We have been interested in studying the roles of two aldehyde dehydrogenases in the biology of lung cancer. In this study, we seek to apply Aldefluor flow cytometry-based assay for the measurement of aldehyde dehydrogenase (ALDH) activity in lung cancer cell lines, which may become a new tool that will facilitate our continued research in this field.
Several established lung cancer cell lines were used, including A549 cell line expressing siRNA against aldehyde dehydrogenase class-1A1 (ALDH1A1). Western blot analysis, spectrophotometry assay, and Aldefluor staining were used to measure protein or enzyme activity in these cell lines. For the purpose of measurement of ALDH activity by Aldefluor in cells with known high ALDH levels, cells were mixed 1:10 with immortalized lung epithelial cell line (Beas-2B), which is known to lack ALDH activity. To delineate dead cells, double staining using Aldefluor and propidium iodide (PI) was done. Double staining was also used to detect changes in ALDH activity in two different cell lines after treatment with 4-hydroperoxycyclophosphamide (4-HC).
Our results show a very good correlation between Aldefluor, Western blot, and spectrophotometry assays. Mixing experiments with Beas-2B cells allowed accurate assessment of ALDH activity in A549 cells at baseline and after siRNA expression, thus establishing an approach that facilitates the measurement of very high ALDH using the Aldefluor assay. Aldefluor staining was able to detect heterogeneity in ALDH expression among as well as within the same cell lines and better assess viability after 4-HC treatment when combined with PI.
Aldehyde dehydrogenase (ALDH) has been reported to be highly expressed in nonsmall cell lung cancer (NSCLC) cell lines (1). ALDH may play an important biological role due to its involvement in the conversion of aldehydes into weak acids, a reaction important in the production of biologically active substances such as retinoic acid. Furthermore, ALDH appears to be highly expressed in primitive hematopoietic cells and developmentally expressed in early embryonic tissue (2, 3). We have been interested in one aspect of ALDH production related to its role in mediating resistance in various tumors against oxazaphosphorines, including cyclophosphamide (CP) and its active metabolites (4–6). Different assays for the measurement of ALDH isozymes have been available including Western blot analysis, spectrophotometric assay for enzyme activity, and immunohistochemistry (4–7).
In recent years, a new flow cytometry-based method that allows measurement of ALDH activity in viable cells has been developed and used to sort hematopoietic cells with high versus low levels of ALDH activity (8–10). It has also been used to better identify viable hematopoietic stem and progenitor cells (10). Taking advantage of this method in the field of hematopoietic stem cell research, it has been shown that the expression of ALDH delineates distinct CD34+ stem cell and progenitor cell compartments (11).
Because of the potential advantages of this flow cytometric method, we have applied it to lung cancer cell lines and compared it to Western blot and spectrophotometric assays. Our observations demonstrate the successful application of this method to solid tumor cell lines that may open the way for other applications, such as the isolation of cells with high levels of ALDH and the measurement of induced changes in ALDH activity in viable cancer cells.
MATERIALS AND METHODS
Several lung cancer cell lines were obtained from ATCC, including A549, H522, H460, H82, H322, H1299, NCI-125, H157, and four others (LCLC-103H, ADLC-5M2, SW210.5, and SCLC-16HC) were obtained from another source (12, 13). H82, SW210.5, and SCLC-16HC were small-cell lung cancer (SCLC) cell lines. These cell lines were chosen because they reflect all the different histological subtypes of lung cancer. One immortalized normal lung epithelial cell line Beas-2B was also obtained from ATCC. All cell lines were thawed, cultured in RPMI-1640 medium (Gibco Invitrogen Corporation, Grand Island, NY) with 10% FBS (Hyclone, Logan, UT) in 5% CO2 cell culture incubator at 37°C, and used within 2–3 passes when in the log phase of growth.
