Sculpting neurotransmission during synaptic development by 2D nanostructured interfaces

0
12


Neurons are continuously exposed to signals generated by the extracellular environment, including genuine physical cues (such as mechanical or topographical ones) at the nanoscale, able to drive key biological tasks.1x1Lutolf, M.P. and Hubbell, J.A. Synthetic biomaterials as instructive extracellular microenvironments for morphogenesis in tissue engineering. Nat Biotechnol. 2005;
23: 47–55
Crossref | PubMed | Scopus (2741)
| Google ScholarSee all References

The majority of current studies on biological membrane stability in response to nanomaterials are focused on the influence of materials’ functionalization or shape/size on cell uptake mechanisms and internalization, to engineer sophisticated drug delivery (nano)-vectors.19x19Lesniak, A., Salvati, A., Santos-Martinez, M.J., Radomski, M.W., Dawson, K.A., and Åberg, C. Nanoparticle adhesion to the cell membrane and its effect on nanoparticle uptake efficiency. J Am Chem Soc. 2013;
135: 1438–1444
Crossref | PubMed | Scopus (248)
| Google ScholarSee all References

Here, by single cell electrophysiology and immunofluorescence microscopy we monitor the dynamics of glutamate receptor-mediated excitatory transmission in cultured hippocampal neurons interfaced to MWCNTs. We specifically addressed whether MWCNTs, once interfaced to neurons, affected synaptic transmission by modulating lipid membrane structure and dynamics. We focused in particular on cholesterol, a largely represented lipid in neuronal membranes known to regulate presynaptic vesicle release.28x28Wasser, C.R., Ertunc, M., Liu, X., and Kavalali, E.T. Cholesterol-dependent balance between evoked and spontaneous synaptic vesicle recycling. J Physiol. 2007;
579: 413–429
Crossref | PubMed | Scopus (87)
| Google ScholarSee all References
,29x29Ramirez, D.M. and Kavalali, E.T. Differential regulation of spontaneous and evoked neurotransmitter release at central synapses. Curr Opin Neurobiol. 2011;
21: 275–282
Crossref | PubMed | Scopus (89)
| Google ScholarSee all References
For a first general assessment we used artificial lipid membranes that, when interfaced to MWCNTs, were more stable, in respect to controls, to a cholesterol depleting-agent, i.e. cyclodextrin.30x30Zidovetzki, R. and Levitan, I. Use of cyclodextrins to manipulate plasma membrane cholesterol content: evidence, misconceptions and control strategies. Biochim Biophys Acta. 2007;
1768: 1311–1324
Crossref | PubMed | Scopus (499)
| Google ScholarSee all References

Methods

Synthesis of MWCNTs

MWCNTs 20-30 nm in diameter (Nanostructured & Amorphous Materials, Inc.) were used as received and substrates were prepared as described previously.12x12Lovat, V., Pantarotto, D., Lagostena, L., Cacciari, B., Grandolfo, M., Righi, M. et al. Carbon nanotube substrates boost neuronal electrical signaling. Nano Lett. 2005;
5: 1107–1110
Crossref | PubMed | Scopus (417)
| Google ScholarSee all References
,15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References
Briefly, MWCNT 20-30 nm were functionalized using 1,3-dipolar cycloaddition with heptanal and sarcosine at 130 °C for 120 h in dimethylformammide (DMF) as solvent. For deposition on the coverslips, the DMF solution of functionalized MWCNTs (0.01 mg/mL) was drop casted to uniformly layer the entire substrate and let evaporated at 80 °C, then, the substrates were heated up at 350 °C under N2 atmosphere to induce the complete re-pristinization of MWCNTs. The uniformity of the deposition was assumed by AFM (Figure 1, B) and by scanning electron microscopy (SEM, in Supplementary Figure S1, D).

