Retrospective electron microscopy: Preservation of fine structure by freezing and aldehyde fixation

ABSTRACT

For many years it has been believed that ultrastructural analysis by transmission electron microscopy (TEM) is not possible using frozen tissues. We have developed a TEM method that allows the evaluation of organelles using pancreatic tissue that was previously liquid nitrogen snap-frozen and stored long-term at −80°C. This method is suitable for the quantitative assessment of mitochondria, rough endoplasmic reticulum (RER), and Golgi structures, as well as organelles originating from autophagy signaling. Frozen pancreatic tissue exhibited no signs of freezing- or storage-related damage and was undistinguishable from fresh material subjected to standard glutaraldehyde fixation. Since pancreatic tissue is the most delicate tissue to work with due to the high expression of digestive enzymes, our method is also suitable for other tissue types such as liver. Thus, by applying proper tissue freezing and fixation techniques, retrospective TEM analysis can be performed on mammalian tissues in a time- and cost-saving manner.

Introduction

Transmission electron microscopy (TEM) has become the state-of-the-art approach for investigating autophagic activity by determining the content of autophagic vesicles in tissues and cells.¹ Autophagy constitutes one of the most rapidly growing research fields in biomedical science, and understanding its relationship to apoptosis and necrosis is a major challenge in translational medical research.

Autophagy, previously described as programmed cell death type II, is a dynamic, evolutionarily conserved process in which parts of the cytoplasm, including organelles and cytosolic components, are sequestered in double-membrane vesicles (autophagosomes) and then delivered to lysosomes where the autophagic cargo is degraded and recycled. This latter step involves the LAMP-2–dependent fusion of autophagosomes with lysosomes to form autolysosomes.^2,3 Autophagy is triggered by nutrient starvation or other environmental stress conditions and appears to be a preponderantly cytoprotective mechanism.^4-6 A widely-used marker of autophagy activity is monitoring the conversion of LC3-I to the electrophoretically more mobile LC3-II by immunoblot analysis. Levels of the autophagy marker LC3-II correlate well with the formation of autophagosomes.⁷ However, accumulation of this marker does not constitute firm proof that autophagosomes are formed to completion and subsequently fuse with lysosomes. Therefore, it is important to investigate the contents of autophagosomes and autolysosomes by TEM.

The pancreas consists of large exocrine and smaller endocrine portions: 85% of the pancreas is occupied by exocrine acinar cells, which synthesize and secrete digestive enzymes including proteases, nucleases, and lipases into the pancreatic duct.⁸ Pancreatitis, a potentially life-threatening disease, has been associated with premature activation of digestive enzymes, particularly trypsinogen, and self-digestion and destruction of parenchymal acinar cells.⁸ The pancreas produces huge quantities of digestive enzymes, more than any other organ of the mammalian body. This enzymatic activity makes it extremely difficult to work with pancreatic tissue. Without intensive precautions, cellular structures as well as nucleic acids and proteins may become rapidly degraded during or shortly after the isolation procedure.

Since autophagy is a rapidly growing research field, it is important to develop protocols for the TEM assessment of previously frozen specimens in a retrospective and time- and cost-effective manner. Here, we describe such a method that is suitable for evaluation of autophagy in pancreatic tissue associated with ethanol abuse.

Methods

Tissue harvest and storage

All experiments were approved by the Institutional Animal Care and Use Committee in accordance with the guidelines of the University of Heidelberg and the Federal Presiding Board of Animal Care Karlsruhe, Germany. Male rats weighing 150–175 g were fed a commercially available liquid diet (Lieber-DeCarli, BioServ, Frenchtown, NJ)^9-11 containing either ethanol (6% final concentration, equivalent to 36% of caloric intake) or maltose-dextrin in isocaloric amounts as described previously.^10,11 After 14 weeks of feeding with the ethanol or control diet, Gram-negative bacterial LPS (E. coli, 026-B6; 1 mg·kg⁻¹ body weight, i.v.) or vehicle control (sterile saline) were injected once intravenously. Rats were divided into 4 groups with 4 animals per group: pair-fed (PF) and ethanol-fed controls (EtOH), and pair-fed or ethanol-fed animals injected with LPS. Animals were euthanized 24 h after LPS injection with pentobarbital (Nembutal, 60 mg·kg⁻¹ body weight, i.p.).

