THE main purpose of chemical fixation is to preserve tissue and subcellular architecture thereby facilitating immunohistochemical detection of epitopes and histochemical detection of chemical substances within cells. Formaldehyde-based fixatives fulfill this requirement to a large extent and have been the preferred compounds for decades in standard pathology procedures. However, a growing demand has emerged for enhanced preservation of proteins and other cellular molecules such as DNA and RNA. In particular, formaldehyde fixation can fragment genomic DNA and hinder quantitative analysis of nucleic acids and generally prohibits PCR amplification of DNA fragments longer than a few hundred base pairs (1). Besides the deleterious effects on DNA, formaldehyde cross-linking alters the tertiary and quaternary structure of proteins and may also mask the secondary structure of globular proteins. Consequently, this can diminish the availability of cellular epitopes to the detriment of immunological-based assays (2). Finally, most formaldehyde fixatives are hazardous and suspected carcinogens, which makes their use dependent on working under a fume hood.
As an alternative to formaldehyde-based fixation, zinc salt-based fixation (ZBF), which does not contain any cross-linking agents, is also effective in histopathology procedures (3–6). In fact, studies have shown that RNA, DNA, and proteins are preserved better in ZBF than in neutral buffered formalin (5, 6). Furthermore, it was shown that the morphological preservation was comparable to aldehyde-based fixation. Recently, it has been shown that enzymes retain their activity in ZBF-fixed tissue enabling analysis of enzymatic activity in tissue sections (7). However, the use of ZBF has not received a wide acceptance in routine pathology where neutral buffered formaldehyde is still the preferred fixative. Many reference antibodies have been selected based on the performance in formalin fixed tissue, which makes the transition difficult.
The aim of this study was to develop a simple, cost-effective and nonhazardous fixation method for cell suspensions that preserves all cellular structures and enables flow cytometric analysis of both surface proteins, intracellular proteins, DNA profiles and pulse-labeling using the thymidine analog EdU in the same cell sample. Since the method seems to preserve the surface epitopes well they can be stained post fixation and simultaneously with the staining of the intracellular compartments.
MATERIALS AND METHODS
Fixation of Cells
ZBF-fixation: Cells were suspended in 1 vol. PBS. Ten vol. ZBF-buffer (0.1 M Tris-HCl pH 7.8, 0.05% calcium acetate (CH3COO)2Ca, 0.5% zinc acetate (CH3COO)2Zn, 0.5% zinc cloride ZnCl2) was added while vortexing. The cells were incubated at 4°C over night. After overnight fixation the cells were either frozen in glycerol (1:1) or washed carefully three times in TBS prior to use (total 30–45 min). For preservation of RNA, cells were immediately frozen in glycerol (1:1) after addition of the ZBF-buffer. PFA-fixation: Cells were fixed in 4% paraformaldehyde for 30 min on ice.
Primary adult epidermal keratinocytes were isolated from seven-week-old K14-EGFP-actin mice kindly provided by Elaine Fuchs, Rockefeller University, New York, NY or C57Bl/6 (Taconic, Ry, Denmark) as previously described (8). Freshly isolated epidermal keratinocytes were suspended in 0.1% BSA/PBS and surface markers stained with PE-conjugated CD49f (BD Biosciences Pharmingen, San Diego, CA), PE-Cy7-conjugated Ly-6A/E (Sca-1, BD Biosciences Pharmingen) and Alexa Fluor 647-conjugated CD34 (eBiosciences, San Diego, CA) antibodies. For intracellular staining, fixed cells were suspended in 0.1% BSA/PBS containing 0.2% saponin and stained using a mouse monoclonal antikeratin 10 antibody (LHP2) (9) followed by staining with a secondary antibody conjugated to Alexa 680 (Molecular Probes, Invitrogen, Eugene, OR). Only the freshly isolated keratinocytes were stained with PI to eliminate dead cells since PI will enter the fixed cells and prevent the use of red fluorochromes. Cells were subjected to flow cytometry analysis using a FACSAria sorter and FACS DiVa 4.1 software (BD Biosciences, Franklin Lakes, NJ). The instrument is equipped with three lasers: Near-UV 375 nm, blue 488 nm, and red 633 nm. Hoechst is exitated by the 375 nm laser and emitted light is collected through a 424/42 filter. FITC, GFP, PE, and PE-Cy7 are exitated by the 488-nm laser and the emitted light is collected through 530/30, 530/30, 585/42, and 780/60 band pass filters, respectively. APC is exitated by the 633 nm laser and the emitted light is collected through a 660/20 band pass filter. In all cases, cells were analyzed using a 100-μm nozzle and 20 psi.
