Experience-Dependence of
Interneuron
Myelination
Scriptie
2016
J. Groeneveld Erasmus MC, Psychiatrie / Avans Hogeschool 8/8/2016i
Experience-Dependence of Interneuron Myelination
Scriptie Versienummer: 1.0 Datum: 1-8-2016 Student:
Jeroen Groeneveld j.groeneveld@student.avans.nl
Begeleiding:
Onderwijsinstelling:
Drs. Ing. M. Lindenbergh – van der Plas m.lindenbergh-vanderplas@avans.nl Stageverlener:
J. Stedehouder j.stedehouder@erasmusmc.nl
Opleiding:
Avans Hogeschool Breda
ATGM: Academie voor Technologie, Gezondheid en Milieu Forensisch Laboratorium Onderzoek
Lovensdijkstraat 61-63 4818 AJ Breda Tel. 076-5250500 Stagebedrijf: Erasmus MC Afdeling Psychiatrie ’s-Gravendijkwal 230 3015 CE Rotterdam Tel. 010-7040704
ii
Foreword
While the statement above may sound rather playful, it is one that has rung true throughout this internship. Working on this project at the lab made me feel more than just an intern because I was actively inspired to think for myself and to voice my opinions and, at times, doubts. Thanks for that, Jeffrey, and for your endless patience with me and your infectious enthusiasm! While spirits were often high, at times a cup of tea and a chocolate-covered, sugar-sprinkled biscuit was just what was needed to focus on the tasks at hand. Thanks, Demi! Working with you guys on this project made me realise that I made the right decision in choosing the profession that I did and that I am not finished studying just yet. Thanks for me helping me realize this.
Thanks family & friends (you know who you are!), for your reassurance and support in whatever way possible. It means a lot to be able to pour your heart out and to hear it’s going to be alright after a tough day at the lab and your genuine interest is greatly appreciated! Because I normally don’t say this out loud, let it be said now.
Voorwoord
Alhoewel de bovenstaande spreuk nogal speels zal klinken, is het er een die waar is gebleken gedurende mijn stage. Tijdens het werken aan dit project heb ik me geen moment een “stagiair” gevoeld. Ik werd aangemoedigd om kritisch na te denken en om mijn meningen en, soms, twijfels uit te spreken. Bedankt daarvoor, Jeffrey, en voor je eindeloze geduld en je aanstekelijke enthousiasme! Desondanks dat de sfeer vaak gemoedelijk was, zorgde een kopje thee en een besuikerd, in chocolade gedipt koekje voor het extra zetje om weer te kunnen focussen op het werk wat op ons stond te wachten. Bedankt, Demi! Het werken aan dit project met jullie heeft ervoor gezorgd dat ik me realiseerde dat ik de juiste beslissing heb gemaakt toen ik besloot om deze richting op te gaan en dat ik tóch nog niet uitgestudeerd ben. Bedankt voor dit inzicht!
Bedankt familie en vrienden (jullie weten wie jullie zijn), voor jullie geruststelling en steun op wat voor manier dan ook. Het is ongelofelijk fijn om na een zware op het lab je hart te kunnen luchten en te horen te krijgen dat het écht wel goed komt en jullie oprechte interesse wordt erg gewaardeerd! Ik weet dat ik dit niet vaak zeg, dus laat het bij deze gezegd zijn.
iii
Summary
Recently it has been established by the department of Psychiatry at the Erasmus MC that inhibitory neurons in the brain are myelinated. The present study provides an insight into the structure PV+ interneuron myelination and describes an experience-dependent effect of PV+ and SST+ interneuron myelination in socially isolated mice. Using biocytin-filled cells coupled with MBP fluorescence we were able to visualize and reconstruct internodes along PV+ interneuron axons and to examine patterns of myelination of PV+ interneurons. We observed relatively short pre-myelinated axonal segments and average internode lengths compared to pyramidal neurons. Internodes along PV+ interneuron axons were observed to be located proximally to the soma and distributed in a patchy fashion. Furthermore, between cells, a variety was observed in the configuration of the initial internode and the first branch point which likely affects action potential initiation. The experience-dependent effect was studied using PV:Ai14 and SST:Ai14 reporter mice coupled with MBP immunofluorescence. We investigated overall areas of myelin in the prelimbic cortex of the mPFC and the myelin content in layer V of the mPFC in both regular and socially impaired mice as well as co-localizations between myelin and PV+ and SST+ interneurons. We found indications that socially isolating mice during a postnatal period (p21 – p35) equally decreases excitatory and inhibitory myelination in the mPFC and its co-localizations with PV+ and SST+ interneurons.
Samenvatting
Recentelijk is het vastgesteld op de afdeling Psychiatrie van het Erasmus MC dat inhibitoire neuronen in het brein gemyeliniseerd zijn. Deze studie weergeeft een inzicht in de structuur van myelinisatie van PV+ interneuronen en beschrijft een ervarings-afhankelijk effect van de myelinisatie van PV+ en SST+ interneuronen in sociaal geïsoleerde muizen. Door middel van cellen gevuld met biocytin in combinatie MBP fluorescentie waren we in staat om internodia op axonen van PV+ interneuronen te visualiseren en te reconstrueren en om patronen van myelinisatie te onderzoeken. We observeerden relatief korte axon initiële segmenten en gemiddelde lengtes van internodia in vergelijking met pyramidale neuronen. Internodia op axonen van PV+ interneuronen waren proximaal gelegen en de distributie van internodia bestond uit korte opeen volgende segmenten afgewisseld door vertakkingen in het axon. Tussen cellen observeerden we een verscheidenheid aan configuraties van het eerste internodium en de eerste vertakking van het axon dat waarschijnlijk van invloed is op de initiatie van actie potentialen. Door het gebruik van PV:Ai14 en SST:Ai14 muizen in combinatie met MBP immunofluorescentie waren we in staat om het ervarings-afhankelijke effect van myelinisatie te bestuderen. We bestudeerden de totale oppervlakte aan myeline in de in de mPFC (PrL) en het myeline gehalte in laag V van de mPFC in zowel normale als sociaal geïsoleerde muizen. Daarnaast zijn co-lokalisaties tussen myeline een PV+ en SST+ onderzocht. We vonden indicaties dat het sociaal isoleren van een muis gedurende een postnatale periode (p21 – p35) excitatoire en inhibitoire myeline en co-lokalisaties met PV+ en SST+ interneuronen in de MPFC in gelijke mate vermindert.
