Sequence-Specific Histone Methylation Is Detectable on Circulating Nucleosomes in Plasma

plasma samples

If H3K9me at heterochromatic ALU elements is detectable in plasma, such detection is not expected to indicate a disease-specific event. Therefore, our study group consisted of only 21 patients with multiple myeloma at diagnosis, and the patients’ clinical characteristics were not considered. We have previously demonstrated that patients with multiple myeloma display variable concentrations of nucleosomal DNA in their blood plasma (25). The study protocol was approved by the local ethics committee of Istanbul Medical Faculty (No. 2006/1029). Peripheral blood was collected into EDTA-containing tubes and centrifuged immediately at 3000g for 30 min. Plasma aliquots were stored at −80 °C until use.

nucleosome elisa

The commercially available Cell Death Detection ELISA^PLUS Kit (Roche Diagnostics) was used according to the manufacturer’s instructions to test for the presence of histone-associated DNA fragments (mononucleosomes and oligonucleosomes) in plasma. This quantitative sandwich enzyme immunoassay contains 2 monoclonal mouse antibodies, the first of which is directed against histones (H1, H2A, H2B, H3, and H4) and binds to the streptavidin-coated microtiter plate via its biotin tail. The second is a peroxidase-labeled antibody that recognizes the DNA component of nucleosomes bound to the first antibody, and its binding is detected colorimetrically. A positive signal is evidence for the presence of both histones and associated DNA in the plasma. Samples were analyzed in triplicate, and mean absorbance values were converted to milliunits, the relative concentrations of the circulating nucleosomes.

quantification of alu-specific free dna in the plasma

We extracted circulating DNA from 200 μL plasma with the High Pure Viral Nucleic Acid Kit (Roche Diagnostics) and quantified ALU-specific cf-DNA by quantitative real-time PCR with the LightCycler (Roche Diagnostics) and the double-stranded DNA–binding dye SYBR Green I as the fluorescent molecule. The primer pair we used amplifies both apoptotic and nonapoptotic fragments representing total DNA and produces a 115-bp amplicon (26). Absolute concentrations (expressed in micrograms per liter) were obtained from a calibration curve generated from serial dilutions of genomic DNA (10 ng to 0.1 pg; r = −0.99). The PCR mix contained 2 μL SYBR Green mix (composed of Taq DNA polymerase, dNTP Mix, SYBR Green I dye, and 10 mmol/L MgCl₂), 2 mmol/L MgCl₂, 200 pg of each primer, and 3 μL of DNA in a total volume of 20 μL. A hot start of 10 min at 95 °C was followed by 45 PCR cycles of denaturation at 95 °C, annealing at 64 °C, and extension at 72 °C (10 s each). The mean concentration for 3 independent PCR amplifications was specified as the amount of free DNA in each patient.