A549 cells transfected with pLXSN retroviral vector containing a short hairpin siRNA under the control of the U6 promoter targeting ALDH1A1 were also used in experiments. The following sequence was inserted into the vector and sequencing verified by GenScript (Scotch Plains, NJ): CTCGAGCGGA CAATGCTGTT GAATTTCCAC ACCAAATTCA ACAGCATTGT CCTTTTTTCC AAGGATCC. Transfection was performed as described before (5) and stably transfected cells were selected using 1mg/ml G-418 (GibcoBRL, Grand Island, NY). Control A549 cells transfected with the empty pLXSN vector were also used. This method of expressing siRNA against ALDH1A1 was reported by us recently to reduce ALDH activity in A549 to (47 ± 9)% (14).
Western Blot Analysis
Cell lysates were obtained as described in many of our previous publications as well as by others (1, 5, 6) and used for both ALDH activity assay (see below) as well as Western blot analysis in order to demonstrate changes in protein levels of ALDH1A1 and ALDH3A1 in the different cell lines. Lysates from each experimental group were size separated in parallel on two 12% denaturing SDS-polyacrylamide gels (Biorad, Hercules, CA), electrotransferred onto nitrocellulose membranes, blocked with 5% milk in TBS, and probed as described before (5), using chicken anti-human ALDH-1A1 and ALDH-3A1 polyclonal antibodies generously provided by Dr. L. Sreerama (St Cloud University, Minneapolis, MN) and Dr. N.E. Sladek (University of Minnesota, Minneapolis, MN). The specificity of these antibodies has been documented by Dr. Sladek's group (7, 15). After washing, the secondary antibody (horseradish peroxidase-labeled rabbit anti-chicken antibody; Sigma Chemical, St. Louis, MO) was used at dilution of 1:4,000 for 1 h. Chemiluminscence method (SuperSignal, Pierce, Rockford, IL) was used for the final visualization of the protein bands on X-ray film (Super Rx, Fuji Photo Film, Tokyo, Japan). After washing and blocking, the same blots were labeled again for visualization of actin as a loading control using anti-actin antibody (Oncogene Research Products, Cambridge, MA). To quantitate the protein bands, the X-ray films were scanned using the ScanJet (Hewlett Packard), and integrated density was measured using ScionImage computer program (Scion Corporation). Relative normalized units were obtained by dividing protein level of ALDH by the actin level.
RT-PCR to Detect ALDH Isozymes in SW210.5 Cell Line
We performed semiquantitative RT-PCR using specific primers for ALDH1A1, ALDH3A1, ALDH2, and GAPDH as a house keeping gene control (14). The sequence of the primers for GAPDH and ALDH3A1 has been previously published (16, 17). Total RNA was extracted using RNeasy Mini Kit (Quiagen), according to the protocol provided by the supplier, from wild type A549 and SW210.5 cells. For the RT reaction, oligo (dT) was used to generate first strand for the antisense transcripts according to the manual Superscript First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA). The RT reaction was carried out separately, followed by PCR in which the specific primers in were used. The following primers were used for GAPDH (forward) 5′-GAA GAT GGT GAT GGG ATT TC-3′ and (reverse) 5′-GAA GGT GAA GGT CGG AGT C-3′; ALDH1A1 (forward) 5′-CGG CGC ATT GTG TTA GCT GAT GCC G-3′ and (reverse) 5′-AGA GAA CAC TGT GGG CTG GAC-3′; ALDH3A1 (forward) 5′ ACT GGG CGT GGT CCT CGT CAT TGG-3′ and (reverse) 5′ GTG AGG ATG GTG GGG GCT ATG TAG-3′; ALDH2 (forward) 5′- CCC GCC GTG GGC CAC GCC TGA-3′ and (reverse) 5′- GTT CTT CTG AGG CAC TTT GAC-3′. The program for PCR was previously published for the ALDH3A1 synthesis (17).
Table 1. Measurement of ALDH Activity and Proteins in Human Lung Cancer Cell Lines
Spectrophotometry (nmol/107 cells per min)
Results in the first two columns of numbers reflect mean ± SD of ≥ 2 measurements, or a single measurement.
Results reflect the mean ratio (from 3 different Western blots) of densitometry measurements of ALDH1A1 or ALDH3A1 divided by that of Actin.
