This protocol enables the optimization and subsequent efficient generation of nuclear and cytoplasmic fractions from primary chronic lymphocytic leukemia cells. These samples are used to determine protein localization as well as changes in protein trafficking that take place between the nuclear and cytoplasmic compartments upon cell stimulation and drug treatment.
Nuclear export of macromolecules is often deregulated in cancer cells. Tumor suppressor proteins, such as p53, can be rendered inactive due to aberrant cellular localization disrupting their mechanism of action. The survival of chronic lymphocytic leukaemia (CLL) cells, among other cancer cells, is assisted by the deregulation of nuclear to cytoplasmic shuttling, at least in part through deregulation of the transport receptor XPO1 and the constitutive activation of PI3K-mediated signaling pathways. It is essential to understand the role of individual proteins in the context of their intracellular location to gain a deeper understanding of the role of such proteins in the pathobiology of the disease. Furthermore, identifying processes that underlie cell stimulation and the mechanism of action of specific pharmacological inhibitors, in the context of subcellular protein trafficking, will provide a more comprehensive understanding of the mechanism of action. The protocol described here enables the optimization and subsequent efficient generation of nuclear and cytoplasmic fractions from primary chronic lymphocytic leukemia cells. These fractions can be used to determine changes in protein trafficking between the nuclear and cytoplasmic fractions upon cell stimulation and drug treatment. The data can be quantified and presented in parallel with immunofluorescent images, thus providing robust and quantifiable data.
The transportation of macromolecules between the nucleus and cytoplasm has long been established to play a key role in normal cellular function and is often deregulated in cancer cells1,2. Such deregulation can result from overexpression/mutation of proteins that control nuclear export. One such protein Exportin-1 (XPO1), is a transport receptor that exports >200 nuclear export signal (NES)-containing proteins into the cytoplasm from the nucleus2. XPO1-cargos include p53, FOXO family members and IB, contributing to their inactivation by inhibiting their mechanism of action1,2,3. Further protein mislocalization can occur when microenvironmental signals impinge upon the cancer cells, leading to the activation of intracellular signaling pathways such as the phosphatidyl-inositol-3-kinase (PI3K)/Akt pathway, resulting in inactivation of FOXO family members and subsequent export from the nucleus4,5. Such mislocalization of tumor suppressor proteins has been implicated in the progression of a number of hematological and solid tumors1,2,6.
The development of small molecule inhibitors for clinical use in hematological malignancies (acute myeloid leukemia (AML)/CLL), which bind to and selectively inhibit XPO1 function, underlines the importance of developing appropriate techniques to address the impact of pharmacological agents on the shuttling of proteins between the nuclear and cytoplasmic compartments6,7,8. Imaging techniques have advanced significantly enabling the identification of proteins in subcellular compartments upon external stimulation of drug treatments, however, the importance of robust and supportive parallel techniques is critical to reliably inform a scientific audience of the validity of a result.
Resting lymphocytes and malignant CLL-B cells isolated from patient blood samples represent a challenge in the generation of nuclear and cytoplasmic fractions due to the high nuclear: cytoplasmic ratio. The optimization of experimental conditions to generate robust and reliable experimental data is of course critical in order to plan future experimental programs. The method described here enables quantification of proteins in the nuclear and cytoplasmic fractions and determines how these proteins can be impacted by cellular stimulation and/or drug treatment.
The use of primary samples from CLL patients described here have been approved by the West of Scotland Research Ethics Service, NHS Greater Glasgow and Clyde (UK) and all work was carried out in accordance with the approved guidelines.
1. Isolation of CLL Cells from Patient Blood Samples
2. Flow Cytometry of CLL Cells
3. Preparation of Subcellular Fractions from CLL Cells
NOTE: When planning the experimental set-up, include a well of unstimulated/untreated cells from which the whole cell extract can be generated.
