The present protocol describes a novel method of identifying a population of enucleating orthochromatic erythroblasts by multi-spectral imaging flow cytometry, providing a visualization of the erythroblast enucleation process.
Erythropoiesis in mammals concludes with the dramatic process of enucleation that results in reticulocyte formation. The mechanism of enucleation has not yet been fully elucidated. A common problem encountered when studying the localization of key proteins and structures within enucleating erythroblasts by microscopy is the difficulty to observe a sufficient number of cells undergoing enucleation. We have developed a novel analysis protocol using multiparameter high-speed cell imaging in flow (Multi-Spectral Imaging Flow Cytometry), a method that combines immunofluorescent microscopy with flow cytometry, in order to identify efficiently a significant number of enucleating events, that allows to obtain measurements and perform statistical analysis.
We first describe here two in vitro erythropoiesis culture methods used in order to synchronize murine erythroblasts and increase the probability of capturing enucleation at the time of evaluation. Then, we describe in detail the staining of erythroblasts after fixation and permeabilization in order to study the localization of intracellular proteins or lipid rafts during enucleation by multi-spectral imaging flow cytometry. Along with size and DNA/Ter119 staining which are used to identify the orthochromatic erythroblasts, we utilize the parameters “aspect ratio” of a cell in the bright-field channel that aids in the recognition of elongated cells and “delta centroid XY Ter119/Draq5” that allows the identification of cellular events in which the center of Ter119 staining (nascent reticulocyte) is far apart from the center of Draq5 staining (nucleus undergoing extrusion), thus indicating a cell about to enucleate. The subset of the orthochromatic erythroblast population with high delta centroid and low aspect ratio is highly enriched in enucleating cells.
Terminal differentiation within the erythroid lineage in mammals concludes with the dramatic process of enucleation, through which the orthochromatic erythroblast expels its membrane-encased nucleus (pyrenocyte)1, generating a reticulocyte2. The exact mechanism of this process, which is also the rate-limiting step of successful, large-scale, production of red blood cells in vitro, is not yet fully elucidated. The localization of key proteins and structures within enucleating erythroblasts relies on the use of fluorescent and electron microscopy3-5. This tedious process typically results in the identification of a limited number of enucleation events and does not always allow meaningful statistical analysis. Expanding on a method of erythroblast identification described previously by McGrath et al.6, we have developed a novel approach of identifying and studying enucleation events by Multi-Spectral Imaging Flow Cytometry (multiparameter high-speed cell imaging in flow, a method that combines fluorescent microscopy with flow cytometry)7, which can provide a sufficient number of observations to obtain measurements and perform statistical analysis.
Here, we describe first two in vitro erythropoiesis culture methods used in order to synchronize erythroblasts and increase the probability of capturing enucleation at the time of evaluation. Then we describe in detail the staining of erythroblasts after fixation and permeabilization in order to study the localization of intracellular proteins or lipid rafts during enucleation by multi-spectral imaging flow cytometry.
Samples are run on an imaging flow cytometer and the collected cells are gated appropriately to identify orthochromatic erythroblasts6. Orthochromatic erythroblasts are then analyzed based on their aspect ratio, as measured in brightfield imaging, versus their value for the parameter delta centroid XY Ter119-DNA, which is defined as the distance between the centers of the areas stained for Ter119 and DNA, respectively. The population of cells with low aspect ratio and high delta centroid XY Ter119/DNA is highly enriched in enucleating cells. Using wild-type (WT) erythroblasts versus erythroblasts with Mx-Cre mediated conditional deletion of Rac1 on Rac2-/- or combined Rac2-/-; Rac3-/- genetic background and this novel analysis protocol of multi-spectral imaging flow cytometry, we recently demonstrated that enucleation resembles asymmetric cytokinesis and that the formation of an actomyosin ring regulated in part by Rac GTPases is important for enucleation progression7.
1. Long-term In vitro Erythropoiesis Culture (Ex vivo Erythroid Differentiation Culture Protocol by Giarratana et al.8, Modified and Adapted for Mouse Cells)
This is a 3-step long-term in vitro erythropoiesis protocol. In the first step (days 0-4) 2 x 105 cells/ml are placed in erythroblast growth medium supplemented with stem cell factor (SCF), interleukin-3 (IL-3), and erythropoietin (Epo). In the second step (days 5-6), cells are resuspended at 2 x 105 cells/ml and co-cultured on adherent stroma cells (MS5) in fresh erythroblast growth medium supplemented only with Epo. In the third step (days 7-9), cells are cultured on a layer of MS-5 cells in fresh erythroblast growth medium without cytokines up to enucleation (Figure 1A).