Table of Contents
Foreword ... ii Voorwoord ... ii Summary ... iii Samenvatting ... iii 1. Introduction ... 1 2. Theoretical Background ... 2 2.1 Interneurons ... 2 2.2 Myelination ... 3 2.3 Mouse-Models ... 4 2.4 Immunofluorescence ... 42.5 Fluorescence Confocal Microscopy ... 5
3. Materials & Methods ... 7
3.1 Breeding ... 7
3.2 Immunofluorescence ... 7
3.2.1 Social Isolation ... 7
3.2.2 Single Cells ... 7
3.3 Confocal Microscopy & Analysis ... 8
3.3.1 Social Isolation ... 8
3.3.2 Single Cells ... 8
4. Results ... 9
4.1 Single Cells ... 9
4.2 Social Isolation ... 12
5. Discussion and Recommendations ... 15
Ch ap te r: In tr o d u cti o n
1
1. Introduction
At the department of Psychiatry of the Erasmus MC, it was recently established that inhibitory interneurons, besides excitatory neurons in the brain, are myelinated. Myelination was observed mainly on the subclass of parvalbumin positive interneurons (PV+). Therefore it is imperative to study the structure of myelination of these PV+ interneurons in order to provide insight into a possible function (e.g. speeding up conduction, metabolic support, modulation of firing or other functions) of myelin related to PV+ interneurons. To investigate the specific pattern of myelination of PV+ interneurons, single neurons in the mouse brain are filled with biocytin followed by a two-step immunohistochemical staining with streptavidin and Alexa488 followed by imaging using fluorescence confocal microscopy. PV+ interneurons target local pyramidal neurons, and PV+ interneurons within one or two layers. [1] Besides the locally projecting nature of the cells, internode length seems to be correlated to axon thickness. [2] Therefore, we expect that axonal internodes are relatively short (<50 µm).
Next, we will examine the experience-dependent plasticity of interneuron myelin in the medial prefrontal cortex (mPFC). This will be investigated by isolating PV:Ai14 and SST:Ai14 mice from p21 until p35, a postnatal critical period for oligodendrocyte maturation in this area. After perfusion, a two-step immunohistochemical reaction will be performed on coronal sections of the mPFC followed by imaging using confocal microscopy. Recent studies have indicated that social isolation in this “critical period” negatively affects oligodendrocyte maturation and subsequently myelination. [3, 4] Also, the mPFC occupies an important role in social cognition and possible changes that occur due to social impairment are likely to be reflected in this region. [5] Therefore, it is expected that a mouse deprived of social interaction will have less overall myelin in the mPFC. Besides less overall myelin it is expected that less myelin will be co-localized with PV+ interneurons in this region. Somatostatin positive (SST+) interneurons will also be investigated to determine whether they are affected in a similar fashion to PV+ interneurons. We expect that both cell types show decreased myelination due to social isolation of the mice. [3]
Chapter two describes the theoretical background of this project. Chapter 3 states the used methods during the project. Chapter 4 presents the results and chapter 5 provides a discussion with possible recommendations. Lastly, bibliography is included in chapter 6.
Ch ap te r: The o retic al Bac kgr o u n d
2
2. Theoretical Background
2.1 Interneurons
Neurons can be divided into two broad classes, the majority being stereotypical excitatory glutamatergic (pyramidal) neurons as opposed to the highly diverse group of gamma-aminobutyric acid-releasing inhibitory (GABAergic) interneurons. [6] Between interneurons a wide variety of subtypes can be distinguished based on electrophysiological and neurochemical features and their connectivity with pyramidal neurons. [1] Whereas some subtypes target the dendrites of pyramidal cells, modulating their input, fast spiking basket cells (Figure 1A), expressing the neuropeptide parvalbumin, target the soma or axonal hillock in order to modulate the output and synchronicity of targeted pyramidal neurons (Figure 1B). PV+ basket cells are also thought to be implicated in gamma-frequency oscillations (30–120 Hz) that are thought to be linked to conscious perception. [1, 6, 7] PV+ interneurons are also known to generate fast, non-accommodating firing in response to depolarizing direct-current injections (Figure 1C). [8] Action potentials are generated at ̴20 µm from the soma. [9] This might play an important role in the morphology of the pre-myelinated axonal segment in relation to myelination. Furthermore, the axonal arborisation of PV+ interneurons might have a significant effect on the length of internodes.
Figure 1: Neuronal structure, wiring and firing of PV+ basket cells. A) Representative arborisation of dendrites (thick) and axons (thin) of a PV+ basket cell in rat. B) Schematic overview of PV+ cell (red) wiring in respect to a pyramidal cell (black). PV+ basket cell axons contact either the soma or the axonal hillock. The pyramidal cell contacts the dendrites of the PV+ cells to allow for modulation of its signal upon firing. C) PV+ basket cell responses to hyperpolarizing, near-threshold and suprathreshold current injections.
GABAergic interneurons constitute between 20% - 30% of the total cortical neurons of which basket cells represent about 50%. However, these percentages may differ from region to region as only 11% of neurons in the hippocampus (CA1) are GABAergic of which 24% is PV+ resulting in only 2.6% PV+ interneurons. PV+ basket cells are prevalent in layers II-III up until V across neocortical layers.
