Primary rat hippocampal neurons were supplied by Dr. George Lykotrafitis and Shiju Gu of the Biomedical Engineering Department at the University of Connecticut (Storrs, CT, USA). The Abberior STED instrument housed in the Advanced Light Microscopy Facility in the Center for Open Research Resources and Equipment was acquired with NIH grant S10OD023618 awarded to Christopher O&#39;Connell. This research was funded by NIH grant R01AG065879 awarded to Nathan N. Alder.

Unveiling Mitochondrial Ultrastructure in Neurons Using STED Microscopy

The authors have nothing to disclose.

This protocol presents the use of human neuroblastoma cell line SH-SY5Y and primary rat hippocampal neurons with the novel IMM-targeting PKMO dye for live cell STED imaging. Due to the novelty of PKMO, there is currently little published using this dye for live STED imaging. Using these cell types for STED imaging poses challenges, specifically because neuronal cells have narrower mitochondria. One limitation of this protocol is the PKMO dye used, as it can be toxic to cells. Different cells and cell lines respond differently to the dye, thus, adjustments to dye concentration and incubation time to optimize results for strong signal without harming cells may be required. A suggested solution is to lower the concentration and increase the staining time63; however, this may result in poorer staining without increasing cell viability.
Similarly to PKMO, the commercial dye Live Orange mito (Table of Materials) also exhibits some cell toxicity. This dye was used for a variety of cultured cells but was unable to exhibit comparable staining in RA-differentiated SH-SY5Y cells successfully with the same parameters as their undifferentiated counterparts (our unpublished observations). However, amenable staining protocols may be optimized for this probe and chosen cell types. With this dye, detector gating times of 1-1.05 to 7-7.05 ns were used, with all other parameters in Table 1 remaining the same. Generally, staining cells with 200-250 nM Live Orange mito for 45 min yielded comparable results as the PKMO results shown. Higher concentration staining for less time or lower concentration staining for the same amount of time or slightly longer can yield different results and may be favorable to other cell types or growth conditions.
Imaging primary rat hippocampal neurons differs from immortalized cells due to the nature of the axon and dendrite projections as well as mitochondrial distribution at the time of imaging. One difficulty in this part of the protocol is that seeding density determines whether the primary cultures will be able to adhere and grow healthily, and at higher densities, the projections tend to overgrow by DIV 10. Therefore, the mitochondria imaged from these primary neurons will likely come from the cell body and not the projections; however, successful growth from a lower starting cell density yields better imaging results at later growth times. The key is to ensure low background and out-of-focus light to have the best contrast for STED. To address concerns regarding cell population, culturing primary hippocampal cells in B27-supplemented neuron growth media prevents the growth of glial cells, and the source reports that &#60;5% of cells are astrocytes and the absence of NbActiv1 supplement in the growth media reduces the number of astrocytes in cultures to &#60;2%87. For both cultured SH-SY5Y cells and primary rat hippocampal neurons, the PDL coating used for growth contributes to background haze in images. Sufficient signal-to-noise is accomplished with the settings reported in (Table 1) and deconvolution removes most of the background observed.
In addition to the imaging covered here, it is also possible to add treatments or stress to cells before or during imaging. For example, adding tert-butyl hydrogen peroxide (tBHP) induces oxidative stress, and it is possible to monitor changes in mitochondria over time after addition. The addition of amyloid-&#946; (A&#946;) with a fluorescent tag allows monitoring of the distribution of this peptide in relation to mitochondria as well as the mitochondrial structure over time. Mitochondrial health has been heavily implicated in AD and is widely supported to play a role in A&#946; toxicity43,71,72. Notably, the differentiation status of SH-SY5Y cells affects A&#946; protein precursor (A&#946;PP) localization85, and experiments using A&#946;PP should be carefully constructed.
As an example of how this protocol can be adapted, it is shown that the fluorescent variant A&#946;(1-42)-HiLyte 647 can be added to PKMO-stained cells 15 min before imaging (Supplementary Figure 1). The imaging parameters are similar (Supplementary Table 2), with the main difference being that a smaller pinhole is needed when imaging narrower mitochondria. Imaging A&#946;-HiLyte647 with STED requires less overall excitation (6%-8%) and STED depletion (10%-12%) laser power and fewer accumulations (six). Detector gating is also extended from 0.1 to 10 ns. Although STED resolution of A&#946; is not necessary, the overall signal-to-noise ratio and A&#946; particle size of raw STED were better than those of the confocal images, and subsequent deconvolution can also be performed. Collecting STED images and deconvolving raw STED z-stack projections of A&#946; appears particularly useful when merging with raw STED or deconvolved STED images of the PKMO stain (Supplementary Figure 1B,C). Both channels were collected in a single frame step. Measurements of time-dependent localization, similar to those listed in Figure 2 and shown in Figure 5, where applicable, and cristae architecture differences can be obtained following stress treatment or other additions.
Other possible methods for dual-labeling in live cell STED of mitochondria not reported here but have been reported by others include the use of SNAP-tagged proteins93, Halo-tagged proteins, and the use of other cell-permeable dyes with generic targets, such as mtDNA63. Notably, the labeling strategy of SNAP and Halo tagging influences the resulting fluorescence signal intensity and longevity when imaging94. Additionally, while this protocol presents several examples of analyses that can be applied to segmented mitochondria, there are many other analyses that software packages can perform on these images.