All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Cincinnati Children’s Hospital Medical Center.
2. Fast Enucleation Assay, According to the Protocol Described by Yoshida et al.12 with Modifications (Figure 1B)
3. Staining of Erythroblasts for Localization of Intracellular Proteins or Lipid Rafts During Enucleation by Multi-spectral Imaging Flow Cytometry
First, cells are analyzed based on their Brightfield Aspect Ratio (the ratio of the length of their minor versus their major axis) and their Brightfield Area (indicative of their size). Events with a Brightfield Area value lower than 20 and higher than 200 are mostly debris and cell aggregates, respectively, and are excluded from the analysis (Figure 2A). Single cells (gate “R1”) are then analyzed based on their value for the Gradient RMS parameter, which indicates sharpness of image. Gate “R2” is created containing cells with Gradient RMS value more than the 50th percentile in order to select the images taken well in focus (Figure 2B). Cells are then gated based on their size as measured by their Brightfield Area, and their positivity for the erythroid marker Ter119 as measured by the Ter119 fluorescent stain-Mean Pixel parameter (gate “Ter119 positive”, Figure 2C). Cells very low or very high for Ter119, are either non-erythroids or remaining cell aggregates, respectively, and are excluded from the analysis. In the next step, cells are selected based on their Draq5 Aspect Ratio Intensity (the ratio of the minor versus the major axis intensity of their nucleus) and the Intensity of Draq5 (Figure 2D). Draq5 negative cells (mostly enucleated cells, such as reticulocytes and RBCs), and cells with a low Draq5 Aspect Ratio (mostly doublets) are excluded from the analysis. Draq5 positive cells (gate “DNA positive”, mostly erythroblasts at this point) are then analyzed based on their Ter119 Area, which indicates the size of the cell, and their Ter119 Mean Pixel/Area (density of Ter119 expression), which indicates the brightness of Ter119 staining. Orthochromatic erythroblasts (gate “OrthoE”) are recognized as small, Ter119hi cells (Figure 2E). Finally, a subpopulation of the orthochromatic erythroblasts highly enriched in enucleating cells is characterized by low Brightfield Aspect Ratio, which is a measurement of cell elongation, calculated by the ratio of minor axis/major axis of the cell image in the Brightfield channel M01, and by high Delta Centroid XY Ter119/Draq5, which is defined by the distance between the center of the incipient Ter119+-reticulocyte and the center of the Draq5+-nucleus (gate “enucleating cells” in Figure 2F).
Along with antibody against Ter119 and the DNA-stain Draq5, the cells have also been stained for filamentous actin (F-actin) with fluorescent phalloidin to evaluate localization of F-actin during erythroblast enucleation7. Of note, a progression of enucleation can be visualized in the fixed cells, as cells with a decreasing aspect ratio (i.e. increasingly elongated) and increasing delta centroid XY Ter119/Draq5 are observed (Figure 3). F-actin is observed to concentrate at the cleavage furrow during enucleation and then dissipate once the nucleus is extruded. Moreover, co-localization of actin and myosin at the cleavage furrow between incipient reticulocyte and nucleus can be demonstrated by multi-spectral imaging flow cytometry after co-staining of WT erythroblasts for pMRLC (phosphorylated myosin regulatory light chain) and F-actin7.
Other proteins and structures of interest can also be stained with appropriate antibodies or fluorescent markers to allow imaging studies of their role during enucleation. Polarized microtubule formation is visible in WT orthochromatic erythroblasts prior to enucleation but not in erythroblasts treated with colchicine (Figure 4A). Inhibition of tubulin polymerization by colchicine diminishes cell polarization, as demonstrated by measuring the parameter delta centroid XY BF/Draq5 between the center of the cell body seen in Brightfield channel and the center of the nuclear staining achieved with Draq5 (Figure 4B).
Utilizing WT or Rac-deficient (after genetic or pharmacologic manipulation) erythroblasts in multi-spectral imaging flow cytometry allowed imaging studies that demonstrate the role of Rac GTPases in enucleation. Rac GTPases regulate at least in part the formation of an actomyosin ring as well as the confluence of lipid rafts in the furrow between incipient reticulocyte and pyrenocyte7.
