A novel method based on click chemistry, which overcomes limitations of cell cycle analysis by classical determination of BrdU incorporation, allowing multiplex antibody staining

Abstract

Quantification of BrdU incorporation into DNA is a widely used technique to assess the cell cycle status of cells. DNA denaturation is required for BrdU detection with the drawback that most protein epitopes are destroyed and classical antibody staining techniques for multiplex analysis are not possible. To address this issue we have developed a novel method that overcomes the DNA denaturation step but still allows detection of BrdU. Cells were pulsed for a short time by 5-ethynyl-2′-deoxyuridine, which is incorporated into DNA. The exposed nucleotide alkyne group of DNA was then derivatized in physiologic conditions by the copper (I)-catalyzed azide-alkyne cycloaddition (CuAAC) using BrdU azides. The resulting DNA-bound bromouracil moiety was subsequently detected by commercial anti-BrdU mAb without the need for a denaturation step. Continuous labeling with EdU showed a slightly increased anti-proliferative activity compared to BrdU. However, using a lower concentration of EdU for labeling can compensate for this. Alkynyl tags could be detected quickly by a highly specific reaction using BrdU azides. Fluorescence quenching by the DNA dye PI using both BrdU azides was negligible. Our labeling method is suitable for FCM and HCA and shows a higher signal to noise ratio than other methods. This method also allowed multiplex analysis by simultaneous detection of EdU-BrdU, caspase-3, and phospho-histone 3 mAbs, proving sensitivity and feasibility of this new technique. In addition, it has the potential for use in vivo, as exemplified for bone marrow studies. We have established a new method to determine the position of cells in the cell cycle. This is superior when compared to traditional BrdU detection since it allows multiplex analysis, is more sensitive and shows less quenching with PI. The method provides new opportunities to investigate changes in protein expression at different cell cycle stages using pulse labeling experiments. © 2008 International Society for Advancement of Cytometry

FCM is one of the most commonly used techniques for studying protein expression at the level of single cells, in particular where multiparametric (1, 2) and multicolor (3) analyses are used as a readout. Integration of FCM into proteomics technologies have played a key role for cell functions and proliferation studies using “cytomics” approches (4, 5). Incorporation of the thymidine analogue BrdU into DNA of proliferating cells is widely used to assess the cell cycle status of cells using different methods, such as image cytometry, HCA, and FCM (6, 7).

A downside of this technique is that detection of BrdU requires DNA denaturation causing many epitopes to become modified or destroyed (8–12), resulting in classical antibody staining methods for multiplex analysis no longer possible (11, 12).

To address this issue we have setup a novel method for identifying S-phase cells that overcomes the DNA denaturation step but still detects BrdU.

The “Click Chemistry” bio-conjugation reaction using the Sharpless cycloaddiction of azides and acetylenes catalyzed by Copper (I) (CuAAC) is widely used in drug discovery, for example for protein tagging (13) or DNA sequencing (14). The method has favorable thermodynamic properties accompanied by high specificity and a high quantitative yield of the end product in aqueous solvents at physiological temperatures, providing 1,4-disubstituited triazoles (15).

In our method, using CuAAC, we were able to detect alkynyl tags from incorporated 5-Ethynyl-2′-deoxyuridine (EdU) during DNA synthesis by reactive BrdU azides under conditions preserving protein epitopes. The resulting bromouracil moiety can then be detected by anti-BrdU mAb. Here we describe the assay development and validation for FCM and image cytometry, including HCA. The combination of the click chemistry reaction with good mAb specificity and an amplification step by AlexaFluor® secondary step are the key points for the high specificity and sensitivity of this new approach.

