Abstract
The synaptic transmission in the mammalian brain is not limited to the interplay between the pre- and the postsynapse of neurons, but involves also astrocytes as well as extracellular matrix (ECM) molecules. Glycoproteins, proteoglycans and hyaluronic acid of the ECM pervade the pericellular environment and condense to special superstructures termed perineuronal nets (PNN) that surround a subpopulation of CNS neurons. The present study focuses on the analysis of PNNs in a quadruple knockout mouse deficient for the ECM molecules tenascin-C (TnC), tenascin-R (TnR), neurocan and brevican. Here, we analysed the proportion of excitatory and inhibitory synapses and performed electrophysiological recordings of the spontaneous neuronal network activity of hippocampal neurons in vitro. While we found an increase in the number of excitatory synaptic molecules in the quadruple knockout cultures, the number of inhibitory synaptic molecules was significantly reduced. This observation was complemented with an enhancement of the neuronal network activity level. The in vivo analysis of PNNs in the hippocampus of the quadruple knockout mouse revealed a reduction of PNN size and complexity in the CA2 region. In addition, a microarray analysis of the postnatal day (P) 21 hippocampus was performed unravelling an altered gene expression in the quadruple knockout hippocampus.
Similar content being viewed by others
Introduction
The functionality of neuronal synapses is warranted by the correspondence of a pre- and a postsynapse that form the synaptic cleft which allows for the transmission of an electrical activity by chemical neurotransmitters. Nowadays, it is obvious that glia cells such as astrocytes1,2,3,4 as well as extracellular matrix (ECM) molecules5,6,7,8,9 contribute to the process of synaptic information transfer. Most of the ECM molecules in the brain are produced and released by astrocytes and neurons themselves10. One specialized structure of the ECM is termed perineuronal net (PNN) and was originally discovered by Camillo Golgi more than a century ago11. PNNs surround inhibitory fast spiking parvalbumin-positive GABAergic interneurons12,13,14, but are also associated with further subpopulations, for example excitatory pyramidal neurons in the cerebral cortex15,16. The core structure of PNNs consists of chondroitin sulfate proteoglycans (CSPGs) of the lectican family17,18, hyaluronan19,20, tenascins, particularly tenascin-R (TnR)21,22 and link proteins23,24 for stabilization. When the composition of PNNs is impaired, the neuronal synaptic plasticity in form of LTP is affected, as has been reported for the tenascin-C (TnC)7, TnR25, brevican26 and neurocan27 knockout mice. In the present study, the interplay between neurons and astrocytes in conjunction with the PNN structure was analysed using the quadruple knockout mouse, which lacks the four ECM components TnC, TnR, neurocan and brevican. The first study describing this knockout mouse revealed an upregulation of the proteins fibulin-1 and fibulin-2 in the brains of one month old mice28. In a previous study our laboratory investigated hippocampal neurons of the quadruple knockout mouse in vitro and found diminished PNNs and an altered expression of synaptic proteins that were accompanied by a decreased frequency of mEPSCs and mIPSCs after 14 and 21 DIV29. This also translates in an altered synaptic plasticity in vivo as the quadruple knockout mouse exhibited a reduced short-term depression connected with changed frequency dependence in the dentate gyrus of the hippocampus subjected to a classical LTP paradigm30.
In the light of these observations we sought to expand our insight into the importance of the ECM for the process of synaptogenesis. To this end we explored the quadruple knockout mouse in conjunction with the indirect co-culture system of cortical astrocytes and hippocampal neurons29,31. We examined the proportion of inhibitory and excitatory synapses in culture, as well as the spontaneous activity of the neuronal networks using the multielectrode array (MEA) approach32. We also monitored the formation of PNNs in the developing hippocampus of quadruple knockout mice in vivo, as it has recently been reported that PNNs are highly enriched in the CA2 region of the murine hippocampus33. Moreover, we compared the gene expression in this brain structure at postnatal day (P) 21 in wildtype and quadruple knockout mice by microarray analysis.
Results
Enhancement of excitatory synapses in the quadruple knockout mouse
Hippocampal neurons and cortical astrocytes of wildtype and quadruple knockout mice were cultivated in the four possible combinations (Nwt/wt/Awt/wt, Nwt/wt/Ak°/k°, Nk°/ko/Awt/wt, Nko/ko/Ako/ko), in the following referred to as conditions. For the detection of excitatory synapses after 14 DIV and 21 DIV (Fig. 1), the presynaptic protein vGlut1 and the postsynaptic protein PSD95 were examined using specific antibodies. We consider situations where pre- and postsynaptic puncta are in register and the corresponding fluorescence signals overlap as structural synapses31,34. In contrast puncta that are not colocalized with their corresponding partner are designated as delocalized, for the purpose of this study. To distinguish between PNN wearing (Fig. 1a–d+a”-d”) and PNN-free neurons (Fig. 1a’–d’+a”’–d”’), aggrecan, a major constituent of the PNN structure, was used35. In case of the formation of structural synapses white puncta indicating the overlap of pre- and postsynaptic puncta were detectable.
Knockout neurons cultivated in the presence of knockout astrocytes (Nko/ko/Ako/ko), which were surrounded by PNNs showed a significant upregulation of PSD95 and colocalized vGlut1 puncta of 30.18 ± 5.79% (p = 0.002) and 38.31 ± 10.19% (p = 0.040) compared to the control (Nwt/wt/Awt/wt) after 14 DIV (Fig. 1e). The delocalized vGlut1 puncta were not significantly increased (16.90 ± 6.78%, p = 0.091). No significant shift of excitatory puncta was observable for knockout neurons grown with wildtype astrocyte (Nko/ko/Awt/wt) where vGlut1 and PSD95 puncta shifted by 13.38 ± 6.55% (p = 0.185) and 19.97 ± 6.47% (p = 0.392), respectively. A significant enhancement of 47.29 ± 12.49% (p = 0.050) could be detected for excitatory synapses. Hence, the astrocytes seemed not capable to rescue the neuronal phenotype. When wildtype neurons were cultivated together with knockout astrocytes (Nwt/wt/Ako/ko), no alteration of the number of synaptic puncta was detectable, (vGlut1 puncta: −3.56 ± 5.30% (p = 0.472), PSD95 puncta: −1.62 ± 6.37% (p = 1.000), colocalized puncta: −4.06 ± 6.73% (p = 1.000), Fig. 1, Supplementary Table S1a,b), suggesting that the distinct genotypes of the astrocytes did not intervene in the formation of excitatory synapses.
An even larger rise of excitatory structural synapses was determined for PNN-free knockout neurons cultivated with knockout astrocytes (Nko/ko/Ako/ko) (Fig. 1f). Here, an increase was measured of 52.24 ± 8.51% (p < 0.001) compared to the control condition (Nwt/wt/Awt/wt). The delocalized vGlut1 and PSD95 puncta were also significantly elevated by 33.69 ± 6.82% (p < 0.001) and 37.64 ± 5.05% (p < 0.001). Analogous results were obtained for knockout neurons grown in the presence of wildtype astrocytes (Nko/ko/Awt/wt) where the vGlut1 puncta were significantly increased by 26.65 ± 5.65% (p = 0.002), as well as the PSD95 puncta (25.14 ± 4.98%, p = 0.002) and the excitatory synapses (49.73 ± 8.01%, p < 0.001). In contrast, when wildtype neurons were co-cultivated with knockout astrocytes (Nwt/wt/Ako/ko), no differences with regard to excitatory synapses were measured (vGlut1: 7.94 ± 4.71% (p = 0.505), PSD95: 5.28 ± 4.92% (p = 1.000), structural synapses: 2.33 ± 4.62% (p = 1.000), Fig. 1f, Supplementary Table S1a,b).
After 21 DIV, PNN wearing knockout neurons co-cultivated with knockout astrocytes (Nko/ko/Ako/ko) (Fig. 1g) also exhibited a strong increase of excitatory synaptic puncta (structural synapses: 31.03 ± 5.9%, p < 0.001; delocalized vGlut1 puncta: 25.29 ± 5.76%, p = 0.008; delocalized PSD95 puncta: 24.38 ± 4.14%, p = 0.001). Knockout neurons combined with wildtype astrocytes (Nko/ko/Awt/wt) displayed a higher formation of excitatory synapses (PSD95 puncta: 28.56 ± 5.34%, p = 0.001; structural synapses: 31.43 ± 7.91%, p = 0.024). In contrast, the delocalized vGlut1 puncta were not significantly altered (14.03 ± 5.50%, p = 0.618). Wildtype neurons co-cultivated with knockout astrocytes (Nwt/wt/Ako/ko) showed no change in their number of synaptic puncta (vGlut1: −2.86 ± 5.83%, p = 1.000; PSD95: 5.59 ± 5.72%, p = 1.000; colocalized puncta: 2.43 ± 7.41%, p = 1.000). In the absence of PNNs, the only significant difference that was observed concerned the number of PSD95 puncta which differed between knockout neuron combinations Nko/ko/Ako/ko compared to Nko/ko/Awt/wt (p = 0.005) (Fig. 1h).
