Spinal cord from body donors is suitable for multicolor immunofluorescence

Immunohistochemistry is a powerful tool for studying neuronal tissue from humans at the molecular level. Obtaining fresh neuronal tissue from human organ donors is difficult and sometimes impossible. In anatomical body donations, neuronal tissue is dedicated to research purposes and because of its easier availability, it may be an alternative source for research. In this study, we harvested spinal cord from a single organ donor 2 h (h) postmortem and spinal cord from body donors 24, 48, and 72 h postmortem and tested how long after death, valid multi-color immunofluorescence or horseradish peroxidase (HRP) immunohistochemistry is possible. We used general and specific neuronal markers and glial markers for immunolabeling experiments. Here we showed that it is possible to visualize molecularly different neuronal elements with high precision in the body donor spinal cord 24 h postmortem and the quality of the image data was comparable to those from the fresh organ donor spinal cord. High-contrast multicolor images of the 24-h spinal cords allowed accurate automated quantification of different neuronal elements in the same sample. Although there was antibody-specific signal reduction over postmortem intervals, the signal quality for most antibodies was acceptable at 48 h but no longer at 72 h postmortem. In conclusion, our study has defined a postmortem time window of more than 24 h during which valid immunohistochemical information can be obtained from the body donor spinal cord. Due to the easier availability, neuronal tissue from body donors is an alternative source for basic and clinical research. Supplementary Information The online version contains supplementary material available at 10.1007/s00418-022-02154-5.


Introduction
Immunohistochemistry is a powerful tool for studying neuronal tissue from humans at the molecular level and this is the foundation for a better understanding of neuronal function. Typically, human neuronal tissue is obtained postmortem from organ donors or body donors who donated their bodies to anatomical departments. The use of fresh neuronal tissue from organ donors for research is limited and due to ethical reasons often prohibited. In contrast, tissue from body donors can be obtained much easier because body donors have already given written consent during a lifetime that his/her body is dedicated to teaching and scientific purposes. For this reason, neuronal tissue from body donors should be tested as an alternative source for immunohistochemical studies.
A prerequisite for immunohistochemical analyses is optimal tissue quality. To avoid loss of quality, it is recommended to refrigerate bodies after death and keep the time between death and the start of tissue fixation (postmortem interval) as short as possible (Werner et al. 2000;Waldvogel et al. 2006;Almulhim and Menezes 2022). However, such a controlled procedure is very challenging, and only possible for donors who die in an intensive care unit.
People who donate their bodies to anatomical departments often die outside a monitoring facility and as there is uncertainty about the exact time of death, inaccuracies, and variability in postmortem interval estimates are unavoidable. Additionally, there is uncertainty about the ambient temperature bodies are exposed to until they are transferred to anatomical departments. Although neuronal tissue (peripheral nerves) from body donors has been analyzed by immunohistochemistry (Rein et al. 2013(Rein et al. , 2015(Rein et al. , 2020Blumer et al. 2016;Pascual-Font et al. 2016;Garcia-Mesa et al. 2017;Lienbacher et al. 2019), it would be crucial to define a period after death during which accurate immunohistochemistry is possible in body donors. Such information is currently lacking but would be relevant for scientists who analyze nerve tissue from body donors at the molecular level.
In the present study, we harvested spinal cord tissue samples from body donors as they routinely arrive at body donor facilities at 24, 48, and 72 h postmortem intervals. We used a series of general and specific neuronal markers as well as glial markers and evaluated the stability of multi-color immunofluorescence and horseradish peroxidase (HRP) immunohistochemistry. Results in body donor spinal cords were compared with immunolabeling assays in fresh spinal cord obtained 2 h postmortem from an organ donor. We demonstrate that 24 h after death, different neuronal elements can be identified with high fidelity in the human spinal cord by immunofluorescence/HRP immunohistochemistry and high contrast images can be used for accurate computerbased quantitative analyses. With prolonged postmortem intervals (48 and 72 h), the staining capacity of the antibodies gradually decreases in an antibody-specific fashion.