Mitochondria are eukaryotic organelles of endosymbiotic origin that are responsible for regulating several key cellular processes, including intermediary metabolism and ATP production, ion homeostasis, lipid biosynthesis, and programmed cell death (apoptosis). These organelles are topologically complex, containing a double membrane system that establishes multiple subcompartments1 (Figure 1A). The outer mitochondrial membrane (OMM) interfaces with the cytosol and establishes direct inter-organelle contacts2,3. The inner mitochondrial membrane (IMM) is an energy-conserving membrane that maintains ion gradients stored primarily as an electric membrane potential (&#916;&#936;m) to drive ATP synthesis and other energy-requiring processes4,5. The IMM is further subdivided into the inner boundary membrane (IBM), which is closely appressed to the OMM, and protruding structures called cristae that are bound by the cristae membrane (CM). This membrane delineates the innermost matrix compartment from the intracristal space (ICS) and the intermembrane space (IMS).
Mitochondria have a dynamic morphology based on continuous and balanced processes of fission and fusion that are governed by mechanoenzymes of the dynamin superfamily6. Fusion allows for increased connectivity and formation of reticular networks, whereas fission leads to mitochondrial fragmentation and enables the removal of damaged mitochondria by mitophagy7. Mitochondrial morphology varies by tissue type8 and developmental stage9 and is regulated to allow cells to adapt to factors including energetic needs10,11 and stressors12. Standard morphometric features of mitochondria, such as the extent of network formation (interconnected vs. fragmented), perimeter, area, volume, length (aspect ratio), roundness, and degree of branching, can be measured and quantified by standard optical microscopy because the sizes of these features are greater than the diffraction limit of light (~200 nm)13.
Cristae architecture defines the internal structure of mitochondria (Figure 1B). The diversity of cristae morphologies can be broadly categorized as flat (lamellar or discoidal) or tubular-vesicular14. All cristae attach at the IBM through tubular or slot-like structures termed cristae junctions (CJs) that can serve to compartmentalize the IMS from the ICS and the IBM from the CM15. Cristae morphology is regulated by key protein complexes of the IMM, including (1) the mitochondrial contact site and cristae organizing system (MICOS) that resides at CJs and stabilizes IMM-OMM contacts16, (2) the optic atrophy 1 (OPA1) GTPase that regulates cristae remodeling17,18,19, and (3) F1FO ATP synthase that forms stabilizing oligomeric assemblies at cristae tips (CTs)20,21. In addition, the IMM is enriched in nonbilayer phospholipids phosphatidylethanolamine and cardiolipin that stabilize the highly curved IMM22. Cristae are also dynamic, demonstrating morphological changes under various conditions, such as different metabolic states23,24, with different respiratory substrates25, under starvation and oxidative stress26,27, with apoptosis28,29, and with aging30. Recently it has been shown that cristae could undergo major remodeling events on a timescale of seconds, underscoring their dynamic nature31. Several features of cristae can be quantified, including dimensions of structures within individual cristae (e.g., CJ width, crista length, and width) and parameters that relate individual crista to other structures (e.g., intra-cristae spacing and cristae incident angle relative to the OMM)32. These quantifiable cristae parameters show a direct correlation with function. For instance, the extent of mitochondrial ATP production is positively related to the abundance of cristae, quantified as cristae density or cristae number normalized to another feature (e.g., cristae per OMM area)33,34,35. Because IMM morphology is defined by nanoscale features, it comprises mitochondrial ultrastructure, which requires imaging techniques that provide resolution greater than the light diffraction limit. As described below, such techniques include electron microscopy and super-resolution microscopy (nanoscopy).
The neural and glial cells of the central nervous system (CNS) are particularly reliant on mitochondrial function. On average, the brain constitutes only 2% of the total body weight, but utilizes 25% of the total body glucose and accounts for 20% of body oxygen consumption, making it vulnerable to impairments in energy metabolism36. Progressive neurodegenerative diseases (NDs), including Alzheimer&#39;s disease (AD), amyotrophic lateral sclerosis (ALS), Huntington&#39;s disease (HD), multiple sclerosis (MS), and Parkinson&#39;s disease (PD), are some of the most extensively studied pathologies to date, with research efforts ranging from understanding the molecular underpinnings of these diseases to seeking potential therapeutic prevention and interventions. NDs are associated with increased oxidative stress originating in part from reactive oxygen species (ROS) generated by the mitochondrial electron transport chain (ETC)37, as well as altered mitochondrial calcium handling38 and mitochondrial lipid metabolism39. These physiological alterations are accompanied by noted defects in mitochondrial morphology that are associated with AD40,41,42,43,44, ALS45,46, HD47,48,49, MS50, and PD51,52,53. These structural and functional defects can be coupled by complex cause-effect relationships. For example, given that cristae morphology stabilizes OXPHOS enzymes54, mitochondrial ROS are not only generated by the ETC, but they also act to damage the infrastructure in which the ETC resides, promoting a feed-forward ROS cycle that enhances susceptibility to oxidative damage. Furthermore, cristae disorganization has been shown to trigger processes such as mitochondrial DNA (mtDNA) release and inflammatory pathways connected to autoimmune, metabolic, and age-related disorders55. Therefore,&#160;analysis of&#160;mitochondrial structure is key to a full understanding of NDs and their molecular underpinnings.
Popular methods of viewing cristae, including transmission electron microscopy, electron tomography and cryo-electron tomography (cryo-ET), and X-ray tomography, in particular cryo-soft X-ray tomography, have revealed important findings and work with a variety of sample types56,57,58,59,60. Despite recent advancements toward better observation of organellar ultrastructure, these methods still come with the caveat of requiring sample fixation and, therefore, cannot capture real-time dynamics of cristae directly. Super-resolution fluorescence microscopy, particularly in the forms of structured illumination microscopy (SIM), stochastic optical reconstruction microscopy (STORM), photoactivated localization microscopy (PALM), expansion microscopy (ExM), and stimulated emission depletion (STED) microscopy, have become popular ways of viewing structures requiring resolution below the diffraction limit that constrains classical methods of optical microscopy. When ExM is used in conjunction with another super-resolution technique, the results are impressive, but the sample must be fixed and stained in a gel61. By comparison, SIM, PALM/STORM, and STED have all been successfully used with live samples, and new and promising dyes that generally stain the IMM provide a novel and easy approach for live imaging of mitochondria cristae dynamics62,63,64,65,66. Recent advancements in live dyes for STED imaging have improved dye brightness and photostability, and these dyes target the IMM at a higher degree of specificity than their predecessors. These developments allow the collection of long-term timelapse and z-stack experiments with super-resolution imaging, opening the door to better live cell analysis of mitochondrial ultrastructure and dynamics.
Herein, protocols for live cell imaging of undifferentiated and differentiated SH-SY5Y cells stained with the PKmito Orange (PKMO) dye using STED63 are provided. The SH-SY5Y cell line is a thrice subcloned derivative from the parental cell line, SK-N-SH, generated from a bone marrow biopsy of metastatic neuroblastoma67,68,69,70. This cell line is a commonly used in vitro model in ND research, particularly with diseases such as AD, HD, and PD, in which mitochondrial dysfunction is heavily implicated10,43,71,72,73. The ability to differentiate SH-SY5Y cells into cells with a neuron-like phenotype through manipulating culture media has proven a suitable model for neuroscience research without relying on primary neuronal cells10,74. In this protocol, retinoic acid (RA) was added to the cell culture medium to induce the differentiation of SH-SY5Y cells. RA is a vitamin A derivative and has been shown to regulate the cell cycle and promote the expression of transcription factors that regulate neuronal differentiation75. A protocol for culturing and live cell imaging of neurons isolated from the rat hippocampus is also provided. The hippocampus has been shown to be affected by mitochondrial degeneration and, along with the cortex, plays an important role in aging and ND76,77,78,79,80.