In conclusion, the elimination of the four extracellular matrix molecules TnC, TnR, brevican and neurocan resulted in a remarkable increase of excitatory synapses in PNN-positive as well as PNN-negative mutant neurons after 14 DIV.
Reduction of inhibitory synapses in the quadruple knockout mouse
For the detection of inhibitory synapses, vGAT was used as a presynaptic and gephyrin as a postsynaptic marker36 after 14 and 21 DIV (Fig. 2a,a”,b,b”,c,c”,d,d”). Compared to the control (Nwt/wt/Awt/wt) and different from the situation of excitatory synapses described above, the number of structural inhibitory synapses appeared reduced in all conditions involving mutant cells, irrespective of the presence or absence of PNNs (Fig. 2 and Supplementary Table S2a,b). For PNN-positive knockout neurons co-cultivated with knockout astrocytes for 14 DIV (Nko/ko/Ako/ko), the gephyrin puncta as well as the inhibitory synapses were significantly reduced (Fig. 2e) by −21.15 ± 2.75% (p = 0.008) and −24.25 ± 3.98% (p = 0.016). The vGAT puncta were not affected (−12.79 ± 7.15%, p = 0.062). Analogous results were obtained for PNN-positive knockout neurons grown with wildtype astrocytes (Nko/ko/Awt/wt) where inhibitory synapses diminished by −23.85 ± 3.98% (p = 0.007). In contrast, the vGAT and gephyrin puncta were not significantly modified, −16.18 ± 6.23%, (p = 0.028) and 1.23 ± 4.05% (p = 1.000), respectively. No significant alteration of synaptic puncta was visible in PNN-positive wildtype neurons co-cultivated with knockout astrocytes (Supplementary Table S2a,b).
When knockout neurons not surrounded by a PNN were cultivated for 14 DIV (Fig. 2f), in the presence of knockout astrocytes (Nko/ko/Ako/ko), the cultures presented a significant decrease of synaptic puncta compared to the control (Nwt/wt/Awt/wt) which amounted to −30.22 ± 4.01% (p < 0.001) for vGAT puncta, −13.70 ± 2.08% (p = 0.049) for gephyrin puncta and −35.11 ± 3.24% (p < 0.001) for structural inhibitory synapses. A significant reduction of vGAT puncta by −27.09 ± 3.78% (p < 0.001) and inhibitory synapses by −32.01 ± 3.64% (p < 0.001) could furthermore be identified on knockout neurons grown in the presence of wildtype astrocytes (Nko/ko/Awt/wt). Interestingly, the percentage of gephyrin puncta appeared unchanged in this condition (1.15 ± 3.45%, p = 1.000). The combination of wildtype neurons with knockout astrocytes (Nwt/wt/Ako/ko) resulted in a decrease of vGAT puncta by −17.83 ± 3.64% (p = 0.035), while the gephyrin puncta as well as the inhibitory synapses remained unaffected (4.99 ± 2.88%, p = 0.372; and −10.12 ± 3.64%, p = 0.667). Thus, after 14 DIV the fraction of inhibitory synapses was reduced both on PNN-positive and PNN-negative knockout neurons.
An analogous analysis was performed after 21 DIV (Fig. 2g,h, Supplementary Table S2a,b). Under this schedule, PNN-positive knockout neurons grown with knockout astrocytes (Nko/ko/Ako/ko) developed significantly lesser gephyrin puncta, while the vGAT puncta and inhibitory synapses were not altered compared to the control (Nwt/wt/Awt/wt); vGAT: −11.66 ± 4.01% (p = 0.287), gephyrin: −20.15 ± 3.50% (p = 0.004) and colocalized puncta: −13.01 ± 3.96% (p = 0.354). The same was examined for the synaptic puncta of knockout neurons with wildtype astrocytes (Nko/ko/Awt/wt). Here, a significant reduction of −25.44 ± 3.24% (p < 0.001) of gephyrin puncta was quantified, whereas the vGAT (3.45 ± 5.95%, p = 1.000) and the structural synapses (−8.40 ± 5.35%, p = 0.396) remained unchanged. Interestingly, wildtype neurons with PNNs co-cultured with knockout astrocytes (Nwt/wt/Ako/ko) exhibited a significant enhancement of the vGAT puncta by 49.47 ± 8.53% (p < 0.001) and inhibitory synapses by 30.97 ± 6.82% (p = 0.005). The gephyrin puncta were not changed in this condition (5.42 ± 3.92% (p = 0.807)).
Neurons without a PNN (Fig. 2h) showed the following expression of synaptic puncta in comparison to the control (Nwt/wt/Awt/wt). No alteration in the percentage increase/decrease of synaptic puncta was found for knockout neurons grown with knockout astrocytes (Nko/ko/Ako/ko) (vGAT: −12.73 ± 3.33% (p = 0.879), gephyrin: −10.77 ± 3.20% (p = 0.098) and colocalized puncta: −7.83 ± 3.45% (p = 1.000). The gephyrin puncta were significantly reduced, by −21.63 ± 2.76% (p < 0.001), when knockout neurons were co-cultivated with wildtype astrocytes (Nko/ko/Awt/wt), but no change could be found for the vGAT and inhibitory synapses in this condition (vGAT: 4.80 ± 4.22% (p = 0.795) and colocalized puncta: 2.38 ± 4.52% (p = 1.000)). Like before seen for neurons with PNNs (Fig. 2g), wildtype neurons combined with knockout astrocytes (Nwt/wt/Ako/ko) exhibited a significantly enhanced percentage of vGAT puncta by 26.78 ± 5.53% (p < 0.001) and of inhibitory synapses by 21.93 ± 4,79% (p = 0.007), whereas the percentage of gephyrin puncta was not modified (−3.30 ± 2.90% (p = 1.000)) after 21 DIV.
Summarizing these observations, the elimination of the four ECM molecules TnC, TnR, Neurocan and brevican led to a reduction of inhibitory synapses on PNN-positive as well as PNN-negative knockout neurons cultivated for 14 DIV. This effect levelled off after three weeks where wildtype neurons maintained in the presence of knockout astrocytes displayed an enhanced number of vGAT-positive afferents and inhibitory synapses, irrespective of the presence or absence of PNNs (Fig. 2g,h).
Higher neuronal network activity in the quadruple knockout mouse
To test the influence of the altered proportion of excitatory and inhibitory synapses at hippocampal quadruple knockout neurons on the spontaneous activity, MEA analysis was conducted as established in our laboratory32. Via the 60 electrodes of the MEA (Fig. 3a,b), the spontaneous activity of the neuronal network developing in vitro was analysed (Fig. 3c,d,e). Each electrode was individually recorded and the traces (Fig. 3g) were represented in small boxes (Fig. 3f,g), reflecting the spontaneous activity as well as the occurrence of bursts in different areas of the network in culture (Fig. 3c–e).