Figure 1 Opens large image

Figure 1

AFM investigation of artificial lipid membranes (SLM). (A) Low magnification topographic reconstruction of an incomplete SLM deposited on a control glass surface (on the top). The height profile corresponding to the highlighted line is shown, revealing SLM height of about 5.0 ± 0.2 nm. Higher resolution AFM reconstruction (on the bottom) demonstrates the high uniformity of the so obtained SLB. B. Low magnification topographic reconstruction, corresponding height profile and higher resolution reconstruction of a MWCNTs carpet deposited on glass via drop-casting. Note the high corrugation of the resulting surface pointed out by the top image and single MWCNTs composing the carpet easily distinguishable in the bottom one. (C) Low magnification topographic reconstruction, corresponding height profile and higher resolution reconstruction of SLBs deposited on a MWCNTs substrate. It is possible to appreciate (white arrows) MWCNT ability to pierce SLMs, indicated by nanotubes emerging from the upper membrane layer. The altimetric profile reveals flat parts, corresponding to superficial SLBs, from which only MWCNTs apexes protrude. (D) Fluorescent decay of SLMs signal after injection of MβCD (500 μM) detected in control (squares) and on MWCNT (circles) membranes. Note that the decay follows a double exponential law characterized by τ values of 1.09 ± 0.02 s and 39.02 ± 9.83 s (gray fitting line) for SLM deposited on glass and 2.02 ± 0.15 s and 63.67 ± 6.92 s for SLM deposited on the MWCNT carpet (black fitting line). Data are averages of 3 independent experiments expressed as mean ± SD. Initial values were normalized before injection of MβCD. E. Raman spectra acquired in the highlighted areas (colored dots) in the bottom panels of A, B and C (glass substrate in yellow, SLM on glass in orange, MWCNTs on glass in green and SLM above MWCNTs in blue). Note that the spectra of MWCNTs with (blue spectrum) and without (green spectrum) SLMs were vertically shifted for illustrative purpose. The reference spectrum of lipids (red spectrum) was acquired on a many-layer membranes sample (MLMs, not shown) in order to minimize Raman signal from the underlying glass surface.

Artificial membrane preparation and characterization

Artificial membranes were prepared by lipid spreading on a supporting glass slide from an organic solvent solution. 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) and cholesterol molecules (both from Avanti Polar Lipids Inc., US) were dissolved in a 2:1 ratio in chloroform (Sigma Aldrich) at a final concentration of 100 μM. 100 μL of solution were deposited on a glass coverslip, used as control, and on MWCNTs substrates supported by the same glass coverslip (Figure 1).15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References

Raman characterization

The Raman measurements were performed in the reflection geometry. A 532 nm continuous-wave laser (Cobolt Samba, 50 mW, bandwidth 1 MHz) was used as excitation source. The beam was focused on the sample by a 100× air objective (NA 0.8, EC EpiPlan, Zeiss) resulting in a diameter of laser spot of about 0.5 μm. A 532 nm RazorEdge Dichroic™ laser-flat beam-splitter and a 532 nm RazorEdge® ultra-steep long-pass edge filter were used to direct the light into microscope and cut Rayleigh scattered light, respectively. The laser power on the sample was controlled by the neutral density filter (Thorlabs) and kept at 100 μW. The acquisition time in all experiments was 60 s. All Raman measurements and analysis were performed by CNR-IOM (TASC Laboratory, Basovizza, Trieste, Italy).

Primary cultures and cell treatment

Hippocampal neurons were obtained from neonatal rats as previously reported.15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References

Electrophysiology

Patch-clamp, whole cell recordings were obtained with glass micropipettes with a resistance of 4 to 8 MΩ. The intracellular pipette solution was the following (mM): 120 K gluconate, 20 KCl, 10 HEPES, 10 EGTA, 2 MgCl2, 2 Na2ATP, pH 7.3. The external standard saline solution contained (mM): 150 NaCl, 4 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, 10 glucose, pH 7.4. All recordings were performed at RT. Cells were voltage clamped at a holding potential set at −56 mV (not corrected for the liquid junction potential, calculated to be 13.7 mV at 20 °C). The uncompensated series resistance had values <8 MΩ. Miniature post-synaptic currents (mPSCs) were recorded in the presence of 1 μM fast-Na+ channel blocker Tetrodotoxin (TTX; Latoxan). In order to block voltage-gated calcium channels we added 3 mM CoCl2 to the external solution. Data were collected using a Multiclamp 700A Amplifier (Molecular Devices, US), and analyzed using either Clampfit 10.3 (Molecular Devices) or Axograph (Axograph Scientific). Glutamate AMPA-receptor and GABAA-receptor mediated post synaptic currents (PSCs) were isolated offline by building two templates with different kinetic parameters: respectively 0.1 ms rise-time; 3 and 30 ms decay time constant (τ); 10 and 100 ms template length. Previous work12x12Lovat, V., Pantarotto, D., Lagostena, L., Cacciari, B., Grandolfo, M., Righi, M. et al. Carbon nanotube substrates boost neuronal electrical signaling. Nano Lett. 2005;
5: 1107–1110
Crossref | PubMed | Scopus (417)
| Google ScholarSee all References
,15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References
indicated that in our experimental conditions, the vast majority of fast-decaying (τ < 5 ms) PSCs are mediated by the glutamate AMPA-receptor type; while the slow-decaying (τ > 20 ms) PSCs are mediated by the GABAA-receptor type.