4. Downstream Analysis of Subcellular Fractions
NOTE: In this protocol, analysis of the generated cell fractions was carried out by Western blotting using standard protocols, loading equal cell numbers/lane (equivalent to ~10 µg of protein) for nuclear and cytoplasmic fractions.
When planning experiments on primary CLL cells, if assays require a large number of cells (>50 x 106 cells), there is a preference to use freshly isolated CLL cells, rather than cryopreserved cells that require thawing, however this is not always possible. This is because the freeze/thaw process can result in the death of up to 50% of the CLL cells, although this is sample dependent. Enrichment of CLL cells with a WCC >40 x 106/mL using density centrifugation as described here (steps 1.3 – 1.5) enables a high cell recovery with high purity (≥ 95%) of primary CLL cells. In the sample shown, the WCC = 177 x 106/mL: from a 30 mL blood sample 5 x 109 cells were recovered, which represents a cell yield of 94% of total cells. Analysis of this sample by flow cytometry revealed a purity of CLL cells of >95% as indicated by the dual surface expression of CLL cell markers CD19 and CD5 after gating on FSC/SSC, single cells that were DAPI negative (viable cells) (Figure 1).
Optimization of the subcellular fractionation procedure was carried out using a range of detergent ratios (1:20 to 1:60) during the preparation of the cytoplasmic fraction (step 3.4). Thereafter, the nuclear fractions and WCLs were prepared (steps 3.5 and 3.6 respectively). Immunoblots were performed on the resultant fractions of the CLL cell line MEC1 (Figure 2A) and primary CLL cells (Figure 2B). The blots were probed for the fraction markers Lamin A/C (nuclear; 74/63 kDa) and β-tubulin (cytoplasmic; 55 kDa) to confirm successful cell fractionation. The fractionation indicates that the optimal detergent level for MEC1 cells is a 1:60 dilution (Figure 2A), compared with a 1:30 dilution being optimal for primary CLL cells (Figure 2B), as indicated by an enrichment of nuclear protein and a lack of cytoplasmic protein in the fractions and vice versa. WCLs represent the total protein and act as a positive control for antibodies used to probe the subcellular fractions. It is important to choose appropriate proteins as fraction markers: Figure 2C shows immunoblots of nuclear/cytoplasmic fractions prepared from MEC1 cells in which RNA polymerase II (Rpb1 CTD; 250 kDa) and Lamin A/C were blotted as markers of nuclear fractions, while β-tubulin and γ-tubulin (50 kDa) were used as cytoplasmic markers. It is clear that γ-tubulin is enriched in the cytoplasm however expression is evident in the nucleus, as shown previously9.
Once experimental conditions are optimized, an experiment can be performed. In the examples shown, the subcellular localization of FOXO1 in nuclear and cytoplasmic fractions was determined upon stimulation of cells with the B cell antigen receptor (BCR) in the presence or absence of the dual mTORC1/2 inhibitor AZD8055, in MEC1 cells (Figure 3A) and primary CLL cells (Figure 3B)5,10. In both examples, the generation of highly enriched nuclear and cytoplasmic fractions was achieved as indicated by the almost exclusive expression of Lamin in the nuclear fraction and β-tubulin in the cytoplasmic fractions. In both cell types, FOXO1 expression was reduced in the cytoplasm following treatment with AZD8055 compared to NDC, accompanied by an increase of FOXO1 expression in the nuclear compartment, thus demonstrating protein translocation (Figure 3). To remove the subjectivity of data interpretation, individual immunoblots from five primary CLL samples were quantified within subcellular fractions (step 4; Figure 4A), using the respective nuclear or cytoplasmic proteins as internal loading controls for each sample, and then normalizing each fraction to the unstimulated (US) no drug control (NDC) control, as indicated. The resultant graph shows trends of FOXO1 movement between the nuclear and cytoplasmic fractions, with AZD8055 reducing the levels of FOXO1 expression in the cytoplasm, while concurrently increasing expression in the nucleus. Furthermore, an elevation in cytoplasmic FOXO1 expression is evident upon BCR crosslinking.