The number of spikes (Fig. 4a) which correspond to action potentials was significantly increased in knockout neurons co-cultured with knockout astrocytes (Nko/ko/Ako/ko) after 14 DIV (4306.84 ± 217.87 (p < 0.001)) and 21 DIV (8762.16 ± 454.98, p < 0.001) compared to the control (Nwt/wt/Awt/wt) (14 DIV: 2348.39 ± 153.26, 21 DIV: 4367.73 ± 261.28). The same applied for the number of bursts (Fig. 4b) in this condition with a value of 101.04 ± 4.78 bursts per 10 min (p < 0.001) after 14 DIV and 195.45 ± 6.82 bursts (p < 0.001) after 21 DIV in comparison to the control values of 57.6 ± 3.22 bursts after 14 DIV and 109.38 ± 5.75 bursts after 21 DIV. Bursts consist of single spikes and occur simultaneously at a variety of electrodes in a short time interval. When knockout neurons were combined with wildtype astrocytes (Nko/ko/Awt/wt), the spike number was nearly the same as for the control, indicating a contribution of the astrocytic phenotype into the process of network activity. Similarly, the spike frequency was upregulated in cultures with knockout neurons co-cultivated with knockout astrocytes at both examined time points (14 DIV: 7.15 ± 0.33, p < 0.001; 21 DIV: 14.60 ± 0.70, p < 0.001) compared to the control (14 DIV: 3.91 ± 0.24; 21 DIV: 7.28 ± 0.43). With progressing cultivation time, the neuronal networks gained complexity, more bursts arose and a higher spike frequency in bursts emerged. Knockout neurons grown in the presence of knockout astrocytes (Nko/ko/Ako/ko) showed a significantly higher spike frequency in bursts of 92.18 ± 4.23 Hz (p < 0.001), which occurred after 14 DIV compared to 66.61 ± 2.44 Hz of the control (Nwt/wt/Awt/wt). When cultivated for 21 DIV, these differences between the two conditions were still present (Nwt/wt/Awt/wt: 86.17 ± 2.97, Nko/ko/Ako/ko: 94.32 ± 2.95 (p = 0.031). Interestingly, knockout neurons co-cultured with wildtype astrocytes (Nko/ko/Awt/wt) showed an opposite effect concerning their spike frequency in bursts, namely a reduction with a value of 48.54 ± 1.38 Hz (p < 0.001) after 14 DIV. After 21 DIV no significant difference could be observed (75.15 ± 1.91 Hz (p = 0.347)).
A higher amount of spikes in bursts appeared in the (Nko/ko/Ako/ko) condition compared to the control condition. Whereas in the control condition only 43.79 ± 2.18% spikes after 14 DIV and 50.37 ± 2.02% spikes after 21 DIV appeared within a burst, 60.62 ± 1,85% (p < 0.001) spikes in bursts after 14 DIV and 71.50 ± 1,90% (p < 0.001) spikes in bursts after 21 DIV occurred in the (Nko/ko/Ako/ko) condition. There was again a significant decrease of this analysed parameter in cultures of knockout neurons cultured with wildtype astrocytes (Nko/ko/Awt/wt) after 14 DIV with a percentage of 23.96 ± 1.31% (p < 0.001) spikes in burst, which attained control levels after 21 DIV, with a percentage of 50.78 ± 1.93% (p = 1.000). The duration of bursts was also measured using MEA analysis (Fig. 4f, Supplementary Table S3a,b). In the control condition (Nwt/wt/Awt/wt) an average burst lasted for 341.88 ± 19.57 ms after 14 DIV and for 284.94 ± 12.83 ms after 21 DIV. No change in the mean duration of a burst was seen for knockout neurons grown with knockout astrocytes (Nko/ko/Ako/ko) after 14 DIV, which showed similar results of 294.96 ± 8.84 ms (p = 0.057) after 14 DIV and 341.72 ± 10.66 ms (p < 0.001) after 21 DIV, which also applies for all other conditions compared to the control condition. After 21 DIV, all conditions showed a significant rise in the burst duration in comparison to the control. The only striking difference of the burst duration was observable after 14 DIV for knockout neurons cultured with wildtype astrocytes (Nko/ko/Awt/wt). Here, the burst duration was significantly protracted to 456.95 ± 26.44 ms (p = 0.032) instead of 341.88 ± 19.57 ms as detected in the control.
In conclusion, the MEA analysis revealed that the deletion of four critical ECM constituents resulted in an increased network activity in the cultures of hippocampal neurons.
Reduction of PNNs in vivo in the quadruple knockout mouse
The ECM molecules TnC, TnR, brevican and neurocan are constituents of PNNs, therefore it appeared of interest to analyze the impact of the elimination of these genes on the formation of PNNs in vivo. The marker WFA was used to visualize PNNs in frontal brain slices of the hippocampus of wildtype (Fig. 5a,c,e,g,i) and knockout mice (Fig. 5b,d,f,h,j).
The analysis of the PNNs in vivo was focused on the hippocampus of 15, 20, 25, 30 and 35 days old mice. Within the CA2 region of the hippocampus, an area consisting of numerous PNNs developed starting at P15 and evolved further along with the maturating hippocampus. This PNN territory was present in the wildtype as well as in the knockout hippocampus. A small proportion of the fluorescence in the CA2 region could be due to the diffuse interstitial matrix. However, it has recently been demonstrated that pyramidal neurons in the CA2 region are a major source of perineuronal nets in the hippocampus and positive for WFA and aggrecan33. This agrees with recent studies that emphasize that the WFA signal in CA2 is PNN-specific37,38. Consistent with this interpretation, a higher magnification image obtained using confocal laser scanning microscopy revealed that the major part of the fluorescence was concentrated in PNNs around neuronal cell bodies, suggesting a specific labeling of these structures (Fig. 5k–n). The PNN areas were determined for the different postnatal stages of the developing hippocampus (P15-P35) by measuring the WFA-positive area (Fig. 6). The quantification revealed a significant downregulation of the PNN area in the quadruple knockout hippocampus at P15, P20, P25, and P30 (Fig. 6e). The PNN area amounted to 42558.46 ± 3223.24 µm2 (p < 0.001) at P15, 62810.40 ± 5409.42 µm2 (p = 0.05) at P20, 63750.31 ± 3300.47 µm2 (p < 0.001) at P25 and 63620.73 ± 5082.00 µm2 (p < 0.001) at P30 compared with the values of the wildtype of 64155.52 ± 3273.71 µm2 at P15, 78005.23 ± 5021.09 µm2 at P20, 90590.67 ± 6055.96 µm2 at P25 and 104404.09 ± 7588.34 µm2 at P30 (Supplementary Table S4a,b). In the wildtype hippocampus, the area size increased over time until P30. No significant difference of the PNN size was detectable between the wildtype and knockout hippocampus at P35. At this stage the PNN area of the knockout hippocampus amounted to 73468.03 ± 8223.63 µm2 and of the area size of the wildtype hippocampus encompassed 90108.04 ± 6145.99 µm2 (p = 0.120), indicating an alignment of their PNN area size towards adulthood.
Next to the area size, the corrected total cell fluorescence (CTCF) of the WFA-positive area in the CA2 region of the hippocampus was measured (Fig. 6c,c’,d,d’,f). As previously seen for the area size, the fluorescence intensity was also significantly diminished in the quadruple knockout hippocampus at P15, P20, P25 and P30 (P15: 1963962.39 ± 150734.82, p < 0.001; P20: 2820496.51 ± 436309.19, p = 0.0018; P25: 2299008.92 ± 160585.48, p = 0.0012; and P30: 3305522.46 ± 395786.81, p = 0.002) compared to the wildtype situation (P15: 4380059.02 ± 403847.23, P20: 5805772.77 ± 691451.33, P25: 5328356.19 ± 702004.79 and P30: 7300182.00 ± 987441.78). At P35, no change with regard to the fluorescence intensity of the PNN area was found in the knockout hippocampus (4217469.63 ± 647443.36 in comparison to the wildtype hippocampus (5529794.88 ± 430019.63, p = 0.108). In conclusion, the maturation of PNNs is significantly delayed in the quadruple knockout hippocampus but normalized after postnatal day 30 to attain the wildtype control levels in adulthood (Supplementary Table S4a,b).
Alterations in the gene expression in the quadruple knockout hippocampus at postnatal day 21
Considering the fact that the PNN maturation was significantly retarded in the P21 hippocampus of the quadruple knockout mouse, this time point was chosen for the comparative analysis of gene expression using microarrays.