Immunocytochemistry

Cultures were fixed in 4% formaldehyde (prepared from fresh paraformaldehyde) in PBS for 20 min, permeabilized with 0.3% Triton-X-100 and subsequently incubated with primary antibodies for 30 min at RT and after washing in PBS incubated with secondary antibodies for 45 min. Cultures were then mounted with the Vectashield (Vector Laboratories) on 1 mm thick microscope glass slides. To visualize neurons and lipid rafts we used the following: rabbit anti-β-tubulin III primary antibody (Sigma T2200, 1:250 dilution) and Alexa 594 goat anti rabbit secondary antibody (Invitrogen, 1:500); Alexa 488 Cholera Toxin subunit-B (CT-B) 1:200 (Molecular Probes) and DAPI, 1:1000 (Invitrogen).

To visualize glutamatergic synapses we co-label neurons with the guinea pig anti-vesicular glutamate transporter 1 (VGLUT1; Millipore, 1:2000) and β-tubulin III (Sigma, 1:250) primary antibodies and Alexa 594 goat anti rabbit (Invitrogen, 1:500) and Alexa 488 goat anti guinea-pig (Invitrogen, 1:500) as secondary antibodies. All images were acquired using an inverted confocal Microscope (Leica Microsystems GmbH, Wetzlar, Germany; 40× oil immersion objective, 1.25 NA).

To quantify VGLUT1 puncta and lipid rafts, n = 20 ± 10 z-stacks (acquired every 0.4 μm) were taken from n = 10 randomly selected fields (240 μm × 240 μm) per coverslip (n = 30, 3 culture series in Control and MWCNTs). To quantify lipid rafts, we selected the CT-B positive objects (<5 μm3) co-localized to the β-tubulin III positive areas. For each image, the volumes of CT-B positive objects were normalized to the β-tubulin III positive volumes. To quantify VGLUT1 puncta, we selected only VGLUT1-positive puncta (<2 μm3) co-localized to the β-tubulin III positive areas and puncta were normalized to the β-tubulin III positive volumes. Images were analyzed using the Volocity software (Perkin Elmer).

For the filipin labeling of membrane cholesterols34x34Hissa, B., Duarte, J.G., Kelles, L.F., Santos, F.P., del Puerto, H.L., Gazzinelli-Guimaraes, P.H. et al. Membrane cholesterol regulates lysosome-plasma membrane fusion events and modulates Trypanosoma cruzi invasion of host cells. PLoS Negl Trop Dis. 2012;
6: e1583
Crossref | PubMed | Scopus (15)
| Google ScholarSee all References

Data analysis

Ethical statement

Results

Artificial lipid membranes interfaced to MWCNTs

In the first set of experiments, we investigated by AFM the appearance of artificial lipid membranes (SLM)35x35Richter, R.P. and Brisson, A. Characterization of lipid bilayers and protein assemblies supported on rough surfaces by atomic force microscopy. Langmuir. 2003;
19: 1632–1640
Crossref | Scopus (87)
| Google ScholarSee all References