chromatin immunoprecipitation

ChIP assays are used to evaluate the association of proteins with specific DNA regions and are the standard technique for analyzing histone methylation. The technique consists of cross-linking proteins with DNA in cells on Petri dishes, harvesting the cells, and sonicating the cells, first to isolate and break down the nuclei and then to fragment the DNA (desired size usually 200–1000 nucleotides). Finally, soluble chromatin is immunoprecipitated with an antibody that recognizes the protein of interest. Because positive ELISA signals indicate the presence of soluble and stable histone complexes and short stretches of associated DNA (mono- and oligonucleosomes) in the circulation and because there is no indication in the literature of a cross-linking requirement in immunologic assays (27), we omitted these steps and started directly with the preanalytical steps of immunoprecipitation. We modified the protocol of the Protein Agarose A Immunoprecipitation Kit (Roche Diagnostics) to immunoprecipitate nucleosomes from plasma. In brief, we mixed 200 μL of plasma with 800 μL of buffer 1 (50 mmol/L Tris-HCl, pH 7.5, 150 mmol/L NaCl, 10 mL/L Nonidet P40, 5 g/L sodium deoxycholate, and protease inhibitors), added 50 μL of a homogeneous protein agarose A suspension (25-μL bed volume) to the samples in a preclearing step designed to reduce the background signal that can occur because of nonspecific adsorption of irrelevant plasma proteins to protein agarose A, incubated the agarose bead suspension overnight at 4 °C on a rocking platform, and eventually pelleted the agarose beads by centrifugation at 12 000g for 20 s. We transferred 1 mL of the supernatants to fresh 1.5-mL plastic tubes, added 1 μg anti-H3K9me1 antibody (Upstate), and incubated at 4 °C for 1 h. We used an antihistone antibody that recognizes all histone proteins to identify the nucleosomal input, and a sample without antibody served as a negative control. To capture the immunocomplex, we added 50 μL of protein agarose A, pelleted the beads, and washed twice with buffer 1. Next, we washed the pelleted beads with 1 mL of high-salt buffer (50 mmol/L Tris-HCl, pH 7.5, 500 mmol/L NaCl, 1 mL/L Nonidet P40, and 0.5 g/L sodium deoxycholate) and then with 1 mL of low-salt buffer (10 mmol/L Tris-HCl, pH 7.5, 1 mL/L Nonidet P40, and 0.5 g/L sodium deoxycholate). Finally, we suspended the pellets in a buffer containing SDS and proteinase K and incubated the suspension for 30 min at 56 °C. We extracted the DNA with the phenol/chloroform method, precipitated the DNA with 0.3 mol/L sodium acetate and 2.5 volumes of absolute ethanol at −20 °C for 4 h, and centrifuged at 13 000g for 30 min. The purified DNA was suspended in 20 μL PCR-grade water.

To address the specificity of histone-methylation detection, we investigated the differences in H3K9me1 concentration between 2 overlapping ALU fragments, ALU115 and ALU247 (Fig. 1⇓ ), which represent apoptotic mononucleosomes and larger oligonucleosomes, respectively. The single-copy, transcriptionally active CDKN2A gene [cyclin-dependent kinase inhibitor 2A (melanoma, p16, inhibits CDK4)]2 was used as a control. We amplified and quantified ALU115 and ALU47 fragments from precipitated DNA as previously described (26). We amplified the CDKN2A gene with the primers previously reported (28) and quantified CDKN2A sequence as described for ALU fragments. We used the mean values of 3 experiments to indicate the H3K9me1 concentration for each sequence.

• Download figure
• Open in new tab

Figure 1.

Map of ALU sequences.

To demonstrate the specificity of histone-methylation detection, we compared 2 overlapping fragments of 115 bp and 247 bp (ALU115 and ALU247) that represent mono- and oligonucleosomes, respectively, with respect to their enrichment in ChIP. Adapted from (35).

We considered as positive samples that had entered the logarithmic phase of the amplification curve during the LightCycler PCR before cycle 41. After amplification, melting-curve analyses indicated melting peaks at approximately 84 °C for ALU115, 86 °C for ALU247, and 85 °C for CDKN2A. Finally, we electrophoresed LightCycler PCR products on a 20-g/L agarose gel stained with ethidium bromide to confirm the presence of the 115-, 247-, and 150-bp amplicons for the ALU115, ALU247, and CDKN2A sequences, respectively.

statistical analyses

Quantitative differences in H3K9me1 concentration between ALU115, ALU247, and CDKN2A were evaluated with the Mann–Whitney U-test. We evaluated quantitative correlations between the variables (e.g., histone methylation, nucleosome ELISA, or free plasma DNA) with the Spearman rank correlation test. P values <0.05 were considered statistically significant.

the concentrations of nucleosomes and free alu-specific dna are correlated

Before H3K9me1 analysis, we demonstrated the presence of circulating histones and DNA in plasma. We quantified nucleosomes by ELISA and analyzed circulating DNA by real-time PCR to quantify ALU repeat–specific sequences. Circulating nucleosomes were present in all samples from the multiple myeloma patients, but there were substantial interpatient differences (Table 1⇓ and Fig. 2A⇓ ). The difference between the lowest and highest values was 43.6-fold, indicating variation in the concentrations of circulating apoptotic nucleosomes in plasma. Interindividual variation in the total free DNA in the plasma was much higher (101.3-fold), with a median of 44.3 μg/L and a mean of 259.3 μg/L (Table 1⇓ and Fig. 2B⇓ ). Samples with lower or higher nucleosome concentrations also displayed low or high concentrations, respectively, of ALU-specific DNA fragments, suggesting that these 2 variables were correlated. A statistical analysis revealed a highly significant positive correlation between the nucleosome amount and total DNA (P < 0.001, Spearman rank test).