Overall 34473 genes were examined of which 438 were significantly altered in the quadruple knockout hippocampus. Because we had detected differences in synapse formation, we were particularly interested in genes associated with the ECM or synaptogenesis. A volcano plot and a clustering heatmap highlight the dysregulated genes of interest of these functional annotations, which are either up- or downregulated (Fig. 7a,b) in the knockout hippocampus. The significantly altered genes comprise Gpc3, Gabrq, Gad2, Wnt7, Syt9, Sod3, Dnm3, Cript, Sema4c, Calr, Cplx3, Col4a3, Col27a1, Adamts13, Grin2d, Col1a2, Spon2 and Rapsn. For example, the three collagens Col1a2 (−1.35-fold, p = 0.0017), Col4a3 (−1.16-fold, p = 0.0014) and Col27a1 (−1.21-fold, p = 0.0098) were significantly downregulated, as well as Spon2 (−1.33-fold, p = 0.0002) (Fig. 7c). Spon2 is known to facilitate neurite outgrowth of hippocampal neurons39,40. In contrast, the ECM genes Gpc3 (1.40-fold, p = 0.0089) and Calr (1.16-fold, p = 0.0038) were significantly upregulated in the quadruple knockout hippocampus. Interestingly, genes related to neuronal development and function (Fig. 7d) were also modified, as shown by a decreased linear fold change of Adamts13 (−1.17-fold, p = 0.0058) and increased values for the linear fold change of Cript (1.05-fold, p = 0.0083), Sod3 (1.10-fold, p = 0.0081) and Wnt7a (1.33-fold, p = 0.0002). The expression of the neuronal genes (Fig. 7f) Cplx3 (−1.17-fold, p = 0.0037) and Grin2d (−1.08-fold, p = 0.0057) was significantly diminished, whereas Dnm3 (1.11-fold, p = 0.0015), Sema4c (1.16-fold, p = 0.0065), Gad2 (1.32-fold, p = 0.0062), Cntnap4 (1.50-fold, p = 0.0198) and Gabrq (1.52-fold, p = 0.0310) were significantly enhanced. However, the expression of mRNA for the synaptic components SLC17A7 (encoding for vGlut: −1.02-fold, p = 0.994), SLC32A1 (encoding for vGAT: 2.13-fold, p = 0.778), Dlg4 (encoding for PSD-95: −1.03, p = 0.898) and Gphn (encoding for Gephyrin: 1.08, p = 0.049) was not significantly altered (Fig. 7e). Chondroitin sulfate proteoglycans such as Cspg4, Cspg5 or Ptprz1 that could potentially compensate the knockout were also analysed, but no significant difference could be observed in their expression levels (Fig. 7b). A tabular overview of the data can be found in the supplement (Supplementary Table S5).
Discussion
Although it is known for a few years that PNNs contribute to the process of synaptogenesis, it is not yet clear to which extent the different molecules are involved therein9,10,41.
Because of the manifold and complex character of the PNN structure, it is not surprising that neurons with PNN, as well as quadruple knockout neurons with a defective PNN composition display considerable changes in their synapse formation. Several studies of the recent years have shown that one of the main features of the ECM and PNNs is to restrict the formation of neurites42 and their synapses with concurrent protection of neurons towards harmful substances or events43. These studies relied on the application of the bacterial derived enzyme chondroitinase ABC (ChABC), which degrades the chondroitin-4-sulfate, chondroitin-6-sulfate, dermatan sulfate and the hyaluronan glycosaminoglycan chains10,44. When hippocampal neurons were treated with ChABC in vitro, synapse formation was boosted6. Different from the ChABC-treated cultures, the quadruple knockout neurons miss the CSPGs neurocan and brevican, but still express aggrecan, the PNN marker used in this study, and possibly other CSPGs such as phosphacan45. These may be expressed similar to the wildtype control because the gene array analysis did not reveal substantial modifications of CSPG genes.
TnR is a pivotal constituent of PNNs and its elimination leads to a change of PNN distribution, composition and function46. A recent study concerning hippocampal neurons from TnR-deficient mice reported a modified PNN organization around dendrites that could be rescued by adding TnR and aggrecan into the culture medium21. With regard to physiological parameters, TnR-knockout mice displayed an enhancement of excitatory synaptic transmission in the hippocampal CA1 region47. This is in agreement with our observation that an increased number of excitatory synapses resulted in enhanced synaptic transmission in the neuronal network of the quadruple knockout mouse, which was recorded via MEA analysis in vitro.
It is conceivable that the missing expression of TnC and TnR is compensated by other glycoproteins. For example, the tenascin family comprises the four genes TnC, TnR, TnW and TnX, but neither TnW nor TnX are elevated in the quadruple knockout mouse48,49. In contrast, the two glycoproteins fibulin-1 and -2 of the ECM are clearly upregulated on the protein level in the quadruple knockout brains. This is accompanied by an altered hyaluronan deposition in the mutant ECM28. Interestingly, fibulin-1 mutations are associated with cognitive deficits in the human and fibulin-2 mediates TGF-beta1 effects on adult neural stem cells50,51.
A former study already reported the synaptogenesis of hippocampal quadruple knockout neurons in an in vitro co-culture system. The loss of TnC, TnR, brevican and neurocan resulted in a significant increase of synaptic density within 14 DIV that transited to a decremented synapse number after 21 DIV. Neurons coated with PNNs were characterized by an attenuated synapse number after both 14 and 21 DIV29.
Unexpectedly, in our approach WT astrocytes were not capable to rescues the PNN formation by knockout neurons. Astrocytes in vitro secrete brevican, neurocan and TnC. However, TnC is not a major component of perineuronal nets52 while TnR is mainly expressed by oligodendrocytes and not by astrocytes10 (and is only present at low concentration in the co-culture of WT astrocytes with KO neurons, if at all. Lecticans do contain binding sites for hyaluronic acid and TnR53. Several reports suggest that CSPGs like aggrecan or brevican are incorporated into perineuronal nets through their interaction with TnR and hyaluronic acid14,53,54. However, TnR is absent from the knockout neurons in co-culture. This may explain why in our model the incorporation of astrocytic CSPGs into PNNs was impaired. This agrees with a study that emphasized the functional importance of TnR for the formation of PNNs21.
For the formation of structurally and functionally intact synapses, specific molecules are necessary. In former studies, it was demonstrated that the astrocyte-secreted factors thrombospondin and hevin lead to the development of structural, but silent synapses55,56. In line with this, Gpc4 and Gpc6 were recently also discovered as astrocytic factors, which stimulate the excitatory synapse formation by regulating glutamate receptor clustering57. The analysis of the gene expression performed in this study revealed an upregulated expression of Gpc3 in the hippocampus of P21 quadruple knockout mice, which is probably an explanation for the higher number of excitatory synapse number and the elevated activity in the quadruple knockout cultures.
A recently published study has demonstrated an altered synaptic plasticity in the dorsal dentate gyrus of the quadruple knockout hippocampus that translated into an enhanced short-term depression in conjunction with a modified frequency-dependence in vivo30. This changed synaptic plasticity in vivo is also detected in in vitro neuronal networks performed in this study. Here, the spontaneous network activity was significantly increased in quadruple knockout cultures. These data are in agreement with the report of Dityatev and colleagues who found a higher excitability of hippocampal neurons in culture after ChABC treatment14. Furthermore, hippocampal neuronal cultures treated with hyaluronidase developed an epileptiform activity due to the occurrence of superbursts58. Thus, it is evident that the loss of pivotal ECM molecules leads to a higher neuronal activity independent of whether the ECM molecules are digested by appropriate enzymes or genetically eliminated.
The initiation of PNN formation in the murine hippocampus detectable by WFA occurred between postnatal stage P10 and P15. In the adult hippocampus, the PNN formation resembled the expression in the wildtype hippocampus, indicating a progressive compensation of the four missing ECM molecules TnC, TnR, brevican and neurocan, possibly by the upregulation of other ECM molecules. The results found for the in vivo PNN formation fit perfectly with the in vitro findings29, both demonstrating a diminished PNN formation in the postnatal quadruple knockout mouse. A recent study demonstrated that especially the CA2 excitatory pyramidal neurons are surrounded by PNNs. When mice were kept in an enriched environment, the PNN expression was enhanced, indicating a correlation between PNN development and environmental stimuli within the juvenile brain33. Interestingly, PNNs have been proposed as indispensable mediators of the modulation of synaptic plasticity by the transcription factor Otx259,60,61. An impaired PNN formation and structure accompanied with altered synaptic plasticity is associated with a couple of psychiatric and neurological diseases such as schizophrenia and epilepsy, as reviewed recently62.
Significant dysregulation of genes of interest of the ECM or genes associated with neuron biology were detected by a comparative transcriptome analysis of the P21 quadruple knockout compared to the wildtype hippocampus using microarrays.
Like the previously discussed Gpc3 also other molecules contribute to synapse formation and synaptic transmission. For example, in the quadruple knockout mouse Sema4c is upregulated, which is known as interaction partner of different PSD95 isoforms in neocortical cultures as well as in the adult brain of mice63. Cplx3 that was significantly downregulated in the quadruple knockout hippocampus regulates the synaptic neurotransmitter release of both excitatory and inhibitory synapses64. The molecule Grin2d, also known as GluN2D, was reduced in the quadruple knockout hippocampus. By using Grin2d-deficient mice, its contribution in synaptic transmission of hippocampal interneurons of young mice as well as of the pyramidal CA1 neurons of newborn pups has been proven65. Furthermore, Grin2d has been found in hippocampal interneurons and enhances their activity due to its involvement in synaptic transmission66.