Figure 1. Schematic demonstration of the erythropoiesis in vitro protocols used in order to produce enucleating erythroblasts for studies. A. Long-term in vitro erythropoiesis culture initiated from LDBM or Lin–cells. B. Fast enucleation assay initiated by splenocytes highly enriched in erythroblasts after stress erythropoiesis induction in vivo by phlebotomy. Please click here to view a larger version of this figure.
Figure 2. Analysis of data utilizing the analysis software specific to the imaging flow cytometer. Successive gating of the populations of interest is shown in panels A-F. The number in parenthesis indicates the approximate percent of the corresponding parent population, in this experiment. A. Initial gating of R1 population removes cell aggregates and cellular debris. B. Gate R2 out of R1 includes the cells that have been imaged clearly, excluding out of focus cells. C. Ter119-positive cells (out of R2) are selected, excluding cells that are either negative for Ter119 or are too intensely stained because of their presence in aggregates. D. DNA-positive cells (out of Ter119-positive cells) are selected, after plotting for Aspect Ratio Intensity versus intensity of the nuclear stain (here Draq5 read in channel 5). E. DNA- and Ter119-positive cells are gated into basophilic, polychromatophilic and orthochromatic erythroblasts based on their location in the Ter119 Mean Pixel versus Ter119 Area (here read in channel 3), as shown previously by McGrath et al5. F. The enucleating cells are those cells out of the orthochromatic erythroblasts, which have low Aspect Ratio (a measurement of cell elongation in brightfield (BF) channel) and high Delta Centroid XY Ter119/Draq5 (distance between center of forming Ter119+-reticulocyte and center of nucleus). This research was originally published in Blood: Konstantinidis DG, Pushkaran S, et al. Signaling and cytoskeletal requirements in erythroblast enucleation. Blood. 2012;119(25):6118-6127 by the American Society of Hematology. Please click here to view a larger version of this figure.
Figure 3. Representative images of enucleating erythroblasts with progressively increasing Delta Centroid XY Ter119/Draq5. WT mouse orthochromatic erythroblasts, stained with Ter119-PECy7, phalloidin-AF488, and Draq5, are gated per their aspect ratio and delta centroid XY Ter119/Draq5. Cells are shown fixed at different, successive stages of enucleation, with a progression that corresponds to decreasing aspect ratio and increasing delta centroid XY Ter119/Draq5 (green-cross within the yellow gate shows the position of the cell imaged on the right). In the cell images from top to bottom, F-actin can be observed during enucleation to concentrate at the cleavage furrow and then dissipate once the nucleus is extruded (as shown in cell #4782 at the lower image). Please click here to view a larger version of this figure.
Figure 4. Formation of a unipolar microtubule assembly and polarization of orthochromatic erythroblasts precedes enucleation. A. Polarized microtubule formation is visible in control WT erythroblasts (stained with anti-b-tubulin–AlexaFluor-488 and the nuclear stain Draq5), whereas b-tubulin is diffusely stained in the erythroblasts incubated with colchicine (5 µM) for 6 hr in the fast in vitro enucleation assay. B. Multi-spectral imaging flow cytometry can offer a quantitative evaluation of cell polarization by analysis of the distribution of the parameter Delta Centroid XY BF/Draq5, which measures the distance between the center of the cell body as seen in bright-field and the center of the nuclear staining achieved with Draq5 (schematic representation in the inset). Delta Centroid BF/Draq5 values of colchicine-treated WT orthochromatic erythroblasts are statistically significantly different than the control Delta Centroid BF/Draq5 values (p<0.001). This research was originally published in Blood: Konstantinidis DG, Pushkaran S, et al. Signaling and cytoskeletal requirements in erythroblast enucleation. Blood. 2012;119(25):6118-6127 by the American Society of Hematology. Please click here to view a larger version of this figure.
In recent years the study of erythroblast enucleation has gained increasing momentum since it is the step in in vitro erythropoiesis cultures that is most difficult to reproduce efficiently in order to achieve successful, large-scale production of red blood cells ex vivo. Up until recently, the study of erythroblast enucleation utilized mainly fluorescence microscopy and flow cytometry methods. Fluorescence microscopy methods, albeit helpful in identifying participating molecules, require days of microscopic observation to identify a small number of orthochromatic erythroblasts undergoing enucleation within hundreds of cells fixed at a particular point in time. Flow cytometry methods, on the other hand, are very helpful in evaluating the rate of enucleation in a culture, as well as the effects of pharmacologic or genetic manipulation of particular molecules on this process, but do not provide any data on the intracellular localization of these molecules.