Tube | Tube Name | Cells/Beads | Antigen | Fluorophore |
1 | Unstained | Cells | NA | NA |
2 | Single Stain | Beads | CD5 | FITC |
3 | Single Stain | Beads | CD19 | PE-Cy7 |
4 | Single Stain | Beads | CD23 | APC |
5 | Single Stain | Beads | CD45 | APC-Cy7 |
6 | Single Stain | Cells | Viability | DAPI |
7 | CLL Stain | Cells | CD5, CD19, CD23, CD45 & viability | FITC, PE-Cy7, APC, APC-Cy7 & DAPI |
Table 1: Table showing the ideal set of sample tubes required for flow cytometry of CLL cells. Each experiment must include all the appropriate controls for accurate analysis of the results obtained.
Figure 1: Representative flow cytometry analysis plot of enriched CLL patients. Mononuclear-CLL cells enriched from the peripheral blood of an individual CLL patient were gated using FSC-A vs. SSC-A, and doublets were then excluded using FSC-A vs. FSC-H (A). Unstained cells (tube 1) and compensation controls (tubes 2-6) were used to set up the flow cytometer to detect cells and compensate between the fluorescent channels, thus ensuring that the fluorescence signals were detected correctly. (B) An example of negative staining (unstained cells; tube 1) in the CD19 and CD5 fluorescence channels. Live (DAPI negative) and CD45 positive cells were gated (C) and the proportion of CD19+CD5+ (95.5%) and CD19+CD23+ (91.2%) cells within the DAPI–CD45+ population was determined (D). Please click here to view a larger version of this figure.
Figure 2: Optimization of nuclear/cytoplasmic fractionation. Cytoplasmic and nuclear fractions, and whole cell lysates (WCL), were prepared from cell pellets (10 – 20 x 106 cells) of the CLL cell line (A) MEC1 or (B) primary CLL cells enriched from the peripheral blood of patients as described in Step 3. Optimization of the subcellular fractionation was carried out by using a range of detergent ratios (1:20 to 1:60) when preparing the cytoplasmic fraction (as described in step 3.4). The resultant samples were immunoblotted and probed with anti-Lamin A/C (nuclear) and anti-β-tubulin (cytoplasmic) antibodies to confirm successful cell fractionation alongside WCL. Molecular weight markers are shown on the left of the blot (M). * indicates the optimal detergent conditions for cell lysis. (C) Immunoblot of nuclear and cytoplasmic fractions from MEC1 cells with control conditions (NDC) or drug treatment (8055) in the presence of absence of stimulation (+ or – BCR crosslinking respectively). Blots were probed with anti-Rbp1 CTD (clone 4H8; recognizing RNA polymerase II subunit B1), anti-Lamin A/C, anti-β-tubulin or anti-γ-tubulin (clone GTU-88) antibodies, to identify subcellular fractions. Please click here to view a larger version of this figure.
Figure 3: Subcellular fractionation demonstrates the shuttling of FOXO1 between the nucleus and cytoplasm in CLL. (A) MEC-1 cells and (B) primary CLL cells were pre-treated for 30 min with 100 nM AZD8055 (8055) or left untreated (NDC) as indicated and then BCR was ligated for 1 h or left US. Nuclear and cytoplasmic fractions were then prepared and immunoblotted. Following confirmation of fractionation by probing with anti-Lamin A/C (nuclear) and anti-β-tubulin (cytoplasmic) antibodies, the effect of both drug treatment and BCR ligation was assessed on FOXO1 protein expression, using an anti-FOXO1 antibody. M indicates molecular weight marker. Please click here to view a larger version of this figure.