Thus, available data concur with our observation that a modified number of excitatory and inhibitory synapses prevail in quadruple knockout neurons in vitro. The excitation versus inhibition balance is considered very important in the context of neuropsychiatric disease67. A shift of the balance can result from neurodevelopmental abnormalities targeting the inhibitory interneurons68 or from structural changes at the synapse69. Our observations suggest that the ECM intervenes in the balance of excitation and inhibition in neural networks. A recent review elegantly summarizes potential mechanisms of action of PNNs70. Brevican for example, which is missing in our mouse model controls the location of AMPA and potassium channels and might therefore affect synapse dynamics71. Neurocan is known to influence the binding of neural cell adhesion molecule (NCAM) and ephrin type-A receptor 3 (EPHA3) at perisomatic synapses72. TnR, another protein missing in the quadruple knockout mouse is known to be associated with HNK-1 and necessary for the maintenance of the excitatory and inhibitory balance in CA1 region47. Beyond these PNN-associated compounds also proteins released by astrocytes influence synaptogenesis, for example thrombospondins and the gabapentin receptor55,73, Hevin and SPARC56, or Pentraxin 3 that promotes excitatory synapse formation by fostering the clustering of AMPA glutamate receptors74. In the light of these observations, the interactions of ECM molecules with neuronal receptors may represent a promising theme for analyzing this issue in future studies.
Methods
Animal housing and ethical standards
The present study was carried out in accordance with the European Council Directive of September 22, 2010 (2010/63/EU) for care of laboratory animals and approved by the animal care committee of North Rhine-Westphalia, Germany, based at the LANUV (Landesamt für Umweltschutz, Naturschutz und Verbraucherschutz, Nordrhein-Westphalen, D-45659 Recklinghausen, Germany). The study was supervised by the animal welfare commissioner of Ruhr-University. Male and female SV129 mice were housed individually with a constant 12-h light-dark cycle and access to food and water ad libitum. All efforts were made to reduce the number of animals in the experiments. Embryos of both sexes were used.
Cell culture
Co-cultivation of astrocytes and neurons
Hippocampal neurons were obtained from embryonic mice and co-cultured with astrocytes in an indirect contact as described previously31 (See supplemental data S6).
Multielectrode array (MEA)
Cultivation of hippocampal neurons on MEAs
Electrophysiological recordings of neuronal networks were performed using multielectrode arrays (MEAs) (60MEA200/30iR-Ti, Multi Channel Systems MCS GmbH) following the previously described protocol32,75.
Electrophysiological recordings
Recordings of the spontaneous activity were performed for 10 min with a sampling frequency of 20 kHz. The data of the individual electrodes were collected using the software MC_Rack (version 3.9.0 by Multi Channel Systems MCS GmbH). The field potentials of raw data were corrected with a high pass filter which had a frequency of 200 Hz. The spontaneous activity was detected with a spike detector only if the value exceeded a threshold value 4.5-fold higher than the standard deviation. The following parameters were set for the measurements: Maximal interval to start burst, 10 ms; Maximal interval to end burst, 100 ms; minimal interval between bursts, 210 ms; minimal duration of burst, 50 ms; minimal number of spikes in bursts, 5.
Immunocytochemistry
Staining of excitatory synapses
Initially, the neurons were fixed with 4% (w/v) PFA (Carl Roth GmbH & Co. KG; Cat. No.: 4235.1) for 10 min at RT. After fixation the cells were washed four times for 5 min each with wash solution (PBS, 10% v/v fetal bovine serum (FBS (Sigma-Aldrich by Merck KGaA; Cat. No.: F7524), 0.1 mM glycine (VWR International GmbH; Cat. No.: 101196X), 0.1% v/v Triton X-100 (AppliChem GmbH; Cat. No.: A4975,0500)). The primary antibodies were diluted in wash solution as follows: vGlut1, polyclonal, guinea pig, 1:500 (Synaptic Systems GmbH; Cat. No.: 135304, RRID: AB_887878); PSD95, IgG2a, mouse, 1:500 (Merck KGaA; Cat. No.: MAB1598, RRID: AB_1121285) and aggrecan, polyclonal, rabbit, 1:500 (Merck KGaA; Cat. No.: AB1031, RRID: 90460). The incubation was carried out for 1 h at room temperature. Afterwards, the secondary antibodies were diluted in wash solution and were incubated for 1 h at RT: anti-guinea pig, IgG (H + L), AF647, 1:500 (Dianova; Cat. No.: 106-605-003, RRID: AB_2337446), anti-mouse, IgM + IgG (H + L), AF488, 1:250 (Dianova; Cat. No.: 115-545-044, RRID: AB_2338844) and anti-rabbit, IgG (H + L), Cy3, 1:500 (Dianova; Cat. No.: 111-165-045, RRID: AB_2338003).
Staining of inhibitory synapses
The staining procedure of inhibitory synapses was similar to the previously described protocol by Dobie and colleagues36. Neurons were fixed after 14 and 21 days in vitro with 4% (w/v) PFA for 10 min at RT. The primary antibodies were diluted in 3% w/v BSA (in PBS) (vGAT, polyclonal, guinea pig, 1:1000 (Synaptic Systems GmbH; Cat. No.: 131 004, RRID: AB_887873), gephyrin, IgG1, mouse, 1:500 (Synaptic Systems GmbH; Cat. No.: 147 011, RRID: AB_887717) and aggrecan, polyclonal, rabbit 1:500 (Merck KGaA; Cat. No.: AB1031, RRID: 90460). The secondary antibodies were diluted in 3% w/v BSA (in PBS), as follows: anti-guinea pig, IgG (H + L), AF647 (Dianova; Cat. No.: 106-605-003, RRID: AB_2337446), 1:500; anti-mouse, IgM + IgG (H + L), AF488 (Dianova; Cat. No.: 115-545-044, RRID: AB_2338844), 1:250 and anti-rabbit, IgG (H + L), Cy3, 1:500 (Dianova; Cat. No.: 111-165-045, RRID: AB_2338003).
Immunohistochemistry
Preparation of cryosections
For the preparation of cryosections, wildtype and quadruple knockout mice were intracardially perfused according to the German animal protection laws and approved by the responsible governmental authorities (Landesamt für Umweltschutz, Naturschutz und Verbraucherschutz (LANUV), Nordrhein-Westfalen, D-45659 Recklinghausen, (AZ 84-02.04.2015.A014)).
With the help of a peristaltic pump (Cole-Parmer GmbH), the blood was exchanged with a 0.9% (w/v) NaCl solution (Carl Roth GmbH & Co. KG) with 0.2% (v/v) of 25 000 I.E./5 ml heparin-sodium (Ratiopharm GmbH). Thereafter, 4% w/v PFA was applied for 10–15 min for fixation of the brain. Decapitation as well as the quick removal of the brain from the skull followed. The brain was transferred into 4% w/v PFA and incubated overnight at 4 °C. On the next day, the brain was placed into 30% w/v sucrose (Thermo Fisher Scientific Inc.). After approximately two days, the brain was embedded with tissue tec freezing medium (Leica Microsystems GmbH; Cat. No.: 14020108926) on dry ice and stored at −20 °C. Afterwards, the frozen brains were cut into 14 µm thick frontal sections (interaural: 1.86 mm, bregma: −1.94 mm) using a cryostat CM 3050 S (Leica Microsystems GmbH).