Tissue cryofixation

The pancreas was carefully extracted at necropsy, immediately immersed in ice-cold PBS, and washed several time in ice-cold PBS. After removing connective and adipose tissue in ice-cold PBS, the pancreatic parenchyma was quickly wrapped in a pre-labeled piece of aluminum foil and snap-frozen in liquid nitrogen, followed by storage in a pre-cooled box at −80°C. More voluminous tissues, such as the liver, were cut into smaller pieces in ice-cold PBS before wrapping and immersion in liquid nitrogen. Long-term storage in aluminum foil prevents freezer burn, minimizes dehydration of the tissue, preserves tissue morphology, and facilitates space-saving storage at −80°C.

Post-aldehyde fixation

For TEM analysis, the tissue samples were taken from the −80°C freezer and kept frozen in liquid nitrogen. Each thin piece of frozen pancreatic tissue in aluminum foil was removed quickly and broken into pieces by hand while still in the foil. Individual frozen tissue fragments approximately 2 mm in diameter were randomly selected and quickly (while still frozen) immersed in freshly prepared ice-cold 1.5% paraformaldehyde in PBS on ice, followed by overnight fixation at 4°C. We tested several paraformaldehyde concentrations (1, 1.5, and 2%) for fixation and obtained the best results at 1.5%.

It should be noted that the tissue must be broken and immersed in paraformaldehyde while the fragments are still frozen. This step is essential, at least for delicate pancreatic tissue.

Tissue fixation methods

Freshly isolated pancreatic tissues were cut into small pieces (2 mm in diameter), fixed in 3% glutaraldehyde, and embedded in Epon. Ultra-thin sections were prepared for TEM according to a routine procedure for sample processing for TEM.

Sample processing for TEM

Tissue processing was conducted under standard conditions. Briefly, the tissue was washed 3 times in 0.2 M Sörensen phosphate buffer (pH 7.2) for 1 h and post-fixed in a freshly prepared solution of 1% aqueous osmium tetroxide in PBS for 1 h at RT. The tissue was rinsed twice for 5 min in Sörensen buffer, dehydrated in an increasing ethanol series (30, 40, 50, up to 100%, 10 min each), further dehydrated twice with propylene oxide for 10 min, and then incubated for 1 h in a 1:1 mixture of propylene oxide and epoxy resin followed by incubation in a 1:3 mixture of the same solvents overnight. The tissue was then incubated in epoxy resin at 40°C for 1 h to evaporate any residual propylene oxide. Resin polymerization was accomplished by immersing the tissue in labeled capsules with freshly prepared resin at 60°C for 48 h. For section overview, semi-thin Epon sections were cut with an ultra-microtome (Reichert-Jung Ultracut) into 1-µm semi-thin sections and stained with methylene blue. From the same tissue block, sections were cut into 50-nm ultra-thin sections and placed on 300 mesh grids. Ultra-thin sections were stained for 5 min each with uranyl acetate followed by lead citrate and subjected to TEM according to a published routine procedure.^10,12

Electron microscopic observation

Quantitative TEM was performed using a Zeiss 902 analytical electron microscope coupled with a Pro-Scan digital camera. Several tissue overview images were captured and evaluated for possible damage induced by the freezing procedure. From the overview, higher magnification images were selected randomly and areas within the captured images of the electron micrographs were measured using the Metamorph tracing tool for semi-automatic counting. In addition to autophagosomes, each mitochondrion and rough endoplasmic reticulum (RER) structure was carefully outlined and the area within the traced region was assessed using Analysis 3.2 Software. All areas were calibrated to scale bars on the micrographs and expressed as area in nm² and as the percentage of the entire tissue area measured, and were digitally evaluated in 4 or 2 animals per group in every other field of 100 randomly selected fields per animal as described recently.^10,12

Since every captured image was calibrated with a scale bar, software such as ImageJ or Adobe Photoshop is also suitable for measuring the area of the organelle profile and can be used to calculate the volume fraction of the organelle of interest.