MATERIALS

Media and Chemical Reagents

FBS, McCoy's 5A and RPMI 1640 medium, PBS pH 7.4, Penicillin/Streptomycin solution, L-glutamine, AlexaFluor® 488 or 647 goat anti mouse and rabbit IgG and ClickIT™ EdU AlexaFluor® 488 were purchased from GIBCO-Invitrogen. Azido deoxy uridine (AdU), 5-bromo-2′-deoxyuridine (BrdU), Camptothecin, Copper (II) sulphate, DMSO, methanol-free-Formaldehyde 37%, Paclitaxel, thymidine, Trypsin/EDTA solution, RNAse A from bovine pancreas, propidium iodide (PI), methanol, (+)-Sodium-L-ascorbate, Tween® 20, Triton® X-100 and chemicals for probe synthesis were purchased from Sigma-Aldrich. The anti-BrdU mAb (clone B44) was purchased from BD Biosciences. Cleaved Caspase-3 (Asp¹⁷⁵) rabbit mAb (9661L) was purchased from Cell Signaling Technology Inc. and phospho-histone H3 (Ser¹⁰) rabbit mAb (06-570) was purchased from Upstate/Millipore.

5-Ethynyl-2′-deoxyuridine (CAS 61135-33-9, EdU) was kindly supplied by Dr. Adam Sniady from Oakland University. For in vivo studies, EdU (PY7562) was purchased from Berry & Associates Inc, MI. Synthesis and chemical identification of 5′-BMA and 3′,5′-BDA (16) are shown in our online Supplementary Materials. 3′-BMA, 5-Bromo-3′-azido-2′,3′-dideoxyuridine (CAS 105784-82-5) was synthesized as published (17) and kindly donated by Dr. Sylvie Pochet from Institut Pasteur, Paris.

Cell Culture

Cell lines were obtained from the European Collection of Cell Cultures (ECACC). U2-OS cells were grown in McCoy's 5A medium, A2780, and HL-60 were grown in RPMI-1640 medium. Media was supplemented with 10% FBS, L-glutamine 2 mM, and 1% penicillin/streptomycin solution and cells were maintained in 5% CO₂ at 37°C. Cells were grown as monolayers in CellStar™ T75 flasks (Greiner Bio-One, Germany) and were routinely sub-cultured twice weekly, detaching them with 1 mL of 0.25% trypsin/0.02% EDTA.

Animals

Ex-vivo experiments were performed in male BALB nu/nu mice housed under pathogen-free conditions in micro-isolator cages, with irradiated rodent chow and water available ad libitum. Animal studies were performed in compliance with Italian Legislative Decree 116, January 27, 1992, enforcing European Communities Council Directive 86/609/EEC on the protection of animals used for experimental or other scientific purposes, and in accordance with institutional policy regarding the care and use of laboratory animals.

METHODS

Cellular Antiproliferation Assay (ATP-lite) for IC₅₀ Determination

Cells were seeded at 5000/well in CulturPlate™-96F 96-wells-plates (Packard-Perkin Elmer) for 24 h. The next day cells were treated for 72 h with different doses (ranging from 0.05 to 500 μM) of EdU, BrdU, and Camptothecin as a standard reference. After incubation, cells were analyzed by CellTiter-Glo® Luminescent Cell Viability Assay (18) from Promega (Promega Corp., Madison) and IC₅₀ values were calculated using the LSW Data Analysis Toolbox for Excel.

Drug Treatments

Cells in the exponential growth phase were treated at 7 h with Paclitaxel at 1 μM or Camptothecin at 5 μM or for 24 h with thymidine at 10 mM. At the indicated times, cells were pulsed for 30 min with EdU (1 μM for HL-60 or 10 μM for A2780 and U2OS cells). For adherent cells, supernatant medium was collected to avoid cell loss and cells were washed using warm PBS, detached with trypsin and added to the previously collected medium. Cells were counted in a Multisizer™ 3 Coulter Counter® (Beckman Coulter Inc., CA) to measure the drug's effects on growth inhibition. Part of the sample (2 × 10⁶ cells) was fixed in cold methanol 70% for cell-cycle analysis and phospho-histone H3 determination and stored at −20°C; another part (2 × 10⁶ cells) was fixed by 3.7% methanol-free Formaldehyde for 15 min, washed twice with PBS and stored at 4°C in 70% methanol before proceeding with the protocol to determine levels of cleaved caspase-3.

Three-Parameter FCM Analysis of Protein Expression, EdU Detection and DNA Content

Protein expression

10⁶ cells were fixed with formaldehyde or methanol, washed with PBS plus 1% FBS solution (PBS-solution) and then permeabilized for 10 min with Triton® X-100, 0.1% in PBS at 4°C.