<table border="1" fo:keep-together.within-page="1" fo:keep-with-next.within-page="always">
	<tbody>
		<tr >
			<td >0.05% Trypsin-EDTA</td>
			<td >Gibco</td>
			<td >25300054</td>
			<td ></td>
		</tr>
		<tr >
			<td >0.4% Trypan blue</td>
			<td >Invitrogen</td>
			<td >T10282</td>
			<td ></td>
		</tr>
		<tr >
			<td >0.5% Trypsin-EDTA, no phenol red</td>
			<td >Gibco</td>
			<td >15400054</td>
			<td ></td>
		</tr>
		<tr >
			<td >100 X antibiotic-antimycotic</td>
			<td >Gibco</td>
			<td >15240062</td>
			<td ></td>
		</tr>
		<tr >
			<td >100 X/1.40 UPlanSApo oil immersion lens</td>
			<td >Olympus</td>
			<td ></td>
			<td >Equipped in Olympus IX83 microscope for STED setup described in Section 4</td>
		</tr>
		<tr >
			<td >All-trans-retinoic acid</td>
			<td >Sigma</td>
			<td >R2625</td>
			<td ></td>
		</tr>
		<tr >
			<td >Amyloid-&#946; (1-42, HiLyte Fluor647, 0.1 mg)</td>
			<td >AnaSpec</td>
			<td >AS-64161</td>
			<td >Other fluorescent conjugates available</td>
		</tr>
		<tr >
			<td >B27 supplement (50 X), serum free</td>
			<td >Gibco</td>
			<td >17504044</td>
			<td ></td>
		</tr>
		<tr >
			<td >Cell Counter (Countess II FL)</td>
			<td >Life Technologies</td>
			<td >AMQAF1000</td>
			<td ></td>
		</tr>
		<tr >
			<td >Centrifuge</td>
			<td >Eppendorf</td>
			<td >5804-R</td>
			<td ></td>
		</tr>
		<tr >
			<td >Counter slides</td>
			<td >Invitrogen</td>
			<td >C10283</td>
			<td ></td>
		</tr>
		<tr >
			<td >Conical tubes (15 mL)</td>
			<td >Thermo Fisher Scientific</td>
			<td >339650</td>
			<td ></td>
		</tr>
		<tr >
			<td >Cuvettes (Quartz Cells)</td>
			<td >Starna Cells, Inc.</td>
			<td >9-Q-10</td>
			<td >Used with Spectrometer as described in Section 1.3</td>
		</tr>
		<tr >
			<td >DMEM (high glucose with sodium pyruvate)</td>
			<td >Gibco</td>
			<td >11995073</td>
			<td >Used for SH-SY5Y cell materials as described in Section 1</td>
		</tr>
		<tr >
			<td >DMEM (high glucose no sodium pyruvate)</td>
			<td >Gibco</td>
			<td >11965092</td>
			<td >Used for primary cell materials as described in Section 2</td>
		</tr>
		<tr >
			<td >DMEM (phenol red-free)</td>
			<td >Gibco</td>
			<td >31053028</td>
			<td >Used for imaging as described in Section 3</td>
		</tr>
		<tr >
			<td >DMSO</td>
			<td >Sigma</td>
			<td >D8418</td>
			<td ></td>
		</tr>
		<tr >
			<td >DNAase I from bovine pancreas</td>
			<td >Sigma</td>
			<td >DN25</td>
			<td >Used for primary cell materials as described in Section 2.2.1 and 2.2.2</td>
		</tr>
		<tr >
			<td >DPBS (no calcium, no magnesium)</td>
			<td >Gibco</td>
			<td >14190144</td>
			<td ></td>
		</tr>
		<tr >
			<td >E18 Rat Hippocampus</td>
			<td >Transnetyx Tissue</td>
			<td >SDEHP</td>
			<td ></td>
		</tr>
		<tr >
			<td >Ethanol (200 proof)</td>
			<td >Fisher Bioreagents</td>
			<td >BP28184</td>
			<td ></td>
		</tr>
		<tr >
			<td >Fetal bovine serum (FBS), not heat-inactivated</td>
			<td >Gibco</td>
			<td >26140079</td>
			<td >For cultured cells, in Section 1</td>
		</tr>
		<tr >
			<td >Fetal bovine serum (heat inactivated)</td>
			<td >Gibco</td>
			<td >10082147</td>
			<td >For primary cell culture, Section 2</td>
		</tr>
		<tr >
			<td >Filter sterilization unit (0.1 &#181;m, 500 mL)</td>
			<td >Thermo Fisher Scientific</td>
			<td >5660010</td>
			<td ></td>
		</tr>
		<tr >
			<td >FIJI (Is Just ImageJ) and Trainable Weka Segmentation (TWS) plug-in</td>
