1. Cell Biology

Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue

  1. Theodore J Zwang  Is a corresponding author
  2. Rachel E Bennett
  3. Maria Lysandrou
  4. Benjamin Woost
  5. Anqi Zhang
  6. Charles M Lieber
  7. Douglas S Richardson
  8. Bradley T Hyman
  1. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, United States
  2. Harvard Medical School, United States
  3. Massachusetts Alzheimer’s Disease Research Center, United States
  4. Department of Chemical Engineering, Stanford University, United States
  5. Department of Chemistry and Chemical Biology, Harvard University, United States
  6. Department of Molecular and Cellular Biology and Harvard Center for Biological Imaging, Harvard University, United States
https://doi.org/10.7554/eLife.84112
Share this article
Cite this article
  1. Theodore J Zwang
  2. Rachel E Bennett
  3. Maria Lysandrou
  4. Benjamin Woost
  5. Anqi Zhang
  6. Charles M Lieber
  7. Douglas S Richardson
  8. Bradley T Hyman
(2023)
Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue
eLife 12:e84112.
https://doi.org/10.7554/eLife.84112
  2. Download BibTeX
  3. Download .RIS

Abstract

Difficulty achieving complete, specific, and homogenous staining is a major bottleneck preventing the widespread use of tissue clearing techniques to image large volumes of human tissue. In this manuscript, we describe a procedure to rapidly design immunostaining protocols for antibody labeling of cleared brain tissue. We prepared libraries of 0.5–1.0 mm thick tissue sections that are fixed, pre-treated, and cleared via similar, but different procedures to optimize staining conditions for a panel of antibodies. Results from a library of mouse tissue correlate well with results from a similarly prepared library of human brain tissue, suggesting mouse tissue is an adequate substitute for protocol optimization. These data show that procedural differences do not influence every antibody-antigen pair in the same way, and minor changes can have deleterious effects, therefore, optimization should be conducted for each target. The approach outlined here will help guide researchers to successfully label a variety of targets, thus removing a major hurdle to accessing the rich 3D information available in large, cleared human tissue volumes.

Editor's evaluation

This valuable Tools and Resources paper presents a solid workflow for testing and comparing variations in tissue clearing, antigen retrieval, and antibody staining methods using thick slices of tissue. Though staining results vary sensitively with processing conditions, results from screening conditions in mouse brain tissue can be carried over to staining in human brain tissue. This solid story will be of broad interest to those carrying out immunohistochemistry experiments in human tissue samples.

https://doi.org/10.7554/eLife.84112.sa0

Introduction

In recent years, there have been substantial improvements in the techniques used to visualize features of brains, organs, and whole organisms in three dimensions. One promising method, tissue clearing, involves lipid removal and refractive index matching to significantly decrease light scattering. Tissue clearing allows for light microscopy to observe the distribution of individual molecules throughout entire organs millimeters to centimeters in size (Ueda et al., 2020b; Richardson et al., 2021). This holds extraordinary promise for understanding human diseases, where the identification and characterization of abnormal structures is often an early step toward the development of clinical treatments (Kim et al., 2016). Difficulties with antibody staining and labeling is common in traditional thin tissue section histology, but is dramatically amplified in large, cleared samples. Although techniques have been developed to improve the homogeneity of staining (SWITCH; Murray et al., 2015) and increase the speed of antibody labeling in large, cleared tissue (eFlash [Yun et al., 2019], EFIC [Na et al., 2021], MDH [Dwyer et al., 2021]), these methods only work if an antibody’s epitope is well preserved during the tissue preparation and clearing process. This is rarely the case and complete, specific and homogenous staining is now the most common bottleneck preventing the widespread use of tissue clearing in samples that cannot be modified to express fluorescent proteins such as human tissue. Therefore, most tissue clearing methods have been optimized for studies in rodents, and only a few recent studies apply clearing techniques to human tissue (Scardigli et al., 2021; Lai et al., 2018; Susaki et al., 2020; Matos et al., 2010; Zhao et al., 2020). Here, we present a simple optimization strategy and automated analysis routine that allows researchers to tune their immunohistochemistry (IHC) protocols for labeling protein targets in cleared and uncleared brain tissue. Further, we demonstrate that our optimized protocols can be extended from mouse to human tissue.

IHC has a history of being difficult, with myriad improvements taking place over decades to allow for its current ubiquity in pathology laboratories and biomarker discovery (Matos et al., 2010). The basic concept of IHC is to use antibodies to label specific antigens in a sample, which can then be imaged to determine the antigen’s distribution. Successful IHC requires four fundamentals: (1) the antibody cannot denature, (2) the epitope cannot denature, (3) the antibody must diffuse completely through the tissue, and (4) the antibody must only bind its intended target. Often researchers develop lengthy protocols that include steps thought to enable successful labeling but are not necessarily based on empirical evidence or are borrowed from others using completely different antibodies or tissues. In general, IHC protocols share several common steps: first components of the tissue must be fixed in place to preserve its structure. This is typically performed by freezing the tissue or placing it in an aldehyde containing solution like formalin. It may then be stored for long periods of time before an investigation is initiated. Next, different methods of antigen retrieval can be deployed to enhance antibody binding. The tissue may also be blocked by incubating with serum or other agents to deactivate endogenous enzymes and reduce non-specific antibody staining. Finally, when labeling and imaging thick tissue, tissue clearing protocols are required that introduce several additional chemical treatments including dehydration, hydrogel embedding, delipidation, and immersion in a solution that is refractive index matched to the processed tissue (Richardson et al., 2021; Richardson and Lichtman, 2015).

It is well understood that differences in staining quality can result from biological heterogeneity such as small differences in tissue composition, individual antibodies, their epitopes, and overall tissue structure. However, further variability can be induced by each step of an IHC protocol (Kim et al., 2016; Matos et al., 2010; Shi et al., 2011). Because of this, IHC protocols can usually be reproduced with the same tissues, antibodies, and fluorophores. Unfortunately, it is common to have difficulties extending these protocols to different antibodies or tissues. The complexity of these multi-step protocols makes it prohibitively costly and time-intensive for researchers to optimize all possible IHC conditions for their specific study. This has slowed the testing and adoption of reported improvements in IHC procedures as many researchers are forced to guess at the best protocol for their samples.

Experiments with human tissue samples are essential for a complete understanding of the pathology of many diseases. However, the necessary optimization of IHC protocols described above is often not possible in human tissue due to its scarcity. Attempting to use IHC protocols optimized for different tissues or antibodies could confound results and limit the insight gained from such precious resources. Ideally, researchers could optimize protocols in a more plentiful model species such as mice and directly transfer these methods to human samples. However, it is not often clear whether IHC improvements in mice (Murray et al., 2015; Wassie et al., 2019; Park et al., 2018) would translate to better staining of human tissue. Methods that can rapidly determine optimal IHC protocols for individual antibodies and assess their applicability to tissues from different species are needed.

Inspired by our previous experience troubleshooting IHC in large volumes of mouse brain tissue, we developed a practical workflow to rapidly optimize IHC protocols for different antibodies. We then applied this methodology to a variety of targets in mouse and human brain tissue to understand (1) how much variation exists if IHC protocols are optimized for individual antibodies and (2) if optimized IHC protocols can be generalized between species. We compare the effect of different IHC protocols on antibodies and a small molecule DNA label, 4′,6-diamidino-2-phenylindole (DAPI). We also test the influence of protocols on different antibodies raised against the same protein target and finally, explore whether IHC protocols optimized in mouse tissue can produce equally high-quality staining in human tissue. Together this work provides a method to empirically determine the optimal procedure for IHC studies in large-volume tissue, assess its generalizability, and support the use of abundant mouse tissue to optimize IHC protocols for scarce human samples.

Results

Strategy to rapidly assess the optimization of protocols

Our workflow for determining an optimal IHC protocol is inspired by the success of quantitative structure-activity relationship (QSAR) modeling for drug discovery. QSAR models are used to predict the biological activities of untested drugs by first testing a library of drugs with similar but varied structures, measuring how each drug affects the activity, then creating a model that relates the structural changes to the observed activities (Muratov et al., 2020). Borrowing from this strategy and making a library of tissues that all went through similar processing steps but for a few key changes allows for us to predict the optimal staining protocol without needing to test every possible combination of conditions.

The use of large intact organs such as whole mouse brains is unnecessary and impractical for optimizing an IHC protocol. However, if the tissue sections are too thin, the impact of antibody diffusion cannot be fully appreciated and compared across protocols; whereas, if the sections are too thick, extensive incubation times are required. Through empirical testing, we found tissue sections 0.5–1.0 mm thick and antibody incubation times of 18–24 hr to be sufficient to assess the influence of our chosen conditions on IHC staining quality (Figure 1, Figure 2, Figure 3—figure supplement 1), and allows dozens of protocols or multiple antibodies to be tested with tissue from a single mouse brain depending on the epitope distribution. It should be noted that these data suggest the signal when staining 24 hr and under may be limited by the depletion of antibodies on their way through the tissue, which should be considered when making comparisons of staining quality.

Figure 1
Download asset Open asset
Process for generating a library of tissue prepared with different protocols.

Start with a baseline protocol (see Materials and methods). Choose conditions to test that can be separated into categories believed to influence staining quality. Prepare tissue by varying only one or two conditions from the baseline at a time. Each protocol should only vary by a single condition from at least one other protocol to allow for a complete understanding of how each change influences staining. So resulting immunohistochemistry (IHC) quality can isolate the effect of each condition change. Stored library of cleared tissues then allows for rapid antibody testing to quickly determine optimal IHC conditions for each antibody as well as protocols that allow for multiplexed IHC. We consistently found conditions in red to have exclusively negative effects on IHC or tissue integrity, so they were excluded from testing with most antibodies.

Figure 2
Download asset Open asset
Quantifying the staining quality for comparison across protocols and antibodies.