Ch ap te r: The o retic al Bac kgr o u n d
3
2.2 Myelination
Whenever an axon is myelinated, it is wrapped in a sheath of myelin by an oligodendrocyte (Figure 2A, 2B). Myelination appears to serve as a form of plasticity to adapt brain function to environmental stimuli and providing metabolic support. Besides metabolic support, myelin enables cells to conduct at a much higher rate through saltatory conduction, the generation of action potentials at the “Nodes of Ranvier” (Figure 2C). Because sodium ions only accumulate at the nodes and only have to be transported to the extracellular space at these locations, less energy is needed to conduct the signal. The myelination of an axon allows it to be smaller in size whilst still conducting at the same speed. To illustrate: an unmyelinated squid axon conducting at 25 m/sec must have a diameter of ~500 µm whereas a mammalian myelinated axon is only a few µm in diameter. [2]
Figure 2: Overview of myelin and oligodendrocyte structure and saltatory conduction. A) Schematic overview of an oligodendrocyte wrapping an axon. Myelin is wrapped around the axon like a blanket and interspersed with Nodes of Ranvier. B) Cross-section of a myelinated axon. C) Schematic overview of saltatory conduction showing depolarization at the Nodes of Ranvier.
With regards to myelination it was found that juvenile social experience influences oligodendrocyte maturation and myelination in the mPFC between postnatal day 21 (P21) and P35, a so called “critical period”. Myelination occurs relatively late in development. In mice, it starts at birth in the spinal cord and is almost completed at P60. [2, 4] Social deprivation during this time results in reduced myelin thickness and a reduced number of myelin gene transcripts in the mPFC. [3] It is believed that myelin has a modulating function in the brain. Besides speeding up the conduction of signals it is thought that myelin helps synchronize action potentials. The myelination of parvalbumin-expressing basket cells could indicate a correlation between the fast-spiking nature of the cells and the myelination of those cells.
Ch ap te r: The o retic al Bac kgr o u n d
4
2.3 Mouse-Models
In order to collectively image PV+ basket cells and myelin using confocal microscopy it is necessary to incorporate fluorescent labels. To obtain a higher specificity and stability in labeling specific cell types we used transgenic Cre-driver lines coupled with a loxP system (Figure 3A) instead of immunohistochemistry. Transgenic mice will be with a line expressing tdTomato, a red fluorescent protein, in PV+ interneurons only (PVAi14), capitalizing on the ability of the Cre/lox system to be used for cell-specific expressions. In this case a mouse with an Ai14 configuration, a tdTomato gene followed by a lox2 STOP cassette (Figure 3B), incorporated into the Gt(ROSA)26Sor (Rosa26) locus is crossed with a mouse that expresses Cre in PV+ interneurons through a cell-specific promotor (PV-IRES-Cre). After breeding, the lox2 STOP cassette that is present after the tdTomato gene will only undergo excision in PV+ interneurons due to the cell-specific expression of Cre. [10] In this case the expression of Cre is regulated by the presence of parvalbumin. The same method will also be applied to all somatostatin neurons (SST+). In this case instead of a PV-Cre mouse, a SST-IRES-Cre mouse will be used to have SST+ interneurons express tdTomato. [11]
Figure 3: The Cre/loxP system and Ai14 promoter. A) Schematic overview of the Cre/loxP system. This system allows for the expression of a transgene in a pre-specified cell type. B) Overview of the location of a promoter within the Rosa26 locus and the configuration of the Cre-reporter Ai14.
2.4 Immunofluorescence
Due to the unavailability at the lab of Cre-driver lines that incorporate fluorescent labels into myelin basic protein (MBP), fluorescent labels are incorporated using a two-step immunohistochemical reaction (Figure 4A) with Alexa- and Cy- conjugated antibodies to compliment the tdTomato already present through breeding. A two-step reaction intensifies the fluorescent signal due to the ability of the primary antibodies to bind multiple secondary antibodies with conjugated fluorophores. By injecting the antigen of interest into an animal (e.g. goat, mouse, rabbit etc.) the primary antibody is generated. By subsequently injecting the formed antibodies in to another animal the secondary antibody is generated which is then conjugated to a fluorophore. For immunohistochemical purposes Alexa and Cy conjugated dyes are widely used due to their optimal excitation by many fluorescence microscopes. These long-wavelength dyes fluoresce at longer long-wavelengths than common sources of cell autofluorescence and have a generally low background. Besides complementary staining of proteins these dyes can also be used to enhance a marker introduced through breeding. [12] In order to obtain detailed images of single
Ch ap te r: The o retic al Bac kgr o u n d
5
cells (Figure 4B) it is not possible to use the aforementioned two-step reaction as that would induce a high background through labelling of other cells. Therefore, during electrophysiology, the cell is injected with 0,5% biocytin which acts as a primary antibody. Streptavidin is then introduced as a secondary that is conjugated with an Alexa or Cy dye. [13]Figure 4: Schematic overview of an immunohistological staining. A) A primary antibody is used that attaches to the tissue. Subsequently, a secondary antibody that carries a fluorescent label (e.g. tdtomato) attaches to the primary antibody. Multiple secondary antibodies are able to bind to a single primary antibody to enable a more pronounced and precise fluorescence. B) Representative image of a cell filled with biocytin (red).