			<td >--</td>
			<td >--</td>
			<td >Free, open-source image analysis software that includes plug-ins including Trainable Weka Segmentation described in Section 5; TWS plug-in from ref. 90 of the main text</td>
		</tr>
		<tr >
			<td >GlutaMAX supplement (100 X)</td>
			<td >Gibco</td>
			<td >35050061</td>
			<td >Glutamine supplement used for primary cell materials described in Section 2.1.2</td>
		</tr>
		<tr >
			<td >Hausser Scientific bright-Line and Hy-Lite Counting Chambers</td>
			<td >Hausser Scientific</td>
			<td >267110</td>
			<td ></td>
		</tr>
		<tr >
			<td >HBSS (no calcium, no magnesium)</td>
			<td >Gibco</td>
			<td >14170120</td>
			<td >Used for primary cell materials described in Section 2.2.1 and 2.2.2</td>
		</tr>
		<tr >
			<td >HEPES</td>
			<td >Gibco</td>
			<td>15630080</td>
			<td ></td>
		</tr>
		<tr >
			<td >Huygens Professional deconvolution software (V. 20.10)</td>
			<td >Scientific Volume Imaging (SVI)</td>
			<td >--</td>
			<td >The deconvolution software used in this protocol and described in Section 5</td>
		</tr>
		<tr >
			<td >IX83 inverted microscope with Continuous Autofocus</td>
			<td >Olympus</td>
			<td >--</td>
			<td >This paper uses a STED Infinity Line system built around an Olympus IX83 inverted microscope, described in Section 4</td>
		</tr>
		<tr >
			<td >Lightbox software (V. 16.3.16118)</td>
			<td >Abberior</td>
			<td >--</td>
			<td >Vendor software used for STED image acquisition, described in Section 4</td>
		</tr>
		<tr >
			<td >Live Orange Mito dye</td>
			<td >Abberior</td>
			<td >LVORANGE-0146-30NMOL</td>
			<td >Live cell imaging IMM-targeting dye described in Discussion</td>
		</tr>
		<tr >
			<td >Neurobasal media</td>
			<td >Gibco</td>
			<td >21103049</td>
			<td >Used for primary cell materials referred to in Section 2.1.2</td>
		</tr>
		<tr >
			<td >Nunc Lab-Tek II 2-well chambered coverglass</td>
			<td >Nunc</td>
			<td >155379</td>
			<td >Can purchase a variety of chambers but make sure the coverglass is #1.5</td>
		</tr>
		<tr >
			<td >Pasteur Pipets (Fisherbrand)</td>
			<td >Thermo Fisher Scientific</td>
			<td >22183632</td>
			<td ></td>
		</tr>
		<tr >
			<td >Penicillin-Streptomycin (10,000 U/mL)</td>
			<td >Gibco</td>
			<td >15140122</td>
			<td ></td>
		</tr>
		<tr >
			<td >PKmito Orange dye</td>
			<td >Spirochrome</td>
			<td >SC053</td>
			<td ></td>
		</tr>
		<tr >
			<td >Poly-D-lysine</td>
			<td >Gibco</td>
			<td >A3890401</td>
			<td ></td>
		</tr>
		<tr >
			<td >SH-SY5Y Cell line</td>
			<td >ATCC</td>
			<td >CRL2266</td>
			<td ></td>
		</tr>
		<tr >
			<td >Sodium pyruvate (100 mM)</td>
			<td >Gibco</td>
			<td >11360070</td>
			<td >Used for primary cell materials described in Section 2</td>
		</tr>
		<tr >
			<td >Spectrometer (GENESYS 180 UV-Vis)</td>
			<td >Thermo Fisher Scientific</td>
			<td >840309000</td>
			<td ></td>
		</tr>
		<tr >
			<td >STED Expert Line microscope</td>
			<td >Abberior</td>
			<td >--</td>
			<td >STED setup can be customized, but at time of purchase instrument was considered Abberior&#8217;s Expert Line; brief description of setup is available in Section 4 of protocol</td>
		</tr>
		<tr >
			<td >T25 flask (TC-treated, filter cap)</td>
			<td >Thermo Fisher Scientific</td>
			<td >156367</td>
			<td >Other culture vessels and sizes available</td>
		</tr>
	</tbody>
</table>