Immunohistochemistry (IHC) quality comparisons can be difficult to determine by eye due to differences in fluorescence intensity and background, as well as the many ways that software can modify the visual representation of data, so we created a pipeline to quickly quantify staining quality in a way that allows rapid comparison. (a) First, a confocal or lightsheet microscope is used to image a Z-stack throughout the entire thickness of the sample. All imaging parameters are kept identical for a given antibody/fluorophore to allow for comparison between protocols. (b) Fluorescence values were normalized across the sample, then a rolling-ball algorithm was used in Imaris software to generate surfaces around fluorescent features with intensity above a threshold value. Alternatively, segmentation masks were generated using Ilastik following their pixel classifier workflow. (c) The voxels within the surface were indexed to define the array of voxels with and without staining that could then be analyzed in Matlab. (d) The staining quality was quantified and visualized by plotting the ratio of voxels with staining over the total number of voxels in a plane at that depth in the image, the average signal intensity within the segmented region, and the average noise intensity outside of the segmented region (Figure 3—figure supplements 2 and 3, Figure 7—figure supplements 1 and 2). A higher proportion of indexed voxels, higher signal, lower noise, and longer depth to reach half-max staining intensity all represent better staining quality. This can also be used to provide visual representation of reproducibility (Figure 3—figure supplements 4–6). It is also possible to reduce these numbers into a figure of merit that can be quantitatively compared between conditions (Figure 3c, Figure 3—figure supplement 7).

Using QSAR modeling as a guide, we selected 17 conditions to test in our proof-of-principle studies (Figure 1). These conditions were divided across four standard IHC/tissue clearing steps: fixation, blocking/unmasking, delipidation, and antibody incubation. Five conditions were later omitted, and multiple 0.5–1.0 mm thick tissue sections were assigned to 12 protocols to create the final tissue library (Figure 1). With the library complete, a tissue sections from each condition could be selected for further head-to-head analysis in our image analysis pipeline that outputs a quantifiable metric of quality (see below). Using these metrics, we were able to predict the optimal combination of IHC conditions for each antibody we tested from initial fixation through to final labeling. During this proof-of-principle study, we only tested high-quality antibodies that were already known to have high specificity and limited off-target binding for IHC in thin tissue slices.

Measuring IHC quality in cleared tissue

To compare the quality of IHC in the cleared tissue libraries we developed a data analysis pipeline to quickly extract the distribution of fluorescence in multiple samples and derive performance parameters. First, we used one of two different software (Imaris or Ilastik) to generate image segmentation masks. For analysis with Imaris, we isolated objects based on the intensity of fluorescence relative to nearby background (Figure 2). For analysis with Ilastik, we leveraged machine learning algorithms that allow users to classify pixels and supervise the generation of the segmentation mask (Berg et al., 2019). Either of these segmentation masks are then run through code in Matlab that extracts the signal, noise, signal-noise ratio, area of staining, and the distance into the tissue at which the staining intensity drops by half. IHC quality can then be made by comparing these features across samples. It should be noted that simply relying on depth to half-max staining does not separate the issues of free antibody penetration speed through the tissue nor the extent of antibody depletion along the way, and it is possible that conditions that reduce the antigenicity may significantly enhance the apparent penetration rate.

This analysis does not require staining of images to be perfect nor precise but requires features that can at least be separated via pixel classification. For images with unintended background staining, Ilastik can be trained to separately classify intended and unintended binding. Note that Imaris is also capable of performing similar pixel classification by integrating Labkit, which we did not compare.

After acquiring segmentation results, turning it into an ideal figure of merit will vary depending on the antigen of interest and its distribution in tissue. For antigens that have relatively constant distribution over the depth of imaging, the area of each plane within a Z-stack that is segmented could be used as a figure of merit for staining quality across conditions. However, if the antigen distribution does not have a constant distribution, it is better to use a figure of merit that is normalized based on the area of segmentation, such as the average staining intensity at each depth plane.

Conditions that influence antibody and epitope preservation

Starting from a comprehensive collection of published IHC protocols (Richardson et al., 2021; Kim et al., 2016; Scardigli et al., 2021; Richardson and Lichtman, 2015), we selected a subset of commonly varied conditions to create 12 test protocols for proof of principle. Each protocol is nearly identical, only varying by a single condition from at least one other protocol to allow for a complete understanding of how each change influences staining. The baseline protocol involved minimal processing and gentle conditions. Specifically, the tissue was fixed in 4% paraformaldehyde (PFA) for 2 days post-perfusion, embedded in hydrogel made of 4% acrylamide, delipidated in sodium dodecyl sulfate (SDS) for 1 day at 37°C, rinsed in phosphate-buffered saline (PBS) with 0.2% Triton X-100 (PBST) at 4°C, and incubated with antibodies at 4°C for 18 hr. We then created a library of tissue that was processed using small changes to this baseline protocol, so we could compare how each of these changes influenced IHC quality. This library included changes in the composition of the hydrogel (with and without 4% PFA), changes in delipidation time (1, 2, 3 days), additional steps before antibody incubation (such as unmasking with 50°C or incubation with glycine), and differential changes in antibody incubation temperature (4°C or room temperature). In addition to this library, we also examined the effects of boiling samples in citrate buffer before or after delipidation, heating to temperatures above 70°C before immunolabeling, and removal or decrease of PFA concentration in the transcardial perfusion or post-perfusion fixing step (Figure 1). However, these conditions were universally observed to negatively impact IHC or tissue integrity, so they were excluded from our final library.

We chose two antibodies commonly used in neuroscience studies of cleared volumes (Glut1 and NeuN) and performed IHC on the same tissue sections from a mouse cortex that were processed according to the 12 protocols (Figure 1, Figure 3). Each antibody was directly conjugated to a fluorophore: Glut1 to Alexa Fluor 488 and NeuN to Alexa Fluor 568. Increasing delipidation time to 2 days had a positive influence on NeuN staining, and no influence on Glut1 staining, but both showed worse IHC when increasing the delipidation time further. Incubation of tissue at 50°C for 1 hr prior to antibody incubation improved staining for Glut1 and NeuN. Glycine treatment had no significant effect on NeuN staining, but inhibited Glut1 staining. Incubating antibodies at room temperature improved NeuN staining and on Glut1 staining. Intriguingly, staining with DAPI appears insensitive to the different protocols (Figure 4). Inclusion of PFA in the hydrogel resulted in decreased IHC quality for both Glut1 and NeuN, however this significantly improved IHC of dopamine receptor 2 (Figure 4—figure supplement 1). Therefore, we ascertained the optimal protocol for simultaneously staining Glut1, NeuN, and DAPI was to delipidate for 2 days, place the tissue at 50°C for 1 hr before incubating with antibodies, and incubating the antibodies at room temperature. Importantly, these same conditions would result in no staining of dopamine receptor 2 (Figure 4—figure supplement 1). Instead, an IHC protocol is required to use PFA in the hydrogel step to show adequate staining of dopamine receptor 2, which can still allow for simultaneous staining of NeuN, Glut1, and DAPI if the tissue is only delipidated for 1 day and incubated with antibodies at 4°C. Together, these results show that different conditions can have opposite effects on the quality of staining and depth of antibody penetration and reinforce the importance of optimizing sample preparation conditions for each antibody.

Figure 3 with 9 supplements see all
Download asset Open asset
Influence of common conditions on immunohistochemistry in mouse brain tissue.

(a) Representative maximum intensity projection views from the top and side of 500 µm thick, 5 mm wide tissue sections. C57Bl/6J mouse isocortex was processed, stained, and imaged simultaneously with NeuN (green, left panels) and Glut1 (purple, right panels). (b) Quantitative comparison of immunohistochemistry labeling across conditions using Imaris. A rolling-ball subtraction was used to determine the area with signal-to-noise above cut-off for every image plane in the Z-stack. The proportion of indexed voxels in each plane was then divided by the maximum value across all conditions to generate a normalized staining quality score, which allows for quick visual assessment of staining quality differences with depth and between conditions. (c) Quantitative comparison of the depth at which the immunohistochemistry (IHC) staining intensity becomes half of the maximum, using segmentation results from Ilastik. Error bars represent standard error. Two-tailed t-test comparing results to baseline were represented with p<0.05 = *, p<0.01 = **, and no significance = unlabeled. N=3 for all samples.

Figure 4 with 1 supplement see all
Download asset Open asset
Influence of common conditions on 4′,6-diamidino-2-phenylindole (DAPI) staining.

Representative maximum intensity projection views from the top and side of 500 µm thick, 5 mm wide tissue sections of C57Bl/6J mouse isocortex stained with DAPI.

Comparison of IHC conditions for different antibodies directed against a single protein

A common strategy following unsuccessful IHC in tissue is to test antibodies from different suppliers or antibodies raised against different epitopes in the desired target. To better understand whether optimal IHC protocols may be shared for antibodies against the same target, we compared the staining of a monoclonal GFAP directly conjugated to a fluorophore and a polyclonal GFAP that was detected with a fluorophore conjugated to a secondary antibody. To enable a direct comparison, both sets of antibodies were used to label the same tissue sections. Immunolabeling GFAP with different antibodies shows that the trends in staining quality are similar for both antibodies across all conditions (Figure 5). This suggests that conditions that enhance GFAP stability and prevent denaturation of its epitopes have a positive effect on all GFAP antibodies provided the antibodies are of high quality, have high specificity and limited off-target binding. This also suggests that IHC of GFAP with these antibodies is limited more by denaturation of the epitope than the antibody itself. We found that both GFAP antibodies successfully labeled their target under similar conditions (Figure 5), and both showed similar conditions where no labeling occurred (Figure 5—figure supplement 1). Some differences in labeling quality are apparent between antibodies, however it is not clear whether the differences in labeling quality are due to the conjugation of the fluorophore or other differences between the monoclonal and polyclonal antibodies. As this approach also reduces the length of the staining protocol, it is recommended to use primary antibodies directly conjugated to fluorophores whenever possible.

Figure 5 with 1 supplement see all
Download asset Open asset
Immunohistochemistry of two different antibodies for GFAP.

Representative maximum intensity projection views from the top and side of 500 µm thick, 5 mm wide tissue sections of C57Bl/6J mouse isocortex stained with two different antibodies for GFAP. Monoclonal GA5 conjugated to Alexa Fluor 488 (green) was stained and imaged in the same tissue as a polyclonal antibody using a goat anti-rabbit secondary conjugated to Alexa Fluor 647 (red). Other protocols show no significant labeling (Figure 5—figure supplement 1).

Multiple antibodies that recognize tau protein were also compared including AT8 anti-tau, HT7 anti-tau, and DAKO total tau antibodies. Tau is a microtubule associated protein that is primarily expressed by neurons and is the main component of neurofibrillary tangles in Alzheimer’s disease (AD). We chose three commonly used antibodies against tau and tested them first in tissue from a transgenic mouse model, hTau (Andorfer et al., 2003), that expresses normal human tau isoforms.