Staining of cryosections
First, the cryosections were incubated in blocking solution for 1 h at RT in a wet chamber, followed by the incubation with primary antibodies. The primary antibodies were diluted in PBT1 (PBS, 0.1% w/v BSA, 0.1% v/v Triton X-100) containing 5% v/v goat serum (Dianova) as follows: neurofilament 200 (NF200), polyclonal, rabbit, 1:300 (Sigma by Merck KGaA; Cat. No.: N4142, RRID: AB_477272), Wisteria floribunda agglutinin (WFA, 1:100 (Sigma by Merck KGaA; Cat. No.: L1516)) and aggrecan, polyclonal, rabbit, 1:300 (Merck KGaA; Cat. No.: AB1031, RRID: 90460). An overnight incubation at 4 °C was carried out and on the next day, sections were washed three times with PBS. Afterwards, the secondary antibodies as well as Hoechst dye (1:100000) were diluted in PBS/A (PBS, 0.1% w/v BSA) as follows: anti-rabbit, IgG (H + L), AF488, 1:250 (Dianova; Cat. No.: 111-545-045, RRID: AB_2338049), anti-streptavidin, Cy3, 1:500 (Dianova; Cat. No.: 016-160-084, RRID: AB_2337244) and anti-rabbit, IgG (H + L), Cy3, 1:500 (Dianova; Cat. No.: 111-165-045, RRID: AB_2338003). Sections were incubated with the secondary antibodies for 3 h at RT in a wet darkened chamber. In a last step, the cryosections were washed three times with PBS and once with MilliQ water and finally mounted with coverslips (24 × 50 mm, Menzel by Thermo Fisher Scientific Inc.) using ImmuMount.
Molecular biology
mRNA isolation
The GenElute Mammalian Total RNA Miniprep Kit (Sigma-Aldrich by Merck KGaA; Cat. No.: RTN350) was used to isolate the mRNA from the hippocampal tissue following the manufacturer’s instructions.
Affymetrix GeneChip® analysis
Using Affymetrix GeneChip® Mouse Gene 2.0 ST Arrays, a whole-transcript expression analysis of >30,000 RefSeq transcripts of P21 hippocampus of wildtype and quadruple knockout mice siblings was performed in triplicates. RNA quantity was assessed using the NanoDrop 1000 (Thermo Scientific Nano Drop Technologies), and RNA quality control was carried out with the RNA 6000 Nano Assay (Agilent 2100 Bioanalyser) to ensure that the samples had an RNA integrity number (RIN) of at least 9. All further processing of total RNA was performed according to the Ambion whole-transcript Expression kit and the Affymetrix GeneChip whole-transcript terminal labeling and control kit manuals. The cDNA fragment size was checked using the 2100 Bioanalyser (fragment size between 50 and 200 bp). Fragmented sense cDNA was biotin-endlabelled with terminal deoxynucleotidyl transferase and samples were hybridised to GeneChip® Mouse Gene 2.0 ST Arrays at 45 °C for 16 hours with 60 rpms. Hybridised arrays were then washed and stained on a Fluidics Station 450 (program: FS450 0002) and scanned on a GeneChip® Scanner 3000 7 G (both Affymetrix). Raw image data were analysed with GeneChip® Command Console® Software (AGCC). Gene expression intensities were normalized (Robust Multichip Average76) and summarized using Affymetrix® Expression Console™ Software. Transcripts that were differently expressed more than 1.5-fold with an Anova p-value less than 0.05 between the analysed samples were addressed as regulated.
The heat-Map was established using AltAnalyse77 version 2.1.0 with following parameters: cosine hopach for column clustering and cosine weighted for row clustering. The rows were normalized relative to the row median. The volcano plot was constructed using Graph Pad Prism 6.
Experimental design and statistical analysis
In general, the data are expressed as mean ± SEM. The statistical significance is given by the p value: p ≤ 0.05 = *, p ≤ 0.01 = ** and p ≤ 0.001 = ***. The program IBM SPSS Statistic (Version 20) was used for statistical evaluation. The exact number of experimental repetitions can be found in the figure legends.
Microscopy
The confocal laser-scanning microscope LSM 510 meta operated with the ZEN2009.Ink software (Carl Zeiss Microscopy GmbH) was used for image acquisition of the immunocytochemical stainings. Different layers were recorded with an interval of 0.25 µm using 630-fold magnification. For generating two-dimensional images, pictures were overlaid. For quantification of synaptic puncta, the program ImageJ was used with the Plugin “Puncta Analyser” from Barry Wark; licensed under http://www.gnu.org/copyleft/gpl.html) and the following settings were adjusted: rolling ball radius = 50 pixel, size (pixel2) = 2-infinity and circularity = 0.00–1.00.
PNNs in vivo
The immunological staining of the cryosections was recorded using the microscope Axio Zoom.V16 (Carl Zeiss Microscopy GmbH). A prominent PNN-positive area was found in the CA2 region of the hippocampus33. To quantify the WFA-intensity and -area, the PNNs in the micrographs were manually bordered, the corrected total cell fluorescence (CTCF) was determined as previously described78 and the area was measured using ImageJ.
Data Availability
All data generated or analysed during this study are included in this published article (and its Supplementary Information files). The microarray data was deposited in the gene Expression Omnibus (GEO accession: GSE127216).
References
Haydon, P. G. GLIA: listening and talking to the synapse. Nat Rev Neurosci 2, 185–193, https://doi.org/10.1038/35058528 (2001).
Bezzi, P. & Volterra, A. A neuron-glia signalling network in the active brain. Curr Opin Neurobiol 11, 387–394 (2001).
Slezak, M. & Pfrieger, F. W. New roles for astrocytes: regulation of CNS synaptogenesis. Trends Neurosci 26, 531–535, https://doi.org/10.1016/j.tins.2003.08.005 (2003).
Dallerac, G., Zapata, J. & Rouach, N. Versatile control of synaptic circuits by astrocytes: where, when and how? Nat Rev Neurosci 19, 729–743, https://doi.org/10.1038/s41583-018-0080-6 (2018).
Faissner, A. et al. Contributions of astrocytes to synapse formation and maturation - Potential functions of the perisynaptic extracellular matrix. Brain Res Rev 63, 26–38, https://doi.org/10.1016/j.brainresrev.2010.01.001 (2010).
Pyka, M. et al. Chondroitin sulfate proteoglycans regulate astrocyte-dependent synaptogenesis and modulate synaptic activity in primary embryonic hippocampal neurons. Eur J Neurosci 33, 2187–2202, https://doi.org/10.1111/j.1460-9568.2011.07690.x (2011).
Evers, M. R. et al. Impairment of L-type Ca2+ Channel-Dependent Forms of Hippocampal Synaptic Plasticity in Mice Deficient in the Extracellular Matrix Glycoprotein Tenascin-C. The Journal of Neuroscience 22, 7177–7194 20026735 (2002).
Dityatev, A. & Schachner, M. Extracellular matrix molecules and synaptic plasticity. Nat Rev Neurosci 4, 456–468, https://doi.org/10.1038/nrn1115 (2003).
Dityatev, A., Schachner, M. & Sonderegger, P. The dual role of the extracellular matrix in synaptic plasticity and homeostasis. Nat Rev Neurosci 11, 735–746, https://doi.org/10.1038/nrn2898 (2010).
Dzyubenko, E., Gottschling, C. & Faissner, A. Neuron-Glia Interactions in Neural Plasticity: Contributions of Neural Extracellular Matrix and Perineuronal Nets. Neural Plast 2016, 5214961, https://doi.org/10.1155/2016/5214961 (2016).
Celio, M. R., Spreafico, R., De Biasi, S. & Vitellaro-Zuccarello, L. Perineuronal nets: past and present. Trends Neurosci 21, 510–515 (1998).
Bruckner, G. et al. Cortical areas are revealed by distribution patterns of proteoglycan components and parvalbumin in the Mongolian gerbil and rat. Brain Res 658, 67–86 (1994).
Hartig, W., Brauer, K. & Bruckner, G. Wisteria floribunda agglutinin-labelled nets surround parvalbumin-containing neurons. Neuroreport 3, 869–872 (1992).
Dityatev, A. et al. Activity-dependent formation and functions of chondroitin sulfate-rich extracellular matrix of perineuronal nets. Dev Neurobiol 67, 570–588, https://doi.org/10.1002/dneu.20361 (2007).
Hausen, D. et al. Pyramidal cells ensheathed by perineuronal nets in human motor and somatosensory cortex. Neuroreport 7, 1725–1729 (1996).
Wegner, F. et al. Diffuse perineuronal nets and modified pyramidal cells immunoreactive for glutamate and the GABA(A) receptor alpha1 subunit form a unique entity in rat cerebral cortex. Exp Neurol 184, 705–714, https://doi.org/10.1016/S0014-4886(03)00313-3 (2003).
Deepa, S. S. et al. Composition of perineuronal net extracellular matrix in rat brain: a different disaccharide composition for the net-associated proteoglycans. J Biol Chem 281, 17789–17800, https://doi.org/10.1074/jbc.M600544200 (2006).