Figure 4: A worked example of quantitative Western blot analysis (densitometry). (A) Densitometry was performed using Western blot analysis software available online. Briefly, within the Analysis ribbon, rectangles were drawn around protein bands in the image to calculate signal intensity. Depicted is densitometry of a representative Western blot image of a CLL patient sample that underwent cytoplasmic/nuclear fractionation. Cytoplasmic and nuclear fractions are distinguished by the expression of cytoplasmic (β-tubulin) and nuclear (Lamin A/C) markers. Normalized expression of FOXO1 for a given condition can be calculated by dividing the signal obtained for FOXO1 by the corresponding signal for β-Tubulin or Lamin A/C, depending on the fraction being analyzed. Relative FOXO1 expression (relative to US vehicle control), can be calculated by dividing normalized FOXO1 expression of a given condition by the normalized FOXO1 expression of the US vehicle control of a given cellular fraction. (B) Graph showing the FOXO1 expression levels in the cytoplasmic (left) or nuclear (right) fractions normalized to US-NDC control within each cellular fraction. The red dot on the graph is the worked example shown. This data shows the average fold change in FOXO1 expression compared to US-NDC ± SEM. P values were determined by two-tailed Students paired t test. n = 5 individual CLL patient samples. Please click here to view a larger version of this figure.
The protocol described provides a fast and efficient method for the generation of nuclear and cytoplasmic fractions from primary CLL cells, and subsequent quantification of protein trafficking between the nuclear and cytoplasmic fractions upon cell stimulation and drug treatment. The data presented demonstrates the ability to detect trafficking of specific proteins, for example, FOXO1, between the nuclear and cytoplasmic fractions, upon treatment with a dual mTOR inhibitor AZD8055 in the presence/absence of BCR crosslinking through F(ab’)2 fragment stimulation (Figure 3 and Figure 4). Coupling these experiments with quantification of Western blots from individual CLL patient samples, enables objective analyses of the data generated and demonstrates the robustness of the assay described to quantify global changes in protein localization in CLL cells isolated from patient cohorts (Figure 4). It is clear from the data that an average of five patient samples in the cytoplasmic fractions reached near significance. Given the clinical heterogeneity of CLL patients11, these analyses would ordinarily be carried out on bigger patient cohorts, and/or focused on specific prognostic subgroups of patients to gain a fuller understanding of the cellular response of CLL cells to specific drug treatments.
The data presented demonstrate the importance of choosing protein markers that exclusively reside in either the cytoplasmic or nuclear fractions, as the purity of the fractionation will be confirmed by these markers. β-tubulin was chosen for cytoplasmic fraction confirmation, and Lamin A/C as a nuclear marker. Additional proteins commonly used are GAPDH and α-tubulin to identify the cytoplasmic fraction or Brg1 (SMARCA4), TFIID and RNA Polymerase II for nuclear fraction purity4,5. However, care must be taken when choosing proteins that are highly enriched in specific fractions, and not present in both fractions (e.g., γ-tubulin) (Figure 2C)9. Indeed, GAPDH and actin while generally considered to be cytoplasmic proteins can localize to the nucleus12,13, highlighting the importance of choosing a fraction marker that does not relocate when stimulation or treatment is applied to the cells. Furthermore, it is important to confirm that the protein marker chosen is expressed in the cell of interest by running the WCL alongside the subcellular fractionations.
In the representative experiment shown, the same number of CLL cells was used for each condition (stimulation/drug treatment), and thereafter the fractionation samples were prepared immediately. Loading 10 µg of fractionated protein/lane provides sufficient material for detection of the proteins of interest. As these samples only underwent a short-term drug treatment and stimulation (up to 4 h), it was assumed that the protein level would remain the same in each sample, and a protein assay was not performed. However, if cell treatments are extended (18 – 72 h), the level of cell death or proliferation in cells may significantly alter the quality and quantity of protein extracted, dependent on the drug/cell stimulation applied, thus altering protein levels in the treated/stimulated samples. In these cases for longer term drug treatments, it is advisable to carry out protein quantification using a Bradford assay or equivalent, prior to Western blotting to ensure the same amount of protein is run in each lane of the immunoblot. The presence of detergents may interfere with specific protein assays14, this interference can be reduced by diluting cell fraction protein samples. In addition, use the complete lysis buffer as the blank, using the same dilution as in the samples being tested.
