This article describes a working protocol to image dendritic spines from hippocampal neurons in vitro using Structured Illumination Microscopy (SIM). Super-resolution microscopy using SIM provides image resolution significantly beyond the light diffraction limit in all three spatial dimensions, allowing the imaging of individual dendritic spines with improved detail.
Dendritic spines are protrusions emerging from the dendrite of a neuron and represent the primary postsynaptic targets of excitatory inputs in the brain. Technological advances have identified these structures as key elements in neuron connectivity and synaptic plasticity. The quantitative analysis of spine morphology using light microscopy remains an essential problem due to technical limitations associated with light’s intrinsic refraction limit. Dendritic spines can be readily identified by confocal laser-scanning fluorescence microscopy. However, measuring subtle changes in the shape and size of spines is difficult because spine dimensions other than length are usually smaller than conventional optical resolution fixed by light microscopy’s theoretical resolution limit of 200 nm.
Several recently developed super resolution techniques have been used to image cellular structures smaller than the 200 nm, including dendritic spines. These techniques are based on classical far-field operations and therefore allow the use of existing sample preparation methods and to image beyond the surface of a specimen. Described here is a working protocol to apply super resolution structured illumination microscopy (SIM) to the imaging of dendritic spines in primary hippocampal neuron cultures. Possible applications of SIM overlap with those of confocal microscopy. However, the two techniques present different applicability. SIM offers higher effective lateral resolution, while confocal microscopy, due to the usage of a physical pinhole, achieves resolution improvement at the expense of removal of out of focus light. In this protocol, primary neurons are cultured on glass coverslips using a standard protocol, transfected with DNA plasmids encoding fluorescent proteins and imaged using SIM. The whole protocol described herein takes approximately 2 weeks, because dendritic spines are imaged after 16-17 days in vitro, when dendritic development is optimal. After completion of the protocol, dendritic spines can be reconstructed in 3D from series of SIM image stacks using specialized software.
A dendritic spine is a small protrusion of the neuron membrane. This characteristic structure is specialized to typically receive input from a single synapse and represents the physical contact area between two neurons. Most functionally mature dendritic spines consist of a globular tip, termed head, and a thin neck that connects the head to the dendritic shaft. However, spines are not static and actively move and change their morphology continuously even in the adult brain 2. Within a 2 week period of time, rat primary hippocampal neuron cultures derived from late embryonic or early postnatal time develop complex dendritic arbors with numerous membrane protrusions that evolve from early filipodia to spine-like structures 3. Based on this dynamic behavior and other characteristics, dendritic spines are thought to provide an anatomical substrate for memory storage and synaptic transmission 4,5.
Given the critical role that dendritic spine size and shape have in synaptic function, it is important to measure their dimensions accurately. Spines vary from around 200 to 2,000 nanometers in length and can be readily identified by confocal laser-scanning fluorescence microscopy. However, spine dimensions other than length are usually below the conventional optical systems’ resolution, theoretically fixed by diffraction around 200 nanometers 6. These resolving powers are insufficient for imaging finer details, such as the width of spine necks and heads. Much work has been dedicated to solve this problem and many relatively new super-resolution microscopy techniques have provided substantial progress. In particular, it is possible to achieve resolution beyond the classical limit without discarding any emission light by using laterally structured illumination microscopy (SIM) in a wide-field, non-confocal microscope 7-10. Using this technique in combination with non-linear microscopy techniques, it is theoretically possible to improve the lateral resolution of the optical microscope by an unlimited factor 11. However, in most experimental circumstances, SIM allows to surpass the resolution limit by a factor of two 1. Other super-resolution optical microscopy techniques such as Stimulated emission depletion (STED) microscopy 12 and photo-activation localization microscopy (PALM) 12 have been applied to imaging of dendritic spines. Localization-based methods such as PALM require very large numbers of raw images to achieve super-resolution and are therefore limited in speed. On the other hand, STED can achieve high imaging speed, although at relatively low photon counts and small fields of view, which may not be the case for SIM 13.
In this article the aim is to provide a working protocol to image dendritic spines from rat primary hippocampal neurons cultured in vitro using SIM. The protocol consists of two distinguishable phases: an initial one consisting of establishment, development, transfection and immunohistochemistry of rat primary hippocampal neuron cultures and a late phase dedicated to sample imaging.
All experimental procedures involving animals were optimized to reduce animal suffering and were approved by the Commission for Animal Experimentation, University of Amsterdam, DEC protocol # DED204 and DED250.
1. Coverslip Preparation
2. Coverslip Coating
The coverslip coating procedure favors neuron attachment to the glass surface and dendritic arborization 14.