Aldefluor staining is done on a mixture of 10% A549 with 90% Beas-2B epithelial cells.
Fresh cell lysates used for the aforementioned Western blot analysis were also used to measure ALDH enzyme activity using the spectrophotometric assay as described previously (5). Briefly, the aliquots of 600 μl lysing buffer were incubated at 37°C in Beckman DLC 64 spectrophotometer cuvettes with the addition of cell lysate, 5 mM NAD+ and 5 mM propionaldehyde as a substrate (both obtained from Sigma). The change in absorbance at 340 nm was measured in three replicates over a 5 min time frame. A control reaction in which the substrate was not added monitored the endogenous rate of NAD+ reduction. The ALDH activity was expressed in nmol/107 cells per min or nmol/mg protein per min.
Use of Aldefluor Reagent to Measure Intracellular Levels of ALDH
The Aldefluor reagent system supplied by Stem Cell Technologies (Vancouver, Canada) was used in all experiments and offers an immunofluorescent method to detect intracellular enzyme activity of ALDH. Aldefluor is supplied in the form of Bodipy-aminoacetaldehyde diethyl acetal (BAAA-DA). BAAA-DA is converted to Bodipy-aminoacetaldehyde (BAAA) by exposure to acid. BAAA is a fluorescent substrate for ALDH. In this study, we treated both small cell and NSCLC cell lines at 0.2–1.0 × 106 cells in 1 ml of Aldefluor assay buffer with BAAA at a concentration of 1.5 μM for 45 min at 37°C. BAAA is uncharged and can diffuse freely across the plasma membrane of intact viable cells. Intracellular ALDH converts BAAA into Bodipy-aminoacetate (BAA), which is retained intracellularly because of its net negative charge, which disallows free diffusion. The assay buffer supplied with the Aldefluor kit contains a transport inhibitor, which prevents efflux of the BAA from the cells. As a result, cells expressing high levels of ALDH retain BAA and thus fluoresce. In each experiment, a sample of cells was stained under identical conditions with 50 μmol/l of the specific ALDH inhibitor diethylaminobenzaldehyde (DEAB; Sigma, St. Louis, MO) to serve as a negative control.
Flow Cytometry Analysis
Flow cytometry was performed using a FACScan instrument equipped with a 488-nm argon laser (B-D Biosciences, San Jose, CA). Aldefluor fluorescence was detected using the green fluorescence channel (530 ± 15). Propidium iodide (Sigma-Aldrich, St Louis, MO) fluorescence was detected using the orange fluorescence channel (585 ± 21 nm). Data for 10,000 cells was collected and analyzed using Cell Quest software, version 3.3 (B-D Biosciences).
4-Hydroperoxycylophosphamide Cytotoxicity Measured by Aldefluor Assay
4-Hydroperoxycyclophosphamide (4-HC) is an active derivative of cyclophosphamide and is oxidized into inactive metabolite by ALDH1A1 and ALDH3A1. A correlation between the cellular levels of these enzymes and toxicity of 4-HC has been extensively reported before (18–22). Therefore, we chose two cell lines with different levels of ALDH activity, including SW210.5 (SCLC) with minimal ALDH activity and H522 (NSCLC) with relatively high ALDH activity, in order to examine the toxicity of 4-HC using Aldefluor flow cytometry assay. We hypothesized that cells with higher ALDH activity will be spared from the toxicity of 4-HC, while cells with the low ALDH activity will show a higher percentage of dead cells. Cultured cells in the log phase of their growth were trypsinized if needed, washed, counted, and aliquoted (5 × 105 cells/4 ml) into 15 ml tubes. After about 2 h in the incubator, cells were treated with freshly prepared 4-HC (NIOMECH, Bielefeld, Germany) for 30 min at 37°C. The cells were then washed twice with chilled culture medium, resuspended in culture medium, and left in the incubator overnight. Next day, cells were harvested, counted, and then stained for flow cytometry using the Aldefluor assay in combination with propidium iodide staining for the quantitation of dead cells.
