The analysis of cellular signaling pathways has largely been limited to biochemical manipulations on large numbers of homogeneous cells either derived in culture or purified by cell sorting techniques. Thus, systems with rare or heterogeneous cell types, such as human peripheral blood mononuclear cells (PBMC), murine splenocytes, bone marrow cells, and other primary cells have largely remained outside the scope of biochemical analysis. In addition, theories on signaling pathways, such as supersensitivity to signaling thresholds and bimodality, require analysis of cells at the single-cell level (1–3). However, recent advances in staining for intracellular epitopes by flow cytometry have opened these previously obscure fields of signaling for clearer analysis (4–6).
Many epitopes have been successfully stained within cells, including viral particles (7, 8), immunoglobulins (9), estrogen receptors (10), cytokines (11, 12), and specific proteins such as Bcl-2 (13) and cyclooxygenase (14). Though staining of static protein molecules can provide insight into cellular responses to stimuli in long-term experiments, it does not yield information concerning the dynamic signaling events that occur rapidly after cell stimulation or stress. It is well known that signaling cascades are often driven by protein phosphorylation on downstream effectors that activate the effectors to carry out their roles. Thus, phospho-specific antibodies that recognize these active proteins might distinguish the “on-off” state of signaling events.
Several groups have demonstrated staining of phospho-epitopes for flow cytometric analysis (6). Among the molecules previously examined are Stat1 (15, 16), Stat4 (17), Akt (18, 19), ERK, MEK (5), cJun, p38 (20), and various others (4). The methods used to prepare cells for staining with phospho-specific antibodies differed in each case, but they generally employed a fixation step with formaldehyde, followed by permeabilization with alcohols, detergents, or saponin. Because many of the epitopes to be recognized are novel and might be sequestered in protected locales within cells, it remains unclear if there exists a general method (or set of methods) by which most phospho-epitopes can be stained and analyzed. For the success of such a protocol, two critical parameters must be met: 1) the initial cell-fixation step must be rapid and effective in “freezing” the phosphorylation status of proteins, and 2) permeabilization steps must allow antibody access to their cognate epitopes, in the proper natured or denatured conformation, for binding. Development of a generally applicable protocol would greatly benefit multicolor, multiparameter analysis. It would also become important as new phospho-specific antibodies are created.
Therefore, we took a systematic approach to examining various intracellular phospho-protein staining protocols in two model systems, the mitogen-activated protein (MAP) kinase cascade in Jurkat T cells (21) and the Janus kinase-signal transducer and activator of transcription (Jak-Stat) pathway in monocyte-like U937 cells (22, 23). Both cascades utilize phosphorylation as a means of activating downstream effectors. In particular, the MAP kinases, extracellular regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38 are doubly phosphorylated (on Thr and Tyr residues); then they translocate into the nucleus to phosphorylate various transcription factors. Stat proteins are activated by growth factors and cytokines, such as IFN-γ, IL-4, and GM-CSF. Upon phosphorylation by Jaks, Stat proteins dimerize and enter the nucleus, where they bind to DNA directly to modulate transcription.
To further develop flow cytometric approaches to determine phospho-epitope levels, we directly compared several current methods of cell preparation, including fixation and permeabilization, staining, and storage. Reagents that are commonly used in both immunofluorescence and flow cytometry were employed: formaldehyde and glutaraldehyde as fixatives; and saponin, alcohols, and detergents as permeabilization reagents (24). We found that each of these reagents had benefits and drawbacks, depending upon the circumstances in which they were employed. We also examined fluorophore choice as a variable in the labeling of phospho-specific mAbs. We found that formaldehyde fixation, followed by permeabilization by methanol, provides an extremely rapid and efficient means to stain a variety of intracellular phospho-epitopes, including ERK, p38, JNK, and Stats 1, 5, and 6. We also observed that phospho-epitopes are largely stable in the conditions defined herein.
The results demonstrate that it will be possible to further extend the technique to analysis of cellular subsets within complex populations or rare cell populations that are difficult to study biochemically, set up diagnostic flow cytometric assays for pathologic human samples based on phospho-protein status, or screen primary cell populations against molecular libraries to discover novel inhibitors and activators of kinase signaling cascades. We expect that variations of the techniques employed will allow for other forms of intracellular staining to be undertaken, including other post-translational modifications (e.g., ubiquitinoylation, glycosylation, methylation, acetylation) or protein-protein interactions.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
We developed a generalized protocol for staining phospho-epitopes for flow cytometry by examining a spectrum of conditions and parameters that might impact staining levels and efficiency (see Materials and Methods). Testing indicated that the optimal protocol was resilient to modest experimental changes that could be expected to occur during staining and in the course of standard laboratory practice (i.e., time and temperature). Formaldehyde concentrations from 0.5% to 3%, fixing times from 5-30 min, permeabilization times from 10 min to 16 h, and staining times from 15 min to 1 h altered staining levels slightly, but did not affect qualitative observations of phosphorylation (i.e., whether or not a particular protein is phosphorylated). This inherent flexibility could promote high reproducibility of staining across experimental conditions. It is clear that flow cytometric analysis of phosphorylation provides similar qualitative results to Western blot, with intermediate changes apparent by both techniques (Fig. 1). Absolute fold changes differed when comparing Western blots to flow cytometry; however, normalized changes correlated closely (Fig. 6). For quantitative work and comparisons between experiments, however, one must keep all of the experimental parameters constant, as 30% (or more) variability is seen when fixation and permeabilization times are changed (e.g., fixing for 30 min instead of 10 min, as seen in Fig. 4). Work is underway in our laboratory to determine the quantitative nature of flow cytometry for the analysis of phospho-epitopes, with the goal of normalizing results to total protein levels.