Tissue collection and preparation
From 15 body donors, the cervical segments 4-6 of the spinal cord were collected and analyzed. Spinal cords were from voluntary body donors of both sexes (seven females and eight males) who provided written consent to donate their dead bodies to be used for teaching and research at the Center of Anatomy and Cell Biology, Medical University of Vienna. The age of the donors ranged between 58 and 103 years (mean and standard deviation 81.2 ± 13.08) and none of them had a history of neurological disorders. Once delivered to the Department of Anatomy, body donors were kept in a mortuary at 4 °C until dissection. For all body donors, the time of death was precisely known, and they were divided into three groups. Each group contained five subjects. In the first group, the time between death and start of tissue fixation (postmortem interval) was 24 h, in the second 48 h, and in the third group 72 h.
Following dissection, the cervical spinal cord was subdivided into three pieces, each piece corresponding to a single spinal cord segment. Then, the issue was immersion fixed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB,pH 7.4) for 24 h at 4 °C and afterward stored in phosphatebuffered saline (PBS) for further processing. In every group, handling procedures for tissue processing were the same.
To evaluate results in the spinal cord from body donors, we analyzed the spinal cord (thoracic segments) of an organ donor (39 years old) used in a previous study (Gesslbauer et al., 2017). The postmortem interval of the organ donor spinal cord was 2 h and because of this short postmortem interval, it was used as a control for all other studied groups. The study has been approved by the local ethics committee (protocol number of the approval No. 1213/2012 and 1373/2021).

Tissue preparation and processing
Spinal cords from the organ donor and body donors were cryoprotected in graded concentrations of sucrose (10, 25, and 40%) in PBS containing 0.05% sodium azide to avoid bacterial and fungal contamination, frozen, and stored at − 80 °C according to our guidelines for cryo-embedding (Blumer et al. 2019). For histological and immunohistochemical evaluation, tissue was cryo-sectioned at 10-μm thickness.

Histology
Human spinal cord sections were labeled with hematoxylin-eosin using a standard protocol (hematoxylin 3 min and eosin 1 min). Thereafter, sections were dehydrated in graded solutions of alcohol and coverslipped in DPX mounting medium (Merck/Millipore, Temecula, CA USA).

Antibodies and antibody characterization
Different neuronal marker antibodies were used to visualize nervous elements in the spinal cord. This included panneuronal markers (anti-neurofilament, anti-synaptophysin, and anti-nuclear neuronal protein), specific neuronal markers (anti-calcitonin gene-related peptide, anti-choline 1 3 acetyltransferase, and anti-vesicular acetylcholine transporter) and glial cell markers (anti-myelin basic protein). The primary antibodies used in the present study including working dilutions, RRID numbers, and suppliers are listed in Table 1.
The neurofilament antibody (AB5539, Merck-Millipore) is a polyclonal antibody and is raised against the bovine neurofilament heavy chain. By immunofluorescence, it has been demonstrated that this antibody visualizes myelinated and non-myelinated axons in humans (Gesslbauer et al. 2017;Blumer et al. 2020) and non-human tissue (Ghilardi et al. 2011;Tereshenko et al. 2021).
The monoclonal mouse antibody against the nuclear neuronal protein (NeuN, MAB377, Merck-Millipore) labels cell nuclei in neurons and is the most widely used neuronal marker in neuroscience research and neuropathological assays (Gusel'nikova and Korzhevskiy 2015).
The polyclonal goat antibody against choline acetyltransferase (ChAT, AB144P, Merck-Millipore) is raised against human placental ChAT. The antibody was validated in the human brain by pre-adsorption (Mesulam et al. 1992). By immunohistochemistry, it has been shown that this antibody visualizes cholinergic neurons in the human spinal cord (Gesslbauer et al. 2017) and brain (Lienbacher et al. 2019).
The monoclonal anti-synaptophysin antibody clone SP15 (MAB329, Merck-Millipore) is validated for use in ELISA, Western blot, and immunofluorescence for the detection of the synaptic vesicle protein synaptophysin (manufacturer's datasheet).
The monoclonal antibody against MBP (MAB386) is raised against bovine myelin basic protein and was tested by Western blot in mouse brain (manufacturer's datasheet). This antibody has been shown to visualize myelinated axons in the human spinal cord (Blumer et al. 2020) and non-human tissue (Nakanishi et al. 2005).
The monoclonal antibody against calcitonin gene-related peptide (CGRP, 14959, Cell Signaling) visualizes CGRP in the central and peripheral nervous system. This antibody was validated in carcinoma patients exhibiting a correlation between perineural invasion and lymph node metastasis (Zhang et al. 2021).
The VAChT antibody (139 103 Synaptic system) is a polyclonal antibody and raised the against vesicular acetyl choline transporter, which is an integral protein of synaptic vesicles. This antibody was knockout-validated in mouse brains (Kljakic et al. 2021) and used to visualize cholinergic neurons in the human pancreas (Campbell-Thompson and Tang 2021).