Cell Cycle Analysis
Cells were resuspended in Hanks Buffered Salt Solution (HBSS) containing 1 μM Hoechst 33342 (Molecular Probes, Invitrogen), 0.2% saponin and 0.1% BSA and incubated at RT for 15–60 min. In some cases Hoechst 33342 was added together with the antibodies and incubated 30 min on ice. After washing off excess antibodies, Hoechst 33342 is added once again to the cell suspension. Pulse processing was used to gate out doublets.
EdU Treatment and Detection
For a short-term pulse, EdU (2 or 4 mg/mouse, Molecular Probes, Invitrogen) was injected s.c. into seven-week-old C57Bl/6 mice once and primary adult epidermal keratinocytes were isolated 30 min or 24 h later as previously described (8). To obtain label-retaining cells, EdU (0.05 mg/g mouse) was injected s.c. every 12 h on Days 10 and 11 and the mice sacrificed when seven weeks old. The isolated cells were subjected to ZBF- or PFA fixation. EdU was detected according to manufactures manual using Alexa Fluor 488 azide and combined with surface and DNA content stainings.
Analysis of DNA and RNA
RNA was isolated from fixed (ZBF or PFA) or nonfixed cells using the High Pure RNA Tissue Kit (Roche, Hvidovre, Denmark). cDNA was synthesized from RNA using the 1st strand synthesis kit for RT-PCR (AMV) (Roche) according to manufacturer's instructions. PCR was performed on DNA or cDNA from fixed (ZBF or PFA) or nonfixed cells using primers recognizing regions of different size in the β-actin gene (1s: 5′-CTGTGCTGTCCCTGTATGCC-3′, 3s: 5′-GCTGGTCGTCGACAACGGCT-3′, 1as: 5′-GTGGTGGTGAAGCTGTAGCC-3′, 3as: 5′-TCCTGTCAGCAATGCCTGGGT-3′, 4as: 5′-CGCAGCTCAGTAACAGTCCGC-3′). Primer combinations, size of products and PCR conditions can be found in Table 1. A PCR reaction was performed on RNA to verify that no contaminating DNA was present in the RNA preparation. Experiments were duplicated in order to verify the results.
Table 1. Combination of primers and PCR conditions used for PCR and RT-PCR of β-actin
94°C 30 s, 59°C 30 s, 72°C 1 min, 30 cycles
94°C 30 s, 59°C 30 s, 72°C 1 min, 30 cycles
94°C 30 s, 64°C 30 s, 72°C 2 min, 30 cycles
94°C 45 s, 60°C 1 min, 72°C 2 min, 30 cycles
Preservation of Scatter and Surface Markers After ZBF-Fixation
We used cutaneous mouse epithelium as a model tissue as it contains several distinct keratinocyte progenitor cell populations that are delineated by their surface marker composition and different labeling characteristics regarding DNA analogs (10). First, we investigated the effect of fixation on the maintenance of scatter and fluorescence parameters on surface-stained cells. Primary adult epithelial keratinocytes were isolated from K14-GFP mouse skin, fixed using either ZBF or PFA and stained with antibodies against the surface markers CD34, Sca-1, and α6-integrin (10). The flow cytometric profile of ZBF-fixed cells was compared with the profile of live cells (Fig. 1). Minor differences were observed in the light scatter signal from ZBF-fixed cells, which may be due to small changes in size and granularity of the cells. However, the staining pattern of the three surface markers in ZBF-fixed cells was similar to both paraformaldehyde-fixed and live cells (Fig. 1).
ZBF fixed leukocyte preparation retain scatter parameters so that lymphocytes, monocytes, and granulocytes can be recognized as separate populations (data not shown).
Preservation of Intracellular Markers After ZBF-Fixation
In addition to preserving surface epitopes it is equally important that a fixative preserves intracellular epitopes after permeabilization of the cells. We used saponin to permeabilize the cells since triton permeabilization resulted in complete loss of certain surface antigens such as CD34 (data not shown). PFA- and ZBF-fixed cells from K14-GFP mice were stained with antibodies against the intracellular marker keratin 10 (Fig. 2) in addition to the surface markers described above (data not shown). The keratin 10 staining pattern was similar following PFA and ZBF fixation methods, and ZBF fixation also preserved GFP epifluorescence (Fig. 2). In addition, effective preservation of other intracellular markers including keratin 14, H2AX, and XRCC1 was also observed after ZBF fixation (data not shown).