After one wash with PBS, 1 μg per sample of rabbit anti-cleaved caspase-3 or phospho-histone H3 mAb in PBS was added and incubated for 1 h at room temperature. Thereafter, samples were washed with PBS solution and incubated with 0.5 μg (1:300) per sample of AlexaFluor® 488 conjugated goat anti-rabbit mAb for 1 h at room temperature. Cells were then washed with PBS solution in preparation for the subsequent step.

Click Chemistry Coupling Via CuAAC

Cells in 1 ml PBS were mixed for 10–15 min with 10 μl sodium ascorbate 1 M and 20 μl Copper (II) Sulphate 0.1 M. This was followed by the addition of either 10 μl BMA or BDA (at 10 mM in DMSO). Direct labeling by ClickIT™ AlexaFluor® 488 was carried out according to the manufacturer's instructions using 5 μl AlexaFluor® 488 Azide.

Cells were then washed with 3 ml PBST (PBS solution and 0.5% Tween® 20) for 15 min and incubated for 30 min at room temperature with 150 μl anti-BrdU mAb diluted 1:50 inPBST. After washing with PBST, cells were incubated for 30 min with AlexaFluor® 647 conjugated F(ab)₂ fragment goat anti-mouse IgG diluted 1:300 in PBST at room temperature in the dark. Cells were incubated for 30 min for direct ligation of AlexaFluor® 488 Click-IT™. When only direct comparison to AlexaFluor® 488 Click-IT™ was required, we replaced the secondary step by AlexaFluor® 488 conjugated mAb.

DNA content analysis

After incubation, cells were centrifuged, washed with PBS, mixed in PI 2 μg/mL in PBS plus RNase A and incubated for 1 h at room temperature or overnight at 4°C in the dark.

FCM setup

Cells were analyzed by dual laser 488/633 nm BD FACSCalibur™. Fluorescence signals were detected using a band-pass filter 530 ± 30 nm and 620 ± 35 nm in combination with a dichroic mirror at 570 nm for the 488 nm excitation and band-pass filter 661 ± 16 nm for the 635 nm excitation. For cell cycle studies, aggregates were excluded from DNA analysis by gating on area/height fluorescences, acquiring at least 10,000 events inside singlets gating.

Image Acquisition by HCA ArrayScan™ 4.5 (Cellomics) Reader

For fluorescence-based high-content analysis, U2-OS cells (6000/well) were grown in ViewPlate™ 96F (PerkinElmer Life Sciences, Boston) for 24 h. After appropriate drug treatments (e.g. 7h or 24h), cells were pulsed with 10 μM EdU for 30 min and were fixed for 15 min by adding formaldehyde to a final concentration of 3.7%. Cells were washed twice and stored in PBS. The following day, cells were permeabilized with 0.3% Triton® X-100 in PBS for 15 min at 4°C and washed twice with PBS. For Click Chemistry coupling, 100 μL of BMA or BDA at 0.1 mM in PBS, 10 μL Sodium ascorbate 100 mM and 20 μL Copper (II) Sulphate 10 mM were added consecutively and cells were exposed for 15 min. For direct labeling by ClickIT™ AlexaFluor® 488, 100 μL/well of reaction buffer containing 30 μL AlexaFluor® 488 Azide in 1 ml Click Chemistry reaction buffer was used. After two washes of 10 min each with PBST, cells were incubated for 30 min with anti-BrdU mAb diluted 1:100 in PBST. Cells were washed twice with PBS and incubated with 100 μL anti-mouse AlexaFluor® 488 or 647 conjugated IgG diluted 1:1000 and 1 μg/mL DAPI in PBST at room temperature for 30 min in the dark. After incubation, cells were washed twice and resuspended in 200 μL PBS until image acquisition. Image acquisition and analysis was performed using a ThermoFisher ArrayScan™ 4.5 HCA reader equipped with a 200 W mercury-xenon lamp. A quadruple band filter XF93 was used to acquire fluorescence in the DAPI channel (365WB50) and green channel (475RDF40) with a 10× objective; for data analysis 700–900 cells were collected for each sample. The nuclear fluorescence intensity was measured with the Cyto-Nuc Translocation Bioapplication Software (Thermo Fisher Scientific).