3. Removal of Brains from E16-E19 Rat Embryos
4. Dissection of the Hippocampi
It is very important that the dissection is done as quickly as possible in sterile conditions to ensure cell viability. Keep the samples cold on ice.
5. Cell Dissociation and Plating
6. Rat Hippocampal Primary Neuron Transfection using Lipofectamine
On DIV 14-15 neurons are transfected using the following protocol:
7. Immunostaining and Mounting of Rat Hippocampal Primary Neurons
To improve fluorescence intensity in transfected cells, perform an immunostaining protocol to enhance GFP detection 48 hr after transfection.
8. Dendritic Spine Imaging using Structure Illumination Microscopy
Dendritic spine imaging using the SIM system described in the materials has a lateral resolution (XY) value of approximately 85-110 nm and an axial (Z) resolution value between 200 – 250 nm, providing a factor of 2 times improvement in resolution compared to wide-field microscopy.
NOTE: Dendritic spine imaging using SIM is done typically 2 days after step 7.22, but could be done up to 3 weeks later if samples are kept in the dark and under a controlled temperature of 22 – 23 °C.
Described here is a standardized working protocol for imaging dendritic spines from rat primary hippocampal neurons in vitro using SIM. The protocol workflow and its crucial steps are shown in Figure 1. Overall, the protocol takes approximately 2 weeks of experimental work separated in a first phase of sample preparation, including culture, development and transfection of rat primary hippocampal neurons and immunohistochemistry, and second phase of sample imaging using SIM. The rat primary hippocampal neurons are fixed approximately 2 weeks after start of the culture, when neurons have developed complex dendritic arbors bearing numerous dendritic spines 3,5. Using the protocol described in detail in sections 1-8, it is possible to systematically image dendritic spines with super resolution. In comparison with a conventional dendritic spine imaging method using confocal fluorescence microscopy that is described before 16,17, the protocol described herein using SIM provides significantly better image resolution and 3D reconstruction, allowing the identification and classification of neuron membrane protrusions ranging from early filipodia to spine-like structures.
Figure 1. The scheme shows a protocol workflow, its steps and timing.
Figure 2. Representative micrographs of dendrites and dendritic spines imaged with confocal (A) and SIM microscopy (B). The representative SIM micrograph was acquired as described in section 8, the representative confocal micrograph was acquired using physical pinhole size: 30 μm laser power 2.6 mW, no averaging. Acquisitions were reconstructed from confocal and SIM images (C and D, respectively) using NeuronStudio software as described before 16. Dendrites were traced and spines were classified automatically with the software after adjusting main parameters such as neck length, neck diameter and head diameter. The boxed areas show one individual dendritic spine imaged with confocal (A) and SIM (B) and reconstructed from their corresponding Z stacks (C and D), depicting differences in resolution and accuracy of the resulting 3D reconstructions. According to SIM’s higher resolution, quantitative analysis of both head (E) and neck (F) diameter reveals that SIM measures significantly smaller dimensions than confocal microscopy for the same dendritic spines, indicating that SIM is capable of detecting smaller changes in dendritic spine morphology. Data in E and F are normalized to the reference confocal measurements. Results are presented as mean ± SD of 3 dendritic spines extracted from 5 dendritic segments imaged in both confocal and SIM microscope modes. For neck diameter there was a significant difference (** p = 0.0049) between confocal (100.0 ± 4.296 normalized units) and SIM (50.61 ± 7.642 normalized units) images, as tested with a Students t-test (E). For head diameter there was a significant difference (* p = 0.0209) between confocal (100.0 ± 6.255 normalized units) and SIM (58.12 ± 9.451 normalized units) images, as tested with a Students t-test.
Figure 3. Screenshot from the Nikon’s NiS Elements 6.14 SIM software package with the settings described in this protocol.
Figure 4. Screenshot from NeuronStudio 3D reconstruction and spine classification software of a representative image of a dendrite.
Table 1. List of reagents. Please click here to view a larger version of this figure.
In this article a working protocol to image dendritic spines from rat primary hippocampal neurons cultured in vitro using SIM is described. The primary hippocampal neuron culture method is an adaptation of the original method described by Kaech and Banker 18. The main differences are the use of Neurobasal/B27 culture medium, which eliminates the requirement of astroglial feeder cultures, and the addition of the mitotic inhibitor FUDR on day 3 which promotes neuronal survival while suppressing glial proliferation, as described by Brewer et al 19.