Combined Analysis of DNA Content, Cell Proliferation and Surface Markers in the Same Cell Sample
In many applications it is useful to be able to combine analysis of surface and intracellular markers with measurements of cell proliferation and DNA content. Classically, measurement of cell proliferation has been performed via BrdU-labeling. However, BrdU detection requires HCl or DNase treatment in order to generate single-stranded DNA that can be accessed by BrdU antibodies. This process is incompatible with maintenance of intact DNA profiles, and many antigens are also sensitive to HCl treatment. Alternatively, measurement of cell proliferation can be performed using the nucleotide analog EdU, which does not require denaturation or digestion of the DNA. To determine if we could combine EdU detection of DNA content together with cell surface marker labeling, we isolated primary keratinocytes from mice injected with EdU 30 min and 24 h before harvesting and fixed the cells in ZBF. Following EdU labeling, the cells were stained using the surface markers CD34, Sca-1, and α6-integrin and DNA was costained using Hoechst 33342. As seen in Figures 3A and 3B, a subset of proliferating cells could be detected together with the DNA profile after 30 min and 24 h EdU pulses. The profiles are, however, different. Twenty-four hours after the injection it appears that the proliferating cells are all in the G1 phase. This indicates that EdU is rapidly removed from the circulation and when the mice are sacrificed at 24 h the labeled cells have completed the cells cycle, divided and returned to the G1 phase.
Importantly, label retaining keratinocyte stem cells could also be detected using this technology (Fig. 3C). As shown in Figs. 3D–I, we were able to delineate proliferating cells (30-min pulse; Figs. 3D–F) and label-retaining cells (Figs. 3G–I) in combination with analysis of multiple cell surface protein (Figs. 3D and 3G). After a short-term pulse, the EdU-labeled cells are mainly found in the CD34−Sca-1+ proliferating population, whereas the label-retaining cells mainly are found in the CD34+Sca-1− stem cell population (10).
Quality of Epitopes After Long-Term Storage
To investigate the stability of the epitopes after long-term storage, cells stored for 14 months at −20°C in glycerol were stained with the surface markers CD34, Sca-1, and α6-integrin (Figs. 4A and 4B). The staining pattern was similar to the pattern in newly fixed cells. No major differences were observed in the preservation of GFP in PFA- and ZBF-fixed cells from K14-GFP mice after four weeks of storage at −20°C (Figs. 4C and 4D). At 4°C the profiles are stable for up to two to three weeks (data not shown).
Preservation of DNA and RNA After ZBF-Fixation
The DNA yield from ZBF fixed cells was three times higher than from PFA fixed cells. Analysis of the DNA by gel electrophoresis show that DNA from ZBF-fixed cells mainly consists of high molecular weight bands whereas DNA from PFA fixed cells are degraded to some extent (Fig. 5A). PCR analysis of the DNA using primers recognizing different regions in the β-actin gene show that fragments from 196 bp to 1890 bp could be amplified from both nonfixed, PFA- and ZBF-fixed samples (Fig. 5B). To determine whether RNA was preserved after ZBF, RT-PCR analysis was performed on RNA from ZBF- and PFA-fixed cells using primers recognizing a region in the β-actin gene. The RNA was degraded in both PFA-fixed cells and cells fixed in ZBF overnight at 4°C (Fig. 5C). If, however, the cells were frozen immediately after addition of the ZBF-fixative at −20°C, the RNA was preserved and RT-PCR possible (Fig. 5D). Evaluation of surface and intracellular markers in the “flash frozen” cells showed that proper FACS profiles could be obtained on these cells too (data not shown).
There has been increased focus on using ZBF for preparation of tissue samples for immunohistochemical analysis (3–6). It has been shown that ZBF-fixed tissues are well preserved, exhibit improved immunolabeling of sensitive epitopes and maintain intact DNA and RNA to a higher extent compared to other fixatives. Furthermore ZBF fixation has the advantage that it is nonhazardous, easy to prepare and inexpensive. There is only sparse data on the molecular mechanisms of zinc fixation. However, for decades zinc salts have been used as additives in formaldehyde-based fixatives in order to improve the morphology. Acetate is a chaotropic anion and may act to destabilize protein structure by interfering with hydrogen-bond formation. By contrast, it is known that Zn ions can stabilize certain parts of the tertiary structure of proteins when subjected to chaotropic reagents (12). We suggest that the combined effects of the zinc and acetate ions work to introduce structural changes or even a mild denaturation of the proteins in the specimen, which e.g. hardens the proteins in the cell membrane into a more rigid layer.