EdU Administration and Bone Marrow Cell Isolation

EdU was prepared in 0.9% NaCl solution containing 5% DMSO at 5 mg/ml and finally diluted in 0.9% NaCl solution from 0.5 to 50 mg/kg prior to administration. BrdU was dissolved directly in physiological solution and administered at 100 mg/kg (1). After 1.5 h, mice were sacrificed and bone marrow cells were isolated from tibias and femurs. Cells were resuspended in PBS solution and kept on ice. Finally, cells were counted by Coulter Counter and fixed in methanol 70% as described previously.

RESULTS

5-Ethynyl-deoxyuridine Shows a Low Cytotoxicity In Vitro and Ethynyl Tags can be Detected With High Sensitivity Using Click Chemistry Coupling by BrdU Azides

The proposed chemical reaction between the ethynyl residue of EdU (Fig. 1A, in comparison to BrdU) incorporated during DNA synthesis and the BrdU azide probe by Click Chemistry via CuAAC is shown (Fig. 1B). The linked bromodeoxyuridine moiety by triazole product provides an easily accessible epitope for the subsequent anti-BrdU mAb staining (Fig. 1B).

Figure 1. (A) EdU and BrdU chemical structures. (B) Click chemistry reaction between the ethynyl tag and azide bearing group BrdU, catalyzed by copper, results in covalently and stable 1,4-triazole product bound BrdU for subsequent mAb detection. (C) ATP antiproliferation assay at 72h treatments (ED₅₀) after varying lengths of EdU exposure. As a positive control CPT was used, which showed an IC₅₀ < 50 nM in all cell lines after 72 hours of incubation (data not shown).

To assess EdU cytotoxicity, we compared the antiproliferative activity of EdU and BrdU in an ATP chemoluminescence assay (Fig. 1C). Cells were exposed to the compounds for 72 h or were pulsed for 1 h and allowed to recover for 72 h prior to assessment of proliferation.

Using this long-term assay, EdU showed a mild antiproliferative activity (from 5 to 63 μM) compared to BrdU (85 to 401 μM). In the pulsed cells, which is a more relevant parameter for our method (1 h plus 72 h drug-free recovery), the IC₅₀ was in the range of 93 to >250 μM for EdU and >250 μM for BrdU in the cell lines tested. When taking into account the 10-fold lower effective dose required to obtain a good signal for EdU compared to BrdU, the margin for tolerability of EdU in vitro is estimated to be at least comparable to that of BrdU.

Titrations were performed to determine the optimal dose of EdU for cellular incorporation. U2-OS and HL-60 cells were pulsed with different concentrations of EdU for 30 min and the incorporation efficiency was determined (Fig. 2A). Depending on the cell line, at concentrations of 1–10 μM of EdU, a S/N plateau was reached comparable to the optimal BrdU concentration in a classical assay. EdU incorporation was found to be rapid, for example, HL-60 cells exposed to 1 μM EdU showed saturation within 20 min. Similar results were obtained using A2780 cells (data not shown). The improved signal to noise ratio at concentrations of EdU higher than 1 μM translates into a higher sensitivity of the assay compared to BrdU incorporation.

Figure 2. (A) EdU incorporation in U2-OS and HL60 cell lines detected by BMA coupling followed by mAb plus AlexaFluor^® 488 secondary step. In the right table are included S/N ratio and percentage of stained cells. (B) BMA titration for FCM assay setup. Shown is the FCM analysis of EdU pulsed A2780 cells after ten minutes reaction time (black bars) along with the percentage of positive cells detected (black line). (C) Measurement of coupling kinetics in A2780 cells in the presence of CuAAC for ligation of EdU to AdU, BMA, BDA and Click-IT™ AlexaFluor^® 488.