There was significantly more variability between tau antibodies than seen with GFAP antibodies, primarily due to the inability for DAKO total tau antibody to effectively penetrate beyond the tissue’s surface (Figure 6, Figure 6—figure supplement 1). However, as the DAKO total tau antibody was not directly conjugated to a fluorophore, we cannot exclude the possibility that the inability to image DAKO-stained tau beyond the surface may be an effect of poor penetration and diffusion of the secondary antibody. Importantly, when used in traditional IHC in microtome slices of human tissue, all three of these antibodies labeled similar structures (Figure 6—figure supplement 2). As expected, similar trends in optimal IHC protocols were observed for AT8 and HT7 anti-tau antibodies in mouse tissue expressing human tau (Figure 6—figure supplement 3) and showed no labeling in a negative control; mice without tau expression (Figure 6—figure supplement 4).

Figure 6 with 4 supplements see all
Download asset Open asset
Tau immunohistochemistry in cleared human tissue.

Representative maximum intensity projection views from the top and side of 500 µm thick, 5 mm wide tissue sections. Other protocols showed worse staining quality (Figure 6—figure supplement 2).

Mouse brain tissue as a model system for optimizing IHC in cleared human tissue

Optimizing IHC in human tissue is especially difficult due to the limited amount of tissue available for experimentation. Therefore, we performed IHC optimization in both mouse and human brain tissue for individual antibodies and analyzed their similarities. Surprisingly, our testing of IHC protocols for NeuN and Glut1 in human tissue showed similar changes in staining quality and depth when compared to mouse tissue (Figure 7, Figure 7—figure supplements 1 and 2). Both human and mouse tissue showed decreases in staining quality with increasing clearing time and fixation with PFA. Room temperature antibody incubation as well as epitope unmasking at 50°C for 1 hr prior to antibody incubation improved Glut1 and NeuN staining in both mouse and human tissue. It is also worth noting that this NeuN antibody appears to accumulate off-target in the human tissue, which is not observed in the mouse tissue.

Figure 7 with 2 supplements see all
Download asset Open asset
Influence of common conditions on immunohistochemistry in human brain tissue.

(a) Representative maximum intensity projection views from the top and side of 500 µm thick, 5 mm wide tissue sections. Tissue was processed, stained, and imaged simultaneously with NeuN (green, left panels) and Glut1 (purple, right panels). (b) Quantitative comparison of immunohistochemistry labeling across conditions using Imaris. A rolling-ball subtraction was used to determine the area with signal-to-noise above cut-off for every image plane in the Z-stack. The area in each plane was then divided by the maximum value across all conditions to provide a staining quality score, which allows for quick visual assessment of staining quality differences with depth and between conditions. (c) Quantitative comparison of the depth at which the immunohistochemistry (IHC) staining intensity becomes half of the maximum, using segmentation results from Ilastik. Error bars represent standard error. Two-tailed t-test comparing results to baseline were represented with p<0.05 = *, p<0.01 = **, and no significance = unlabeled. N=3 for all samples.

The optimal protocol that allows NeuN, Glut1, and AT8-tau staining was then determined by finding the conditions that allow for the simultaneous IHC of all three antibodies in a single tissue section, and then finding the combination of conditions among those that maximizing the staining quality. It is worth noting that this was not necessarily the best protocol for each individual antibody. It is more important to have adequate staining of each epitope than to have fantastic staining of one at the cost of another. As described above, we determined the optimal protocol for multiplexed labeling by combining the protocol changes that quantitatively improve the staining quality, defined by half-max depth, for each antibody above the baseline protocol and do not significantly degrade performance of other antibodies. For NeuN, Glut1, and AT8 antibodies this protocol involves clearing for 2 days, using 50°C to unmask epitopes, and incubation of antibodies at room temperature. The optimal protocol was then validated in 1 mm thick pieces of human brain tissue, approximately 100-fold thicker than typical paraffin embedded sections. Antibodies for tau in mice expressing human tau show similar variability across conditions for the labeling of labeling tau in human tissue (Figure 6). We concurrently incubated the human tissue with three antibodies conjugated to three different fluorophores (NeuN-A568, Glut1-A488, and AT8-A647). These antibodies were used to stain tissue that was first embedded in a 4% acrylamide hydrogel, cleared for 2 days and thoroughly washed. The antibodies were incubated with the tissue for 14 days to ensure complete staining through the whole sample in both humans (Figure 8) and mice (Figure 8—figure supplement 1). It should be noted that we found 3 days of antibody incubation sufficient for consistent and thorough staining across brain regions in 500 µm mouse tissue (Figure 3—figure supplements 8 and 9), and there were no notable downsides to leaving the antibody incubating longer. By carefully following the optimal IHC protocol determined in this work, we were able to achieve complete penetration and simultaneous labeling of the 1 mm thick human tissue with all three antibodies. Both the Glut1-labeled vasculature and AT8-labeled tau are extremely well defined and could be easily quantitated. The NeuN reflects the same quality that it had during the troubleshooting but does not appear to be an ideal antibody for assessment of human neurons under these conditions.

Figure 8 with 1 supplement see all
Download asset Open asset
Immunohistochemistry in 1 mm thick human tissue.

The tissue was crosslinked without paraformaldehyde (PFA), cleared for 2 days, rinsed, then the antibodies NeuN-A568 (red), Glut1-A488 (green), and AT8-A647 (white) were incubated with the tissue for 14 days. (a) Top-down view of fluorescence from 1 mm section human brain. (b) Side view of each antibody fluorescence through entire section depth. (c) A 200 µm virtual section of human tissue, from a 1 mm thick sample. (d) Zoomed-in regions from the full images show that AT8-labeled tau tangles are present at multiple locations throughout the human brain sections.

Discussion

Although it is well known that alterations to IHC protocols can have detrimental effects to the staining results, many researchers accept the dogma that a universal IHC protocol exists that is robust enough to support labeling with a wide range of antibodies. When a subset of antibodies fail in a multi-target IHC protocol, often the antibodies themselves are blamed. Here, we have found that small changes to an IHC protocol can have profound effects on labeling efficiency. Rather than the antibody, our data suggests that poor staining is most often due to denaturation or blockage of the epitope. These procedural differences do not influence every antibody-antigen pair in the same way, so the optimal protocols for IHC should be determined for each individual target. Using combinations of antibodies in the same tissue may require using a protocol with conditions that are acceptable for all antibodies, rather than conditions that are optimal for each individual antibody. Unfortunately, as we found with the dopamine receptor 2 (Figure 4—figure supplement 1) and GFAP (Figure 5, Figure 5—figure supplement 1) IHC, a protocol may not exist to allow for both targets to be effectively stained simultaneously. However, the methods described here allow a researcher to rapidly determine (1) an optimal protocol for single antibody staining, (2) an acceptable protocol that allows simultaneous multi-antibody staining, or (3) that cyclical single-antibody staining and stripping will be required to match conditions necessary for each antibody.

Importantly, temperature, fixation conditions, and the other conditions investigated here have greater influence, the larger the tissue volume becomes. For many of the tested protocols, the difference in staining is not apparent at the surface which suggests that the use of tissue sections that are too thin does not allow for appropriate optimization. Many of the protocols we tested for NeuN or Glut1 produce some degree of labeling at the surface of the sample that would have appeared no different between samples thinner than 50 µm, despite obvious differences in staining quality at depth.

Most surprising to us was that mouse tissue is a good model system for determining large-volume IHC protocols for human samples. However, these results show that minor procedural changes can disrupt the quality of staining, so it is important that tissue preparations are identical.

IHC protocols in large-volume tissue that requires clearing can take weeks or months to complete depending on the initial state and size of the tissue (Susaki et al., 2020; Ueda et al., 2020a). Previous studies have suggested that human tissue requires processing for longer than comparable-sized rodent tissue. Our results suggest that this is not due to intrinsic differences between human and mouse brain tissue, but rather a result of differences in the state of the tissue when clearing experiments are started. Human tissue is often stored in fixative for extended periods of time and, as our data shows, even a small increase in PFA exposure can have significant influence on IHC quality and depth. To our surprise, tissue that was delipidated for 3 days showed less rapid IHC penetration than tissue that was cleared for 2 days in all cases except for DAPI labeling. This suggests that samples can quickly become over-cleared and lose antigenicity which may extend the amount of time previously reported as necessary for IHC. It follows that future experiments should explore labeling tissue with antibodies with even shorter or milder fixation conditions and, where possible, carried out prior to any tissue clearing procedures.

Our methodology enables rapid optimization of IHC protocols, including protocols that require tissue clearing and provides a simple path to imaging tissue blocks that are 100 times thicker than conventional tissue slices. First, a library of cleared tissue samples that were each processed using different conditions is generated that can be stored for future experiments. Then, optimal IHC conditions are determined for new antibodies within a few days (approximately 24 hr for staining, 24 hr for rinsing, 12–24 hr for refractive index matching/clearing, 3 hr to image the samples, and a similar amount of time to quantify and compare the staining quality). The variability in staining quality across protocols suggests that researchers should use this process to test how changes may affect antibodies or epitopes that they plan to study before committing many weeks or months to a procedure in larger volumes. Additionally, these results speak to the necessity for scientists to publish both conditions that improve and degrade specific antibody performance. Understanding which procedures and antibodies are appropriate for a given sample will be essential for tissue clearing techniques to gain widespread use in human tissue. We are optimistic that our optimization process will allow researchers use these techniques to successfully label cleared human tissue, which will improve the adoption of this technique and the ability to unlock new insight into human pathology and neuroanatomy.

Materials and methods

Human tissue samples

Request a detailed protocol

Human tissue was provided by the Massachusetts Alzheimer’s Disease Research Center (ADRC) with approval from the Mass General Brigham IRB (1999P009556). Three human participants with AD were selected from the Massachusetts Alzheimer’s Disease Research Center. Sex, age at death, Braak staging, which hemisphere, and comorbidities are listed in Supplementary file 1. Autopsy tissue from human brains were collected at Massachusetts General Hospital, with informed consent of patients or their relatives and approval of local institutional review boards. Human tissue was received unfrozen immediately after autopsy and fixed (sample 2620) or frozen immediately after autopsy then thawed prior to fixation (samples 1581 and 1762). Tissue was placed in 4% PFA for 48 hr at 4°C, transferred to PBS for 24 hr at 4°C before slicing. This tissue was treated identically to the mouse tissue described below in the tissue slicing procedure and onward.