using live cell sted imaging to visualize mitochondrial inner membrane ultrastructure in neuronal cell models

Mitochondria play many essential roles in the cell, including energy production, regulation of Ca2+ homeostasis, lipid biosynthesis, and production of reactive oxygen species (ROS). These mitochondria-mediated processes take on specialized roles in neurons, coordinating aerobic metabolism to meet the high energy demands of these cells, modulating Ca2+ signaling, providing lipids for axon growth and regeneration, and tuning ROS production for neuronal development and function. Mitochondrial dysfunction is therefore a central driver in neurodegenerative diseases. Mitochondrial structure and function are inextricably linked. The morphologically complex inner membrane with structural infolds called cristae harbors many molecular systems that perform the signature processes of the mitochondrion. The architectural features of the inner membrane are ultrastructural and therefore, too small to be visualized by traditional diffraction-limited resolved microscopy. Thus, most insights on mitochondrial ultrastructure have come from electron microscopy on fixed samples. However, emerging technologies in super-resolution fluorescence microscopy now provide resolution down to tens of nanometers, allowing visualization of ultrastructural features in live cells. Super-resolution imaging therefore offers an unprecedented ability to directly image fine details of mitochondrial structure, nanoscale protein distributions, and cristae dynamics, providing fundamental new insights that link mitochondria to human health and disease. This protocol presents the use of stimulated emission depletion (STED) super-resolution microscopy to visualize the mitochondrial ultrastructure of live human neuroblastoma cells and primary rat neurons. This procedure is organized into five sections: (1) growth and differentiation of the SH-SY5Y cell line, (2) isolation, plating, and growth of primary rat hippocampal neurons, (3) procedures for staining cells for live STED imaging, (4) procedures for live cell STED experiments using a STED microscope&#160;for reference, and (5) guidance for segmentation and image processing using examples to measure and quantify morphological features of the inner membrane.