Horseradish peroxidase immunohistochemistry
HRP immunohistochemistry was performed in ChAT and CGRP staining. Sections were pretreated with 1% H 2 O 2 to suppress the endogenous peroxidase activity following incubation with 10% normal rabbit serum (ChAT staining) or 10% goat serum in PBS-T (CGRP staining). Thereafter, sections were incubated with the primary antibodies as described above. After washing in PBS, sections were incubated for 2 h at 37 °C with the biotinylated secondary antibodies rabbit anti-goat for the ChAT staining or goat antirabbit for the CGRP staining. After a further washing step, sections were incubated with avidin-neutravidin conjugated with HRP for 2 h at 37 °C. The antigenic site was visualized using 0.01% diaminobenzidine (DAB) containing 0.06% H 2 0 2 in 0.1 PBS for 5 to 10 min. Sections were coverslipped with aqueous mounting medium (Dako Cytomation). The biotinylated secondary antibodies including working dilutions, RRID numbers, and suppliers are listed in Table 1.
As a positive control, sections from the organ donor spinal cord were included in every staining combination and staining run. This approach allowed us to verify the validity of the staining process and exclude misinterpretation of results in the body donor spinal cord. Omission of the primary antibodies and using the secondary antibodies alone resulted in a complete lack of immunostaining. This confirmed that the secondary antibodies bind specifically to the primary antibodies.

Light microscopy
Hematoxylin-eosin staining sections were analyzed with a slide scanner microscope (Olympus VS120, Olympus Europa SE & Co. KG, Hamburg, Germany).
Sections labeled by the HRP immunohistochemistry technique were examined with a light microscope (Axioscope, Zeiss, Jena, Germany) and photo-documented using a digital camera (Nikon Eclipse 600 camera, Nikon, Tokyo, Japan) connected to the microscope.

Confocal laser scanning microscopy
Fluorescently labeled sections were analyzed with a confocal laser scanning microscope (CLSM Olympus FV3000). Images were captured with dry lenses of 4x and 20x magnification and oil lenses of 60x magnification. A series of virtual CLSM sections of 1-µm thickness were cut through the structures of interest. Each section was photo-documented with a 1024 × 1024 pixel resolution and 3D projections were rendered using ImageJ software (NIH, Bethesda, MD, USA). Single-colored images (Fig. 2) were generated using a laser with excitation wavelength 633. Double-colored images (Figs. 3,4,5,6,7,8,9) were generated using lasers with excitation wavelengths of 488 and 568 nm, and triplecolored images (Figs. 10,11,12) using additionally a laser with excitation wavelengths of 405 or 640 nm. In certain cases, bright-field images were recorded in the CLSM to correlate immunolabeled structures to morphological structures (Fig. 5a'',b'',c'',d'').

Scanning strategy
In all staining combinations and all postmortem groups, fluorescently labeled sections were analyzed and photodocumented immediately (within 1 day) after the staining experiments were completed. For every staining combination, individual CLSM settings were manually adjusted and care was taken to avoid oversaturation or undersaturation of images. Once found, settings parameters were kept constant for image acquisition of all samples in each staining combination and postmortem interval. This scanning strategy allowed us to determine changes in the signal intensity at different postmortem intervals.

Axon counts
To evaluate the accuracy of our immunofluorescence, we separately counted neurofilament positive axons and neurofilament/ChAT-positive axons in the anterior roots of the human spinal cord (N = 3). This was done in the spinal cords at 24 and 48 h postmortem intervals. Whole cross sections of the anterior roots were scanned with a slide scanner [Axio Observer Z.1 microscope (Carl Zeiss Meditec AG, Jena, Germany)] followed by automated counts of axons using StrataQuest version 5.1.249 and Tissue Quest version 4.0.1.0128 (TissueGnostic Vienna, Austria).

Quantification of the staining intensity
The intensity of the antibody signals was quantified in human spinal cords with 2-h (N = 1), 24-h (N = 3), 48-h (N = 3), and 72-h (N = 3) postmortem intervals. The antibodies studied in each postmortem interval were anti-neurofilament, anti-ChAT, anti-CGRP, and anti-synaptophysin. For the quantification of the staining intensity, images of the specimens were captured at 20x magnification using identical CLSM settings. In each image, 25 randomly selected squares of 6 × 6 μm were used to measure the intensity of the antibodies signal (i.e., optical density). In all cases, for background correction, ten optical density values of the same area were taken. Then, the background value was subtracted, and the values of the antibody optical density were normalized with respect to the 2-h postmortem spinal cord. The antibody staining intensity was set at 100% in the 2-h postmortem spinal cord and values for spinal cords with 24-, 48-, and 72-h postmortem intervals were calculated as the percentage of the 2-h postmortem control value.