In this study we show that the ZBF is in addition very well suited for fixation of single cells for flow cytometry. ZBF fixation enables analysis of multiple parameters on the same cell sample because both proteins, DNA and RNA are preserved. We demonstrated that the FACS profile of surface markers was comparable to the profile of live cells. Staining of surface markers in the ZBF-fixed cells could be combined with analysis of intracellular markers, DNA profiles, and cell proliferation enabling analysis of various parameters in different cellular subsets. One major advantage of this method is the long-term stability at −20°C. This enables collection of cell samples at different time points over long periods of time and allows us to process and analyze all collected samples at the same time. Thereby errors such as day-to-day variation in cell staining and data collection can be avoided. No upper storage-limit has yet been observed when the cells are stored at −20°C. It is well known that formaldehyde-based fixation leads to fragmented DNA. Here we show that the DNA from ZBF-fixed cells are preserved in a high molecular weight form and that fragments of at least 1890 bp could be amplified by PCR. Previously it has been shown that PCR fragments larger than a few hundred base pairs in size were difficult to obtain using DNA from NBF-fixed tissue (1, 5). In this study, we were able to amplify DNA fragments of at least 1890 bp using DNA from PFA-fixed cells. This difference may be due to the fact that PFA degrades DNA to a minor degree than NBF. Replacement of zinc acetate with zinc trifluoroacetate in ZBF buffer is reported to improve DNA, RNA and protein preservation of fixed tissue (5); however, in our analysis on cell suspensions, we were not able to measure any differences between the two ZBF fixatives (data not shown). FACS profiles were similar and the RNA was still degraded when incubated in the modified ZBF buffer over night and preserved when frozen immediately after fixation. However, we did not analyze the RNA using a Bioanalyzer, which may have showed a difference in the RNA quality between the two fixatives.
One reported disadvantage of ZBF is incomplete penetration of tissues compared to NBF (5, 11) resulting in well-fixed edges but poorly-fixed tissue interiors. ZBF fixation caused shrinkage in all organs tested so far, which is confirmed by a slight change in scatter parameters when used for flow cytometry. However, the fixative is able to penetrate deep enough into normal pathological samples to allow sectioning and immunohistochemical processing. We would therefore argue that the issue raised by both Lykidis and Rothaeusler (5, 11) is of minor relevance to single cell suspensions. This is supported by the fact that DNA and other nuclei components such as H2AX and XRCC1 are preserved in our samples (data not shown). We therefore conclude that complete penetration of the ZBF-fixative must occur in single cell suspensions. In this study we used EdU instead of BrdU to label proliferating cells. In addition to the advantage that the DNA does not need to be denatured or digested, the mice seemed to tolerate EdU better. For long-term chase studies mice injected with BrdU gained weight slower and the hair was lost in patches of the back in contrast to EdU-treated mice that gained weight faster and did not loose the hair (data not shown). Therefore, EdU may be more appropriate for analyzing DNA profiles in epithelial progenitor cells residing in hair follicle appendages.
It has been shown that detergents such as NP40 and organic solvents such as 95% ethanol commonly used to permeabilize cell membranes cause cell clumping and loss of APC signals (11). All the fluorochromes we have tested (Alexa488, Alexa647, Alexa680, Alexa700, FITC, PE, PE-Cy7, APC-Cy7) worked well on ZBF-fixed and saponin-permeabilized cells. In addition, ZBF fixation combined with saponin permeabilization preserved some surface antigens better than permeabilization using triton. Cell clumping can occur using ZBF; however, it is easily avoided by gentle vortexing of the cell pellet upon addition of ZBF.
Hicks et al. (4) have reported incomplete inactivation of pathogens and therefore precautions should be taken when using the ZBF on potentially infectious material.
We have shown that ZBF fixation of cells is well suited for flow cytometry. Proteins and DNA are well preserved enabling analysis of DNA profiles in combination with surface and intracellular proteins. Immediate freezing of the samples in the ZBF fixative also enables analysis of the RNA. The use of a nonhazardous fixative and the possibility of long-term storage of cell samples allow more simple and flexible routines.
The authors are grateful to Anette Thomsen for excellent technical assistance.