Mice strains

Request a detailed protocol

Male and female C57Bl/6J mice (Jackson Laboratory 000664) were used for most tissue experiments. For tissue to test IHC of tau, 10- to 12-month-old male hTau mice bred from a local colony (B6.Cg-Mapttm1(EGFP)Klt Tg(MAPT)8cPdav/J, Jackson Laboratory 005491) that express only human tau isoforms were used. hTau knockout mice bred from a local colony (Mapttm1(EGFP)Klt/J, Jackson Laboratory 004779) were used as controls for tau IHC experiments. Mouse experiments were performed with approval of local institutional review boards (IACUC protocol # 2019N000026).

Perfusion

Request a detailed protocol

Mice were transcardially perfused with ice-cold 40 ml PBS, followed by 40 ml 4% PFA (Electron Microscopy Sciences). Brains were removed from the cranium and placed in 4% PFA for 48 hr at 4°C. Brains were then transferred to PBS for 24 hr at 4°C before slicing.

Tissue slicing

Request a detailed protocol

Brains were embedded in 4% agarose hydrogel. To do this, extracted brains were placed individually in a small Petri dish and let dry slightly. The agarose solution was dissolved by boiling in a microwave, then left on the counter to cool before pouring. Agarose was poured to completely cover the top of the brain, which allows it to provide a solid support to prevent deformation during cutting. A block of agarose containing the brain was placed in a vibratome (VT1000S vibrating blade microtome, Leica, Wetzlar, Germany) then cut into 500 or 1000 µm coronal or sagittal slices. Each slice was then removed from the agarose through gentle manipulation with blunt forceps or paintbrushes before crosslinking.

Tissue hydrogel crosslinking

Request a detailed protocol

Each slice was incubated with a hydrogel crosslinking solution for 1 day at 4°C to let the molecules diffuse through the tissue. Solutions should be kept cold before and after the addition of VA-044 to prevent premature initiation of polymerization. Different crosslinking solutions were used to determine the effect of including PFA on IHC. Crosslinking solution 1: PBS with 4% (wt/vol) acrylamide (Sigma-Aldrich, St Louis, MO) and 0.25% (wt/vol) VA-044 thermal polymerization initiator (Thermo Fisher Scientific, Pittsburgh, PA). Crosslinking solution 2: PBS with 4% (wt/vol) acrylamide, 0.25% (wt/vol) VA-044 thermal polymerization initiator, and 4% PFA. After incubation with the crosslinking solution for 1 day, the tissue was left in solution and placed in vacuum at 37°C for 3 hr. After polymerization with X-CLARITY polymerization system (Logos Biosystems, South Korea), the slices were rinsed with 50 ml PBS three times over 3 hr. Each slice was then cut into small pieces with width and length at least twice the thickness. This was done to increase the number of conditions that could be tested with small amounts of tissue.

Delipidation

Request a detailed protocol

After hydrogel crosslinking, the tissue was placed in a delipidation solution (SDS 200 mM, sodium sulfite 20 mM, sodium borate 20 mM, sodium hydroxide to pH 8.5–9, filtered if necessary) at 37°C for different amounts of time (1, 2, or 3 days), to determine the effect of delipidation in IHC. After delipidation, the brain slices were rinsed with 50 ml PBST (PBS with 0.2% Triton X-100, Thermo Fisher Scientific) five times over 1 day.

Pre-IHC tissue processing

Request a detailed protocol

Different conditions were tested at this step to determine their influence on IHC. Each tissue section was tested using only one of the following conditions. Baseline: Tissue was left at 4°C in PBST. 50°C Condition: Tissue was heated to 50°C for 1 hr in PBST. Glycine: Tissue was incubated with 0.1 M glycine at 4°C for 12–18 hr. Glycerol: Tissue was placed in a low osmolarity solution of 80% glycerol/20% water for 12–18 hr. Blocking: Tissue was placed in 5% BSA in PBS for 12–18 hr.

Storage of tissue library

Request a detailed protocol

The tissue library was then used immediately or stored at 4°C in PBS with 0.02% (wt/vol) sodium azide. Mouse and human tissue was processed following all of the steps above and then used up to 3 months later. Additional mouse and human tissue was stored following delipidation, then successfully used up to 6 months later by resuming from the pre-IHC tissue processing state. IHC conditions for a given antibody were always compared using batches of tissue that were stored for the same amount of time.

Immunohistochemistry

Request a detailed protocol

  • The following antibodies were purchased then stored according to the vendors specifications: NeuN antibody conjugated to Alexa Fluor 568 (Abcam, Ab207282).

  • GFAP antibody unconjugated (Abcam, ab16997) and conjugated to Alexa Fluor 488 (Thermo Fisher, 53-9892-82).

  • Glut1 antibody conjugated to Alexa Fluor 488 (EMD Millipore, 07-1401-AF488).

  • Phospho-tau antibody AT8 conjugated to biotin (Thermo Fisher, MN1020B).

  • Tau monoclonal antibody HT7 conjugated to Alexa Fluor 647 (Thermo Fisher, 51-5916-42).

  • DAKO total tau antibody unconjugated (Agilent, A0024).

  • Dopamine receptor 2 antibody (Santa Cruz Biotechnology, SC5303)

  • Secondary antibody Goat Anti-Rabbit IgG H&L conjugated to Alexa Fluor 405 (Abcam, ab175652) and Alexa Fluor 647 (ab150079).

  • Streptavidin conjugated to Alexa Fluor 568 (Thermo Fisher, S11226).

Each brain slice was placed in a 2 ml Eppendorf tube (or similar vessel) that could hold the slice so its large, flat sides could be exposed to solution. PBST was added to just cover the top of the samples (~300 µl). One to 5 µl of stock antibody solution was added, depending on the concentration of the stock. The final concentration of each antibody in solution was 5 µg/ml. Fluorophores or host species were chosen so multiple different antibodies could be tested at the same time in a single slice of tissue. The tissue was then incubated with the antibodies at 4°C or room temperature for 18 hr. The tissue was then rinsed in PBST 3×50 ml overnight (18–24 hr) at room temperature. If using unconjugated primary antibodies, the secondary antibodies were then added to give a final concentration of 5 µg/ml and incubated for 48 hr at 4°C. The tissue was then rinsed in PBST 3×50 ml overnight at room temperature.

IHC of human microtome sections

Request a detailed protocol

Prior to preparation of cleared tissue volumes, fresh/frozen human samples were tested for the presence of AT8, HT7, and DAKO. Microtome sections were fixed in 4% PFA for 15 min, rinsed in Tris-buffered saline (TBS) three times, then blocked in TBS with 5% bovine serum albumin. Primary antibodies were incubated at a concentration of 1:500 overnight. Tissue was rinsed in TBS 3×, then incubated in secondaries at room temperature for 60 min, and rinsed in TBS 3×. Immumount was then used to seal the tissue prior to imaging.

Refractive index matching

Request a detailed protocol

The samples were incubated with 80% glycerol, 15% DI water, 5% PBS (RI-matching solution) for at least 3 hr at room temperature with gentle shaking. The RI-matching solution was then replaced with fresh solution and left to incubate for at least 3 more hr before imaging. Samples were stable stored in this solution at room temperature and could be re-imaged for at least 3 months.

Imaging

Request a detailed protocol

Samples were imaged using Olympus Inverted Confocal FV3000 with a 10× objective except for samples using dopamine receptor antibodies which were imaged using the Zeiss Lightsheet Z.1. Image Z-stacks were then reconstructed and visualized using Imaris microscopy image analysis software.

Post-imaging analysis

Request a detailed protocol

Prior to analysis, images were assigned a random number so the analyzing scientist was blind to tissue clearing conditions but still aware of the antibodies used in the tissue, which was necessary to identify staining.

Analysis with Imaris

Request a detailed protocol

Imaris software was used to assess the staining quality score. First, the distribution of fluorescence values across all slices in the Z-stack were normalized. Then, a rolling-ball algorithm was applied to define the edges of fluorescent shapes with signal intensity greater than background by a threshold value. The area within each slice of a Z-stack that contained staining above this threshold was then divided by the maximum across all tissue stained with that fluorophore to determine the staining quality throughout the tissue. This metric was only applied to stains like NeuN or Glut1 in the cortex where we can make the assumption that there is a uniform and consistent distribution of these cells across sections. All samples that were imaged as part of the library described in Figure 1 were included in this study, with at least three independent samples per condition to determine reproducibility.

Analysis with Ilastik

Request a detailed protocol

Raw data was converted to HDF5 format using ilastik’s ImageJ plugin plus custom macro to batch convert. The staining was then segmented using their pixel classifier workflow (Berg et al., 2019). In short, a paintbrush was used to draw over the signal and background to help train the classifier on how to segment each image. Real-time segmentation results were then improved using a minimum of 600 image planes from six different staining conditions that reflect the distribution of staining quality in the entire set. All images were then batch processed through the trained pixel classifier and exported as multipage tiff.

Post-image segmentation analysis in Matlab

Request a detailed protocol

The image segmentation output from either Imaris or Matlab was saved as a single-channel multipage Tiff and placed into a folder, and all the raw images were saved as a single-channel multipage Tiff and placed into a different folder. Image conversion was done using ImageJ macros. Quantification and generation of data for plotting can then be done by specifying the data path, ensuring the Matlab code is on the path, then running the code. The main code ‘QuantifyIHC.mat’ should be run first, which goes through the files in each folder in order and determines the segmentation area, staining within the segmentation, and noise outside of the segmentation. Next, running the ‘FindHalfHax.mat’ code will find the starting plane for plotting, the half-max depth, place all data for plotting into arrays, allow for the re-ordering of blinded data, and plotting of the heatmaps.

Code availability

Request a detailed protocol

All Matlab code and ImageJ macros used to quantify data can be accessed at https://github.com/tjzwang/IHC, (copy archived at swh:1:rev:64acafda08ebdda1021edfe47e3d328b21218aa6; Zwang, 2022).