We obtained the best results by first fixing cells with formaldehyde, then permeabilizing with an alcohol such as methanol. Formaldehyde performs two critical roles in the protocol. First, it freezes cellular processes by cross-linking proteins to one another and to themselves, likely creating a lattice of static proteins within the cell and preventing further signaling. Second, it stabilizes cell structure, a process which could be considered critical for experiments involving human peripheral blood or murine splenocytes, where phenotypic gating by light scatter is an important component. As shown in Figure 3, the formaldehyde/methanol combination maintains scatter properties of splenocytes as well as the commonly used formaldehyde/saponin combination does. However, after formaldehyde fixation, one would expect that intracellular proteins are in a mostly natured (or “native”) state. Permeabilization with reagents such as saponin and Triton X-100 is not expected to produce significant denaturation. One could expect that denaturation is a critical component of the staining process, because phospho-specific mAbs are often generated against linear peptides corresponding to the regions surrounding the various phospho-epitopes. Thus, certain classes of mAbs may be particularly suited to recognize denatured antigens, as in Western blots. Methanol, through its dehydrating effects, may lead to some protein denaturation. In addition, methanol permeabilizes cells very effectively and allows efficient and rapid access to nuclear antigens, such as the Stat transcription factors and activated MAP kinases tested herein. Therefore, the combination of formaldehyde and methanol might be expected to give optimal results for phospho-epitope staining with peptide-specific mAbs, particularly for nuclear antigens. Further work is required to determine whether this combination will be as effective for cytoplasmic and plasma membrane-bound epitopes.
The availability of multiple fluorophores for labeling the phospho-mAbs make multiparameter flow cytometry plausible (Table 1 and Fig. 6). Alexa 647 tended to give better results in our experiments, most likely because of greater cellular autofluorescence near the emission of Alexa 488. However, this fluorophore requires excitation at 633 nm, and therefore requires a two-laser cytometer. For researchers using cytometers with only a 488 laser, the Alexa 488 conjugates provided adequate fold changes, but ensuring optimal staining may require a more careful experimental technique. We are currently looking to expand the number of colors available for analysis. Work with PE conjugates of the mAbs showed that these reagents work in some cases, but not all (e.g., see Fig. 5). The PE fluorophore is bulky and may slow staining times or block antigen binding sites.
When setting up experiments to analyze phospho-epitopes intracellularly via flow cytometry, one must keep several parameters in mind: 1) stimulation conditions must be found that efficiently induce phosphorylation of the target protein; 2) fixation must be rapid enough to freeze intracellular events; 3) fixation time must be optimized for the epitope of interest (though 10 min appears to be a good starting point); 4) stability of the epitope of interest (in methanol) must be ensured prior to long-term storage; 5) if using saponin, the amount of saponin used (typically 0.1% to 0.5%) must be optimized; 6) optimal antibody titration on stimulated versus unstimulated cells must be determined; 7) the antibodies must be labeled with optimal FTP ratios (typically FTP ratios of 2-4 are sufficient); and 8) staining time must be optimized. Though the protocol outlined in this paper serves as a starting point for other phospho-epitopes, small changes may lead to greatly enhanced signals for as yet undeveloped antibodies. In addition, one must keep in mind the difference between qualitative and quantitative work. Determining whether or not a protein is phosphorylated in the qualitative sense is not critically dependent on the parameters listed above, because on a logarithmic scale 10-fold shifts do not appear much different than 7-fold shifts. However, determining quantitative fold changes requires careful control of fixation, permeabilization, and staining times, as well as antibody labeling and titration, which can all affect fold change up to 30% or more.
Although the flow cytometric methods outlined in this article give similar results to Western blotting, analysis of phospho-epitope levels by flow cytometry possesses many advantages versus Western blotting, including: 1) a large dynamic range of data collection (up to 10,000-fold); 2) rapid protocols that take about 2 h instead of the 1-2 days needed for Western blots; 3) simultaneous analysis of multiple epitopes in the same cell; 4) possibility for analysis in complex populations such as peripheral blood or murine splenocytes; and 5) possibility for rapid quantitative experiments. Western blots do, however, have the advantages of: 1) showing antigen size/MW; 2) indicating antibody specificity; and 3) the possibility of subcellular fractionation prior to analysis.
Work is currently underway to combine the intracellular staining methods with surface staining. Initial results suggest that surface staining can readily be combined with the optimized intracellular staining protocols described here, though we have early indications that certain surface epitopes are compromised by methanol. We expect that it will also be possible to read out non-phospho-epitopes of a variety of proteins within cells, further expanding the utility of the protocol. Application of intracellular staining protocols will allow for the examination of pathways in detailed time courses and pathway-specific manners that have previously been unavailable. We expect that the provided protocol for phospho-epitope staining will provide an opportunity to study signaling in cells that were not accessible by biochemical techniques. In addition, because of the speed of analysis, large screening experiments can be performed to readily identify specific signaling pathway modulators at the cellular level, avoiding the pitfalls of in vitro biochemical screens.