BrdU Azide Coupling Shows a Rapid Kinetic and Reduced Quenching When Using PI in Combination with EdU Detected by AlexaFluor® 488

To study the kinetics of BrdU coupling we incubated EdU pulsed A2780 cells with 0.001-5 × 10⁻⁷ mol of BMA and BDA in 1 ml click chemistry buffer for various times ranging from 1 to 60 min. Cells were then washed with 5 ml of 2% EDTA pH 7.4 to stop the copper activity, followed by a second wash with 10 ml of PBST. Finally, cells were incubated with anti-BrdU mAb and anti mouse AlexaFluor® 488 as described in material and methods (Fig. 2B).

A sigmoid dose response was observed for the detection of the ethynyl tags with BMA (Fig. 3A). At concentrations above 0.05 mM BMA, saturation of the signal can be observed corresponding to 0.5 × 10⁻⁷ mol/million cells. Similar results were obtained when using the 3′-BMA isomer or BDA, but with a greater absolute fluorescence intensity.

Figure 3. (A) S/N ratio after CuAAC ligation by Click-IT™ AlexaFluor^® 488 or BMA/BDA followed by mAb plus AlexaFluor^® 488 secondary step and percentage of positive cells measured before and after PI staining. (B) Relative EdU incorporation and DNA content dot-plot. Histograms displayed inside dot-plots show EdU-AlexaFluor^® 488 fluorescence emission before PI (upper right) and after PI staining (lower right).

We also compared the time kinetics of detection of the ethynyl tags by direct ligation with AlexaFluor® 488 Click-IT™ or by indirect ligation with BMA or BDA azides. BMA and BDA show a faster reaction (peaked at 2 min) kinetic compared to AlexaFluor® 488 Azide used for the Click-IT™ method which reaches optimal signal intensity after 15 min. Interestingly, a bright fluorescence peak was observed at very short times with both BrdU azide derivatives, while at longer time points a small decrease in fluorescence was observed. The signal stabilized after 15 min but still with a high S/N ratio.

Since quenching of PI is often an issue for classical BrdU incorporation, we compared AlexaFluor® 488 Click-IT™ and our method in which the anti-BrdU mAb is detected by AlexaFluor® 488 before and after PI staining at 2 μg/ml (Fig. 3A and 2B). AlexaFluor® 488 fluorescence histogram analysis shows the S/N ratio and percentage of positive cells before and after PI staining, revealing that PI strongly quenched green fluorescence from Click-IT™ direct ligation whereas little quenching was observed for the same fluorochrome associated with the BrdU azides (Fig. 3A). This difference is probably due to the wider distance of the fluorescent dyes resulting from the indirect staining approach used in our method. Interestingly, cells treated with BDA showed a greater S/N ratio after staining than cells treated with 3′ or 5′-BMA (Fig. 3B).

BrdU Azides Can Also be Efficiently Used for HCA Applications

To study if BrdU azide derivatives could be used in HCA applications we pulsed U2-OS cells by EdU and labeled by AlexaFluor® 488 Click-IT™ ligation or indirectly with BMA or BDA followed by mAb detection and finally counterstained with DAPI (Fig. 4).

Figure 4. (A) Azide titration for assay setup in HCS Analysis. Data analysis were based on the average of 8-well replicate. (B) Representative fields of asynchronous growing U-2OS cells (ct) or cells treated with thymidine 10 mM (th) pulsed using EdU. Images were acquired with ArrayScan HCA system by exposing fields for fixed times using a XF93 excitation filter in the DAPI and green channel with a 10x objective. (C) ArrayScan analysis distributions compared with FCM histogram overlays (continuous line represents control cells; dotted line represents thymidine treated cells). The percentage of positive cells were calculated using the shown threshold and are displayed in the upper right of the histograms, with numbers for untreated cells in bold text and treated cells in normal text.

In Figure 4A, a dose titration was performed in order to validate the HCA assay. Results were calculated based on the average of an 8-well replicate. With our method, we saw a strong correlation between MFI and % of positive cells and saturation was observed at >0.1 nmol/well.

As shown in Figure 4B, we observed higher fluorescences for BMA and BDA compared to the Click-IT™ direct ligation. BDA showed an approximately two-fold increased fluorescence compared to BMA, as reported for the FCM studies. At concentrations greater than 1 nmol/well we were also able to completely stain EdU pulsed-cells.