Data availability

All data used in publication can be accessed at the BioImage Archive https://www.ebi.ac.uk/biostudies/studies/S-BIAD479. All Matlab code and ImageJ macros used to quantify data can be accessed at https://github.com/tjzwang/IHC, (copy archived at swh:1:rev:64acafda08ebdda1021edfe47e3d328b21218aa6).

The following data sets were generated
    1. Zwang TJ
    2. Bennett RE
    3. Lysandrou M
    4. Woost B
    5. Zhang A
    6. Lieber CM
    7. Richardson DS
    8. Hyman BT
    (2022) ebi
    ID S-BIAD479. Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue.
    https://www.ebi.ac.uk/biostudies/bioimages/studies/S-BIAD479

References

Decision letter

  1. Albert Cardona
    Reviewing Editor; University of Cambridge, United Kingdom
  2. Suzanne R Pfeffer
    Senior Editor; Stanford University, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting the paper "Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Comments to the Authors:

We are sorry to say that, after consultation with the reviewers, we have decided that this work will not be considered further for publication by eLife at this stage.

Specifically, the reviewers noted and detailed a lack of clarity in key points of the protocol, and a rather large variability in the staining outcomes. Additionally, the reviewers were concerned about a few key assumptions regarding the properties of the subset of antibodies tested, and premature generalization of the results from analysis of these antibodies.

Overall, the work has a lot of potential, and we encourage resubmission of the story when the reviewer comments can be adequately addressed.

The complete and detailed reviewer evaluations are below.

Reviewer #1 (Recommendations for the authors):

Immunohistochemistry (IHC) encompasses a complex set of methods with many possible variations on conditions for tissue fixation, clearing, antigen retrieval, blocking, and staining. Choices of these conditions that are optimized for one antibody/target pair often do not generalize to others, or to other tissue types. As high throughput microscopy (particularly lightsheet microscopy) and clearing methods allow for imaging thicker blocks of tissue, these problems are increasingly a limiting factor in experimental studies. A more systematic understanding of how these various aspects of IHC methods choices should be made is of great interest. Optimizing IHC conditions for a given protein target in a given tissue is currently somewhat of an art. A more systematic approach to carrying out this optimization is also potentially of great interest. This would be particularly helpful for human tissue, as it is too precious to routinely test many IHC method variants.

The authors set out to create a streamlined, standardized tissue processing and image analysis workflow for comparing different tissue preparations in terms of IHC performance. They create a library of mouse brain tissue specimens processed with several variants of each processing step including fixation, clearing, and antigen retrieval. They vary only one parameter at a time to facilitate direct comparison between conditions. They set up a data analysis pipeline that is extremely simple and easy to use but probably underpowered for comparing quantitatively between conditions because it keeps only a signal mask and discards signal intensity information. It is also not clear how the pipeline should be modified to accommodate targets with different types of staining pattern. The results are presented as a 1D heatmap that varies with depth in the specimen, rather than a simple line plot, making it difficult to compare quantitatively between conditions.

Each of 12 process variants is carried out for two antibodies on three replicate specimens to assess repeatability. The authors do not discuss this experiment in detail, but there is high variability between replicates. For example, in Figure S1, the 2 days cleared case has fairly high signal in two replicates and almost none in the middle one. The staining solutions were apparently also not well mixed, as the antibody signals on the two sides of each specimen are very different in all cases, and the amount of difference between the two sides is also highly varying between specimens. From Figure 1, it is also clear that the signal varies across each specimen within the imaging plane. In this comparison of 12 conditions, one result pops out: incubating the antibodies at room temperature (as opposed to the default of 4oC) improves penetration. This effect is dramatic for NeuN. It is probably also present for Glut1, although it is necessary to look at Figure S2, showing all three replicates, to see this clearly. Every other method variant appears to result in worse performance except that in the case of 2 day clearing with Glut1 staining, one replicate has a particularly high signal area (Figure S2).

They then test two antibodies against GFAP and show that varying processing conditions affects both antibodies similarly, though they do not include any quantification. In contrast, they find that two antibodies against tau do vary differently with processing conditions, but again without quantification. Next, they carry out the same screen in human brain tissue that had previously been done in mouse brain tissue in order to assess the suitability of mouse tissue as a proxy for IHC quality in human tissue. The trends of IHC quality as a function of method variant appear to be qualitatively preserved between mouse and human. They finally show that three antibodies (primaries directly conjugated with fluorophores) against Glut1, tau, and NeuN fully penetrate through a 1 mm slice of human brain tissue after incubating for 14 days.

The goal of putting IHC on a more quantitative footing is an admirable and timely one, but the approach presented here requires some more work to deliver on this substantial promise.

Tissue library: Tissue thickness and staining time choices should be justified more carefully. Of particular concern is the antibody staining time, as the optimal time depends strongly on the temperature of incubation. Staining should be quantified at several time points to determine the extent of saturation at the tissue surface. The assumption that antigens have a "relatively constant distribution with the depth of imaging" should also be supported with data. The same goes for lateral variation. Variation between replicates also would ideally be reduced. From the methods, it seems as though the antibody is simply added to the staining buffer with the tissue, which might indicate there is insufficient mixing before and during antibody incubation.

Image analysis: A more sophisticated but still highly user-friendly segmentation strategy such as the Ilastik software package might be a more reliable and versatile choice for analyzing image data. Furthermore, the signal intensity distribution within the masked signal area can be extracted (e.g. in Matlab) to quantitatively compare processing conditions, in addition to the area occupied by the mask as used here. The area outside the mask could potentially be treated as a background segment, whose intensity distribution can also be quantified. This way, quantitative measures of both signal intensity and background staining levels could be used for a more full comparison between method variants. The results should be viewed as a 1D line plot for more quantitative comparison between conditions. Finally, this lineplot could be used to measure derived performance parameters such as the distance into the tissue at which the staining intensity drops by half.

These improvements in quantitation are necessary to draw strong conclusions from any further experiments, but with them this work stands to make a strong contribution to the field.

Reviewer #2 (Recommendations for the authors):

This work presents a systematic approach to evaluation and optimization of antibody labeling in cleared human brain tissue. The authors outline the problems very clearly and present strategies to solve them. First, libraries of tissue prepared using different protocols are generated to easily test the performance of different antibodies and different staining conditions. Importantly, only one condition at a time is varied, which allows to evaluate the contribution of individual variables to the final staining quality. Second, a strategy for finding the optimal conditions for simultaneous staining with several different antibodies is outlined. And third and very important results is that mouse tissue can be used for optimization of antibody labeling of human tissue, thus allowing efficient use of human tissue which is a scarce and very valuable resource.

Strengths

Antibody labeling of tissues, and especially large volume cleared tissues involves a multitude of steps and reagents. Variations in any of the tissue preparation or immunolabeling steps can affect antibody performance and finding the optimal conditions can be a very tedious process. This paper provides clear examples of how much variability there can be even when using the same antibody on the same tissue prepared in different ways or stained under different conditions. It gives a clear solution of how this problem can be addressed by using libraries of tissues prepared by varying a single condition in each instance.

The presented approach results in a huge saving of time as it can be completed much faster (several days) compared to the complete immunolabeling protocol for large volume cleared tissue which can take several weeks.

The demonstration that mouse tissue can be used for optimization of staining conditions for human tissue is very important, allowing the best use of human tissue.

Weaknesses

Results from a small number of antibodies are extrapolated to apply to all antibodies in general. For example, the results from the comparison between 2 different GFAP antibodies, showing that the antibody conjugated to a fluorophore performs better that the antibody that is not conjugated, is used to generalize that primary antibodies directly conjugated to a fluorophore are preferrable to use compared to antibodies that are visualized using indirect immunolabeling with a secondary antibody. Even when considering the 2 GFAP antibodies this conclusion cannot be made, because it is not clear whether the difference in performance is due to the primary antibodies themselves (monoclonal vs. polyclonal) or to the method of their visualization (direct vs indirect).

The authors conclude that "IHC is limited more by denaturation of the epitope than the antibody itself". This conclusion is based on the experiments that show that varying the conditions of tissue preparation and staining can dramatically change antibody performance from no label to excellent label. However, the authors do not take into account that they are using antibodies that have already been shown to work reasonably well for immunolabeling.

Important methodological information is missing in some instances, for example how was the human tissue fixed or how are the tissue libraries stored.

Recommendations in order of appearance in the manuscript:

Introduction: It is very well written! It is important to specify here, as done further below in the paper (in Comparison of IHC conditions for different antibodies directed against a single protein) that the proposed strategies apply when "the antibodies are of high quality, have high specificity and limited off-target binding".

Results, Measuring IHC quality in cleared tissue: It is unclear how the threshold was determined. Another concern is that this "measure of IHC quality" is essentially a measure of what fraction of the tissue is labeled. Thus, it does not account for specific vs non-specific staining.

Conditions that influence antibody and epitope preservation: It is unclear whether the baseline protocol described here applies only to mouse tissue or to human tissue as well. The mention of perfusion leads me to believe this is specifically for mouse, but I couldn't find information on how the human tissue was fixed.

There is no mention here of whether the antibodies (GluT1 and NeuN) are directly conjugated to fluorophores. I assumed that they were, based on the info in the Methods section. But then in the last paragraph of Results the authors say that "Therefore, we directly conjugated three antibodies to three different fluorophores (NeuN-A568, Glut1- A488, and AT8-A647)." Which to me implies that they weren't conjugated in previous experiments.

Comparison of IHC conditions for different antibodies directed against a single protein: As explained in the Public review section, I disagree with the conclusion "it is recommended to use primary antibodies directly conjugated to fluorophores whenever possible." Unless the authors add more experiments that compare direct vs. indirect immmunolabeling with the same primary antibody (for several different antibodies), this conclusion is not justified.

"In order to examine how much of the observed variability in antibody labeling is due to targets being more or less accessible depending on cell type or subcellular location, multiple antibodies were also chosen that recognize tau protein." I don't understand what the comparison of the tau antibodies will tell us about the accessibility of targets depending on cell type or subcellular location.

Mouse brain tissue as a model system for optimizing IHC in cleared human tissue: Determining the optimal protocol for simultaneous staining of 3 antibodies is an important point. It will be very useful if the authors elaborate more on how this optimal protocol was determined. How was the combination of conditions that maximize staining quality established? Minor point: how was the 14 days duration determined?

Discussion: "Rather than the antibody, our data suggests that poor staining is most often due to denaturation or blockage of the epitope." I disagree with this conclusion, as explained above.