Galtrey, C. M. & Fawcett, J. W. The role of chondroitin sulfate proteoglycans in regeneration and plasticity in the central nervous system. Brain Res Rev 54, 1–18, https://doi.org/10.1016/j.brainresrev.2006.09.006 (2007).
Yasuhara, O., Akiyama, H., McGeer, E. G. & McGeer, P. L. Immunohistochemical localization of hyaluronic acid in rat and human brain. Brain Res 635, 269–282 (1994).
Giamanco, K. A. & Matthews, R. T. Deconstructing the perineuronal net: cellular contributions and molecular composition of the neuronal extracellular matrix. Neuroscience 218, 367–384, https://doi.org/10.1016/j.neuroscience.2012.05.055 (2012).
Morawski, M. et al. Tenascin-R promotes assembly of the extracellular matrix of perineuronal nets via clustering of aggrecan. Philos Trans R Soc Lond B Biol Sci 369, 20140046, https://doi.org/10.1098/rstb.2014.0046 (2014).
Sorg, B. A. et al. Casting a Wide Net: Role of Perineuronal Nets in Neural Plasticity. J Neurosci 36, 11459–11468, https://doi.org/10.1523/JNEUROSCI.2351-16.2016 (2016).
Kwok, J. C., Carulli, D. & Fawcett, J. W. In vitro modeling of perineuronal nets: hyaluronan synthase and link protein are necessary for their formation and integrity. J Neurochem 114, 1447–1459, https://doi.org/10.1111/j.1471-4159.2010.06878.x (2010).
Carulli, D. et al. Composition of perineuronal nets in the adult rat cerebellum and the cellular origin of their components. J Comp Neurol 494, 559–577, https://doi.org/10.1002/cne.20822 (2006).
Bukalo, O., Schachner, M. & Dityatev, A. Modification of extracellular matrix by enzymatic removal of chondroitin sulfate and by lack of tenascin-R differentially affects several forms of synaptic plasticity in the hippocampus. Neuroscience 104, 359–369 (2001).
Brakebusch, C. et al. Brevican-Deficient Mice Display Impaired Hippocampal CA1 Long-Term Potentiation but Show No Obvious Deficits in Learning and Memory. Molecular and Cellular Biology 22, 7417–7427, https://doi.org/10.1128/mcb.22.21.7417-7427.2002 (2002).
Zhou, X. H. et al. Neurocan Is Dispensable for Brain Development. Molecular and Cellular Biology 21, 5970–5978, https://doi.org/10.1128/mcb.21.17.5970-5978.2001 (2001).
Rauch, U., Zhou, X.-H. & Roos, G. Extracellular matrix alterations in brains lacking four of its components. Biochemical and Biophysical Research Communications 328, 608–617, https://doi.org/10.1016/j.bbrc.2005.01.026 (2005).
Geissler, M. et al. A. Primary hippocampal neurons, which lack four crucial extracellular matrix molecules, display abnormalities of synaptic structure and function and severe deficits in perineuronal net formation. The Journal of Neuroscience 33, 7742–7755, https://doi.org/10.1523/JNEUROSCI.3275-12.2013 (2013).
Jansen, S., Gottschling, C., Faissner, A. & Manahan-Vaughan, D. Intrinsic cellular and molecular properties of in vivo hippocampal synaptic plasticity are altered in the absence of key synaptic matrix molecules. Hippocampus 27, 920–933, https://doi.org/10.1002/hipo.22742 (2017).
Gottschling, C., Dzyubenko, E., Geissler, M. & Faissner, A. The Indirect Neuron-astrocyte Coculture Assay: An In Vitro Set-up for the Detailed Investigation of Neuron-glia Interactions. J Vis Exp, https://doi.org/10.3791/54757 (2016).
Geissler, M. & Faissner, A. A new indirect co-culture set up of mouse hippocampal neurons and cortical astrocytes on microelectrode arrays. J Neurosci Methods 204, 262–272, https://doi.org/10.1016/j.jneumeth.2011.11.030 (2012).
Carstens, K. E., Phillips, M. L., Pozzo-Miller, L., Weinberg, R. J. & Dudek, S. M. Perineuronal Nets Suppress Plasticity of Excitatory Synapses on CA2 Pyramidal Neurons. J Neurosci 36, 6312–6320, https://doi.org/10.1523/JNEUROSCI.0245-16.2016 (2016).
Pyka, M., Busse, C., Seidenbecher, C., Gundelfinger, E. D. & Faissner, A. Astrocytes are crucial for survival and maturation of embryonic hippocampal neurons in a neuron-glia cell-insert coculture assay. Synapse 65, 41–53, https://doi.org/10.1002/syn.20816 (2011).
Rowlands, D. et al. Aggrecan Directs Extracellular Matrix-Mediated Neuronal Plasticity. J Neurosci 38, 10102–10113, https://doi.org/10.1523/JNEUROSCI.1122-18.2018 (2018).
Dobie, F. A. & Craig, A. M. Inhibitory synapse dynamics: coordinated presynaptic and postsynaptic mobility and the major contribution of recycled vesicles to new synapse formation. J Neurosci 31, 10481–10493, https://doi.org/10.1523/JNEUROSCI.6023-10.2011 (2011).
Lensjo, K. K., Christensen, A. C., Tennoe, S., Fyhn, M. & Hafting, T. Differential Expression and Cell-Type Specificity of Perineuronal Nets in Hippocampus, Medial Entorhinal Cortex, and Visual Cortex Examined in the Rat and Mouse. eNeuro 4, https://doi.org/10.1523/ENEURO.0379-16.2017 (2017).
Hayani, H., Song, I. & Dityatev, A. Increased Excitability and Reduced Excitatory Synaptic Input Into Fast-Spiking CA2 Interneurons After Enzymatic Attenuation of Extracellular Matrix. Front Cell Neurosci 12, 149, https://doi.org/10.3389/fncel.2018.00149 (2018).
Feinstein, Y. et al. F-spondin and mindin: two structurally and functionally related genes expressed in the hippocampus that promote outgrowth of embryonic hippocampal neurons. Development 126, 3637–3648 (1999).
Feinstein, Y. & Klar, A. The neuronal class 2 TSR proteins F-spondin and Mindin: a small family with divergent biological activities. Int J Biochem Cell Biol 36, 975–980, https://doi.org/10.1016/j.biocel.2004.01.002 (2004).
Kwok, J. C., Dick, G., Wang, D. & Fawcett, J. W. Extracellular matrix and perineuronal nets in CNS repair. Dev Neurobiol 71, 1073–1089, https://doi.org/10.1002/dneu.20974 (2011).
McKeon, R. J., Schreiber, R. C., Rudge, J. S. & Silver, J. Reduction of neurite outgrowth in a model of glial scarring following CNS injury is correlated with the expression of inhibitory molecules on reactive astrocytes. J Neurosci 11, 3398–3411 (1991).
Karetko, M. & Skangiel-Kramska, J. Diverse functions of perineuronal nets. Acta Neurobiol Exp (Wars) 69, 564–577 (2009).
Prabhakar, V. et al. Biochemical characterization of the chondroitinase ABC I active site. Biochemical Journal 390, 395–405, https://doi.org/10.1042/bj20050532 (2005).
Zimmermann, D. R. & Dours-Zimmermann, M. T. Extracellular matrix of the central nervous system: from neglect to challenge. Histochem Cell Biol 130, 635–653, https://doi.org/10.1007/s00418-008-0485-9 (2008).
Bruckner, G. et al. Postnatal development of perineuronal nets in wild-type mice and in a mutant deficient in tenascin-R. J Comp Neurol 428, 616–629 (2000).
Saghatelyan, A. K. et al. Reduced perisomatic inhibition, increased excitatory transmission, and impaired long-term potentiation in mice deficient for the extracellular matrix glycoprotein tenascin-R. Mol Cell Neurosci 17, 226–240, https://doi.org/10.1006/mcne.2000.0922 (2001).
Tucker, R. P. et al. Phylogenetic analysis of the tenascin gene family: evidence of origin early in the chordate lineage. BMC Evol Biol 6, 60, https://doi.org/10.1186/1471-2148-6-60 (2006).
Adams, J. C., Chiquet-Ehrismann, R. & Tucker, R. P. The evolution of tenascins and fibronectin. Cell Adhes Migr 9, 22–33, https://doi.org/10.4161/19336918.2014.970030 (2015).
Bohlega, S., Al-Ajlan, H. & Al-Saif, A. Mutation of fibulin-1 causes a novel syndrome involving the central nervous system and connective tissues. Eur J Hum Genet 22, 640–643, https://doi.org/10.1038/ejhg.2013.210 (2014).