In Figure 1, E the reference Raman spectrum of lipids (in red) is plotted and is characterized by peaks associated with C–N stretch (715 cm−1), C–C stretch (1090 cm−1), CH2 deformation (1305 cm−1 and 1440 cm−1), and C = C stretch (1655 cm−1) vibrations.36x36Stone, N., Kendall, C., Smith, J., Crow, P., and Barr, H. Raman spectroscopy for identification of epithelial cancers. Faraday Discuss. 2004;
126: 141–157
Crossref | PubMed | Scopus (311)
| Google ScholarSee all References
,37x37Bergholt, M.S., Zheng, W., Ho, K.Y., Teh, M., Yeoh, K.G., So, J.B.Y. et al. Fiber-optic Raman spectroscopy probes gastric carcinogenesis in vivo at endoscopy. J Biophotonics. 2013;
6: 49–59
Crossref | PubMed | Scopus (43)
| Google ScholarSee all References
The Raman spectrum of MWCNTs alone on glass substrate (Figure 1, E, in green) shows two broad peaks centered at 1350 cm−1 and 1590 cm−1 that are commonly assigned to the presence of disorder in graphitic materials and to the tangential vibrations of the carbon atoms, respectively.38x38Osswald, S., Havel, M., and Gogotsi, Y. Monitoring oxidation of multiwalled carbon nanotubes by Raman spectroscopy. J Raman Spectrosc. 2007;
38: 728–736
Crossref | Scopus (350)
| Google ScholarSee all References
,39x39Lehman, J.H., Terrones, M., Mansfield, E., Hurst, K.E., and Meunier, V. Evaluating the characteristics of multiwall carbon nanotubes. Carbon. 2011;
49: 2581–2602
Crossref | Scopus (390)
| Google ScholarSee all References
The distinguishable shoulder at 1620 cm−1 is a double-resonance Raman feature induced by disorder, defects or ion intercalation between the graphitic walls.38x38Osswald, S., Havel, M., and Gogotsi, Y. Monitoring oxidation of multiwalled carbon nanotubes by Raman spectroscopy. J Raman Spectrosc. 2007;
38: 728–736
Crossref | Scopus (350)
| Google ScholarSee all References

Next, we incorporated a fluorescent lipid to SLMs (see Methods) and we compared the efficacy of MβCD treatment (500 μM; 1 h; used to bind and extract cholesterol from the membrane32x32Christian, A.E., Haynes, M.P., Phillips, M.C., and Rothblat, G.H. Use of cyclodextrins for manipulating cellular cholesterol content. J Lipid Res. 1997;
38: 2264–2272
PubMed
| Google ScholarSee all References
,33x33Steck, T.L., Ye, J., and Lange, Y. Probing red cell membrane cholesterol movement with cyclodextrin. Biophys J. 2002;
83: 2118–2125
Abstract | Full Text | Full Text PDF | PubMed | Scopus (158)
| Google ScholarSee all References
) in depleting cholesterol in control and MWCNT SLMs. The plot in Figure 1, D summarizes the decay in fluorescence monitored during MβCD application. In order to have comparable measurements, we sampled single SLM layers in both groups. In MWCNTs, SLM lipids appeared more stabilized, as indicated by their significantly slower depletion with respect to controls (decay time constant (τ) average values: τ1 = 1.09 ± 0.02 s and τ2 = 39.02 ± 9.83 s for control SLMs; τ1 = 12.02 ± 0.15 s and τ2 = 63.67 ± 6.92 s for SLM deposited on MWCNTs; P < 0.01 for τ1 values; P < 0.05 for τ2; n = 3 different samples each group).

Such results suggest that interfacing with MWCNTs might affect lipid, in particular cholesterol, dynamics in biological membranes.

Acute cholesterol removal in control and MWCNTs hippocampal cultures

To further explore the potential role of MWCNTs in membrane lipid dynamics when interfacing living cells, and in particular in altering cholesterol homeostasis in neurons, we cultured hippocampal cells on control substrates or on meshworks of MWCNTs (control and MWCNT, respectively). We monitored the synaptic networks before and after cell exposure to MβCD (1 mM; 1 h)30x30Zidovetzki, R. and Levitan, I. Use of cyclodextrins to manipulate plasma membrane cholesterol content: evidence, misconceptions and control strategies. Biochim Biophys Acta. 2007;
1768: 1311–1324
Crossref | PubMed | Scopus (499)
| Google ScholarSee all References

Figure 2 Opens large image

Figure 2

Cholesterol removal by MβCD application in hippocampal cultures. (A) Representative traces of spontaneous synaptic activity in control (left) and MWCNT (right) neurons before (top) and after (bottom) MβCD application. (B) Box plots summarize pooled data of PSC frequencies (top) and amplitudes (bottom) recorded from control and MWCNTs neurons prior and after MβCD. Note the higher PSC frequency displayed by MWCNT neurons in standard saline and the opposite effects brought about by MβCD in control and MWCNTs.