Multi-spectral imaging flow cytometry combines the benefits of flow cytometry and immunofluorescence microscopy since it allows rapid acquisition of both flow cytometric and morphologic data on several thousands of cells. This is a significant advantage versus classic flow cytometry in erythropoiesis studies, since the different stages of erythroblast differentiation (proerythroblasts, basophilic, polychromatophilic, and orthochromatic) have been defined using morphological criteria6. However, multi-spectral imaging flow cytometry is optimal for visualization of cellular structures rather than relative quantitation comparisons in population numbers. For such comparisons, routine flow-cytometry that does not require the permeabilization step necessary for immunofluorescence of intracellular structures performs better. For example the relative percentage of BasoE:PolyE:OrthoE in Figure 2E does not correspond to the physiologic ratio of 2:4:8, since the mature erythroblasts are more sensitive to the permeabilization step and are preferentially lost during the staining process.
The imaging data can be processed in association with the flow cytometry data using the analysis software specific to the imaging flow cytometer, allowing the collection of cells with certain morphologic characteristics within gates. Approximately one hundred enucleating cells can be identified within a population of 10,000 erythroblasts with the analysis method described above in a fast and efficient manner7, allowing more meaningful observations and statistical evaluation.
Moreover, image processing is facilitated with the analysis software specific to the imaging flow cytometer allowing quantitative analysis of such characteristics as the Delta Centroid XY to study e.g. the relative position of the nucleus to the cytoplasm that can be used as a measure of polarization.
Critical steps within the protocol and troubleshooting
It is well known that even gentle pipetting can result in the separation of reticulocyte and pyrenocyte12. This has the potential to severely limit the number of enucleation events imaged. As a result, care must be taken, particularly when lifting erythroblasts bound to MS5 cells in culture.
Fixation with formaldehyde solution can cause alterations to extracellular regions of surface markers resulting in decreased specific antibody binding and/or increased non-specific antibody binding. An advantage of the imaging flow cytometer is that surface staining is visualized and its quality can thus be evaluated. Dotted, instead of uniform, staining for abundant surface markers such as Ter119 indicates overfixation and should be tackled through a mix of lower formaldehyde concentration and shorter duration of fixation.
Following fixation, keeping cells on ice for at least 15 min is vital in order to prevent excess cell breakage during the permeabilization process due to temperature differences. Although acetone permeabilization maintains the fragile late erythroids better than detergent-mediated permeabilization, the step where 100% acetone is required will result in a noticeable, but not detrimental, loss of cells. At the end, after a wash with cold FACS buffer, cells are allowed at room temperature for the antibody incubation steps.
The imaging flow cytometer has a set rate of flow (cells/sec) depending on the lens used (rate is decreased as the magnification increases). Following final wash before measurement, it is recommended that cell pellets are resuspended in a small volume (50-60 μl), in order to accelerate processing of the sample. For a large number of samples that require long duration of run (over 30 min), samples should be kept on ice.
The long-term enucleation assay offers the benefit of an expanded erythroid cell production in order to produce enough cells that can be collected at the enucleation stage for biochemical evaluation like immunoblotting and pull-downs. The fast enucleation assay gives the benefit of time requiring only 2 days to perform, although a mouse needs to be phlebotomized 4 days prior to the experiment. We have not observed a significant difference in enucleation efficiency between the two methods. Of note, flow cytometry evaluation, which does not require the permeabilization step necessary for immunofluorescence of intracellular structures, as described before7, is most appropriate for the quantitative evaluation of enucleation efficiency.
Limitations of the technique
The method described here utilizes induced stress erythropoiesis in vivo and culture in medium containing hydrocortisone in vitro in order to amplify erythroid populations and synchronize them at the stage of enucleation. Both of these conditions likely imitate erythroblast-enucleation under stress. In addition, hydrocortisone significantly increases erythroid yield and survival. To obtain similar yield of enucleation events from bone marrow cells derived from wild type mice with steady-state erythropoiesis and cultured without hydrocortisone (therefore avoiding synchronization), we would need to culture cells from multiple mice and run and process a lot more events through Multi-Spectral Imaging Flow Cytometry. Although multi-spectral imaging flow cytometry accelerates collection of morphological data immensely in comparison to immunofluorescence microscopy using magnification lens up to 60x, comparison of the observations obtained by imaging flow cytometer with classic, Z-stack, and confocal microscopy images is valuable to obtain better detail, since the objective lens in a microscope can provide magnification up to 100x and Z-stack images give a 3-dimensional impression of the structures, which is not attainable by imaging flow cytometer in the same cell. The relative high number of similar cells collected in a multi-spectral imaging flow cytometry experiment, at a random orientation, partially compensates this limitation allowing observations from a different view-point.