"Unfortunately, as we found with the dopamine receptor 2 and GFAP IHC, a protocol may not exist to allow for both targets to be effectively stained simultaneously." – I didn't see this presented in the Results section.

Materials and methods:

How were the human samples fixed? Were the slicing and all the following steps performed in an identical way? I assume they were, but this should be stated clearly in the methods. How were the tissue libraries stored? For how long can they be used after preparation?

Figures

Figures 1, 2 and others. How is the top of the side view determined? Is this the surface of the tissue and therefore conditions where the label in the side view starts lower that in other conditions means that the tissue surface was not labeled? And why would that be? If that is not the case, it will be best to align the side views and have the labeling start at the same level.

Figure S3. This is a very clear way of summarizing the differences in labeling. It would be very helpful to have this summary next to the actual immunostaining images in the main figures.

Figure S9 Representative microtome images taken from the human tissue prior to staining: Do you mean prior to clearing?

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue" for further consideration by eLife. Your revised article has been evaluated by Suzanne Pfeffer (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

Essential revisions:

Regarding the comparison between conjugated antibodies and primary plus secondary antibodies, it seems that it's possible to interpret the data in a different way, as outlined below by the reviewers, given that the comparison is not between the same primary antibodies with one conjugated and one not. Thus the observed differences can also be due to the primary antibodies themselves. Also, it is not clear whether a different secondary antibody would give a similar result.

On the question of antibody penetration, it is possible to interpret some of the results not as a problem of an impediment to physical penetration but of antibody depletion as the latter binds the tissue as it diffuses across.

Reviewer #1 (Recommendations for the authors):

The revised manuscript has addressed most of the reviewers' concerns and is significantly improved.Regarding the comparison between directly conjugated antibodies and use of secondary antibodies. I agree that directly conjugated antibodies save time and can be recommended for that purpose. I still have a problem with claims like: "As shown in Figure 3, the use of a primary antibody directly conjugated to a fluorophore greatly reduced background fluorescence over traditional primary/secondary antibody systems." The comparison is not between the same primary antibodies with one conjugated and one not; thus the observed differences can also be due to the primary antibodies themselves. Also, it is not at all clear whether a different secondary antibody would give a similar result. There is a great variety of secondary antibodies commercially available and, generally, the preadsorbed secondary antibodies that have undergone an extra purification step have higher specificity and give less background labeling; however the secondary antibodies used here were not of the "preadsorbed" kind.

The authors now include more details on the human tissue handling, which clarifies many questions that I had. A few remaining ones are:

Materials and methods: Human tissue samples: what was the postmortem interval before fixation? Even a rough estimate will be helpful.

Supplementary Table 1. Hemisphere column: "Frozen left". Does this mean that the human tissue was stored frozen? And then fixed? If it was indeed frozen before fixation, which was not mentioned elsewhere in the methods, that could explain why it looks like the antibodies penetrate deeper into the human tissue compared to mouse, (e.g. Figure 1 vs. Figure 5), but it is also an important difference to keep in mind.

Reviewer #2 (Recommendations for the authors):

The additions to the manuscript re-submission are greatly appreciated, in particular the updated analysis approach and 1d line profiles of signal versus tissue depth. There is one additional point that I believe is important and can be handled in the text. From the screen of antibody staining times (Figure S1), it appears that up through 24 hours of incubation the effect of increased incubation time is to increase the signal in a ~120 µm surface layer whose thickness does not change appreciably. After this, up to 48 hr (at least for RT incubation) the thickness of the surface layer changes strongly. This suggests that the limitation on signal for staining 24hr and under is depletion of antibodies on the way through the tissue. It should be pointed out that in the present work, the focus on depth to half-max staining does not separate the issues of (1) free antibody penetration speed through the tissue and (2) the extent of depletion along the way. Reducing antigenicity will improve the apparent penetration rate by reducing the effect of depletion due to epitope binding. This may be what is happening with the 50oC unmasking treatment, which increases penetration but greatly reduces the staining level (Figure 1 b,c, column 8). This limitation should be discussed explicitly.

https://doi.org/10.7554/eLife.84112.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1 (Recommendations for the authors):

[…]

Tissue library: Tissue thickness and staining time choices should be justified more carefully. Of particular concern is the antibody staining time, as the optimal time depends strongly on the temperature of incubation. Staining should be quantified at several time points to determine the extent of saturation at the tissue surface.

We have conducted additional experiments to compare room temperature and 4oC incubation of NeuN at 5 time points, plotted in figure S1. These results show both the increase in saturation at the tissue surface and within the tissue. For both incubation temperatures, the tissue appears to saturate at the surface within 24 hours but there are significant differences in the staining intensity with depth. Neither shows significant signal at a depth of 250 microns, suggesting that 500 micron thick tissue is adequate for comparing these conditions with <24 hour incubation times.

The assumption that antigens have a "relatively constant distribution with the depth of imaging" should also be supported with data. The same goes for lateral variation. Variation between replicates also would ideally be reduced.

We conducted additional experiments to compare NeuN and Glut1 staining across three different brain regions in longitudinal slices from each hemisphere of three different mice. We incubated these antibodies for 3 days to achieve saturation, then quantified the average signal, percent segmented area, average noise, and signal/noise ratio in each imaging plane. Our original metric of percent segmented area shows some variation with depth depending on the brain region, but we find that normalizing the signal intensity per segmented area allows for a quantifiable metric for staining quality that is independent on the change in distribution of antigen, so long as there is some antigen in each plane. These data show insignificantly different results across tissue samples and brain regions. Additionally, as we will describe below, Ilastik segmentation is more robust which decreased a significant amount of variation between replicates. We feel the variation between replicates is reasonable, as quantified by the Signal-Noise ratio standard error of mean values in the hippocampus (10.2±2.4%) anterior cortex (11.0±3.5%) and posterior cortex (10.3±1.8%), especially when considering these tissue samples came from different mice.

From the methods, it seems as though the antibody is simply added to the staining buffer with the tissue, which might indicate there is insufficient mixing before and during antibody incubation.

We have clarified in the methods section that the antibody was mixed in the staining buffer, and that the tissue was placed on a gentle shaker during antibody incubation for all conditions.

Image analysis: A more sophisticated but still highly user-friendly segmentation strategy such as the Ilastik software package might be a more reliable and versatile choice for analyzing image data.

We attempted to use Ilastik, as recommended, and found it to be an extremely robust solution for analyzing our data. Before, using Imaris, we did not have access to the batch processing module due to its cost. This meant that we had to go through each file and individually generate the segmentation mask for each channel, which was a significant amount of time for the hundreds of imaging files used in this work. Ilastik allowed for us to train the segmentation on a subset of our data, then apply the segmentation to all of the similar images. Additionally, this did not require the normalization step we had applied to our data in Imaris, further improved the segmentation of poorly stained samples, and allowed for the isolation of nonspecific staining that may have confounded analysis. We are extremely grateful for this suggestion.

Furthermore, the signal intensity distribution within the masked signal area can be extracted (e.g. in Matlab) to quantitatively compare processing conditions, in addition to the area occupied by the mask as used here. The area outside the mask could potentially be treated as a background segment, whose intensity distribution can also be quantified. This way, quantitative measures of both signal intensity and background staining levels could be used for a more full comparison between method variants. The results should be viewed as a 1D line plot for more quantitative comparison between conditions. Finally, this lineplot could be used to measure derived performance parameters such as the distance into the tissue at which the staining intensity drops by half.

We appreciate these very constructive and insightful suggestions. As suggested, we performed this analysis on four of our sets of data (Mouse NeuN, Mouse Glut1, Human NeuN, and Human Glut1) consisting of 142 image volumes as a proof-of-principle. These quantifications provide some additional insight into the difference between conditions. Most interesting is the significant difference in the percent staining area in each imaging plane was more changed than the signal intensity across conditions, which provides support for our claim that IHC is significantly influenced by denaturation or masking of the epitope. The antibody is having the same overall effectiveness for signal intensity per unit area, but is binding to less. Additionally, the performance parameter suggested by the reviewer (which we call half-max depth in the manuscript) showed reproducible differences across samples and translated well between mouse and human tissue. We included this plot in Figure 3 and Figure 7.

Reviewer #2 (Recommendations for the authors):

[…]

Weaknesses

Results from a small number of antibodies are extrapolated to apply to all antibodies in general. For example, the results from the comparison between 2 different GFAP antibodies, showing that the antibody conjugated to a fluorophore performs better that the antibody that is not conjugated, is used to generalize that primary antibodies directly conjugated to a fluorophore are preferrable to use compared to antibodies that are visualized using indirect immunolabeling with a secondary antibody. Even when considering the 2 GFAP antibodies this conclusion cannot be made, because it is not clear whether the difference in performance is due to the primary antibodies themselves (monoclonal vs. polyclonal) or to the method of their visualization (direct vs indirect).

Although we have used relatively few antibodies compared to the many thousands used in research, we rigorously tested each of the antibodies in this study. We have collected a large amount of data for each antibody, which includes at least three replicates from independent brains for 12 different conditions. Additionally, each replicate includes approximately 100 imaging planes with 2-4 channels, depending on the stains that were tested. We feel confident that these data strongly support our main claims that small changes in conditions lead to significant variation in IHC quality, which we see among the antibodies we tested. Additionally, the variation in IHC quality for a dozen different conditions is consistent between mice and human tissue that are prepared similarly. These results are what allowed for us to find conditions that produce the beautiful images in figure 8 that show simultaneous staining of multiple antibodies in human tissue. That said, we did not rigorously test the use of conjugated primary antibodies compared to use with secondary antibodies with multiple antibodies, so we have removed those claims which we feel are a minor point in the original manuscript.

The authors conclude that "IHC is limited more by denaturation of the epitope than the antibody itself". This conclusion is based on the experiments that show that varying the conditions of tissue preparation and staining can dramatically change antibody performance from no label to excellent label. However, the authors do not take into account that they are using antibodies that have already been shown to work reasonably well for immunolabeling.

We believe that the new analyses suggested by reviewer 1 help support our claim that IHC may be limited more by the denaturation of the epitope than the antibody itself. We agree that we are using antibodies that have already been shown to work reasonably well for immunolabeling, and have included mention of this in the second paragraph of the Results section.