To provide supporting evidence for findings described here, parallel experiments could be performed using fluorescence microscopy to analyze the location of FOXO1 within CLL cells to enable visualization of these findings5. Furthermore, the subcellular fractions generated can also be used for enzyme activity assays or proteomics analysis in further downstream analyses.
The authors have nothing to disclose.
The authors would like to thank Dr. Natasha Malik for critically reviewing the manuscript. This study was funded by a Bloodwise project grant awarded to AMM (18003). FACS analysis facilities were funded by the Howat Foundation. MWM was funded by a PhD studentship from Friends of Paul O’Gorman Leukaemia Research Centre, JC was funded by the Friends of Paul O’Gorman Leukaemia Research Centre and JH was funded by a Bloodwise project grant (18003).
1.5 mL microcentrifuge Tubes | Griener Bio one | 616201 | |
3 mL Pasteur Pipettes | Griener Bio one | 612398 | |
12 mm x 75 mm FACS Tubes | Elkay | 2052-004 | |
15 mL Tube | Griener Bio one | 188271 | |
50 mL Tube | Griener Bio one | 227261 | |
anti-CD5 FITC antibody | BD Biosciences | 555352 | phenotypic surface marker |
anti-CD19 PE Cy7 antibody | BD Biosciences | 557835 | phenotypic surface marker |
anti-CD23 APC antibody | BD Biosciences | 558690 | phenotypic surface marker |
anti-CD45 APC Cy7 antibody | BD Biosciences | 557833 | phenotypic surface marker |
anti-β-Tubulin antibody | Cell Signaling | 2146 | cytoplasmic marker |
anti-γ-Tubulin Mouse antibody (clone GTU-88) | Sigma-Aldrich | T5326 | |
anti-FoxO1 (C29H4) Rabbit antibody | Cell Signaling | 2880 | |
anti-Lamin A/C antibody | Cell Signaling | 2032 | nuclear marker |
anti-mouse IgG, HRP-linked Antibody | Cell Signaling | 7076 | Secondary antibody |
anti-rabbit IgG, HRP-linked Antibody | Cell Signaling | 7074 | Secondary antibody |
anti-Rpb1 CTD antibody (clone 4H8) | Cell Signaling | 2629 | nuclear marker |
BDFACS Canto II | BD Biosciences | By Request | Flow Cytometer |
DAPI Solution | BD Biosciences | 564907 | live/dead discriminator |
DMSO | Sigma | D2650 | |
EDTA | Sigma | EDS | |
Fetal Bovine Serum | Thermofisher | 10500064 | |
GraphPad Prism 6 | GraphPad Software | Software | |
Histopaque1077 density gradient media | Sigma | H8889 | |
HyperPAGE 10 – 190 kDa protein marker | Bioline | BIO-33066 | Molecular weight marker |
Image Studio Lite (version 5.2.5) | LI-COR | www.licor.com | Software |
Labnet VX100 | Fisher Scientific | Vortex | |
Nucelar Extract Kit | Active Motif | 40010 | |
OneComp eBeads | Thermofisher | 01-111-42 | |
Trypan Blue Solution | Thermofisher | 15250061 | |
PageRuler Plus Prestained Protein Ladder | Thermo Fisher Scientific | 26619 | Molecular weight marker |
PBS Tablets | Fisher Scientific | BR0014G | |
RosetteSep Human B Cell Enrichment Cocktail | Stem Cell Technologies | 15064 | |
Sigma 3-16P | SciQuip | Centrifuge | |
Sigma 1-15PK | SciQuip | Centrifuge |