There are many critical steps in the protocol. The thickness of the coverslips used to plate the cells is crucial for an accurate SIM experiment. Sterility during coverslip preparation and coating is important. Do not let poly-l-lysine-treated coverslips dry during 2.3 and 2.4. The diameter of the flame-polished pipette used in step 5.8 is crucial. Too narrow a tip will result in low cell viability at later stages. Isolation of the hippocampi as quickly as possible will ensure high cell viability. The timing of the incubation and trypsin concentrations are crucial to ensure high cell viability. Loss of trypsin enzymatic activity may also affect cell viability. Finally, the addition of FUDR in step 5.14 is crucial to promote neuron survival and inhibit glial proliferation.
The sample imaging phase is straightforward when performed following strictly the protocol described here and results in the acquisition of super resolution images that can be readily reconstructed in 3D to analyze and classify dendritic spines according to their morphological features. As shown in Figure 2, the quality of the images acquired in the SIM mode is substantially better than images of exactly the same dendritic segments and individual spines acquired using the confocal mode of the same microscope. This result suggests that the use of SIM could provide an excellent opportunity to image more than subtle changes in dendritic spine morphology, as quantified in Figures 2E and 2F.
So far, mostly (fluorescent) wide-field microscopy has been used to image live cells, due to its low phototoxicity. Similarly, due to its low phototoxic effects and good combination with conventional (genetic) fluophores, SIM allows live cells imaging and identification of dendritic spines in low fluophore expressing cells at super-resolution. In comparison to other super-resolution microscopy methods such as STED or PALM, SIM provides a quick and affordable method for the imaging of dendritic spines from rat primary hippocampal neurons in vitro. Although in practice SIM only increases resolution by two fold compared to conventional confocal microscopy.
One example of a limitation of the SIM is that it relies on fluorescence, which in some experimental setups can be difficult to apply. To this end, microscopy techniques which do not rely on fluorescence such as electron microscopy may provide a possible solution. Nevertheless, electron microscopy in particular is a tedious, expensive an slow method. Moreover, electron microscopy can only be carried out on fixed samples. Therefore, SIM is more suitable for super-resolution imaging of live cells. The rationale for applying a fluorescent protein encoding plasmid transfection is that it results in a scarce, yet reproducible cytoplasmic labeling of isolated cells, preventing overlap of dendrites from different cells and the identification of individual dendritic spines. Combination of transfection with immunostaining has been shown previously to enhance fluorescence 17. Nevertheless, other fluorescent techniques could also be applicable to the imaging of dendritic spines with SIM, for example sufficient fluorescent staining could be acquired using recently developed actin binding probes 20.
Since recent technical developments have allowed the application of SIM to dynamic cell imaging, demonstrating that high-speed structured-illumination microscope is capable of 100-nm resolution at frame rates up to 11 Hz 13, a very logical future application of the protocol described herein could be its application to time-lapse SIM of live rat primary hippocampal neurons and the analysis of fast dynamic changes in dendritic spine morphology. A next challenge for this application could be the expected cumulative phototoxicity associated with long time intervals of live microscopy.
The authors have nothing to disclose.
This work was financed by a VIDI grant number H64.09.016 from The Netherlands Organization for Scientific Research (NWO) to CPF. CPF is grateful to Dr. Silvina A. Fratantoni for her critical comments and corrections on the final manuscript. GMRDL/EMMM are supported by the Dutch Technology Foundation STW (project 12151 and 11350), which is part of the NWO, and which is partly funded by the Ministry of Economic Affairs. We thank the Catherine Kitts and Peter Drent of Nikon Instruments Europe BV for assistance and support. HX was supported by the Royal Dutch Academy of Arts and Sciences (grant 11CDP10) and WT was supported by grant the Netherlands Organization for Scientific research (grant 820.02.006).
Fine forceps | |||
Big forceps | |||
Fine scissor | |||
Big scissor | |||
Blunt spatula | |||
Dissecting microscope with illumination | |||
Light microscope | |||
37 °C water bath | |||
Laminar flow cell culture hood | |||
High-temperature dry-oven | |||
Bunsen burner | |||
Cell culture incubator (5% CO2, 37 °C) | |||
Microcentrifuge | |||
Orbital shaker | |||
Nikon structured illumination microscope setup consisting of: | |||
Nikon Eclipse Ti research inverted microscope with Perfect Focus System | |||
Nikon CFI Apo TIRF 100x oil objective lens (N.A. 1.49) | |||
4 Coherent Sapphire Lasers (458, 488, 514 and 561 nm exitation wavelength) | |||
SIM Illuminator | |||
Nikon Stage Controller | |||
MCL Nano-Drive piezo controller | |||
Nikon Intensilight C-HGFIE mercury lamp | |||
SIM Microscope Enclosure temperature control | |||
Andor EM-CCD Camera iXon DU897 | |||
PC with Microsoft Windows 7 Home Edition | |||
Nikon’s NiS Elements 6.14 SIM software package | |||
Nikon type A immersion oil |