ALDH1A1 and ALDH3A1 Expression in Lung Cancer Cell Lines
Figure 1A and 1B shows the protein levels of ALDH1A1 and ALDH3A1 SCLC and NSCLS cell lines. The results show that higher levels of both enzymes are more frequently detected in NSCLC cell lines when compared to SCLC cell lines (H-82, SW210.5, and SCLC-16HC), with heterogeneity within each group. Relative normalized units of the ratio ALDH1A1/actin and ALDH3A1/actin are shown in Table 1 along side the results of the Aldefluor and spectrophotometry assay, both measuring ALDH activity. In general, a very good correlation between these ratios and ALDH activity can be seen. Beas-2B is an immortalized lung epithelial cell line that was used in our flow cytometry experiments and it has no detectable protein bands on Western blot (blot is not shown). The Aldefluor results reflect the percentage of fluorescent cells without taking into account the intensity of the staining at the individual cell level. This limitation of the Aldefluor staining may explain the differences seen between the two methods in regards to the ALDH activity in SW210.5 SCLC cell line (Table 1). To verify the status of the expression of these enzymes in the SW210.5 cells, we performed RT-PCR on total RNA extracted from these cells and detected ALDH1A1 and ALDH2 but not ALDH3A1 (Fig. 2). All these enzymes use propionaldehyde as a substrate. Thus, the only explanation we have for the slight discrepancy between ALDH activity by Aldefluor and that by spectrophotometry in SW210.5 cells is the possibility that the two in this case reflect different aspects: the Aldefluor assay reflects the % of fluorescent cells, while the spectrophotometry reflects the total amount of activity. In view of that, we can state that the Aldefluor assay may have a new advantage of identifying ALDH expression in cell lines that register as negative when assayed using other standard technologies.
ALDH Activity by Aldefluor Flow Cytometry Assay
Our preliminary results show that cell lines with high level of ALDH activity, such as A549, may create artifactual problems due to the shear amounts of ALDH molecules in the cells that may require impractically large amounts of the fluorescent substrate or the inhibitor in order to achieve accuracy in this flow cytometric analysis. After we failed to improve the results by increasing the fluorescent substrate by 4-folds and the inhibitor by 2-folds, we chose to perform mixing experiments in which lung cancer cells are mixed at 1:10 ratio with immortalized lung epithelial cells (Beas-2B) that are known to be lacking any significant ALDH activity (Table 1). The amount of Aldefluor positive cells (%) is determined by placing the R3 gate in the region based on the shift of fluorescent cells seen after DEAB (inhibitor of ALDH activity) treatment. Figure 3 shows that the ALDH activity in A549 cells is >80% (or 8 out of 10 A549 cells displayed Aldefluor fluorescence) by the dilution assay (middle panel), while it would be only 3% (or 3 of 100 A549 cell) without the dilution (right panel). The same data is also shown in the form of histograms (Fig. 3, bottom row). Such approach can similarly be used with cell lines that have lower ALDH activity, such as SW210.5 (Fig. 4), but it may not be needed because ALDH activity measurements were close, 83% without mixing with Beas-2B cell (panel A) and 7% of 10% SW210.5 cells mixed 1:10 with Beas-2B cells or 70% (panel B).
We next tested this approach to measure the effect of retrovirally mediated expression of siRNA [pL(AD1)SN] against ALDH1A1 on overall ALDH activity in A549 cells. A549 cells were chosen specifically for the siRNA experiments because of the high levels of expression of both enzymes. We thought that if the siRNA was found to be effective in these cells, then there will be no doubt that it will be effective with other cell types that have lower levels of these enzymes (14). Stably transfected A549 cells again mixed 1:10 with Beas-2B epithelial cells were analyzed. Figure 5 shows significant decrease in ALDH activity in the cells with pL(AD1)SN in comparison to the control cells expressing pLXSN vector only.