2.5 Fluorescence Confocal Microscopy
In order to image the incorporated labels, we use fluorescence confocal microscopy which offers the ability to view complete brain regions within the specimen as well as specific layers within a region at a high resolution. Furthermore, the view can be toggled between the different fluorescent labels and composite images of the labels can be obtained to distinguish co-localizations. The fluorescent labels are visualized by focusing a beam of light on a small section inside the tissue which can be as small as 0,5 µm. The objective through which the beam of light is focused gathers the reflected light coming back from the tissue and projects it. In order to prevent blurring of the image by scattered light, a pinhole aperture is used to only allow light emitted from the desired focal point to pass through. At the other side of the pinhole a photomultiplier is used to detect the incoming confocal light which allows the specimen to be imaged one point at a time (Figure 5). Using confocal microscopy it is possible to do “tile-scans” of wide areas of specimens and to make “Z-Stacks” in the Z plane of the specimen. To investigate global myelination, tile scans will be used to image the different layers of the mPFC. To obtain images of single cells filled with bioctyin, Z-Stacks coupled with tile scans will be used. These images will be traced and then 3D rendered to investigate the pattern of myelination of these cells. [14]
Ch ap te r: The o retic al Bac kgr o u n d
6
Figure 5: Schematic overview of confocal microscopy. A laser at a specific wavelength is directed a focal plane in the tissue. The emitted light by the fluorophore is then refracted into a light detector by a dichroic mirror. The dichroic mirror enables the laser at the initial wavelength to go through unrefracted so the beams do not cross yielding a clear image. The objective targets the light into a focal plane of approx. 1µm thick enabling the user to scan the tissue at different depths.
When the excitation laser hits the target, it generates a high intensity of fluorescence. In order to separate the incoming and outgoing light a dichroic mirror is used which reflects the higher-wavelength fluorescent light into the light detector whilst enabling the shorter wavelength excitation light to go through unaltered into the tissue. [15] In order to obtain a complete image multiple dyes are combined (Figure 6A). In order to stain nuclei DAPI is used which binds strongly to A-T rich regions in DNA and emits at 461 nm (blue). When excited, tdTomato emits at 581 nm (red). In order to stain myelin a secondary antibody conjugated with A488 is used that binds to the primary antibody (MBP). A488 emits 519 nm (green) when excited. [12]
Figure 6: Representative confocal images with multiple dyes. A) Overview photo of mPFC with DAPI (blue), MBP (green), PV (red) and a merge. B) Representative example of a MBP (green) + PV (red) co-localization. The axon is shown in red with an outer contour of green.
Ch ap te r: M at erial s & M et h o d s
7
3. Materials & Methods
3.1 Breeding
For all experiments, heterozygous mice were obtained from The Jackson Laboratory, Maine, USA. PV+ interneurons were genetically labeled by crossing B6;129S6-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J (Ai14) (The
Jackson Laboratory, 007908) [10] with B6;129P2-Pvalbtm1(cre)Arbr/J (PVcre) (The Jackson Laboratory, 008069). SST+ interneurons were labelled by breeding Ai14 mice with Ssttm2.1(cre)Zjh/J (Sst-IRES-Cre) (The
Jackson Laboratory, 013044). [16] Mice were maintained on a 12 h light/dark cycle with food and water available ad libitum. 8 male PV:Ai14 were divided into two groups of 4 mice each and 6 male SST:Ai14 mice were divided into two groups of 3 mice each. Of both mouse lines one group was group-housed and maintained regularly up until perfusion. The other group was socially isolated and each mouse was housed individually starting 3 weeks after birth (p21) up until perfusion. All mice were perfused at 5 weeks (p35).
3.2 Immunofluorescence
After deep anesthesia induced by intra-peritoneal injection of pentobarbital (50 mg kg− 1), mice were transcardially perfused with saline, followed by 4% paraformaldehyde. Brains were dissected and post-fixed in 4% paraformaldehyde for 2 h at 4 °C. After post-fixation, the brains were transferred into 10% sucrose phosphate buffer (PB 0.1 M, pH 7.3) and stored overnight at 4 °C. Embedding was performed in a 12% gelatin/10% sucrose block, with fixation in 10% paraformaldehyde/30% sucrose solution for 2 h at room temperature and immersed in 30% sucrose at 4 °C until cutting. Forty micrometer coronal sections were collected using a freezing microtome (Leica, Wetzlar, Germany; SM 2000R) and stored in 0.1M PB.
3.2.1 Social Isolation
Free-floating sections were pre-incubated with PBS 0.1 M containing 0.5% Triton X-100 and 10% normal horse serum (NHS; Invitrogen, Bleiswijk, The Netherlands) for 1 h at room temperature. Sections were incubated in a mixture of primary antibodies, in PBS containing 0.4% Triton X-100 and 2% NHS for 72 h at 4 °C. Two primary antibodies were used: polyclonal goat anti-mbp (1:300, Santa Cruz sc-13914, Heidelberg, Germany) and monoclonal mouse anti-PV (1:1000, Swant 235, Marly, Switzerland). Sections were washed with PBS 0.1 M and incubated with the corresponding Alexa-conjugated secondary antibodies (1:300, Invitrogen, Waltham, USA) in PBS 0.1 M containing 2% NHS and 0.4% Triton X-100 for 2h at room temperature. Sections were washed with PBS 0.1 M and PB 0.1 M. Nuclear staining was performed using DAPI (1:100, Invitrogen) in 0.1M PB. Sections were washed with PB 0.1 M and mounted on slides, cover slipped with Vectashield H1000 fluorescent mounting medium (Vector Labs, Peterborough, UK) and sealed.
3.2.2 Single Cells
During electrophysiology on 300 µm thick sections 10 neurons in the mPFC were filled with 0.5% biocytin (Sigma-Aldrich 576-19-2, Zwijndrecht, Netherlands). Afterwards sections were cut serially into 40 micrometer sections on a freezing microtome. Immunofluorescence was done as previously described. Sections were incubated with a Streptavidin-coupled Alexa-conjugated secondary antibody (1:300, Invitrogen) and cyanine dyes (1:300, Sanbio, Uden, The Netherlands) in PBS containing 2% NHS and 0.4% Triton X-100 for 3h at room temperature.