In Figure 4C, we compared HCA and FCM analysis of U2-OS cells either growing asynchronously or following 24-h treatment with 10 mM thymidine to block them in G1-S phase of the cell cycle. U2-OS cells were pulsed by 10 μM EdU for 30min and were then washed, collected, and fixed as described above. For HCS Analysis, cells were stained by Click-IT™ AlexaFluor® 488 or BMA/BDA in saturation conditions. Nuclei were counterstained with DAPI. For FCM analysis, we compared green fluorescence histograms before PI staining in order to avoid signal quenching. As expected, 10 mM thymidine treatment abrogated EdU incorporation since thymidine induces a G1-S cell cycle block (1). Results from HCA and FCM analysis were both in agreement and showed a clear decrease in the signal for treated cells (dotted line).

Coupling of BrdU Azides Allows Multiparametric Analysis for Cell Cycle studies

To evaluate if our method is compatible with multiplex staining, we analyzed HL-60, A2780, and U2-OS cells to assess simultaneously the cell cycle profile, the presence of mitotic cells or induction of apoptosis in the presence of different cytotoxic drugs (Fig. 5). Cells in mitosis were determined by following phosphorylation of histone H3 Ser¹⁰ while cleavage of caspase-3 was used as a marker for induction of apoptosis.

Figure 5. (A) Dot-plot of EdU incorporation vs. DNA content after Camptothecin or Paclitaxel treatments. Cells were gated for the different cell cycle phases in yellow (sub-G1), green (G1), red (EdU-BrdU positive cells), black (EdU-BrdU negative cells) and blue (G2/M). Numbers indicate the relative percentages of cells in each phase of the cell cycle. Histograms displayed inside dot-plots shown relative DNA content profiles. (B) EdU incorporation and detection of cleaved caspase 3 or phospho histone H3 protein epitopes in untreated or drug treated cells.

We gated the populations in different phases of the cell cycle using PI staining: sub-G1 (yellow), G1 (green), S-phase (red), and G2/M (blue). This allowed us to follow the cell cycle specific changes of our markers (Fig. 5A) which were detected using AlexaFluor® 488, whereas incorporated EdU-BrdU was detected using AlexaFluor® 647 as a fluorochrome linked to the secondary Ab.

HL-60, A2780, and U2-OS cells were treated for 7 h with camptothecin (CPT) 5 μM or paclitaxel (PTX) 1 μM. At the end of the treatments, cells were exposed to EdU for 30 min at 1 μM for HL-60 and 10 μM for A2780 and U2OS and processed as described in material and methods.

As expected, the topoisomerase I inhibitor CPT induced an S-phase block and apoptosis, while the microtubule binder PTX caused a decrease of cells in S-phase and an increase of cells in G2/M phase. The biparametric analysis for EdU-BrdU staining versus cleaved caspase-3 (Fig. 5B) shows a substantial increase of cleaved caspase-3 after CPT treatments in all cell lines tested, with the exception of HL-60.

Phosphorylation of histone H3 was also followed during different phases of the cell cycle after treatment with the various cytotoxic drugs (Fig. 5C). Although CPT treatment blocked cells in S-phase with low phospho histone H3, PTX caused a clear increase of cells, which were positive for the mitotic marker phospho histone H3.

EdU Could Be Administrated in vivo for Bone Marrow Studies

Finally, we explored the possibility of EdU usage in vivo. Nude mice were injected i.p. with EdU from 0.5 to 50 mg/kg or with BrdU at 100 mg/kg. Isolated cells were fixed and processed as described above. Click chemistry coupling was carried out with BDA and compared to the classical BrdU incorporation method (Fig. 6). Representative overlay analysis of bone marrow cells from EdU-BrdU or BrdU treated mice are shown in (Fig. 6A) along with relative S-phase analysis in terms of positive and negative cells identified by double staining versus PI staining (Fig. 6B). Cell cycle analysis showed that a plateau was reached above 5 mg/kg EdU comparable to BrdU at 100 mg/kg.