In some instances, we have also noticed heterogeneity of ALDH expression within the same cell line with a spectrum of expression from low to very high such as is seen with SW210.5 and H522 cells. Because cells with low ALDH activity could potentially be dying cells, we used propidium iodide (PI) to differentiate dead cells from cells with true low ALDH expression. Figure 6, shows the two cell lines both before and after PI staining. The side scatter panel with just Aldefluor staining shows two populations, but when PI is added, gating on the R3 fraction reveals mostly ALDH+ cells, while gating on the non R3 fraction reveals 61.99% of SW210.5 cells (Fig. 6, panel A) and 58.5% of H522 (Fig. 6, panel B) to be dead while the rest is true ALDH negative cells. Thus, the results in Figure 6 show that there are three cell populations within each of these two cell lines including those with positive ALDH activity (ALDH+), those with negative ALDH activity and PI− (ALDH−), and cells that are dead (PI+) and have no ALDH activity. After 4-HC treatment, the three cell populations are affected in both cell lines (Fig. 7A and 7B). The data is presented differently in this figure to better display the three cell populations. The results show that after 20 μg/ml 4-HC most of the SW210.5 cells are either dead (PI+) or dimly fluorescent for ALDH (>97%) (panel A), while about 78% of the H522 cells continue to exhibit brightly fluorescent ALDH staining, indicating relatively higher resistance to 20 μg/ml 4-HC (middle box, panel B), most likely due to higher level of ALDH activity in these cells in comparison to SW210.5 cells. H522 cells were also treated with 40 μg/ml 4-HC, which resulted in a dose response increase in the dead cells with decrease in the brightly positive ALDH cells to 60% (right box, panel B). These results again demonstrate the ability of the Aldefluor staining to measure the effects of 4-HC on the three cell populations mentioned earlier, and demonstrate dose response in H522 cells treated with two different doses of 4-HC. It is possible that ALDH− cells are dying but remain with intact membranes, which prevent them from becoming PI+. Some support for that comes from manually counting the viable cells using trypan blue after overnight incubation and before performing this specific experiment shown in Figure 7. The calculated portion of H522 cells recovered (number of viable cells after 4-HC treatment/number of viable cells without treatment at the same time point X 100 = % of recovered cells) at 1 and 5 days after 20 μg/ml 4-HC was 90% and 72%, respectively, and 8% and 1.9%, respectively, for SW210.5 cells. The recovered portion of cells treated with 40 μg/ml 4-HC after 5 days was 12% for H522 and 1% for SW210.5. SW210.5 cells had similar high sensitivity to 4-HC even at concentrations as low as 5 μg/ml. Overall, these findings were observed with repeated experiments and confirm that some dying cells will initially maintain membrane integrity to be PI−, but will be leaky enough to become negative for Aldefluor fluorescence. Thus, combining PI and Aldefluor staining may give a better assessment for 4-HC cytotoxicity in a specific cell type.
Aldefluor staining and flow cytometry analysis have been used in stem cell research on the basis of high ALDH activity in stem cells, mainly hematopoietic stem cells (23–29). In this study, we demonstrate for the first time that Aldefluor staining can be used successfully to assess ALDH activity in lung cancer cells, even when the ALDH activity is very high. Because the identification of ALDH positive cells (ALDH+) depends on a companion sample of control cells stained with BAAA in the presence of 10 μM DEAB, a potent inhibitor of ALDH, the presence of very high ALDH activity such as in A549 cells exposed a new limitation to the assay. This limitation was most likely related to the limited ability of DEAB to inhibit the activity of large amounts of ALDH1A1 and ALDH3A1. Thus it was necessary to dilute these A549 cells with cells lacking ALDH activity, such as the Beas-2B bronchial epithelial cells, in order to obtain accurate activity measurement which, in these cells, was virtually 100%. Since this is the first time such analysis has been reported in solid tumor specimens, we hope that our report and suggested modifications of the assay will open the door for similar analysis in other tumor types.
Furthermore, in order to show a practical use of the Aldefluor assay in cancer cells, we stained A549 cells that were transfected with pLXSN retroviral vector containing a siRNA against ALDH1A1, and showed a significant decrease in the ALDH activity in comparison to cells transfected with empty pLXSN.