Ch ap te r: M at erial s & M et h o d s
8
3.3 Confocal Microscopy & Analysis
3.3.1 Social Isolation
Stained images of the prelimbic area of the mPFC were acquired at Bregma 1,98 mm up until Bregma 1,70 mm using a mouse brain atlas. [17] A Zeiss LSM 700 confocal microscope (Carl Zeiss, Oberkochen, Germany) equipped with Zeiss Plan-Apochromat × 10/0.45 and × 63/1.4 (oil immersion) objectives was used. Native tdTomato, Alexa488 and DAPI were imaged using the excitation wavelengths of 555, 488 and 405 nm, respectively. Of the mPFC a 10x 2x2 tile scan was acquired spanning layers I up until layer V in both hemispheres. Layers were identified using a DAPI staining to observe the density of cells in each layer. Layer I was identified as a sparsely populated layer, layer II-III was observed through a distinct increase in density of cells and layer V was subsequently observed to be more densely populated than layer II-III. Within layer V, several 63x images were acquired at randomly sampled locations. Native tdTomato and A488 within 2x2 tile scan images were quantified using ImageJ (NIH, 1.42q) using image thresholding and particle analysis. A488 (MBP) and its co-localizations with native tdTomato (PV+/SST+) within 63x images were quantified manually within ImageJ by counting MBP fragments and visually identified co-localizations with tdTomato. Significance of observations was established by using an unpaired Student’s t-test.
3.3.2 Single Cells
Stained cell images were acquired at 63x using a 5x5 tile scan and a 40 micrometer Z-stack with 0,5 micrometer intervals. Image analysis, 3D reconstruction and tracing of internodes were done using Neurolucida 360 (MBF Bioscience, Williston, USA, v2.60). Cells were analyzed for the pre-myelinated axonal segment (PMAS), number of internodes, average internode length and total myelin length per cell and general morphological features were described.
Ch ap te r: Results
9
4. Results
4.1 Single Cells
To examine the structure of myelination along PV+ interneuron axons, we used fast-spiking PV+ cells filled with biocytin during electrophysiology to study the pattern of myelination of PV+ interneurons. Detailed (63x) stack images of cells located in the prelimbic cortex of the mPFC (n = 10 cells) (Figure 7A) were analyzed and internodes were reconstructed (Figure 7B, 7C) to investigate the structure of myelin along axons and to obtain an indication of myelin content per cell. All 10 cells were myelinated. The average PMAS was 31,82 ± 2,62 µm (mean ± standard error) (n = 10 cells) in length which is similar to pyramidal neurons in layers V/VI. [18] Average observed internode length was (n = 51 analyzed internodes, 10 cells) 23,15 ± 2,84 µm. The range of possible internode lengths between pyramidal neurons and PV+ interneurons appears to be largely similar although overall pyramidal interneuron internode length is higher (from 20 µm to 200 µm). [18] PV+ interneurons do have internodes shorter in length than 20 µm as opposed to pyramidal neurons. (Figure 8B). The average total observed length of myelin was 139,84 ± 28,08 µm (n = 10 cells) (Figure 8A).
Internodes observed along PV+ interneurons axons were located proximally to the soma and were often interspersed by branch points (Figure 9A, 9B). While pyramidal neurons have been described to have long unmyelinated segments between internodes (up to 55 µm), this appears not to be the case in PV+ interneurons where internodes are alternated by branch points. [18] Between PV+ interneurons, abundant morphological varieties of the PMAS were observed. While the locations of internodes were all proximal to the soma, we observed different configurations of initial internodes in relation to the PMAS. Out of all cells observed, few cells had dendrite originating axons. Furthermore, in some cells the PMAS appeared have a branch point first, which was then followed by internodes. Other cells, however, did have an internode before the first branching of the axon. The discrepancies observed in the configurations of the PMAS in different PV+ cells could play a role in action potential initiation. Previous studies estimated that the action potential in PV+ interneurons is generated at approximately 20 µm from the soma. [9] It is possible that the length of the PMAS determines whether a branch point is introduced before the initial internode and that the distance to the first branch point occurs at a roughly fixed distance from the soma. The length of the initial internode could then determine whether that internode could be placed before or after the first branch point.
Ultimately, the observed structure of myelination along PV+ interneurons axons is not likely to contribute to a significant increase in the speed of action potential propagation due to the proximal coverage of the myelin and the relatively short length of internodes. On the other hand, it is suggested that myelin provides metabolic support [19, 20] which could be relevant for the fast spiking of the PV+ interneurons. Besides metabolic support, the relatively short internodes might contribute to optimizing signal timing which could contribute to the modulation role of PV+ interneurons. [21]
Ch ap te r: Results
1
0
Figure 7: Schematic overview of locating and tracing co-localizations in an image of a biocytin-filled cell. A) Image of a biocytin-filled (red) cell indicating a traced region. B) Schematic overview of a tracing on a single cell image indicating a traced co-localization. (biocytin = red, MBP = green, Soma = tan) C) Indicated co-localization showing a typical visual cue of MBP and biocytin co-localization (biocytin =red, MBP = green).
Figure 8: Overview of data from single cell analysis within the prelimbic cortex of the mPFC. A) (n = 10 filled cells) Shown data includes: number of analysed internodes, average internode length, PMAS and total myelin length including average and standard error for all parameters. B) Graph displaying the distribution of average internode length of all measured internodes including a polynomial fitted curve. (n = 51 internodes, 10 cells)
Ch ap te r: Results
1
1
Figure 9: Schematic overview of a 3D reconstructed PV+ interneuron. A) A complete representation of the axonal arborisation of the cell (axon + soma = tan, myelin = white) B) Detailed representation of proximal myelin on the axon. In between branch points, the axon is almost completely covered in myelin.