Figure 6. (A) Overlay analysis of EdU incorporation compared to BrdU incorporation in bone marrow samples from nude mice. (B) S-Phase analysis based on cell cycle analysis. (C) Biparametric analysis of EdU incorporation at 50 mg/kg and p-histone H3 in bone marror. Dot-plots show isotype controls (IgG) for AlexaFluor^® 488 and 647, isotype control for AlexaFluor^® 647 and p-Histone 3 (IgG+pH3), BrdU incorporation at 100 mg/kg by acid denaturation and p-Histone 3 (BrdU+pH3), EdU-BrdU incorporation by CuAAC coupling and p-Histone 3 (EdU+pH3). Relative dot-plots for measurement of DNA content and EdU/BrdU incorporation are also shown.

Figure 6C shows a comparison of the dot plots for phospho-histone-H3 expression detected by AlexaFluor® 488 or BrdU and EdU-BrdU incorporation detected by AlexaFluor® 647. Cells prestained with phospho histone H3 mAb were denatured by acid treatments when BrdU was used or incubated with BDA in appropriate buffer when EdU was used. BrdU expression was then analyzed by anti-BrdU mAb and AlexaFluor® 647 secondary Ab. As expected, while BrdU incorporation was detected with sufficient sensitivity, phospho histone H3 expression clearly decreased in cells exposed to acid denaturation when stained prior to DNA denaturation. In contrast, when the click chemistry reaction was applied no decrease in fluorescence was observed.

Discussion

Since the introduction of the BrdU technique in 1982 by Gratzner, BrdU immunostaining procedures have required denaturation of chromatin to enable mAb binding to incorporated BrdU (6, 7). Usually, the denaturation step is performed by pretreating samples with chemicals, but such harsh procedures can alter cell morphology and interfere with the simultaneous immunostaining of other cellular antigens (11, 12). Other approaches used to solve this issue, such as enzymatic treatment by nucleases or UV-B photolysis, have disadvantages in terms of cost, reproducibility, and sensitivity (8–10).

In this article we have shown an interesting alternative approach for proliferation assays without a denaturation step. In our method, we coupled the terminal ethynyl tag (19, 20) of incorporated EdU to BrdU azide derivatives using a click chemistry reaction, followed by detection using an anti-BrdU mAb and fluorescence amplification in the secondary Ab staining step. Since our azide probes are still not commercially available, we also describe a rapid and convenient chemical protocol for producing BMA/BDA azides from any BrdU source. (see Online Supplementary Materials).

EdU showed a slightly increased but mild anti-proliferative activity compared to BrdU at continuous treatments (72 h) in a classical anti-proliferation assay. However, when cells were pulsed for 1 h and kept in drug-free medium in the same conditions, growth inhibition was negligible.

Previous studies (21–23) have shown that EdU increases growth of thymidine synthetase deficient cells as it acts as a thymidine analog, but conversely antiviral activity similar to other pyrimidine nucleoside analogs was observed due to its activity against deoxynucleoside kinases and thymidine synthetase. EdU modified primers showed an increase of Tm in RT-PCR reactions and an increased stabilization of the DNA duplex (24). This could explain a postreplicative toxicity after EdU incorporation during DNA synthesis. However, the precise mechanism of action of these modified nucleotides is still not well documented. Since our method is very sensitive with respect to the detection of ethynyl residues, very low concentrations of EdU (1 μM) used for a short pulse (20 min) were sufficient for the readout and reduce the risk of side effects. The BrdU azide reaction was found to be very quick and saturation was reached in less than 10–15 min. One of the major advantages of the method is that fluorescence from AlexaFluor® 488 linked antibodies after BMA/BDA ligation was minimally quenched by the commonly used DNA dye PI. On the contrary, fluorescence emission from direct ligation of AlexaFluor® 488 Click-IT™ was significantly quenched, making FCM applications in combination with PI more difficult. The greater molecular distance between the PI intercalated into DNA and the AlexaFluor® 488 fluorochrome linked to the secondary Ab in the indirect staining method might be the reason for the reduced quenching. This could decrease the risk of energy-transfer due to the spectral behaviors of the fluorescent dyes. Another advantage is that our method raises the possibility of using a broad spectrum of conjugated secondary antibodies for multicolor analysis (3, 4), which overcomes the limitation of fluorescence dye azide availability. This could also be advantageous when using HCA methods such as ArrayScan Cellomics readers, which are highly sensitive and robust when used with our method.