Author Spotlight: Decoding Mitochondrial Aging 

Using Live Cell STED Imaging to Visualize Mitochondrial Inner Membrane Ultrastructure in Neuronal Cell Models

Our research program focuses on structural and functional aspects of mitochondria, particularly as they relate to aging related dysfunction. To this end, we use a broad array of biophysical, biochemical, and structural techniques to explore the basic mechanisms of mitochondrial aging and to develop therapeutic interventions to preserve and restore mitochondrial health. Mitochondrial function and structure are inextricably related.Standard light microscopy techniques are suitable for measuring large scale mitochondrial features like basic morphology and network structures. Analyzing mitochondrial ultra structure or features that require higher resolution than afforded by traditional optical microscopy is typically done using electron microscopy or tomography on fixed samples. Super resolution microscopy is a fluorescence based imaging technique that measures features beyond the diffraction limit.For mitochondria, this includes measurements of nano scale protein distribution and cristae architecture. The stead imaging protocol described here allows researchers to measure the detailed structure of the inner membrane in living cells. Live cell imaging allows us to observe and quantitatively analyze complex features of cristae under physiologically relevant conditions without the need for sample fixation.This will give researchers novel insights into the dynamic changes that occur in ultra structural features and their temporal responses to stressors and pharmacological compounds.