Important methodological information is missing in some instances, for example how was the human tissue fixed or how are the tissue libraries stored.

We have included more details of our methodology in the Human tissue samples and Immunohistochemistry sections, as well as added a section on the Storage of tissue library

Recommendations in order of appearance in the manuscript:

Introduction: It is very well written! It is important to specify here, as done further below in the paper (in Comparison of IHC conditions for different antibodies directed against a single protein) that the proposed strategies apply when "the antibodies are of high quality, have high specificity and limited off-target binding".

Results, Measuring IHC quality in cleared tissue: It is unclear how the threshold was determined. Another concern is that this "measure of IHC quality" is essentially a measure of what fraction of the tissue is labeled. Thus, it does not account for specific vs non-specific staining.

We believe that the new methodology using ilastik addresses reviewer 2’s concerns as well. Using this new methodology there is no need for setting a threshold. It does require the users to properly train the segmentation software, however the results are updated live during training which allows for users to quickly assess its quality. Furthermore, you can train ilastik to separate non-specific staining from the signal and background which allows for quantification of non-specific staining.

Conditions that influence antibody and epitope preservation: It is unclear whether the baseline protocol described here applies only to mouse tissue or to human tissue as well. The mention of perfusion leads me to believe this is specifically for mouse, but I couldn't find information on how the human tissue was fixed.

We have included more details of this methodology in the Human tissue samples and Immunohistochemistry sections.

There is no mention here of whether the antibodies (GluT1 and NeuN) are directly conjugated to fluorophores. I assumed that they were, based on the info in the Methods section. But then in the last paragraph of Results the authors say that "Therefore, we directly conjugated three antibodies to three different fluorophores (NeuN-A568, Glut1- A488, and AT8-A647)." Which to me implies that they weren't conjugated in previous experiments.

We have included more details of this methodology in Immunohistochemistry section, as well as in the figure 3 and 7 captions.

Comparison of IHC conditions for different antibodies directed against a single protein: As explained in the Public review section, I disagree with the conclusion "it is recommended to use primary antibodies directly conjugated to fluorophores whenever possible." Unless the authors add more experiments that compare direct vs. indirect immmunolabeling with the same primary antibody (for several different antibodies), this conclusion is not justified.

As recommended by reviewer 2, we have withdrawn the conclusion that our data shows significant advantages of primary over secondary antibodies for reducing nonspecific staining. This would take a significant amount of effort for a relatively minor point, which we may be able to address in the future, however we are currently exclusively using conjugated antibodies in all our experiments. We have still included a recommendation to use conjugated antibodies where possible because it reduces the number of steps and total time for IHC, sometimes on the order of weeks depending on tissue thickness.

"In order to examine how much of the observed variability in antibody labeling is due to targets being more or less accessible depending on cell type or subcellular location, multiple antibodies were also chosen that recognize tau protein." I don't understand what the comparison of the tau antibodies will tell us about the accessibility of targets depending on cell type or subcellular location.

We have removed our statement on subcellular location, but hope that we can address this in future work.

Mouse brain tissue as a model system for optimizing IHC in cleared human tissue: Determining the optimal protocol for simultaneous staining of 3 antibodies is an important point. It will be very useful if the authors elaborate more on how this optimal protocol was determined. How was the combination of conditions that maximize staining quality established?

We have clarified our process for determining the optimal protocol from the collected tissue library data in the main text in multiple places, including the following text under Mouse Brain tissue as a model system for optimizing IHC in cleared human tissue in the Results section: “The optimal protocol that allows NeuN, Glut1, and AT8-tau staining was then determined by finding the conditions that allow for the simultaneous IHC of all three antibodies in a single tissue section, and then finding the combination of conditions among those that maximizing the staining quality. It is worth noting that this was not necessarily the best protocol for each individual antibody. It is more important to have adequate staining of each epitope than to have fantastic staining of one at the cost of another. As described above, we determined the optimal protocol for multiplexed labeling by combining the protocol changes that quantitatively improve the staining quality, defined by half-max depth, for each antibody above the baseline protocol and do not significantly degrade performance of other antibodies. For NeuN, Glut1, and AT8 antibodies this protocol involves clearing for 2 days, using 50oC to unmask epitopes, and incubation of antibodies at room temperature.”

Discussion: "Rather than the antibody, our data suggests that poor staining is most often due to denaturation or blockage of the epitope." I disagree with this conclusion, as explained above.

We believe that the new analyses suggested by reviewer 1 help support our claim that IHC may be limited more by the denaturation of the epitope than the antibody itself. We agree that we are using antibodies that have already been shown to work reasonably well for immunolabeling, and have included mention of this in the second paragraph of the Results section.

"Unfortunately, as we found with the dopamine receptor 2 and GFAP IHC, a protocol may not exist to allow for both targets to be effectively stained simultaneously." – I didn't see this presented in the Results section.

It is described in the section Conditions that influence antibody and epitope preservation in the second paragraph, specifically: “Inclusion of PFA in the hydrogel resulted in decreased IHC quality for both Glut1 and NeuN, however this significantly improved IHC of dopamine receptor 2 (figure 4-supplement 1).”

Materials and methods:

How were the human samples fixed? Were the slicing and all the following steps performed in an identical way? I assume they were, but this should be stated clearly in the methods. How were the tissue libraries stored? For how long can they be used after preparation?

We have added a new section “Storage of tissue library” with the following text: The tissue library was then used immediately or stored at 4oC in PBS with 0.02% (w/v) sodium azide. Mouse and human tissue was processed following all of the steps above and then used up to 3 months later. Additional mouse and human tissue was stored following delipidation, then successfully used up to 6 months later by resuming from the pre-immunohistochemistry tissue processing state. IHC conditions for a given antibody were always compared using batches of tissue that were stored for the same amount of time.

Figures

Figures 1, 2 and others. How is the top of the side view determined? Is this the surface of the tissue and therefore conditions where the label in the side view starts lower that in other conditions means that the tissue surface was not labeled? And why would that be? If that is not the case, it will be best to align the side views and have the labeling start at the same level.

We have aligned the side views as recommended by reviewer 2. In these images, sometimes the imagine plane started slightly above the top plane of the tissue. For the purposes of quantification and presentation, our matlab code now starts the plot from the first plane with 80% of the max staining intensity within a sample. This is clarified in the Materials and methods where we describe the code for analysis. The top view is chosen arbitrarily as a section that faithfully represents the quality of the rest of the image.

Figure S3. This is a very clear way of summarizing the differences in labeling. It would be very helpful to have this summary next to the actual immunostaining images in the main figures.

We have added the half-max intensity values to figure 3 near the actual immunostaining images to summarize the differences in labeling.

Figure S9 Representative microtome images taken from the human tissue prior to staining: Do you mean prior to clearing?

In figure S9 (now figure 6-supplement 2) we do mean that the images were taken prior to clearing, and have changed the text to clarify that point.

[Editors’ note: what follows is the authors’ response to the second round of review.]

Essential revisions:

Reviewer #1 (Recommendations for the authors):

The revised manuscript has addressed most of the reviewers' concerns and is significantly improved.Regarding the comparison between directly conjugated antibodies and use of secondary antibodies. I agree that directly conjugated antibodies save time and can be recommended for that purpose. I still have a problem with claims like: "As shown in Figure 3, the use of a primary antibody directly conjugated to a fluorophore greatly reduced background fluorescence over traditional primary/secondary antibody systems." The comparison is not between the same primary antibodies with one conjugated and one not; thus the observed differences can also be due to the primary antibodies themselves. Also, it is not at all clear whether a different secondary antibody would give a similar result. There is a great variety of secondary antibodies commercially available and, generally, the preadsorbed secondary antibodies that have undergone an extra purification step have higher specificity and give less background labeling; however the secondary antibodies used here were not of the "preadsorbed" kind.

As recommended by the reviewer, we have removed any claims that suggest an improvement in the staining from using conjugated antibodies. Specifically, we have replaced the following sentences from the previous draft:

“Importantly, primary GFAP antibodies directly conjugated to a fluorophore showed significantly better penetration, signal-to-noise ratio, and less off-target staining for both GFAP antibodies under all conditions where IHC successfully labeled its target (Figure 3) and both showed similar conditions where no labeling occurred (Figure s10).”

To these new sentences in the revised draft:

“We found that both GFAP antibodies successfully labeled their target under similar conditions, (Figure 3) and both showed similar conditions where no labeling occurred (Figure s10). Some differences in labeling quality are apparent between antibodies, however it is not clear whether the differences in labeling quality are due to the conjugation of the fluorophore or other differences between the monoclonal and polyclonal antibodies.”

We also removed the following sentence: “As shown in Figure 3, the use of a primary antibody directly conjugated to a fluorophore greatly reduced background fluorescence over traditional primary/secondary antibody systems.”

Last paragraph of Results before Discussion: "Therefore, we directly conjugated three antibodies to three different fluorophores (NeuN-A568, Glut1-A488, and AT8-A647)." This seems to be incorrect as these are the commercially available antibodies that are sold already directly conjugated.

Thank you for catching this mistake. We changed the sentence to the following: “We concurrently incubated the human tissue with three antibodies conjugated to three different fluorophores (NeuNA568, Glut1-A488, and AT8-A647).”

Materials and methods: Human tissue samples: what was the postmortem interval before fixation? Even a rough estimate will be helpful.

We have amended supplementary table 1 to include the postmortem interval before fixation.

Supplementary Table 1. Hemisphere column: "Frozen left". Does this mean that the human tissue was stored frozen? And then fixed? If it was indeed frozen before fixation, which was not mentioned elsewhere in the methods, that could explain why it looks like the antibodies penetrate deeper into the human tissue compared to mouse, (e.g. Figure 1 vs. Figure 5), but it is also an important difference to keep in mind.

To help clarify that some of the human tissue was stored frozen then thawed prior to fixation we have added the following sentence to the Human tissue samples section under Materials and methods: “Human tissue was received unfrozen immediately after autopsy and fixed (sample 2620) or frozen immediately after autopsy then thawed prior to fixation (samples 1581 and 1762).”