Ch ap te r: Results
1
2
4.2 Social Isolation
To investigate the experience-dependence of myelination, we used the PV:Ai14 and SST:Ai14 reporter lines and split them into two groups. One group was socially-isolated from p21 till p35 (PV:Ai14: n = 4, SST:Ai14, n = 3), the other group was housed regularly (PV:Ai14: n = 4, SST:Ai14, n = 3) up until perfusion (p35). To visualize myelin and its co-localizations with PV+ and SST+ interneurons the reporter lines were coupled with MBP immunofluorescence. Overview images (10x) and detailed images (63x) of MBP immunofluorescence were analyzed for an overall effect of social isolation on myelination (Figure 10A). We found that the overall area of myelin in the mPFC in PV:Ai14 mice was not significantly decreased (Figure 10B) (GH: 2,38E+08 ± 3,57E+07 vs. SI: 1,76E+08 ± 4,18E+07 total area of myelin, P=0.3). Detailed images (63x) obtained in layer V in the mPFC of PV:Ai14 mice (Figure 10C) did show a significant decrease in myelination (GH: 223 ± 8,9 vs. SI: 175 ± 12,3 MBP fragments, *P<0.05). The observed decrease in overall myelination in the mPFC of SST:Ai14 mice (Figure 11A) was more distinct (GH: 1,32E+08 ± 2,06E+07 vs. SI: 3,93E+07 ± 1,50E+07 total area of myelin, *P<0.05). As opposed to the PV:Ai14 mice, the decrease in myelination observed in detailed images (63x) within layer V obtained from SST:Ai14 mice is less pronounced (Figure 11B) (GH:280 ± 38,8 vs. SI: 221 ± 19,0 MBP fragments, P=0.25). The observed decrease in both mouse lines at both overview (10x) and detailed (63x) level indicate that myelination is indeed an experience-dependent process. At a detailed level (63x) in layer V both mouse types seem to be similarly affected by social impairment.
Next we examined the myelination of PV+ and SST+ interneurons in socially isolated PV:Ai14 and SST:ai14 mice respectively. We found that PV:Ai14 mice show a distinct decrease in co-localizations between MBP and PV+ interneurons (Figure 10D) (GH: 20,8 ± 1,11 vs. SI: 13,8 ± 2,02 co-localizations, *P<0.05). In SST:Ai14 mice less co-localizations between MBP and SST+ interneurons were observed (Figure 11C) (GH: 16,3 ± 2,48 vs. SI: 7,53 ± 0,41 co-localizations, *P<0.05). Co-localizations in both mouse types appear to decrease at roughly the same rate under influence of social isolation.
Together, this indicates that a clear indication exists that myelination decreases due to disturbances in social development which corresponds with previous studies. The observed myelin decrease in these studies amounted to roughly 50% compared to control groups. [3, 4] Additionally, both the MBP count in layer V and the co-localizations with PV+ and SST+ appear to decrease at a similar rate. This suggests that both excitatory and inhibitory are similarly affected by social impairment.
Ch ap te r: Results
1
3
Figure 10: Socially impaired PV:Ai14 mice show decreased overall myelination and myelination of PV+ interneurons in the mPFC. A) Schematic overview of the locations of collected images in the mPFC. (n = 4 group-housed (GH) mice and 4 socially isolated (SI) mice) B) Representative images (10x) of the mPFC showing a decrease in overall MBP area in socially isolated mice, P = 0.3. C) Representative images (63x) of layer V in the mPFC show a decrease in MBP fragments in socially isolated mice, *P < 0.05. D) Representative images showing co-localization staining for MBP and PV. Socially impaired mice show a decrease in myelination of PV+ interneurons, *P < 0.05.
Ch ap te r: Results
1
4
Figure 11: Socially impaired mice SST:Ai14 show decreased overall myelination and myelination of SST+ interneurons in the mPFC. (n = 3 group-housed (GH) mice and 3 socially isolated (SI) mice) A) Representative images (10x) of the mPFC showing a decrease in overall MBP area in socially isolated mice, P < 0.05. B) Representative images (63x) of layer V in the mPFC show a decrease in MBP fragments in socially isolated mice, P = 0.25. C) Representative images showing co-localization staining for MBP and SST. Socially impaired mice show a decrease in myelination of SST+ interneurons, *P < 0.05.
Ch ap te r: D iscu ssi o n and R eco m m end atio n s
1
5
5. Discussion and Recommendations
This study examines myelination of PV+ interneurons by analyzing single biocytin-filled fast-spiking PV+ interneurons in the mPFC and reconstructing patterns of myelination of these cells. Furthermore, the experience-dependence of myelination is examined by isolating PV:Ai14 and SST:Ai14 mice during p21-p35 and comparing them with regularly housed mice. In doing this, this study provides an initial indication of the structure of myelination of PV+ interneurons and describes an observed effect on myelination due to social impairment. Uncovering the structure and plasticity of myelination of PV+ interneurons could potentially result in novel insights regarding myelin function. PV+ interneurons are non-accommodating fast-spiking cells that occupy a strong regulatory function in local networks in the brain including self-inhibition, the inhibition of pyramidal neurons and inhibition of nearby interneurons. [21] It is likely that myelination of these cells compliments this particular function.