This protocol describes cell growth conditions for cultured and primary cells with a focus on live cell STED imaging and subsequent analyses of mitochondrial cristae. Projections made with ImageJ of mitochondria from undifferentiated SH-SY5Y (Figure 3A) and RA-differentiated SH-SY5Y (Figure 3B) cells can be collected as z-stacks with traditional confocal and STED, and the raw STED images can then be deconvolved. Timelapse imaging can also be performed and subsequently deconvolved (Figure 3C,D). Using slightly different imaging parameters for primary rat hippocampal neurons (Table 1), confocal and raw STED images can be acquired as z-stacks, and the raw STED images can be deconvolved (Figure 4A). Timelapse imaging of mitochondria from primary neurons is also possible (Figure 4B). In general, the time-lapse images should be able to show mitochondrial dynamic events.
When raw STED and deconvolved STED z-stack projections from the samples used for segmentation appear consistent, quantitative measurements are performed. The TWS plugin uses the deconvolved STED image to segment to make a probability mask, which is then used to create a binary mask of the cristae to obtain size and shape parameters (Figure 5A). The regions from this mask are saved in the ROI manager and can be applied to the raw STED image if desired to measure differences in relative intensity. The deconvolved STED projections can also be used to determine the cristae periodicity and density in a given area (Figure 5B).
<img alt="Figure 1" class="xfigimg" src="/files/ftp_upload/65561/65561fig01.jpg" /> 
Figure 1: Mitochondrial morphology. Mitochondria have a two-membrane system that defines different subcompartments (A). Cristae are infolds of the inner membrane with defined features (B). Abbreviations: OMM, outer mitochondrial membrane; ICS, intracristal space; IMS, intermembrane space; CM, cristae membrane; IBM, inner boundary membrane; IMM, inner mitochondrial membrane; CT, cristae tip; CJ, cristae junction. <a href="https://www.jove.com/files/ftp_upload/65561/65561fig01large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 2" class="xfigimg" src="/files/ftp_upload/65561/65561fig02.jpg" /> 
Figure 2: Schematic of workflow. SH-SY5Y cells or primary rat hippocampal neurons are grown on a PDL-coated coverglass. SH-SY5Y cells are grown in parallel to remain undifferentiated or subjected to RA differentiation over the course of six days. Primary rat hippocampal neurons were grown on a PDL-coated coverglass after being isolated from hippocampal sections for seven days. Once ready to be imaged, cells were stained with PKMO and imaged with STED. Raw STED images are then deconvolved, and the deconvolved images are processed in FIJI to obtain size and shape measurements, such as cristae density, area, perimeter, circularity, and aspect ratio. <a href="https://www.jove.com/files/ftp_upload/65561/65561fig02large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 3" class="xfigimg" src="/files/ftp_upload/65561/65561fig03.jpg" /> 
Figure 3: Imaging of mitochondria in SH-SY5Y cells. Representative confocal (left), raw STED (center), and Huygens deconvolved STED image z-stack projections (right) of mitochondria from non-differentiated (A) and&#160;RA-differentiated (B) SH-SY5Y cells with PKMO staining are shown. A timelapse with 30 s intervals and 5 iterations of RA-differentiated SH-SY5Y cells is shown (C) with selected regions (white boxes) expanded upon (D) using scaled images of those regions without interpolation. Scale bars: A,B, 250 nm; C,D, 1 &#181;m. <a href="https://www.jove.com/files/ftp_upload/65561/65561fig03large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 4" class="xfigimg" src="/files/ftp_upload/65561/65561fig04.jpg" /> 
Figure 4: Imaging of mitochondria in primary rat hippocampal neurons. Representative confocal (left), raw STED (center), and Huygens deconvolved STED (right) image z-stack projections of mitochondria from primary rat hippocampal neurons are shown (A). A timelapse with 25 s intervals and 5 iterations of mitochondria in these neurons is shown (B). Scale bars: A, 250 nm; B, 1 &#181;m. <a href="https://www.jove.com/files/ftp_upload/65561/65561fig04large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 5" class="xfigimg" src="/files/ftp_upload/65561/65561fig05.jpg" /> 