The Aldefluor staining was used as recommended by the manufacturer in cells with low levels of ALDH activity. Our overall findings revealed low levels of ALDH activity in SCLC cell lines, while higher levels were detected in some, but not all, NSCLC cell lines. These results correlated very well with protein and enzyme activity as measured by Western blot and spectrophotometry, respectively, and we conclude that heterogeneity in ALDH activity exists among and within the different lung cancer cell types. There was one exception as discussed in the results in which SW210.5 cell line registers small amount of ALDH activity by spectrophotometry and no detectable protein bands for ALDH1A1 and ALDH3A1 by Western blot, while the Aldefluor assay show about 50% of them with ALDH activity. By semi-quantitative RT-PCR, this cell line expresses mRNAs for ALDH1A1 and ALDH2, but not ALDH3A1, all use the same substrate for activity assay. The fact that these cells are highly sensitive to 20 μg/ml of 4-HC suggests that these cells do not have high level of ALDH activity. Thus, in our opinion, Aldefluor assay reflects the % of fluorescent cells but not the total amount of activity and therefore it is likely more sensitive in identifying ALDH activity in cell lines when other technologies fail, but less quantitative.
Aldefluor staining was initially marketed as a sensitive method for detecting viable cells, mainly viable hematopoietic progenitor cells, because dead cells will not retain fluorescence (Manufacturer Brochure). At least two cell lines, SW210.5 (SCLC) and H522 (NSCLC), show two cell subpopulations with Aldefluor staining and therefore it was important to use propidium iodide (PI) and DEAB in order to be able to separate dead cells from truly low ALDH (ALDH−) expressing cells. These experiments confirmed the presence of an ALDH− population in both cell lines. Moreover, treatment with 4-HC increased the dead cells (PI+) but did not eliminate the ALDH− cells; however, trypan blue staining and long term culture of 4-HC-treated cells indicate that the only remaining viable cells were the ALDH+ cells. Thus, our studies confirm the potential usefulness of Aldefluor staining as a more accurate method for determining true cell viability rather than the measurement of dead cells. Even then, the quantitation of viable cells after chemotherapy treatment using Aldefluor may vary depending on the timing after treatment. Further studies, which are beyond the scope of this study, will be needed to better define the use of Aldefluor assay in this context.
There is increasing evidence supporting the cancer stem cell hypothesis (30, 31). It has also been proven that tumor cells are heterogeneous comprising rare tumor initiating cells and abundant nontumor initiating cells. Tumor initiating cells are presumed to be the cancer stem cells that have the ability for self-renewal and proliferation, are resistant to drugs, and express typical markers for stem cells. The existence of cancer stem cells was suggested in human SCLC cell lines NCI-H146 and H345 as defined by detecting the Hoechstlow side population (SP) using flow cytometry (32). Bronchioalveolar stem cells and alveolar type II cells have been suggested to be the pluripotent stem cells that give rise to the NSCLC mainly in animal models (33). Whether these cancer stem cells will express high levels of ALDH, as shown for hematopoietic stem cells, is an important research question that now can be addressed with the use of Aldefluor staining. The heterogeneity in expression of ALDH activity that we have established in this study suggests the existence of a spectrum of cells expressing low, intermediate, or very high ALDH levels among the different lung cancer cell lines as well as within some cell lines. If indeed lung stem cells express high levels of ALDH activity, then the finding that some cell lines express high ALDH activity homogeneously while other cell lines express ALDH in less than 1% of the cells may indicate that some cell lines originate purely from undifferentiated stem cells while others originate from stem cells able to differentiate into progenitors with low ALDH activity. The availability of the Aldefluor assay may be critical in investigating such hypothetical scenario in the future.
In summary, the Aldefluor assay can be adopted to measure ALDH activity in live solid tumor cells and it opens new horizons in studying ALDH role in the biology of lung cancer such drug resistance, tumor behavior, as well as the cell origin of lung cancer.