Regarding the examination of single cells, it must be carefully considered that the filling of the cells with biocytin during electrophysiology is a critical step that determines whether a successful analysis can be performed. The injection of biocytin directly into the soma is especially vulnerable due to the potential of the cell membrane rupturing after injection effectively spilling the contents of the cell. Secondly, when maneuvering the needle around the soma, axons might be cut since these are not clearly visible at that time. This in turn prevents proper reconstruction of internodes during analysis. However, during analysis no abnormal axonal patterns were observed that could indicate a possible mishap of sorts. Besides electrophysiology, the reconstruction of single cells using confocal microscopy has proven to be a significant challenge. Detailed Z-stack images are of considerable size and are therefore difficult to handle digitally. This has rendered us unable to visualize a complete cell in one image and to reliably trace internodes past a certain radius. Were it not for this limitation, analysis of single cells could have included farthest myelin segments and total myelin segments of single cells among other important structural details besides providing a 3D representation of the proximal myelin structure. Nevertheless, by acquiring and analyzing multiple images per cell it was observed that internodes were located proximally and that the majority of internodes could be analyzed within a single image. Therefore, these limitations did not hamper our ability to correctly observe the structure of myelination.
Functionally, myelin is often described as an insulation that helps speed up action potentials. [2] However, this is unlikely to play a major role in PV+ interneurons due to the proximal coverage. Furthermore, average internode length of PV+ interneurons (around 25 µm) also appears to be lower than that of pyramidal neurons. Internodes of that length are unlikely to improve signal conduction by a significant amount. [2] Besides signal conduction there has been evidence that myelin provides metabolic support [20] to axons which might be necessary to accommodate the fast-spiking nature of the cells. [19] In regard to the fast-spiking nature of the cells, myelin might play a role in signal timing. PV+ interneurons modulate local networks in conjunction with other (PV+) interneurons. Therefore optimal signal timing across neurons is crucial for the synchronization of action potentials. [21] As opposed to a significant increase in conduction speed by longer internodes, smaller internodes might be able to adjust the timing of action potentials in a precise manner, perhaps fractions of milliseconds. Even though in this case conduction speed is not noticeably changed, the small “corrections” in timing
Ch ap te r: D iscu ssi o n and R eco m m end atio n s
1
6
applied to action potentials across axons of a single interneuron or even multiple interneurons might help achieve synchronicity in a larger local network. [21]Due to the fast-spiking nature of the cells, action potential generation in the PMAS fulfills an important role and the morphology of the PMAS is known to regulate the excitability of neurons, shape of action potentials, frequency of firing and the amount of neurotransmitter released from synapses. Therefore, the observed variety of the placement of the initial internode regarding the first branch point on the axons of PV+ interneurons might have a significant effect. [22] The observed length of the PMAS of PV+ interneurons (around 30 µm average) which is relatively short compared to certain pyramidal neurons (can be up to 80 µm long) might indicate a fast initiation and the typical short length [1] of action potentials due to a shorter distance to the trigger point located at the distal end of the PMAS. [21] Besides providing indications to PV+ interneuron myelination structure, this study also indicates that social impairment leads to a global decrease of myelination in the mPFC and subsequently, a decrease in the myelination of PV+ and SST+ interneurons. Regarding the data presented, it is important to note that a decrease in myelination is uniformly observed in both mouse lines at both an overall (10x) and detailed (63x) level including co-localizations with PV+ and SST+ interneurons. This consistent observed effect suggests that the social impairment does result in a decrease of myelination. Furthermore, biological replications of the same experiment complemented the displayed results regarding the observed effect. The lack of significance of observations could be ascribed to the small sample size. Statistics were applied to both mice and photos within each group. The data displayed is based on the statistics done on both groups of mice (PV:Ai14: n = 4 GH mice and 4 SI mice, SST:Ai14: n = 3 GH mice and 3 SI mice). The statistics applied to the obtained images of each group (PV:Ai14: n = 45 GH images, 57 SI images, SST:Ai14: n = 12 GH images and 12 SI images) result in a higher significance of observations. In this case a threshold of P < 0,01 is met in all cases but the images in layer V mPFC of SST:Ai14 mice. In the case of a technical replication, it is recommended to increase the sample size of the mouse groups.
In the data presented the overall level myelination seems to be more prominently affected in SST:Ai14 mice despite undergoing the same procedure regarding social isolation. Therefore, it must be considered that when working with transgenic mice, there is a possibility of biological factors influencing the observations as a result of the inbred transgenes that might play a role in each mouse line. That said, the observations done in layer V of the mPFC correlate more closely between both mouse lines which also holds true for the observed co-localizations of the respective interneurons with MBP. Ultimately, biological replications of the experiment with SST:Ai14 mice could indicate whether the observed effect is consistently as prominent in these mice as suggested by the data presented. A possible explanation for the aforementioned effect in SST:Ai14 mice could be attributed to passenger genes: genes that flank the gene of interest and are also transferred to the offspring. In an ideal situation only the donor gene would be present against a background which is completely comprised of genes from the recipient. However, passenger genes “pollute” the background so that it includes genes from the donor. [23] By comparing PV:Ai14 and SST:Ai14 mice this might result in a different genomic background in both mouse lines that can affect experimental outcomes. In this case, since Ai14 is the shared gene, possible passenger genes might persist in either the PV-Cre or SST-Cre line.