Different behaviors were seen for BMA and BDA, with a brighter signal observed for BDA. This might be due to improved chemical characteristics. In fact, it was shown that the 1,3-bis (azido) chemical class is more reactive than the corresponding azide (25). This could result in an increased activity against ethynyl tags in less accessible or more compact chromatin producing more alkynyl tags.

Since the major advantage of this new method, when compared with BrdU incorporation, is the option for performing multiplex analysis by click coupling of BMA or BDA, we analyzed HL-60, A2780 and U2-OS cells simultaneously for cell cycle progression and phosphorylation of histone H3 on Ser¹⁰ or induction of apoptosis by following cleavage of caspase-3 in the presence of different cytotoxic drugs. We showed differential staining patterns of these cell lines, which are in line with the expected phenotypes.

Finally, our method is also applicable for in vivo applications. We have shown that EdU administration in vivo can be used for FCM studies in isolated bone marrow cells by staining different intracellular epitopes with high sensitivity. EdU is very well tolerated and even at high doses (tested up to 50 mg/kg) no signs of toxicity were seen. The possibility of using low concentrations of EdU (down to 5 mg/kg) for detection of the signal assures a robust window for tolerability in vivo, much wider than for BrdU.

Finally, the major properties of the different methods in comparison to the classical BrdU assay are summarized in Table 1. The click chemistry methods using EdU do not require DNA denaturation since the chemical reaction was performed at physiological conditions and was compatible with multiple mAb staining and multicolor analysis including HCA applications. In addition BrdU coupling by CuAAC was faster, exhibited a higher signal to noise ratio and was less affected by PI quenching than direct ligation by fluorochrome azide.

Table 1. Summary table for comparison of BrdU classical assay and Click Chemistry approches based on direct ligation using Click-IT™ technology (AlexaFluor^® 488) and indirect coupling by BrdU azides

	BrdU INDIRECT	EdU + CLICK-IT™ ALEXA 488 “DIRECT”	EdU + BrdU AZIDE “INDIRECT”
(1) EdU shows a distinct antiproliferative effect compared to BrdU, as reported in discussion section; (2) epitope modifications have been observed after acid denaturation; (3) dependent on fluorochrome azide availability; (4) higher S/N ratio observed for BDA: ++ using BMA, +++ Using BDA; (5) N/A, not applicable for the assay.
Antiproliferative activity in vitro	Negligible	Weak (1)	Weak (1)
DNA incorporation in vivo (BM)	Yes	Yes	Yes
DNA denaturation	Required	No	No
Assay time setup	+++	+	++
CuAAc reaction time	N/A	>30 min	Short
Quenching after PI staining	+/−	++	+/−
Multiplex analysis by MAb	+/−(2)	Yes	Yes
Polychromatic analysis	++	+ (3)	++
Signal to noise ratio (sensitivity)	++	++	++/+++ (4)
Suitable for imaging/HCA	−	+	++

Our approach based on BrdU coupling by click chemistry ligation followed by mAb detection could potentially be interesting not only for cell cycle studies but also for a number of additional readouts. Putative applications envisaged are the use of ethynyl modified oligonucleotide (14, 24) for siRNA interference studies, localization of engineered proteins bearing pEthynyl-Phe aminoacids (13) or for detection of alkynyl chemical tags in modified substrates (20, 26) expanding FCM and image cytometry analytical capability (27–29).

Acknowledgements

We thank Dr. Roman Dembinsky and Adam Sniady from the Department of Chemistry and Centre for Biomedical Research, Oakland University, MI for the kind gift of EdU and Dr. Sylvie Pochet from the Institut Pasteur Paris for the gift of 3′-BMA. We also acknowledge Dr. Vanessa Marchesi for critical reading of the manuscript and the pharmacology group for care and treatment of the animals.