The authors have nothing to disclose.
The authors thank the Research Flow Cytometry Core at Cincinnati Children’s Hospital Research Foundation and Richard Demarco, Sherree Friend, and Scott Mordecai from the Amnis Corporation (part of EMD Milllipore) for expert technical support. This work was supported by the National Institutes of Health grants K08HL088126 and R01HL116352 (T.A.K.) and P30 DK090971 (Y.Z.).
αMEM medium | CellGro | 15-012-CV | |
IMDM medium | Hyclone (Thermo Scientific) | SH30228.01 | |
Stempro-34 SFM | GIBCO (Life Tech) | 10640 | |
Stempro-34 nutrient supplement | GIBCO (Life Tech) | 10641-025 | |
Fetal Bovine Serum (FBS) | Atlanta Biologicals | 512450 | |
BIT9500 | Stemcell Technologies | 09500 | |
Bovine Serum Albumin (BSA) | Fisher Scientific | BP-1600-100 | |
Phosphate buffered saline (PBS) | Hyclone (Thermo Scientific) | SH30028.02 | |
Penicillin/Streptomycin | Hyclone (Thermo Scientific) | SV30010 | |
L-glutamine | Hyclone (Thermo Scientific) | SH30590.01 | |
Isothesia (Isoflurane) | Butler-Schein | 029405 | |
Histopaque 1.083mg/ml | Sigma | 10831 | |
BD Pharmlyse (RBC lysis buffer) | BD Biosciences | 555899 | |
Acetone | Sigma-Aldrich | 534064 | |
Formaldehyde | Fisher Scientific | BP 531-500 | |
Hydrocortisone | Sigma | H4001 | |
Stem Cell Factor (SCF) | Peprotech | 250-03 | |
Interleukin-3 (IL-3) | Peprotech | 213-13 | |
EPOGEN Epoetin Alfa (Erythropoietin, EPO) | AMGEN | available by pharmacy | |
CD44-FITC antibody | BD Pharmingen | 553133 | |
CD71-FITC antibody | BD Pharmingen | 553266 | |
Ter119-PECy7 antibody | BD Pharmingen | 557853 | |
Phalloidin-AF488 | Invitrogen (Life Technologies) | A12379 | |
β-tubulin-AF488 antibody | Cell Signaling | #3623 | |
anti-rabbit AF488-secondary antibody | Invitrogen (Life Technologies) | A11008 | |
anti-rabbit AF555-secondary antibody | Invitrogen (Life Technologies) | A21428 | |
AF594-cholera toxin B subunit | Invitrogen (Life Technologies) | C34777 | |
pMRLC (Ser19) antibody | Cell Signaling | #3671 | |
γ-tubulin antibody | Sigma | T-3559 | |
Syto16 | Invitrogen (Life Technologies) | S7578 | |
Draq5 | Biostatus | DR50200 | |
Ferrous sulfate | Sigma | F7002 | |
Ferric nitrate | Sigma | F3002 | |
EDTA | Fisher Scientific | BP120500 | |
15ml tubes | BD Falcon | 352099 | |
50ml tubes | BD Falcon | 352098 | |
6-well plates | BD Falcon | 353046 | |
24-well plates | BD Falcon | 351147 | |
Flow tubes | BD Falcon | 352008 | |
Tuberculin syringe | BD | 309602 | |
Insulin syringe | BD | 329461 | |
Syringe needle 25G5/8 | BD | 305122 | |
Capped flow tubes | BD | 352058 | |
40μm cell strainer | BD Falcon | 352340 | |
Scalpel (disposable) | Feather | 2975#21 | |
FACS Canto Flow Cytometer | BD | ||
ImagestreamX Mark II Imaging Flow Cytometer | AMNIS (EMD Millipore) | ||
Image Data Exploration and Analysis Software (IDEAS) version 4.0 and up. | AMNIS (EMD Millipore) | ||
Hemavet 950 Cell Counter | Drew Scientific | CDC-9950-002 | |
NAPCO series 8000WJ Incubator | Thermo scientific | ||
Allegra X-15R Centrifuge | Beckman Coulter | 392932 | |
Mini Mouse Bench centrifuge | Denville | C0801 |