Statistical analysis

Statistical analysis was performed by ANOVA, followed by Student's t-test for each experimental group treated with LPS or vehicle. Results were considered significant when the p-value was <0.05 (indicated with an asterisk) and were reported as mean ± standard error of the mean (SEM), as indicated in the figure legends. Statistical calculations were performed using GraphPad Prism 5 software.

Results

Methylene blue-stained sections

By light microscopy, semi-thin pancreatic tissue sections stained with methylene blue showed no notable signs of any freezing-related damage (Supp. Fig. 1A–C). All methylene blue-stained semi-thin sections prepared for TEM showed very similar tissue morphology to our previously formalin-fixed, paraffin-embedded tissue sections stained with hematoxylin and eosin (not shown),¹⁰ and no notable damage-related changes in tissue morphology were discernible by light microscopy.

General observations at low magnification

None of the pancreatic ultra-thin tissue sections that were prepared from the semi-thin sections and processed according to our method exhibited any signs of unusual alterations in the control tissue such as tissue holes, displacement of organelles, or any other possible freezing-related damage (Fig. 1A). In animals exposed to the combination of alcohol treatment and endotoxemia, a model of mild acute pancreatitis, we observed signs of tissue damage related to the treatment, including cellular vacuolization and pericellular fibrosis at low magnification (Fig. 1B). This tissue damage was not found in controls, implying that the pancreatic injury was caused by the in vivo treatment.

Figure 1.

TEM of pancreatic acinar cells at low magnification with cryo-post-aldehyde fixation. (A) Representative TEM image of pancreatic acinar cells from controls (scale bar = 10 µm). (B) Representative TEM image of pancreatic acinar cells from animals with pancreatitis induced by alcohol plus LPS treatment (scale bar = 10 µm). Note that there are no signs of tissue damage inflicted by freezing or long-term storage. Some degree of fibrosis within the extracellular space and zymogen granules can be observed.

General observations at high magnification

We next evaluated the integrity of acinar cell organelles from control pancreata at higher magnification. We found that our new method did not destroy any organelles including mitochondria, the RER, and the Golgi apparatus, all of which showed intact structures without discontinuities of intracellular membranes or other signs of deformation or destruction (Fig. 2A–C).

Figure 2.

TEM of pancreatic acinar cells from controls at higher magnification with cryo-post-aldehyde fixation. (A) Representative TEM image of intact mitochondria in pancreatic acinar cells (scale bar = 250 nm). (B) Representative TEM image of an intact RER in pancreatic acinar cells (scale bar = 500 nm). (C) Representative TEM image of an intact Golgi complex in pancreatic acinar cells (scale bar = 250 nm). None of the images show signs of tissue damage inflicted by freezing or long-term storage.

Morphometric quantitation of mitochondria, RER, and autophagosomes

Mitochondria and RER structures were carefully outlined and the volume fraction of mitochondria and the RER were determined digitally. First, we determined the fraction of ultrastructurally defective mitochondria as a percentage of the total volume or as a percentage of the total number of mitochondria in control pancreata as well as organs retrieved from animals that had been treated with alcohol and/or LPS (Fig. 3A, B). The volume fractions (Fig. 3B, left) and the numbers of mitochondria (Fig. 3B, right, Suppl. Table 1) were found to be very similar, suggesting that both parameters are suitable for accurate determination of tissue mitochondrial damage by TEM morphometry. All three treatments (LPS alone, alcohol alone, and the combination) significantly increased injury to acinar cell mitochondria by 30% to 40%. This result confirmed that pancreatitis is associated with severe mitochondrial damage.

Figure 3.

