-1, and progesterone
The findings of the present study strongly imply that Purkinje cells initially inhibit OPC differentiation during their maturation by releasing Shh and then subsequently promote OPC differentiation by producing vitronectin. Purkinje cells thus appear to orchestrate OPC differentiation in cerebellar organotypic cultures.
All procedures were submitted and approved by the Regional Ethics Committee in Animal Experiment N°3 of Ile-de-France region (p3/2009/020). Cerebellar organotypic cultures were established from newborn (P0) Swiss mice (Mus musculus, Janvier, Le Genset St Isle, France), as previously described
For each result, At least 5 animals and 25 slices were studied in 3 independent experiments. In the figures, “N” is the number of animals. The experimental plan was designed in accordance with the European Union Guidelines for the care and use of experimental animals.
In experiments evaluating the effect of a decreased number of Purkinje cells, tissue slices were dissected along the midline between the dorsal and ventral halves, with two needles, under a dissection microscope, after 2 days in vitro (DIV). The two parts were gently separated to ensure complete axotomy and were then re-apposed.
We added Shh (3 µg/ml, R&D Systems, Lille, France), cyclopamine (5 µM, Toronto Research Chemicals, North York, Ontario, Canada), vitronectin (VN; 5 µM, Oxford Chemical Research, Euromedex, Souffelweyersheim, France), goat anti-VN antibody (0.2 µg/ml, Abcam, Cambridge, UK), or a control goat antibody (goat anti-Brn3b, 0.2 µg/ml, Santa Cruz Biotechnology, Santa Cruz, CA) to the medium. The medium and the added drug were replaced every two to three days. Untreated slices were used as controls.
Cerebellar slices were depleted of mature oligodendrocytes by killing OPCs through the addition of 5-bromo-2-deoxyuridine (BrdU; 150 µM; Sigma, Saint Louis, MO) in NaCl solution (9 g/l) to the nutrient medium for the first 3 DIV (P0 slices with a 3-day BrdU treatment: P0–3DBT
To study the effect of Shh on OPC proliferation, both control and treated (5 µM Shh) slices were incubated with BrdU (20 µM) for 3 h before fixation.
P0 slices maintained in culture until 7 DIV contained very few MBP (myelin basic protein) immunoreactive oligodendrocytes (
A slice ratio of 2 matures: 1 immature (
In some experiments, immature slices were cultured on conditioned medium (CM). CM was obtained from cultures of P0 slices, between the 8th and the 14th days of culture (
Mouse monoclonal antibodies against MBP (diluted 1/500, Chemicon, MAB381, Millipore) were used to visualize mature oligodendrocytes and myelin
After 7, 10, or 14 DIV, the cultures and co-cultures were fixed by incubation in 4% paraformaldehyde in phosphate buffer (0.1 M, pH 7.4) for 1 h at room temperature. Slices incubated with anti-MBP antibody were washed in PBS (Invitrogen) and immersed in Clark's solution (95% ethanol/5% acetic acid) for 20 min at 4°C to extract some of the lipids —increasing the accessibility of MBP antigens— and were then washed several times with PBS. All slices were incubated for 1 h in PBS containing 0.2% gelatin, 0.1% sodium azide (PBSGA) and 0.1 M lysine. This was followed by an overnight incubation at room temperature with primary antibody diluted in PBSGA. The primary antibodies were detected with the following secondary antibodies: CY3-conjugated goat anti-mouse (1/200 dilution; Jackson ImmunoResearch Laboratories Inc, West Baltimore Pike, PA), FITC-conjugated sheep anti-rabbit (1/200 dilution, Chemicon), CY3-conjugated donkey anti goat (1/200 dilution, Jackson Immunoresearch), Alexa Fluor 488-conjugated donkey anti mouse (1/400 dilution, Invitrogen), AMCA-conjugated donkey anti rabbit (1/50 dilution, Jackson Immunoresearch), CY3-conjugated goat anti-rabbit (1/200 dilution; Jackson ImmunoResearch). Slices incubated for 2 h with the secondary antibodies in PBSGA were washed several times in PBS and mounted in Mowiol (Calbiochem). The sections were analyzed under a Leica DMR microscope equipped with a Coolscan camera (Princeton Instruments, Evry, France).
MBP immunoreactive oligodendrocytes and internodes (myelin segments) were quantified by acquiring an image of each slice using the software, Metaview. The images were then analyzed using the software, Metamorph (Universal Imaging Corporation). Contours of the slices were drawn —the threshold was adjusted based on the MBP immunostaining— and the density of MBP immunostaining was determined for each slice. Means and standard errors of the mean (SEM) were calculated for each type of slice.
To distinguish between oligodendrocyte differentiation and myelination, we performed a semi-quantification of MBP+ internodes which did not consider the number of differentiated oligodendrocytes. The ×5 pictures were enlarged on the computer screen (zoom 4) to evaluate the number of internodes, and the first 26 internodes were pointed and counted using ImageJ software. Two groups were specified: Group I included slices that either did not contain MBP+ internodes (myelin segments) or contained less than 25 internodes; group II included slices containing more than 26 internodes.
On slides with Calb1 labeling of Purkinje cells, pictures of areas centered on white matter close to the deep nuclear neurons were acquired for OLIG2 and APC using the Metaview software. Then, number of cells positive for both OLIG2 and APC per 81,000 µm2 were determined manually using ImajeJ software. Means and standard errors of the mean (SEM) were calculated for each type of slice.
A proliferation index for OPCs was calculated by determining the number of cells positive for both OLIG2 and BrdU. On slides with Calb1 labeling of Purkinje cells, single areas of white matter close to the deep nuclear neurons were chosen, 1 µm thick confocal sections were acquired for OLIG2 and BrdU labeling (Leica confocal microscope; Plateforme d'Imagerie, IFR83). Using the software Metamorph, images containing pixels labeled for both OLIG2 and BrdU were generated and the number of double-stained cells per 150,000 µm2 was determined.