Ch ap te r: D iscu ssi o n and R eco m m end atio n s
1
7
Besides incorporating fluorescent labels in transgenic mice, an immunohistochemical staining is paramount in providing the ability to properly image MBP in conjunction with inbred tdTomato in PV:Ai14 and SST:Ai14 mice. However, an immunohistochemical staining is comprised of many individual laborious steps. Including sectioning of the tissue, this leaves room for many errors regarding the MBP immunofluorescence. Besides using the proper antibodies it must also be considered with conjugate dyes are used. Ideally, the fluorescence of conjugate dye has as little as possible overlap with the inbred fluorescent protein. [24] Furthermore, the immunohistochemical staining was performed on free-floating sections. In this case, it is desired for the sections not to clump and to be unattached from the container to have as much surface area as possible for the antibody to incorporate into the tissue. Although these factors must be taken into consideration when performing and immunohistochemical staining, the technique is a powerful method to label proteins and can even be used to visualize more co-localizations within tissue. For SST:Ai14 mice, it is known that some cells are false positive due to maturing PV+ cells also being SST+. Hence, it is beneficial to introduce a PV+/SST+ co-localization when performing immunohistochemistry. In this study, such a co-localization could have been used to identify true SST+/MBP co-localizations in SST:Ai14 mice while accounting for PV+/MBP co-localizations.Using confocal microscopy we were able to reconstruct internodes along interneuron axons, observe an overall decrease in MBP and analyze co-localizations between MBP and PV+/SST+ interneurons. On the other hand, analysis of confocal images remains a critical point since analyses are subject to resolution of the images and the quality of the immunohistochemical staining that was performed. The automatic analysis of the overview (10x) images was done using thresholding. Thresholding analysis is dependent on the settings with which the images were obtained during confocal microscopy. It is important to use similar settings as much as possible between acquiring images to guarantee a more consequent automatic analysis. [25] The manual analysis done on the detailed (63x) images is largely unaffected by these factors due to the ability to critically evaluate the quality of the acquired images on a per image basis. Even though a manual analysis is somewhat subjective, this is remedied by carefully establishing guidelines for the analysis. Although a manual analysis is not sufficient for a strict and stringent quantitative analysis, it did prove to be suitable to observe an overall effect regarding myelination in this study.
Though this study has provided an indication into PV+ interneuron myelination and experience-dependence much is left to be discovered. In conjunction with the experience-dependent effect of myelination of PV+ interneurons it is interesting to investigate the integrity of myelin on PV+ interneuron axons under these circumstances. Besides the fact that maturation of oligodendrocytes is affected by social experience [4], damage to oligodendrocytes is shown to have negative consequences on myelination such as slowing conduction velocity and morphological degeneration of axons. [19] Whether an experience-dependent defect in oligodendrocytes is comparable to cell damage and whether these effects are also seen in regard to PV+ interneuron myelination are interesting thoughts worth pursuing. Besides using confocal microscopy, electron microscopy is able to provide an even higher level of detail which would be optimal to investigate the myelination in depth along a single axon and to observe myelinated axons coronally. [18] Using this method it can be investigated how the structural integrity of myelin along PV+ interneuron axons is affected in a similar social experiment. In
Ch ap te r: D iscu ssi o n and R eco m m end atio n s
1
8
order to test whether and how myelin affects the synchronicity in a local network with PV+ interneurons paired recordings can be utilized.To conclude, we have debated that the structure of myelination of PV+ interneurons likely provides metabolic support and improves signal timing and that the data presented provides indications that myelination is negatively affected by a lack of social stimuli. While confocal microscopy coupled with immunofluorescence has enabled us to investigate myelin structure in detail and at an overall level, we must remain mindful of possible limitations and keep in mind advantages of other techniques that can be applied in possible studies that might follow. Lastly, we have suggested ideas concerning PV+ interneuron myelination that might be examined in future studies which could provide valuable insight.
Ch ap te r: Liter atur e
1
9
6. Literature
1.
Jiang, X., et al., Principles of connectivity among morphologically defined
cell types in adult neocortex. Science, 2015. 350(6264).
2.
Snaidero, N. and M. Simons, Myelination at a glance. Journal of Cell
Science, 2014. 127(14): p. 2999-3004.
3.
Liu, J., et al., Impaired adult myelination in the prefrontal cortex of socially
isolated mice. 2012. 15(12): p. 1621-1623.
4.
Makinodan, M., et al., A Critical Period for Social Experience–Dependent
Oligodendrocyte Maturation and Myelination. Science, 2012. 337(6100):
p. 1357-1360.
5.
Bicks, L.K., et al., Prefrontal Cortex and Social Cognition in Mouse and
Man. Frontiers in Psychology, 2015. 6: p. 1805.
6.
Druga, R., Neocortical inhibitory system. Folia biologica, 2009. 55(6): p.
201.
7.
Hu, H., J. Gan, and P. Jonas, Interneurons. Fast-spiking, parvalbumin(+)
GABAergic interneurons: from cellular design to microcircuit function.
Science, 2014. 345(6196): p. 1255263.
8.
Freund, T.F., Interneuron Diversity series: Rhythm and mood in
perisomatic inhibition. Trends in Neurosciences, 2003. 26(9): p.
489-495.
9.
Li, T., et al., Action Potential Initiation in Neocortical Inhibitory
Interneurons. PLoS Biology, 2014. 12(9): p. 1-16.
10. Madisen, L., et al., A robust and high-throughput Cre reporting and
characterization system for the whole mouse brain. 2010. 13(1): p.
133-140.
11. Besser, S., et al., A Transgenic Mouse Line Expressing the Red Fluorescent
Protein tdTomato in GABAergic Neurons. PLoS ONE, 2015. 10(6): p.
e0129934.
12. Berlier, J.E., et al., Quantitative comparison of long-wavelength Alexa
Fluor dyes to Cy dyes: fluorescence of the dyes and their bioconjugates. J
Histochem Cytochem, 2003. 51(12): p. 1699-712.
13. Thomson, A.M. and W.E. Armstrong, Biocytin-labelling and its impact on
late 20th century studies of cortical circuitry. Brain Research Reviews,
2011. 66(1): p. 43-53.
14. Nwaneshiudu, A., et al., Introduction to confocal microscopy. The Journal
of investigative dermatology, 2012. 132(12): p. e3.
15. Ragazzi, M., et al., Fluorescence confocal microscopy for pathologists.
Modern pathology : an official journal of the United States and Canadian
Academy of Pathology, Inc, 2014. 27(3): p. 460-471.
Ch ap te r: Liter atur e