Reviewer #2 (Recommendations for the authors):The additions to the manuscript re-submission are greatly appreciated, in particular the updated analysis approach and 1d line profiles of signal versus tissue depth. There is one additional point that I believe is important and can be handled in the text. From the screen of antibody staining times (Figure S1), it appears that up through 24 hours of incubation the effect of increased incubation time is to increase the signal in a ~120 µm surface layer whose thickness does not change appreciably. After this, up to 48 hr (at least for RT incubation) the thickness of the surface layer changes strongly. This suggests that the limitation on signal for staining 24hr and under is depletion of antibodies on the way through the tissue. It should be pointed out that in the present work, the focus on depth to half-max staining does not separate the issues of (1) free antibody penetration speed through the tissue and (2) the extent of depletion along the way. Reducing antigenicity will improve the apparent penetration rate by reducing the effect of depletion due to epitope binding. This may be what is happening with the 50oC unmasking treatment, which increases penetration but greatly reduces the staining level (Figure 1 b,c, column 8). This limitation should be discussed explicitly.

We would like to thank Reviewer #2 for the insight into our data, as well as their assessment of its limitations. We have added the following text to explicitly reflect their insight:

We added the following text under the “Strategy to rapidly assess the optimization of protocols” section in paragraph 2 “It should be noted that these data suggest the signal when staining 24 hours and under may be limited by the depletion of antibodies on their way through the tissue, which should be considered when making comparisons of staining quality.”

We added the following text under the “Measuring IHC quality in cleared tissue” It should be noted that simply relying on depth to half-max staining does not separate the issues of free antibody penetration speed through the tissue nor the extent of antibody depletion along the way, and it is possible that conditions that reduce the antigenicity may significantly enhance the apparent penetration rate.

https://doi.org/10.7554/eLife.84112.sa2

Article and author information

Author details

  1. Theodore J Zwang

    1. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, United States
    2. Harvard Medical School, Boston, United States
    3. Massachusetts Alzheimer’s Disease Research Center, Charlestown, United States
    Contribution
    Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    tzwang@mgh.harvard.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6815-4482

  2. Rachel E Bennett

    1. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, United States
    2. Harvard Medical School, Boston, United States
    3. Massachusetts Alzheimer’s Disease Research Center, Charlestown, United States
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Investigation, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared

  3. Maria Lysandrou

    Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, United States
    Contribution
    Data curation, Formal analysis, Investigation, Writing - review and editing
    Competing interests
    No competing interests declared

  4. Benjamin Woost

    Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, United States
    Contribution
    Data curation, Formal analysis, Investigation, Writing - review and editing
    Competing interests
    No competing interests declared

  5. Anqi Zhang

    Department of Chemical Engineering, Stanford University, Stanford, United States
    Contribution
    Conceptualization, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6121-8095

  6. Charles M Lieber

    Department of Chemistry and Chemical Biology, Harvard University, Cambridge, United States
    Contribution
    Conceptualization, Resources, Supervision, Methodology, Project administration, Writing - review and editing
    Competing interests
    No competing interests declared

  7. Douglas S Richardson

    Department of Molecular and Cellular Biology and Harvard Center for Biological Imaging, Harvard University, Cambridge, United States
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3189-2190

  8. Bradley T Hyman

    1. Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, United States
    2. Harvard Medical School, Boston, United States
    3. Massachusetts Alzheimer’s Disease Research Center, Charlestown, United States
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Methodology, Project administration, Writing - review and editing
    Competing interests
    Reviewing editor, eLife
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7959-9401

Funding

National Institute on Aging (NIA-K99AG068602)

  • Theodore J Zwang

Massachusetts General Hospital (Harrison Gardner, Jr Innovation Award)

  • Theodore J Zwang

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This work was supported by the NIH/NIA-K99AG068602 (TJZ) and the Harrison Gardner, Jr Award (TJZ).

We would like to thank Massachusetts General Hospital Center for comparative Medicine and Analiese Fernandes for help with animal husbandry and care. We would also like to thank the Harvard Center for Biological Imaging including Christian Hellriegal for the use of their imaging facility and helpful conversations. Autopsy tissue from human brains were collected at Massachusetts General hospital, with informed consent of patients or their relatives and approval of local institutional review boards.

Ethics

Autopsy tissue from human brains were collected at Massachusetts General hospital, with informed consent of patients or their relatives and approval of local institutional review boards. Human tissue was provided by the Massachusetts Alzheimer's Disease Research Center (ADRC) with approval from the Mass General Brigham IRB (1999P009556).

This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the mice were handled according to approved institutional animal care and use committee (IACUC protocol # 2019N000026 and 2020N000069).

Senior Editor

  1. Suzanne R Pfeffer, Stanford University, United States

Reviewing Editor

  1. Albert Cardona, University of Cambridge, United Kingdom

Version history

  1. Preprint posted: May 26, 2022 (view preprint)
  2. Received: October 11, 2022
  3. Accepted: January 18, 2023
  4. Accepted Manuscript published: January 19, 2023 (version 1)
  5. Version of Record published: January 31, 2023 (version 2)
  6. Version of Record updated: February 10, 2023 (version 3)
  7. Version of Record updated: March 3, 2023 (version 4)

Copyright

© 2023, Zwang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,966
    Page views
  • 165
    Downloads
  • 2
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Theodore J Zwang
  2. Rachel E Bennett
  3. Maria Lysandrou
  4. Benjamin Woost
  5. Anqi Zhang
  6. Charles M Lieber
  7. Douglas S Richardson
  8. Bradley T Hyman
(2023)
Tissue libraries enable rapid determination of conditions that preserve antibody labeling in cleared mouse and human tissue
eLife 12:e84112.
https://doi.org/10.7554/eLife.84112

Categories and tags

Research organism

Further reading

    1. Cell Biology

    Rab12 is a regulator of LRRK2 and its activation by damaged lysosomes

    Xiang Wang, Vitaliy V Bondar ... Anastasia G Henry
    Short Report Updated

    Leucine-rich repeat kinase 2 (LRRK2) variants associated with Parkinson’s disease (PD) and Crohn’s disease lead to increased phosphorylation of its Rab substrates. While it has been recently shown that perturbations in cellular homeostasis including lysosomal damage can increase LRRK2 activity and localization to lysosomes, the molecular mechanisms by which LRRK2 activity is regulated have remained poorly defined. We performed a targeted siRNA screen to identify regulators of LRRK2 activity and identified Rab12 as a novel modulator of LRRK2-dependent phosphorylation of one of its substrates, Rab10. Using a combination of imaging and immunopurification methods to isolate lysosomes, we demonstrated that Rab12 is actively recruited to damaged lysosomes and leads to a local and LRRK2-dependent increase in Rab10 phosphorylation. PD-linked variants, including LRRK2 R1441G and VPS35 D620N, lead to increased recruitment of LRRK2 to the lysosome and a local elevation in lysosomal levels of pT73 Rab10. Together, these data suggest a conserved mechanism by which Rab12, in response to damage or expression of PD-associated variants, facilitates the recruitment of LRRK2 and phosphorylation of its Rab substrate(s) at the lysosome.

    1. Cell Biology

    Genome-wide screen reveals Rab12 GTPase as a critical activator of Parkinson’s disease-linked LRRK2 kinase

    Herschel S Dhekne, Francesca Tonelli ... Suzanne R Pfeffer
    Research Advance Updated

    Activating mutations in the leucine-rich repeat kinase 2 (LRRK2) cause Parkinson’s disease. LRRK2 phosphorylates a subset of Rab GTPases, particularly Rab10 and Rab8A, and we showed previously that these phosphoRabs play an important role in LRRK2 membrane recruitment and activation (Vides et al., 2022). To learn more about LRRK2 pathway regulation, we carried out an unbiased, CRISPR-based genome-wide screen to identify modifiers of cellular phosphoRab10 levels. A flow cytometry assay was developed to detect changes in phosphoRab10 levels in pools of mouse NIH-3T3 cells harboring unique CRISPR guide sequences. Multiple negative and positive regulators were identified; surprisingly, knockout of the Rab12 gene was especially effective in decreasing phosphoRab10 levels in multiple cell types and knockout mouse tissues. Rab-driven increases in phosphoRab10 were specific for Rab12, LRRK2-dependent and PPM1H phosphatase-reversible, and did not require Rab12 phosphorylation; they were seen with wild type and pathogenic G2019S and R1441C LRRK2. As expected for a protein that regulates LRRK2 activity, Rab12 also influenced primary cilia formation. AlphaFold modeling revealed a novel Rab12 binding site in the LRRK2 Armadillo domain, and we show that residues predicted to be essential for Rab12 interaction at this site influence phosphoRab10 and phosphoRab12 levels in a manner distinct from Rab29 activation of LRRK2. Our data show that Rab12 binding to a new site in the LRRK2 Armadillo domain activates LRRK2 kinase for Rab phosphorylation and could serve as a new therapeutic target for a novel class of LRRK2 inhibitors that do not target the kinase domain.

    1. Cell Biology

    Differential regulation of hair cell actin cytoskeleton mediated by SRF and MRTFB

    Ling-Yun Zhou, Chen-Xi Jin ... Hao Wu
    Research Article Updated

    The MRTF–SRF pathway has been extensively studied for its crucial role in driving the expression of a large number of genes involved in actin cytoskeleton of various cell types. However, the specific contribution of MRTF–SRF in hair cells remains unknown. In this study, we showed that hair cell-specific deletion of Srf or Mrtfb, but not Mrtfa, leads to similar defects in the development of stereocilia dimensions and the maintenance of cuticular plate integrity. We used fluorescence-activated cell sorting-based hair cell RNA-Seq analysis to investigate the mechanistic underpinnings of the changes observed in Srf and Mrtfb mutants, respectively. Interestingly, the transcriptome analysis revealed distinct profiles of genes regulated by Srf and Mrtfb, suggesting different transcriptional regulation mechanisms of actin cytoskeleton activities mediated by Srf and Mrtfb. Exogenous delivery of calponin 2 using Adeno-associated virus transduction in Srf mutants partially rescued the impairments of stereocilia dimensions and the F-actin intensity of cuticular plate, suggesting the involvement of Cnn2, as an Srf downstream target, in regulating the hair bundle morphology and cuticular plate actin cytoskeleton organization. Our study uncovers, for the first time, the unexpected differential transcriptional regulation of actin cytoskeleton mediated by Srf and Mrtfb in hair cells, and also demonstrates the critical role of SRF–CNN2 in modulating actin dynamics of the stereocilia and cuticular plate, providing new insights into the molecular mechanism underlying hair cell development and maintenance.