Statistical analyses
The axon quantification data of double-positive axons (ChAT and neurofilament) and single-positive neurofilament axons were statistically compared using Student's t test at a level of significance of P < 0.05. The data of staining intensity (anti-neurofilament, anti-ChAT, anti-CGRP, and antisynaptophysin) of the different postmortem intervals (2, 24, 48, and 72 h) were statistically compared using the one-way-ANOVA test followed by post hoc multiple comparisons (Holm-Sidak method) at a significance level of P < 0.05. All values were expressed as the mean ± standard error of the mean (SEM). Statistical analysis was carried out in Sigma Plot, version 11 (Systat Software, San José, CA, USA).

Histology
The histology of the spinal cord from body donors was analyzed using hematoxylin-eosin staining. The spinal cord at 24-h postmortem interval exhibited strong staining and the tissue architecture appeared normal. The two parts of the spinal cord, the inner H-shaped gray matter, and the outer white matter were clearly identifiable (Fig. 1a). In line with a previous study, we observed that the anterior horn of the gray matter has a wide lateral expansion at the cervical level of the spinal cord (Kameyama et al. 1996). Forty-eight hours after death, small gaps were observed readily in the white and gray matter of the spinal cord ( Fig. 1b) and they increased in size and number 72 h postmortem, indicating advanced autolysis of the tissue (Fig. 1c). Moreover, 72 h postmortem the staining intensity decreased and there was no sharp boundary between the spinal cord's gray and white matter.

Lipofuscin autofluorescence
Lipofuscin is a lipid-containing degradation product and increases with age in neuronal and non-neuronal tissue (Brody 1960;Shiers et al. 2021). In the human brain, lipofuscin deposits were observed intra-as well as extracellularly (Waldvogel et al. 2006). Although lipofuscin is virtually present in every type of neuron, it is most abundant in neurons initiating movements including motoneurons of the spinal cord (Gray and Woulfe 2005). Typically, lipofuscin exhibits strong red and green autofluorescence and can be recognized in a fluorescence microscope when using appropriate filters (Gray and Woulfe 2005). Concordant with this, we observed high lipofuscin autofluorescence in the motoneurons of the anterior horn of the organ donor and body donor spinal cords. Due to the older age, the lipofuscin accumulation was higher in body donors (mean age 81.2 ± 13.08 years) than in the 39-year-old organ donor. The inset in Fig. 2a highlights the anterior segment of the spinal cord, which is shown in immunofluorescent staining in Fig. 2b' visualizes lipofuscin inside motoneuron cells bodies and processes in the neuropil. In motoneurons, the lipofuscin accumulation was graded and while few motoneuron cell bodies lacked lipofuscin, most cell bodies were filled to various degrees with this age-related cellular inclusion . Because of the lipofuscin autofluorescence, care is necessary when interpreting immunofluorescence staining in the spinal cord and, at appropriate places, we referred to this point in the manuscript.