Figure 5: Processing of deconvolved STED images in ImageJ. Representative use of the Trainable Weka Segmentation plugin to measure cristae size and shape is shown (A). From left to right, the following images are shown: the deconvolved STED image, the probability map based on segmentation from the TWS plugin, the mask from thresholding in FIJI using the probability map as input, the mask with the ROIs outlined, and the ROIs overlayed onto the original deconvolved STED image. The resulting area, perimeter, circularity and aspect ratio measurements corresponding to these objects can be found in Supplementary Table 1.&#160;A line plot using the deconvolved STED image to measure peak-to-peak distances as a readout for cristae density is shown (B). Scale bars: 0.5 &#181;m. <a href="https://www.jove.com/files/ftp_upload/65561/65561fig05large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<table border="1" fo:keep-together.within-page="1" fo:keep-with-next.within-page="always">
	<tbody>
		<tr>
			<td></td>
			<td>Pixel size (nm)</td>
			<td>Dwell time (&#181;s)</td>
			<td>Line acc.</td>
			<td>561 nm excitation during STED acquisition (%)</td>
			<td>775 nm STED depletion power (%)</td>
			<td>Step size (nm)</td>
			<td>Pinhole (AU)</td>
			<td>Timelapse interval (s)</td>
			<td>Timelapse iterations</td>
		</tr>
		<tr>
			<td>Undiffer-entiated SH-SY5Y</td>
			<td>20</td>
			<td>4</td>
			<td>10</td>
			<td>15</td>
			<td>20</td>
			<td>200</td>
			<td>1.0</td>
			<td>30</td>
			<td>5</td>
		</tr>
		<tr>
			<td>RA-Differe-ntiated SH-SY5Y</td>
			<td>20-25</td>
			<td>4</td>
			<td>10-12</td>
			<td>15</td>
			<td>20-22</td>
			<td>150-200</td>
			<td>1.0</td>
			<td>30</td>
			<td>5</td>
		</tr>
		<tr>
			<td>Primary Neurons</td>
			<td>20-25</td>
			<td>4</td>
			<td>10</td>
			<td>10</td>
			<td>25</td>
			<td>200</td>
			<td>0.7</td>
			<td>30</td>
			<td>5</td>
		</tr>
		<tr>
			<td colspan="10">NOTE: Pixel size can vary based on imaging requirements and intent to deconvolve images. Proper sampling is required for deconvolution. Pixel sizes for raw STED images without deconvolution can go up to 30 nm.</td>
		</tr>
	</tbody>
</table>
Table 1: Summary of STED acquisition parameters. The settings used for 2D STED imaging for each cell type, undifferentiated SH-SY5Y, RA-differentiated SH-SY5Y, and primary rat hippocampal neurons, are displayed. For all time-lapses, 5 iterations were taken with varying intervals based on ROI size.
Supplementary Figure 1: Imaging of SH-SY5Y cells with amyloid-&#946; (A&#946;) addition. Representative confocal (left), raw STED (center), and deconvolved STED (right) images of RA-differentiated SH-SY5Y cells with PKMO stain (top) and A&#946;-HiLyte647 (bottom) are shown (A). Merged z-stack projections of raw PKMO STED (green) with raw&#160;A&#946; STED (magenta) (B) or deconvolved PKMO STED (green) with deconvolved A&#946; STED (magenta) (C) are shown. Scale bars: 0.5 &#181;m. <a href="https://www.jove.com/files/ftp_upload/65561/Suppl Fig.pdf" target="_blank">Please click here to download this File.</a>
Supplementary Table 1: Size and shape measurements of segmented cristae. The size and shape measurements of area (&#181;m2), perimeter (&#181;m), circularity, and aspect ratio, corresponding to the objects outlined in Figure 5A from segmented mitochondria, are shown. <a href="https://www.jove.com/files/ftp_upload/65561/65561_Supplementary Table 1.xlsx" target="_blank">Please click here to download this File.</a>
Supplementary Table 2: Summary of acquisition parameters with amyloid-&#946; samples. The settings used for 2D STED imaging of PKMO and A&#946;-HiLyte647 in undifferentiated and RA-differentiated SH-SY5Y cells are displayed. Confocal of A&#946;-HiLyte647 may be used alone as there is no specific structure to resolve; STED images of A&#946;-HiLyte647 are shown here for smaller particle sizes. <a href="https://www.jove.com/files/ftp_upload/65561/65561_Supplementary Table 2.xlsx" target="_blank">Please click here to download this File.</a>
Supplementary File 1: Amyloid-&#946; treatment protocol. <a href="https://www.jove.com/files/ftp_upload/65561/65561_Supplementary File 1.docx" target="_blank">Please click here to download this File.</a>

Watch this Scientific Journal Video about Decoding Mitochondrial Aging at JoVE.com

This protocol presents a workflow for the propagation, differentiation, and staining of cultured SH-SY5Y cells and primary rat hippocampal neurons for mitochondrial ultrastructure visualization and analysis using stimulated emission depletion (STED) microscopy.

Mitochondria play many essential roles in the cell, including energy production, regulation of Ca2+ homeostasis, lipid biosynthesis, and production of ...

Using Live Cell STED Imaging to Visualize Mitochondrial Inner Membrane Ultrastructure in Neuronal Cell Models

Abstract