Quantitative assessment of the volume fraction and number of defective mitochondria in pancreatitis from cryo-post-aldehyde fixation tissue. (A) Representative TEM image of defective mitochondria in a pancreatic acinar cell induced by chronic alcohol exposure and endotoxemia (scale bar = 1 µm). (B) Quantitative assessment of the volume fraction (left) and number (right) of defective mitochondria measured in pancreatitis tissue. Volume fraction and number were obtained from Suppl. Table 1. Data are plotted as means ± SEM of 2 individuals per group. *, p < 0.0001 versus pair-fed 24 h after LPS; **, p < 0.0005 vs. ethanol-fed rats; ***, p < 0.0028 versus ethanol-fed rats 24 h after LPS; **** p < 0.0028 vs. ethanol-fed rats 24 h after LPS.

We next investigated the effects of alcohol exposure and endotoxemia on the abundance of RER by TEM morphometry. RER structures significantly decreased by approximately 40% upon alcohol exposure, LPS injection, and combined treatment (Fig. 4A, B; Suppl. Table 2).

Figure 4.

Quantitative assessment of the volume fraction of the rough endoplasmic reticulum in pancreatitis in cryo-post-aldehyde fixation tissue. (A) Representative TEM image of pancreatic acinar cell rough endoplasmic reticulum induced by chronic alcohol exposure and endotoxemia (scale bar = 1 µm). (B) Quantitative assessment of the volume fraction of the RER measured in pancreatitis tissue. Volume fractions were obtained from Suppl. Table 2. Data are plotted as means ± SEM of 2 individuals per group. *, p < 0.0001 versus ethanol-fed rats 24 h after LPS; **, p < 0.0005 vs. ethanol-fed rats 24 h after LPS. Note that one defective and one intact mitochondria are visible together with the disrupted RER.

Next, we determined the volume fraction (Fig. 5B, left) and the number of autophagosomes in the entire area (Fig. 5B, right), obtaining similar results (Suppl. Table 3). Autophagosomes were identified as vesicles containing cytoplasmic material, including zymogen granules, within a ribosome-free double membrane structure.¹³ We found that most autophagosomes contained engulfed zymogen granules (Fig. 5B). Such a process has been referred to as “zymophagy” and has been recently described in pancreatitis.¹⁴ The volume fraction of autophagosomes/zymophagosomes significantly increased in all treatment groups compared with controls (Fig. 5A/B). LPS, alcohol, and the combination increased the volume fraction of autophagosomes by 6-, 3- and 9-fold, respectively. Thus, all of these treatments initiated autophagy signaling, although to different extents. Ethanol plus LPS increased the autophagosomal fraction by 1.4-fold compared with LPS alone (Fig. 5A), indicating that the combination treatment induced the highest levels of autophagy/zymophagy. Counting the numbers of autophagosome profiles is a widely used approach in autophagy research.¹ In our investigation, counting the numbers of autophagosome profiles correlated well with the autophagosomal volume fraction (Fig. 5B), consistent with the fact that autophagosomes can be heterogeneous in size and shape. Therefore, we conclude that dysregulated zymogen granule trafficking and premature trypsinogen activation, which are both known to be induced in pancreatitis, are associated with the induction of zymophagy.

Figure 5.

Quantitative assessment of the volume fraction and number of autophagosomes in pancreatitis in cryo-post-aldehyde fixation tissue. (A) Representative TEM image of a pancreatic acinar cell autophagosome induced by chronic alcohol exposure and endotoxemia (scale bar = 0.5 µm). (B) Quantitative assessment of the volume fraction (left) and number (right) of autophagosomes measured in pancreatitis tissue. Volume fractions and numbers were obtained from Suppl. Table 3. Data are plotted as means ± SEM of 2 individuals per group. *, p < 0.0001 versus pair-fed 24 h after LPS; **, p < 0.0005 vs. ethanol-fed rats; ***, p < 0.0028 versus ethanol-fed rats 24 h after LPS.