Double immunofluorescence and peroxidase immunohistochemistry
We performed four double-labeling immunofluorescence assays in the spinal cord of the organ donor and body donors at different postmortem intervals. Double immunofluorescence included labeling with anti-neurofilament plus (1) anti-ChAT, (2) anti-synaptophysin, (3) anti-CGRP, and (4) anti-NeuN. Alternatively, HRP immunohistochemistry for anti-ChAT and anti-CGRP was performed. Anti-ChAT is a marker for cholinergic neurons, anti-synaptophysin is a marker for synapses, anti-CGRP is a marker for putative nociceptor fibers, and anti-NeuN is a marker for the nuclei of neuronal cells. In our immunolabeling assays, some antibodies exhibited stronger signal intensity than others and we therefore adapted the CLSM settings for every individual staining combination. As a result of this approach, the red and green lipofuscin autofluorescence is visible in some Fig. 1) is provided showing autofluorescence signal dependent on CLSM settings.
Following anti-neurofilament/anti-ChAT incubation, neurofilament signals were observed in the gray and white   neurofilament/ChAT signals, but this was more frequent 72-h postmortem (arrows in Fig. 5c-c'', d-d''). Additionally, by 72 h, most axons revealed less clear delineation coupled with incomplete labeling (Fig. 5d-d'').
We quantified the staining intensity in the 2-h postmortem spinal cord of the organ donor (N = 1) and in the spinal cords of the body donors at every postmortem interval (N = 3 for each postmortem interval). Measurements are percentages relative to the 2-h spinal cord value, which was the reference value and set at 100% (Fig. 4f, left bar; 4 g left bar). At time point 24 h, the neurofilament staining intensity was 90% of the 2-h reference value (Fig. 4f, second bar). At the time point of 48 h, the neurofilament signal was at about 70% (Fig. 4f, third bar) and at 72 h at around 45% of the reference value (Fig. 4f, fourth bar). Relative to the reference value, the ChAT signal was 93% at time point 24 h (Fig. 4g, second bar), 40% at time point 48 h (Fig. 4g, third bar), and less than 20% at 72 h (Fig. 4g, fourth bar).
We evaluated the precision of the neurofilament/ChAT immunolabeling by counting separately neurofilament and neurofilament/ChAT-positive axons in the anterior root of the 24-and 48-h spinal cord (N = 3 for each point of time). Because the anterior roots of the human spinal cord carry exclusively cholinergic (motor) axons, all neurofilamentpositive axons in the anterior root should express ChAT as well. Automated quantification revealed that at time point 24 h, 99.5% of the axons in the anterior root expressed ChAT (Fig. 4h, left bars) and this was confirmed by high-resolution images, as there was a complete match of neurofilament/ ChAT signals (Fig. 5b-b''). The small gap to 100% could be explained that the ChAT signal varies along the axons (Zimmermann et al. 2013) and it is possible that axons with weaker ChAT signals were missed by the analysis software. At time point 48 h, significantly fewer ChAT-positive axons (90.3%) were counted (Fig. 4h, right bars). This finding proves that immunolabeling was very precise 24 h postmortem but no longer 48 h postmortem.
Although the ChAT signals were strong and almost equal in the motor axons 2-h and 24-h postmortem, ChAT immunoreactivity in motoneuron cell bodies was observed 2 h but not 24 h postmortem (Fig. 6a-a'', b-b''). Interestingly, the ChAT intensity was lower in motoneuron cell bodies than in motoneuron axons. It is important to note that the ChAT signal in motoneuron cell bodies was only visible in the green fluorescence but not in the red fluorescence channel of the CLSM (Fig. 6a-a''), thus excluding being lipofuscin autofluorescence, which would be detectable in both fluorescence channels. We therefore conclude that the ChAT immunoreactivity in motoneuron cell bodies 2 h postmortem was specific. Different from that pattern, motoneuron cell bodies in the 24-h postmortem spinal cord exhibited signals in the green fluorescence and red fluorescence channel, indicating lipofuscin autofluorescence instead of a specific ChAT signal ( Fig. 6b-b''). Interestingly, cholinergic synapses, which were observed on motoneurons, exhibited bright ChAT signals 2 and 24 h postmortem (Fig. 6a'-a'', b-b''). The lack of ChAT signals in motoneuron cell bodies 24 h postmortem might be explained by the fact that the ChAT enzyme became dissociated or alternatively was at a level too low for detection by immunofluorescence. We therefore performed HRP immunohistochemistry because there are examples in the literature that this technique is more sensitive than immunofluorescence (Molne et al. 2005). By using this approach, we observed robust ChAT immunoreactivity in the motoneurons of the 2-h donor spinal cord (Fig. 6c, e), whereas in the 24-h body donor spinal cord, weak signals were found in some but not all motoneurons (Fig. 6d, f). Analog to immunofluorescence, cholinergic synapses exhibited a high signal intensity in immunoperoxidase staining (arrows in Fig. 6c, d, f).