Exposure to MβCD did not affect neuronal viability and the overall integrity of membranes as estimated by comparing the input resistance values of the recorded neurons before and after treatment (for control: 655 ± 73 MΩ, n = 21, and 694 ± 167 MΩ, n = 23, P = 0.834; for MWCNT 487 ± 37 MΩ, n = 22, and 625 ± 90 MΩ, n = 23, P = 0.117; prior and after MβCD incubation, respectively) and the values for the cell capacitance (for control: 75 ± 32 pF, n = 21, and 80 ± 7 pF, n = 23, P = 0.536; for MWCNT 92 ± 8 pF, n = 22, and 74 ± 7 pF, n = 23,P = 0.120; prior and after MβCD incubation, respectively).

Synaptic cholesterol balances spontaneous and evoked neurotransmission by inhibiting spontaneous vesicle turnover and, conversely, by promoting evoked exo-endocytosis.28x28Wasser, C.R., Ertunc, M., Liu, X., and Kavalali, E.T. Cholesterol-dependent balance between evoked and spontaneous synaptic vesicle recycling. J Physiol. 2007;
579: 413–429
Crossref | PubMed | Scopus (87)
| Google ScholarSee all References

We measured synaptic activity in these two groups of cultures after cholesterol depletion and, surprisingly, addition of MβCD resulted in two macroscopic, but opposite, changes, illustrated by the sample recordings in Figure 2, A (bottom tracings). In control neurons MβCD treatment led to a significant (P < 0.01) increase (by 228%) in PSC frequency, without affecting PSC amplitude values; in MWCNTs, on the contrary, MβCD incubation induced a significant reduction in both PSC frequency (by 62%; P < 0.05) and PSC amplitude (by 55% P < 0.01; summarized in the box plots of Figure 2, B). Before analyzing in more details these synaptic changes (see below) we ascertain whether the tight interfacing14x14Cellot, G., Cilia, E., Cipollone, S., Rancic, V., Sucapane, A., Giordani, S. et al. Carbon nanotubes might improve neuronal performance by favouring electrical shortcuts. Nat Nanotechnol. 2009;
4: 126–133
Crossref | PubMed | Scopus (297)
| Google ScholarSee all References

Figure 3 Opens large image

Figure 3

MβCD efficiently removed membrane cholesterol without disrupting lipid rafts. (A) Fluorescence micrographs of control and MWCNT hippocampal cells labeled by filipin prior (top) and after (bottom) MβCD treatment. Scale bars 50 μm. (B) Bar plot of filipin-derived fluorescence intensity in cultured neurons, note the similar values between control and MWCNT conditions and the comparable reduction upon cholesterol depletion by MβCD. (C) Left, 3D and 2D confocal reconstructions of hippocampal cultures immune-labeled with the neuronal marker β-tubulin III (in red) and the lipid-raft marker CT-B (in green), in blue DAPI labeling for nuclei. Scale bars 15 μm. Right, bar plot quantifies the CT-B volume in respect to the β-tubulin III one, note that no differences were observed between controls and MWCNTs before and after the MβCD treatment.

Despite the similar effects with respect to cholesterol, MβCD affects in an opposite manner PSCs frequency in control and MWCNTs, suggesting the presence of presynaptic terminals where vesicle release is differentially regulated by cholesterol.44x44Moulder, K.L., Jiang, X., Taylor, A.A., Shin, W., Gillis, K.D., and Mennerick, S. Vesicle pool heterogeneity at hippocampal glutamate and GABA synapses. J Neurosci. 2007;
27: 9846–9854
Crossref | PubMed | Scopus (40)
| Google ScholarSee all References

Effects of acute cholesterol depletion on glutamatergic synapses

Spontaneous synaptic activity in our recording conditions15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References

Figure 4 Opens large image

Figure 4

Depletion of cholesterol with MβCD alters fast PSCs occurrence. (A) Offline differential analysis of PSC decays (τ) identifies fast and slow events (insets average tracings from the same neurons as in Figure 2, A) both in control and MWCNTs. Bar plot summarizes the frequency of fast and slow PSCs in controls and MWCNTs. (B) Spontaneous synaptic activity recorded in the presence of TTX; controls and MWCNT mPSC frequency and amplitudes are summarized in the box plots before and after MβCD treatment. Note that mPSC amplitude values are unaffected.