Comparison of the new fixation method with the standard protocol

Finally, we compared TEM images obtained by classic glutaraldehyde fixation of fresh tissues with those obtained using the novel method described above. Using the classic method, acinar cells from control pancreata contained more electron-dense zymogen granules (Fig. 6A) than we observed with the cryo-post aldehyde fixation method. In pancreatitis induced by alcohol plus LPS, a void space became visible between zymogen granules and RER, comparable to results with the cryo-post aldehyde fixation method (Fig. 6B). Also similar to the cryo-post fixation method, the classic method revealed disrupted intracellular areas with the loss of organelles and the presence of extracellular collagen fibers in the context of pancreatitis (Fig. 6C). In addition, injured mitochondria became visible (Fig. 6D), further confirming that the cryo-post aldehyde fixation and the classic glutaraldehyde fixation methods yield comparable results.

Figure 6.

TEM of pancreatic tissue using standard glutaraldehyde fixation. (A) Representative TEM image of pancreatic acinar cells from an untreated control animal at low magnification (scale bar = 5 µm). (B) Representative TEM image of pancreatic acinar cells from an animal with pancreatitis (scale bar = 200 nm). Note the void spaces between zymogen granules and the rough endoplasmic reticulum as indicated by arrows. (C) Representative TEM image of pancreatic acinar cells from an animal with pancreatitis (scale bar = 200 nm). Note the loss of rough endoplasmic reticulum (arrow) and the presence of collagen fibers indicating fibrosis. (D) Representative TEM image of pancreatic acinar cells from an animal with pancreatitis (scale bar = 500 nm). Note the damaged mitochondria indicated by arrows.

Discussion

The method described here provides an alternative approach to the widely used fixation of fresh tissue with glutaraldehyde and is compatible with the retrospective study of previously frozen pancreas tissues. Our study shows that snap-frozen cryo-fixation of tissue pieces followed by 1.5% paraformaldehyde post-fixation is suitable for TEM analysis. No signs of freezing- or storage-related damage were observed in any pancreatic tissue samples.

Using the post-cryo fixation method we were able to quantitatively evaluate mitochondrial and RER integrity as well autophagic activity by determining the volume fraction of these organelles in proportion to the entire tissue area. The volume fractions correlated well with numbers for both mitochondria and autophagosomes, in line with the observation that the mean size of these organelles showed little, if any, variation in the context of the 2 stimuli used to induce pancreatitis, namely ethanol and LPS, either alone or in combination.

Although the post-cryo fixation method yielded a lower percentage of electron-dense zymogen granules than the standard glutaraldehyde fixation, it appears that pathologic alterations of such granules leading to zymophagy are readily detectable by the new method, concordant with the overall notion that damaged zymogen granules play a critical role in pancreatitis.

In summary, our new tissue fixation method yields high-quality TEM organelle images with the advantage of being simple and economical. The post-cryo fixation method is compatible with micromorphological and ultrastructural evaluation of organelles.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We would like to acknowledge excellent electron microscopy support from Peter Rieger, Zlata Antoni, and Ingrid Hausser-Schiller (University of Heidelberg electron microscopy core facility). This study was supported by a DFG Excellent Initiative Frontier award (F. F.) and institutional funding (M.W.B. & T.H.). GK is supported by the Ligue contre le Cancer (équipe labelisée), Agence National de la Recherche (ANR) – Projets blancs, ANR under the frame of E-Rare-2, the ERA-Net for Research on Rare Diseases, Association pour la recherche sur le cancer (ARC), Cancéropôle Ile-de-France, Institut National du Cancer (INCa), Institut Universitaire de France, Fondation pour la Recherche Médicale (FRM), the European Commission (ArtForce), the European Research Council (ERC), the LeDucq Foundation, the LabEx Immuno-Oncology, the SIRIC Stratified Oncology Cell DNA Repair and Tumor Immune Elimination (SOCRATE), the SIRIC Cancer Research and Personalized Medicine (CARPEM), and the Paris Alliance of Cancer Research Institutes (PACRI).

Author Contributions

Study concept, analysis and statistic: G.K. & F.F. Supervision and funding: M.W.B., T.H., G.K. & F.F. Manuscript writing: M.W.B., G.K., T.H. and F.F.