We next performed labeling with anti-neurofilament and anti-synaptophysin. Synaptophysin is a protein present in synaptic contacts (De Camilli et al. 1988) and synaptophysin was observed throughout the spinal cord's gray matter as demonstrated in a panoramic image of the body donor spinal cord 24 h postmortem (Fig. 7a). Synaptophysin was also found along cell processes (dendrites) extending from the gray into the white matter (arrows in Fig. 7a). We observed synaptophysin signals at the periphery of the spinal cord and in the posterior part of the spinal cord along the midline (Figs. 7a, 10a). Microscopic analyses showed that the fluorescence signal corresponded with plaques in the white matter.
To evaluate the quality of the synaptophysin labeling along postmortem intervals, we examined synaptic contacts on spinal cord motoneurons. We observed no differences between 2 and 24 h postmortem, and synapses expressing high levels of synaptophysin covered the surface of the motoneuron cell bodies (Fig. 7b-b'', c-c''). Forty-eight hours after death, the signal brightness was reduced but the synaptic density on motoneuron cell bodies appeared unchanged (Fig. 7d-d''). At time point 72 h, the signal intensity further decreased and the synaptic density was reduced (Fig. 7e-e''). Morphologically, large tissue-free gaps surrounded the motoneurons 72 h postmortem (Fig. 7e-e'').
Measurements of the staining intensity demonstrated that the synaptophysin signal was almost equal in the 2-h postmortem spinal cord of the organ donor and in the body donor spinal cord 24 h postmortem (Fig. 7f). At time point 48 h, the signal intensity of synaptophysin was reduced to about 60% of the 2-h value (Fig. 7f) and at time point 72 h to approximately 25% of the 2-h value (Fig. 7f).
After labeling the spinal cords of the organ donor and body donors with anti-neurofilament/anti-CGRP, we observed CGRP expression in the superficial layer of the posterior horn with decreasing CGRP signals in deeper 1 3 layers. A panoramic image showing high CGRP expression in the posterior horn 24 h postmortem is presented in Fig. 8a. These findings are in line with data obtained from other organ donor spinal cords (Tschopp et al. 1984;Shiers et al. 2021). Quantitative analyses showed that in the 24-h body donor spinal cord, the CGRP brightness was slightly decreased and the value was 85% of the reference value measured in the 2-h organ donor spinal cord (Fig. 8a, b, c). CGRP intensity was only half of the 2-h value after 48 h and about 30% after 72 h (Fig. 8d, e, f). Structurally, large gaps were present in the posterior horn 72 h postmortem (Fig. 8e).
In the spinal cord of the organ donor, some motoneuron cell bodies exhibited a punctate pattern of CGRP immunoreactivity ( Fig. 9a-a'') which, however, was much lower than the CGRP reactivity in the superficial layer of the posterior horn. The signal was only seen in the green fluorescence but not in the red fluorescence channel, thereby excluding lipofuscin autofluorescence which would be present in both channels. We observed that other motoneurons lacked CGRP signals (arrow in Fig. 9c). These findings are analogous to studies in the rodent spinal cord where it has been shown that a limited number of motoneurons express CGRP and the expression varies with age (Caldero et al. 1992;Blasco et al. 2020). No CGRP signals were detected in the motoneurons of the body donor spinal cords by immunofluorescence ( Fig. 9b-b''). Instead, in many motoneuron cell bodies, we observed lipofuscin autofluorescence in the green and red fluorescence channels (Fig. 9b-b''). We tested HRP immunohistochemistry and by this technique, CGRP signals were observed in some motoneurons 24 h postmortem (Fig. 9d).
We next tested anti-neurofilament and anti-NeuN staining in the spinal cord. NeuN is a neuron-specific nuclear protein (Gusel'nikova and Korzhevskiy 2015). Hoechst 33342, a fluorescent dye that binds to the DNA, was used to counterstain cell nuclei. This approach allowed us to distinguish between neuronal cells expressing NeuN and Hoechst 33342 and nonneuronal cells expressing Hoechst 33342 alone. Because we observed a drop-down in the signal intensity over prolonged postmortem intervals in the staining combinations anti-neurofilament with (1) anti-ChAT, (2) anti-synaptophysin, and (3) anti-CGRP, the neurofilament/anti-NeuN staining experiment was exclusively done in the 2-h spinal cord of the organ donor and the 24-h spinal cord of the body donors.
Following fluorescent labeling, we visualized motoneurons in the anterior horn of the spinal cord. The inset in Fig. 10a indicates the region visualized in the spinal cord. In the organ donor and body donor spinal cord, the nucleus of the motoneuron cell body exhibited strong NeuN immunoreactivity ( Fig. 10b-b", c-c", d-d", e-e"). The NeuN signal in motoneurons co-localized with the Hoechst 33342 nuclear stain thereby confirming that the NeuN antibody is neuron-specific and nuclear-localized (Fig. 10c-c", e-e"). We did not observe a difference in the anti-NeuN signal intensity 2 and 24 h postmortem. In addition to NeuN/Hoechst 33342-positive motoneurons, we found small nuclei which expressed NeuN/Hoechst 33342 as well (arrows in Fig. 10c, e). Interestingly, the NeuN signal was weaker in the organ donor spinal cord. We assume that these cells are interneurons, which for identification would require further neuronal markers.
Following immunolabeling with anti-neurofilament/antisynaptophysin/anti-CGRP, we observed co-localization of CGRP and synaptophysin in the superficial layer of the posterior horn of the gray matter (Fig. 11a, b-b'').