Cholesterol depletion in neurons is known to increase the rate of spontaneous transmission but it impairs evoked neurotransmission28x28Wasser, C.R., Ertunc, M., Liu, X., and Kavalali, E.T. Cholesterol-dependent balance between evoked and spontaneous synaptic vesicle recycling. J Physiol. 2007;
579: 413–429
Crossref | PubMed | Scopus (87)
| Google ScholarSee all References
,29x29Ramirez, D.M. and Kavalali, E.T. Differential regulation of spontaneous and evoked neurotransmitter release at central synapses. Curr Opin Neurobiol. 2011;
21: 275–282
Crossref | PubMed | Scopus (89)
| Google ScholarSee all References
and synapses usually segregate, at the presynaptic terminals, the distinct vesicle pools responsible for spontaneous or evoked release.29x29Ramirez, D.M. and Kavalali, E.T. Differential regulation of spontaneous and evoked neurotransmitter release at central synapses. Curr Opin Neurobiol. 2011;
21: 275–282
Crossref | PubMed | Scopus (89)
| Google ScholarSee all References
,46x46Melom, J.E., Akbergenova, Y., Gavornik, J.P., and Littleton, J.T. Spontaneous and evoked release are independently regulated at individual active zones. J Neurosci. 2013;
33: 17253–17263
Crossref | PubMed | Scopus (65)
| Google ScholarSee all References

We characterized the effects of cholesterol depletion on spontaneous release by recording miniature PSCs (mPSCs) in the presence of TTX (1 μM). mPSCs, in dissociated hippocampal cultures at this age, comprise virtually only fast events (τ = 4.5 ± 0.6 ms in control, n = 9, and τ = 5 ± 0.9 ms in MWCNT, n = 9; Figure 4, B, top tracings) thus representing only excitatory mPSCs.15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References

Figure 5 Opens large image

Figure 5

Co-localization of VGLUT1 and β-tubulin III immunostaining in control and MWCNT hippocampal cultures. (A) Right, confocal images of hippocampal neurons at 8-10 DIV stained for β-tubulin III (in red) and VGLUT1 (in green). In the insets high magnifications are shown corresponding to the white square areas. Scale bars 20 μm. Left, bar plot summarizes the quantification of VGLUT1 positive puncta, significantly higher in MWCNTs. (B) Right, summary graph of the PSCs frequency in standard saline solution and after TTX and Co++ application in controls (squares) and MWCNTs (circles). Right, bar plot summarizes the residual activity, in respect to standard saline, after application of TTX and Co++.

These results suggest that control and MWCNTs glutamatergic synapses express different release machineries. Namely, control glutamatergic synapses are dominated by spontaneous fusions, on the contrary MWCNT ones preferentially release neurotransmitter in response to action potentials.

To address the expression of heterogeneous populations of presynaptic vesicles, and the different partitioning of synaptic vesicles between the two pools, we estimate the residual activity when PSCs were recorded prior and after the application of TTX and Co++ (3 mM; n = 7 control and n = 7 MWCNTs). The block of voltage-gated Ca++-channels impairs the fusion of calcium-dependent vesicles,47x47Ermolyuk, Y.S., Alder, F.G., Surges, R., Pavlov, I.Y., Timofeeva, Y., Kullmann, D.M. et al. Differential triggering of spontaneous glutamate release by P/Q-, N- and R-type Ca2+ channels. Nat Neurosci. 2013;
16: 1754–1763
Crossref | PubMed | Scopus (50)
| Google ScholarSee all References

This different ratio suggests that synapses in control cultures rely for their basal activity mostly on low calcium-dependent vesicle release.25x25Baoukina, S., Monticelli, L., and Tieleman, D.P. Interaction of pristine and functionalized carbon nanotubes with lipid membranes. J Phys Chem B. 2013;
117: 12113–12123
Crossref | PubMed | Scopus (30)
| Google ScholarSee all References

The effects of acute cholesterol removal in control and MWCNT hippocampal cultures at later stages of synaptic development