Mann-Whitney tests were used to compare two groups and Kruskal-Wallis tests were used for multiple comparisons. P values≤0.05 were considered to significant. In the figures, the values are represented as the percentage of the control.
Affymetrix microarrays were used to assess patterns of gene expression in murine cerebellar cortical areas. Total RNA was extracted from dissected areas centered on the layer of Purkinje cells from vermal lobules 5 and 6 of Swiss mice of various ages: P0, P3, P5, P7, and P10 (4 replicates = 4 independent measurements for each stage; different litters were used for each measurement). RNA was extracted using RNAeasy Mini Kit (Qiagen, Courtaboeuf, France). Following reverse transcription, cDNAs were hybridized with Affymetrix microarrays (MOE430 GeneChip, Affymetrix Platform, Institut Curie, Paris, France). All the experimental procedures and results have been loaded on ArrayExpress database (E-MEXP-3444, Experiment name: PC transcriptome,
Matlab was used for mathematical manipulations and statistical analysis. The aim was to identify transcripts displaying a significant increase or decrease over time from P0 to P10.
The four replicates were used to estimate the intrinsic variability of expression, or pure error, for each transcript. We used the pure error in F-tests of lack of fit
A constant model was retained (i.e. no lack of fit) for 26,912 of the 31,818 transcripts, an affine model was retained (lack of fit of the constant model, but not of the affine model) for 4,474 transcripts, and a quadratic model was retained (lack of fit of the constant and affine models, but not of the quadratic model) for the remaining 432 transcripts. All “constant” transcripts were discarded. We retained all of the “affine” transcripts as potential candidates. We also searched for “quadratic” transcripts displaying continual increases or decreases. One quadratic transcript was identified with a continually decreasing pattern over time. Two lists were therefore generated: 2,262 affine transcripts showing continuous increases in expression and 2,213 transcripts with continuous decreases in expression (2,212 quadratic, 1 affine). Among the latter, we selected transcripts with at least a two-fold difference in expression (higher or lower) at P10 compared with P0. This gave a list of 1190 transcripts, from which we further discarded those with a very low level of expression (below 45), to give a final list of 627 candidate transcripts.
Gene annotations were expanded and upgraded with NCBI Entrez Gene ID, Unigene, and PubMed for the remaining 627 candidate probesets. Transcribed sequences and expression sequence tags (ESTs) with no identified function were eliminated from the reported lists. Gene annotations for the MOE430 GeneChip obtained from Webgestalt (web-based gene set analysis toolkit
RT-PCR quantified in real time was carried out on two independent sets of mRNA for each age (P0, P7, P10), to assess Shh and VN mRNA levels using a LightCycler 480 real-time PCR system (Roche). First-strand cDNAs were synthesized from 500 ng total RNA (Thermo Scientific, Surrey, UK), in accordance with the manufacturer's instructions. The reaction mixture contained 1× SYBR Green I Master (Roche Diagnostics, Mannheim, Germany), 1.5 ng of the first-strand cDNA and 240 nM of each forward and reverse primer, in a total volume of 13 µl. Thermal cycle parameters were: 95°C for 5 min, followed by 40 cycles of 95°C for 10 s, 56°C for 20 s, and 72°C for 20 s. All primers were synthesized by Eurofins MWG Operon (Ebersberg, Germany). The following primer sequences were used: VN forward:
Purkinje cells were isolated from BacL7-GFP mice as described previously
Total RNA was extracted from approximately 3.104 purified Purkinje cells at indicated stages with Trizol reagent (Invitrogen). The cDNA was prepared by reverse transcription of 100 ng RNA using SuperScript III First-Strand Synthesis System (Invitrogen) with an oligo-dT primer according to the manufacturer's instructions. The resulting cDNA was used as a template for real time PCR using a Light Cycler 480 thermocycler (384 plates, Roche Diagnostics) with a home-made SYBR Green QPCR master mix
Cerbellar slices were lysed in RIPA buffer (50 mM Hepes, 150 MM NaCl, 5 mM EDTA, 1% NP-40, 0.5% SDS; pH 7.7). Lysates were clarified by centrifugation. The protein pellet was collected by centrifugation, washed with cold acetone and recentrifuged. The pellet was then allowed to air dry and resuspended in 10 mM Tris pH 8. The DCA protein assay (Biorad, Hercules, California) was used to determine protein levels. Samples (2 µg of CM proteins) were denatured by heating for 3 min at 95°C in Laemmli buffer, separated by 10% SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (PVDF, Amersham Biosciences, GE Healthcare, UK). TBS supplemented with 0.1% Tween-20 and 5% dry milk powder was used for blocking and antibody incubations. Primary (rabbit polyclonal anti-VN, 1/1000) and secondary Alkaline phosphatase conjugated goat anti-rabbit polyclonal antibody (1/7500, Promega, Madison, WI). Membranes were incubated overnight at 4°C with the primary antibody and then for 1 h at room temperature with the secondary antibody. They were then washed three times in TBS supplemented with 0.1% Tween-20 and once in TBS. The alkaline phosphatase-conjugated antibody was detected by NBT/BCIP (Promega). A prestained protein ladder (Fermentas, Euromedex, Souffelweyersheim, France) was used as the size marker.
We thank Isabelle Caille, Beatrice Durand and Rachel Sherrard for helpful discussion, Richard Schwarzman, Plateforme imagerie IFR83, for assistance with confocal microscopy and Julien Dahan for his work on microarray analysis.