Radice, P. D. et al. Fibulin-2 is a key mediator of the pro-neurogenic effect of TGF-beta1 on adult neural stem cells. Mol Cell Neurosci 67, 75–83, https://doi.org/10.1016/j.mcn.2015.06.004 (2015).
Irintchev, A., Rollenhagen, A., Troncoso, E., Kiss, J. Z. & Schachner, M. Structural and functional aberrations in the cerebral cortex of tenascin-C deficient mice. Cereb Cortex 15, 950–962, https://doi.org/10.1093/cercor/bhh195 (2005).
Lundell, A. et al. Structural basis for interactions between tenascins and lectican C-type lectin domains: evidence for a crosslinking role for tenascins. Structure 12, 1495–1506, https://doi.org/10.1016/j.str.2004.05.021 (2004).
Hagihara, K. et al. Immunohistochemical evidence for the brevican-tenascin-R interaction: colocalization in perineuronal nets suggests a physiological role for the interaction in the adult rat brain. J Comp Neurol 410, 256–264 (1999).
Christopherson, K. S. et al. Thrombospondins are astrocyte-secreted proteins that promote CNS synaptogenesis. Cell 120, 421–433, https://doi.org/10.1016/j.cell.2004.12.020 (2005).
Kucukdereli, H. et al. Control of excitatory CNS synaptogenesis by astrocyte-secreted proteins Hevin and SPARC. Proc Natl Acad Sci USA 108, E440–449, https://doi.org/10.1073/pnas.1104977108 (2011).
Allen, N. J. et al. Astrocyte glypicans 4 and 6 promote formation of excitatory synapses via GluA1 AMPA receptors. Nature 486, 410–414, https://doi.org/10.1038/nature11059 (2012).
Vedunova, M. et al. Seizure-like activity in hyaluronidase-treated dissociated hippocampal cultures. Front Cell Neurosci 7, 149, https://doi.org/10.3389/fncel.2013.00149 (2013).
Beurdeley, M. et al. Otx2 binding to perineuronal nets persistently regulates plasticity in the mature visual cortex. J Neurosci 32, 9429–9437, https://doi.org/10.1523/JNEUROSCI.0394-12.2012 (2012).
Spatazza, J. et al. Choroid-plexus-derived Otx2 homeoprotein constrains adult cortical plasticity. Cell Rep 3, 1815–1823, https://doi.org/10.1016/j.celrep.2013.05.014 (2013).
Bernard, C. et al. Graded Otx2 activities demonstrate dose-sensitive eye and retina phenotypes. Hum Mol Genet 23, 1742–1753, https://doi.org/10.1093/hmg/ddt562 (2014).
Soleman, S., Filippov, M. A., Dityatev, A. & Fawcett, J. W. Targeting the neural extracellular matrix in neurological disorders. Neuroscience 253, 194–213, https://doi.org/10.1016/j.neuroscience.2013.08.050 (2013).
Inagaki, S. et al. Sema4c, a transmembrane semaphorin, interacts with a post-synaptic density protein, PSD-95. J Biol Chem 276, 9174–9181, https://doi.org/10.1074/jbc.M009051200 (2001).
Xue, M. et al. Complexins facilitate neurotransmitter release at excitatory and inhibitory synapses in mammalian central nervous system. Proc Natl Acad Sci USA 105, 7875–7880, https://doi.org/10.1073/pnas.0803012105 (2008).
von Engelhardt, J. et al. GluN2D-containing NMDA receptors-mediate synaptic currents in hippocampal interneurons and pyramidal cells in juvenile mice. Front Cell Neurosci 9, 95, https://doi.org/10.3389/fncel.2015.00095 (2015).
Perszyk, R. E. et al. GluN2D-Containing N-methyl-d-Aspartate Receptors Mediate Synaptic Transmission in Hippocampal Interneurons and Regulate Interneuron Activity. Mol Pharmacol 90, 689–702, https://doi.org/10.1124/mol.116.105130 (2016).
Sohal, V. S. & Rubenstein, J. L. R. Excitation-inhibition balance as a framework for investigating mechanisms in neuropsychiatric disorders. Mol Psychiatry, https://doi.org/10.1038/s41380-019-0426-0 (2019).
Lopatina, O. L. et al. Excitation/inhibition imbalance and impaired neurogenesis in neurodevelopmental and neurodegenerative disorders. Rev Neurosci, https://doi.org/10.1515/revneuro-2019-0014 (2019).
Howell, B. W. & Smith, K. M. Synaptic structural protein dysfunction leads to altered excitation inhibition ratios in models of autism spectrum disorder. Pharmacol Res 139, 207–214, https://doi.org/10.1016/j.phrs.2018.11.019 (2019).
Fawcett, J. W., Oohashi, T. & Pizzorusso, T. The roles of perineuronal nets and the perinodal extracellular matrix in neuronal function. Nat Rev Neurosci 20, 451–465, https://doi.org/10.1038/s41583-019-0196-3 (2019).
Favuzzi, E. et al. Activity-Dependent Gating of Parvalbumin Interneuron Function by the Perineuronal Net Protein Brevican. Neuron 95, 639–655 e610, https://doi.org/10.1016/j.neuron.2017.06.028 (2017).
Sullivan, C. S. et al. Perineuronal Net Protein Neurocan Inhibits NCAM/EphA3 Repellent Signaling in GABAergic Interneurons. Sci Rep 8, 6143, https://doi.org/10.1038/s41598-018-24272-8 (2018).
Eroglu, C. et al. Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. Cell 139, 380–392, https://doi.org/10.1016/j.cell.2009.09.025 (2009).
Fossati, G. et al. Pentraxin 3 regulates synaptic function by inducing AMPA receptor clustering via ECM remodeling and beta1-integrin. EMBO J 38, https://doi.org/10.15252/embj.201899529 (2019).
Hales, C. M., Rolston, J. D. & Potter, S. M. How to culture, record and stimulate neuronal networks on micro-electrode arrays (MEAs). J Vis Exp, https://doi.org/10.3791/2056 (2010).
Irizarry, R. A. et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249–264, https://doi.org/10.1093/biostatistics/4.2.249 (2003).
Emig, D. et al. AltAnalyze and DomainGraph: analyzing and visualizing exon expression data. Nucleic Acids Res 38, W755–762, https://doi.org/10.1093/nar/gkq405 (2010).
McCloy, R. A. et al. Partial inhibition of Cdk1 in G 2 phase overrides the SAC and decouples mitotic events. Cell Cycle 13, 1400–1412, https://doi.org/10.4161/cc.28401 (2014).
Acknowledgements
The authors thank Marion Voelzkow for preparing cryosections of postnatal mouse brains and M.Sc. Michael Kostka for providing high magnification images of immunohistochemical WFA and aggrecan labelling of the hippocampus. The authors are grateful for grant support (German Research Foundation (DFG), SPP 1172/3 FA 159/13-3, and FA 159/22-1 to AF). We acknowledge support by the DFG Open Access Publication Funds of the Ruhr-Universität Bochum.
Author information
Authors and Affiliations
Contributions
C.G. designed the study, performed the experiments and quantified the data. A.F. planned and designed the study, interpreted the data and provided administrational and grant support. B.D. carried out and evaluated the gene array analysis and contributed the interpretation of the results. C.G., D.W. and A.F. wrote and revised the paper.
Corresponding author
Ethics declarations
Competing Interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Gottschling, C., Wegrzyn, D., Denecke, B. et al. Elimination of the four extracellular matrix molecules tenascin-C, tenascin-R, brevican and neurocan alters the ratio of excitatory and inhibitory synapses. Sci Rep 9, 13939 (2019). https://doi.org/10.1038/s41598-019-50404-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-019-50404-9
- Springer Nature Limited
This article is cited by
-
The circulating proteome and brain health: Mendelian randomisation and cross-sectional analyses
Translational Psychiatry (2024)
-
A disinhibitory microcircuit of the orbitofrontal cortex mediates cocaine preference in mice
Molecular Psychiatry (2024)
-
The extracellular matrix and perineuronal nets in memory
Molecular Psychiatry (2022)
-
Perineuronal nets affect memory and learning after synapse withdrawal
Translational Psychiatry (2022)
-
Extracellular Matrix in Neural Plasticity and Regeneration
Cellular and Molecular Neurobiology (2022)