Action potential evoked and spontaneous neurotransmitter release modalities involve different molecular machineries regulating segregated vesicle pools at the presynaptic site. Distinct forms of neurotransmission, involving these two modes of release, were reported to change during hippocampal synaptic development in vitro.49x49Andreae, L.C., Fredj, N.B., and Burrone, J. Independent vesicle pools underlie different modes of release during neuronal development. J Neurosci. 2012;
32: 1867–1874
Crossref | PubMed | Scopus (24)
| Google ScholarSee all References

Figure 6 Opens large image

Figure 6

Controls and MWCNTs spontaneous synaptic activity in long-term cultures. (A) Left, representative recordings of synaptic activity at 21 DIV in controls and MWCNTs before (top) and after (bottom) MβCD. Box plots (right) summarize the values of PSCs frequency (top) and amplitude (bottom); note the similar values measured in all conditions. (B) Confocal micrographs of cultures labeled by β-tubulin III (in red) and VGLUT1 (in green). In the insets high magnifications are shown corresponding to the white square areas. In long-term cultures controls and MWCNTs neurons display similar amount of VGLUT1 puncta, summarized in the bar plot (right). Scale bar, 20 μm.

At 18-21 DIV incubating with MβCD induced a slight and non-significant increase (Figure 6A) in the frequency of PSCs in both cultures groups, without altering PSCs amplitudes, suggesting that in vitro aging leads to an overall balance between synapses expressing different release modes.49x49Andreae, L.C., Fredj, N.B., and Burrone, J. Independent vesicle pools underlie different modes of release during neuronal development. J Neurosci. 2012;
32: 1867–1874
Crossref | PubMed | Scopus (24)
| Google ScholarSee all References

These results strengthen the hypothesis that MWCNTs boost the overall network maturation, in terms of number of synapses and efficacy,15x15Cellot, G., Toma, F.M., Varley, Z.K., Laishram, J., Villari, A., Quintana, M. et al. Carbon nanotube scaffolds tune synaptic strength in cultured neural circuits: novel frontiers in nanomaterial-tissue interactions. J Neurosci. 2011;
31: 12945–12953
Crossref | PubMed | Scopus (76)
| Google ScholarSee all References

Discussion

Carbon nanotubes are increasingly incorporated in the development of novel two-dimensional biomaterials designed to interface tissue reconstruction and signaling.16x16Fabbro, A., Prato, M., and Ballerini, L. Carbon nanotubes in neuroregeneration and repair. Adv Drug Deliv Rev. 2013;
65: 2034–2044
Crossref | PubMed | Scopus (50)
| Google ScholarSee all References
,52x52Harrison, B.S. and Atala, A. Carbon nanotube applications for tissue engineering. Biomaterials. 2007;
28: 344–353
Crossref | PubMed | Scopus (667)
| Google ScholarSee all References
In material science, MWCNTs are adopted in composites to strengthen biomaterial mechanical properties, electrical conductivity or microenvironment-defining moieties.53x53Bosi, S., Ballerini, L., and Prato, M. Carbon nanotubes in tissue engineering. Top Curr Chem. 2014;
348: 181–204
Crossref | PubMed | Scopus (11)
| Google ScholarSee all References

The principal finding of the present report is that two-dimensional MWCNT interfaces do not alter the homeostasis of membrane lipids, in particular cholesterol, in neurons. In fact, neurons cultured interfaced to MWCNTs, display a similar membrane cholesterol distribution and, when we used a traditional tool to remove membrane cholesterol, MβCD,26x26Lopez, C.F., Nielsen, S.O., Moore, P.B., and Klein, L.M. Understanding nature’s design for a nanosyringe. Proc Natl Acad Sci U S A. 2004;
101: 4431–4434
Crossref | PubMed | Scopus (225)
| Google ScholarSee all References

The present data show that, in both culture groups, the treatment used to remove cholesterol did not affect cell viability, as sustained by the values of cell input resistance and membrane capacitance.55x55Carp, J.S. Physiological properties of primate lumbar Motoneurons. J Neurophysiol. 1992;
68: 1121–1132
PubMed
| Google ScholarSee all References
,56x56Gao, Y., Liu, L., Li, Q., and Wang, Y. Differential alterations in the morphology and electrophysiology of layer II pyramidal cells in the primary visual cortex of a mouse model prenatally exposed to LPS. Neurosci Lett. 2015;
591: 138–143
Crossref | PubMed | Scopus (8)
| Google ScholarSee all References
In addition, membrane micro-domains enriched in cholesterol such as lipid rafts,57x57Allen, J.A., Halverson-Tamboli, R.A., and Rasenick, M.M. Lipid raft microdomains and neurotransmitter signalling. Nat Rev Neurosci. 2007;
8: 128–140
Crossref | PubMed | Scopus (497)
| Google ScholarSee all References

However, we found that at 8-10 DIV, release at glutamatergic synapses in control and MWCNTs was regulated in an opposite manner by cholesterol reduction.