'', d-d'', e-e'' Labeling with anti-neurofilament (NF) and anti-ChAT of the 2-h organ donor spinal cord (a-a''), the body donor spinal cord from two different subjects 24 h (b-b'', c-c''), and from one subject 48 h (d-d''), and 72 h postmortem (e-e''). A reduction in the signal intensity is observed in the
After staining combination anti-neurofilament/anti-MBP/anti-CGRP, we visualized myelinated nerve by anti-MBP and putative nociceptor fibers by anti-CGRP in the posterior roots. We observed that most axons in the posterior root were covered by a myelin sheath (Fig. 12c) and few axons expressed CGRP (Fig. 12c'). All CGRP-positive axons were of small diameter, and they were with or without a myelin sheath (Fig. 12c'').
Taken together, by multicolor immunolabeling with a series of neuronal marker (antibodies against neurofilament, ChAT, synaptophysin, CGRP, NeuN, VAChT, and MBP), we have visualized different neuronal elements in fresh spinal cord from organ donor and body donors at different postmortem intervals and demonstrated that the image quality is almost equal in the spinal cord from organ donor and body donors 24 h postmortem. By HRP immunohistochemistry, the absence of ChAT and CGRP fluorescence signals in Fig. 6 Immunofluorescence with anti-neurofilament (NF) and anti-ChAT, and HRP immunohistochemistry with anti- ChAT. a-a'', b-b'' Immunofluorescence of motoneurons stained with anti-neurofilament (a, b) and anti-ChAT (a', b') from the organ donor (a-a'') and 24 h body donor spinal cord (b-b'').Dashed lines demarcate the motoneuron cell bodies (a-a'') including the proximal dendrite (b-b''). Some motoneurons exhibit red and green autofluorescence due to lipofuscin deposits (LP). A weak ChAT signal is observed 2 h postmortem (a') but absent 24 h postmortem (b'). Synapses  (48 and 72 h), the signal intensity of neuronal markers diminished but this downward trend was antibody-specific and moderate in the neurofilament antibody and strong in the ChAT antibody whereas values for CGRP and synaptophysin antibodies were in between. Additionally, we showed that automated quantification of axons is possible with high accuracy 24 h after death. The results of the present study are summarized in Fig. 13.
At this point, it is also important to note that in our immunofluorescence assays, only minimal staining variability was observed between individuals of the same postmortem group except for one case in the 72-h group where stronger immunofluorescence was observed than in other cases of this group.

Discussion
Immunohistochemistry is a powerful tool in analyzing human neuronal tissue at the molecular level. Typically, neuronal tissue is harvested from organ donors (Tschopp et al. 1984;Eftekhari and Edvinsson 2011;Gesslbauer et al. 2017;Mioton et al. 2019;Stakenborg et al. 2020;Dominguez-Alvaro et al. 2021;Shiers et al. 2021) or body donors (Rein et al. 2013(Rein et al. , 2015(Rein et al. , 2020Pascual-Font et al. 2016;Lienbacher et al. 2019). Harvesting neuronal tissue from organ donors requires a high administrative workload and is often prohibited. Contrary, tissue from body donors is more easily available and can be an alternative source for research.
The present study aimed to define a postmortem time window during which accurate immunohistochemical analyses are possible in the spinal cord of body donors. Here we demonstrated by immunofluorescence/HRP immunohistochemistry that precise identification of multiple, molecularly different neuronal elements was possible in the spinal cord of body donors until 24 h postmortem, and generated image data were qualitatively almost equal to those obtained from the organ donor spinal cord with 2-h postmortem interval. Although present data are related to tested molecules, our findings provide a useful guide for immunolabeling assays in neuronal tissue from organ donors.
Today there is increasing interest in analyzing human neuronal tissue including the brain (Waldvogel et al. 2006;Blair et al. 2016), spinal cord (Tschopp et al. 1984;Eftekhari and Edvinsson 2011;Shiers et al. 2021), and peripheral nerves (Gesslbauer et al. 2017;Mioton et al. 2019;Stakenborg et al. 2020) by immunohistochemistry to better understand neuronal function. As there is a need for high tissue quality, fresh tissue from organ donors is often preferred, although tissue collection is very challenging. In the present study, we showed that robust neurochemical information can be obtained from body donor spinal cords 24 h postmortem. By multi-color immunofluorescence, up to three molecules could be visualized 24 h after death in the same tissue sample and this offered the opportunity to detect colocalization of molecules and to perform counts of multiple neuronal elements. The high accuracy of the neurochemical information 24 h postmortem was confirmed by automated quantification. Specifically, 99.5 out of 100% of the axons leaving the spinal cord via the anterior root exhibited ChAT immunoreactivity and the same value (more than 99%) of cholinergic axons was counted in fresh spinal cord obtained from heart-beating organ donors (Gesslbauer et al. 2017). This important information suggests that neuronal tissue from body donors is suitable for qualitative and quantitative analyses of axonal components in mixed peripheral nerves and could substitute tissue from organ donor tissue, which was used recently for such analyses (Gesslbauer et al. 2017;Mioton et al. 2019;Stakenborg et al. 2020). Altogether, the present findings underline that accurate interpretation of immunohistochemical data is possible in neuronal tissue harvested from body donors.
It has been shown in rodents that the cell bodies of the anterior horn motoneurons express ChAT (Friese et al. 2009) and at least a subpopulation of them express CGRP (Caldero et al. 1992;Blasco et al. 2020). By immunofluorescence, no ChAT and CGRP signals were detected in the motoneuron cell bodies of the 24-h body donor spinal cord whereas weak ChAT and CGRP immunoreactivity was detected in the organ donor. By HRP immunohistochemistry, we visualized ChAT and CGRP in some but not all motoneuron cell bodies 24 h postmortem. These findings indicate that ChAT and CGRP, although at a low level, were still present 24 h postmortem in motoneuron cell bodies and it is recommended to consider HRP immunohistochemistry when performing postmortem studies.
As expected, immunofluorescence signals decreased over postmortem intervals, yet the staining intensity in the 48-h spinal cord was acceptable and high enough for qualitative information. Similar postmortem changes were observed in organ donor brains and reduced but stable immunoreactivity was present in neurons visualized with anti-tubulin and anti-neurofilament 48 h postmortem (Blair et al. 2016). It is however important to note, that the indicating that in the missing axons, ChAT signals were at a level too weak for being detected by the analysis software. These findings along different postmortem intervals indicate that qualitative immunohistochemical information can be obtained from body donor neuronal tissue over a Based on our observations, a postmortem interval of 72 h is too long for proper immunohistochemistry in neuronal tissue from body donors. At this time point, spinal cords exhibited substantial morphological changes due to autolysis and with the exception of the neurofilament antibody, immunohistochemical signals in the other antibodies (anti-ChAT, anti-synaptophysin, and anti-CGRP) were very low i.e., between 20 and 40% of the initial value. These postmortem changes are consistent with findings in the brain, where the synaptophysin signal significantly decreased 72 h after death (Liu and Brun 1995), and do not support data in autonomic nerve fibers of the heart where the signal drop was observed not until the 7th day postmortem (Chow et al. 1998). The low tissue quality coupled with strong immunohistochemical signal loss questions the scientific value of information obtained from human neuronal tissue 72 h after death.
The antibody-specific signal reduction observed in the present study is consistent with two older publications evaluating the postmortem effects on the immunohistochemistry of autonomic nerve fibers in humans (Gu et al. 1985; Chow  et al. 1998) and a recent study evaluating this effect in the brain (Waldvogel et al. 2006;Blair et al. 2016). After death, tissue proteins degrade and this reduces the binding capacity of antibodies and consequently the signal strength (Ramos-Vara and Miller 2014). Because the degradation of tissue proteins is temperature dependent, it is important to refrigerate body donors after arrival at body donor facilities. The stability of different tissue proteins varies in postmortem tissue (Waldvogel et al. 2006) and there are indications that structural proteins like neurofilament (Karlsson et al. 1991) and cell membrane proteins like synaptophysin (Wiedenmann and Franke 1985) are better preserved than soluble enzymes (ChAT) (Bruce and Hersh 1987). This was confirmed in the present study because the signal loss was less dramatic in neurofilament and synaptophysin staining compared to ChAT staining. CGRP a neuromodulator protein is similar well preserved as synaptophysin as shown here.
An additional discovery of the present study was that the signal intensity of the ChAT antibody varied in different compartments of the motoneuron. Specifically, in the 24-h spinal cord high levels of ChAT were present in the axons of the anterior root and in the synapses of the motoneuron cell bodies whereas low or no ChAT signal was inside the motoneuron cell bodies. This spatial variation of the staining intensity suggests that the ChAT enzyme disintegrates postmortem with variable speed in different motoneuron compartments and based on the low antibody signal, this process is quicker in cell bodies than in axons and synapses. Taken together, information on the stability of different neuronal proteins and variations of the protein stability in neuronal compartments is crucial and must be considered when evaluating immunohistochemistry in human postmortem tissue.