To understand the reason for the observed difference, we examined the main variable that might conceivably affect release tuning by cholesterol. We had already excluded the possibility of differences in membrane cholesterol distribution and depletion. We thus turned our attention to pre-synaptic process that may be regulated by cholesterol. Removal of cholesterol variably affects spontaneous or evoked neurotransmitter release in cultured neurons, improving spontaneous vesicle fusion and decreasing evoked vesicle recycling.28x28Wasser, C.R., Ertunc, M., Liu, X., and Kavalali, E.T. Cholesterol-dependent balance between evoked and spontaneous synaptic vesicle recycling. J Physiol. 2007;
579: 413–429
Crossref | PubMed | Scopus (87)
| Google ScholarSee all References

We thus hypothesized that control synapses display a higher relative amount of spontaneous vesicle pools with respect to MWCNT ones.

In the presence of TTX we observed that in control conditions the frequency of mPSCs was still affected by MβCD, whereas no changes occurred in MWCNT neurons. Our hypothesis was further supported by the block of voltage dependent Ca++-channels by Co++ applications.59x59Mitterdorfer, J. and Bean, B.P. Potassium currents during the action potential of hippocampal CA3 neurons. J Neurosci. 2002;
22: 10106–10115
PubMed
| Google ScholarSee all References

Recent reports have shown that the maturation of neurotransmission is accompanied by changes in pre-synaptic release modes; in particular spontaneous vesicle pools are predominant on evoked ones during early stages of development, and these two populations are gradually rebalanced during the synaptic maturation process.49x49Andreae, L.C., Fredj, N.B., and Burrone, J. Independent vesicle pools underlie different modes of release during neuronal development. J Neurosci. 2012;
32: 1867–1874
Crossref | PubMed | Scopus (24)
| Google ScholarSee all References
,60x60Andreae, L.C. and Burrone, J. Spontaneous neurotransmitter release shapes dendritic arbors via long-range activation of NMDA receptors. Cell Rep. 2015;
10: 873–882
Abstract | Full Text | Full Text PDF | Scopus (10)
| Google ScholarSee all References

It is tempting to speculate that MWCNTs accelerate synaptic network maturation, improving synapse formation and favoring more mature release modes, an effect that is homeostatically regulated upon prolonged interfacing. In fact, in our experiments, control and MWCNTs displayed functional and anatomical similarities after three weeks of growth. We cannot distinguish whether this was due to a progressive shielding of the MWCNTs by extracellular matrix proteins,61x61Barros, C.S., Franco, S.J., and Müller, U. Extracellular matrix: functions in the nervous system. Cold Spring Harb Perspect Biol. 2011;
3: a005108
Crossref | Scopus (71)
| Google ScholarSee all References

In conclusion, the main finding of the present study is that MWCNTs when used in interfacing neurons can regulate synapse formation and function in a dynamic manner, tuning exquisite neurobiological mechanisms, such as neuronal maturation.62x62Fabbro, A., Sucapane, A., Toma, F.M., Calura, E., Rizzetto, L., Carrieri, C. et al. Adhesion to carbon nanotube conductive scaffolds forces action-potential appearance in immature rat spinal neurons. PLoS One. 2013;
8: e73621
Crossref | PubMed | Scopus (25)
| Google ScholarSee all References

Brain interfaces of the future require the application of nanomaterial-related technologies to target some of the current ambitions in interfacing neurons: improving the stability, flexibility and durability of the interface, improving the efficacy of the charge transfer to and from neurons, and minimizing the reactivity in the surrounding tissue.63x63Fattahi, P., Yang, G., Kim, G., and Abidian, M.R. A review of organic and inorganic biomaterials for neural interfaces. Adv Mater. 2014;
26: 1846–1885
Crossref | PubMed | Scopus (140)
| Google ScholarSee all References



Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here