• Download figure
• Open in new tab

Figure 2.

Free nucleosome and DNA concentrations in plasma.

(A), The relative concentrations of histone-associated DNA fragments (mono- and oligonucleosomes) were measured with an ELISA kit for cell-death detection. A positive signal indicates the presence of both histones and associated DNA in the circulation. (B), The presence of free DNA in the plasma was analyzed by amplifying the ALU115 fragment via real-time PCR. Each patient value is the mean of 3 experiments. The graphs display the lowest and highest concentrations, the median (broken line), and 95% confidence intervals (boxes).

volume-dependent detection of histone methylation in plasma

After demonstrating that histones and intact DNA are present in the circulation, we investigated the presence of histone methylation in plasma. Because plasma samples may have different H3K9me1 concentrations, we used a 200-μL volume of a plasma mixture to demonstrate the characteristics of histone-methylation detection. In this primary analysis, we analyzed for the presence of H3K9me1 on the ALU115 fragment because this fragment covers both mono- and oligonucleosomes. The ALU115 fragment was amplified from ChIP-precipitated DNA by real-time PCR. With the use of positive and negative controls, we demonstrated the presence of H3K9me1-specific signals, which were validated by the amplification curves, the specific melting peaks at approximately 84 °C, and the presence of the 115-bp fragment on agarose gels. To demonstrate volume dependence and linearity of detection, we set up a plasma-dilution series. With decreasing amounts of plasma (200, 20, 2, 0.2, and 0 μL), H3K9me1 was linearly detectable to 2 μL (r = −0.98).

h3k9me1 detection is sequence dependent

In subsequent analyses, we addressed the issue of the specificity of histone-methylation detection in plasma by comparing 2 overlapping fragments from ALU elements for their enrichment by the anti-H3K9me1 antibody (Fig. 1⇑ ). The ALU115 signal represents DNA bound to mono- and oligonucleosomes, and the ALU247 signal indicates DNA only from oligonucleosomes. H3K9me1 was detected on ALU115 in 17 (81%) of 21 samples and on ALU247 in only 47.6% of the samples. More importantly, ALU115 showed significantly higher concentrations of antibody-precipitated DNA than ALU247 (Table 1⇑ and Fig. 3⇓ ; P = 0.004). This result is in accord with our observation that the concentration of free mononucleosomal DNA in plasma is substantially higher than that of oligonucleosomes (data not shown). Our findings show that the 2 overlapping multiple-copy sequences that differ only in nucleosomal size can have different detection rates and amounts of histone methylation, indicating that the detection of histone methylation is sequence specific.

• Download figure
• Open in new tab

Figure 3.

Sequence-dependent detection of histone methylation.

The concentrations of H3K9me1 on ALU115, ALU247, and CDKN2A gene fragments were measured. Data represent the mean of 3 experiments. Statistically significant differences are indicated. Shown are lowest and highest concentrations, the median, and 95% CIs (boxes).

A comparison of H3K9me1 detection for ALU115 and the single-copy CDKN2A sequence provided further demonstration of the sequence dependence of histone-methylation detection. We detected H3K9me1 at the CDKN2A promoter at a rate of only 33.3% (Table 1⇑ ) and found, as expected, that ALU115 and CDKN2A showed highly significant differences in H3K9me1 concentrations among samples (P < 0.001, Mann–Whitney U-test; Fig. 3⇑ ). Again, these results indicate that the detection of H39me1 in plasma is sequence dependent.

On the other hand, we found no correlation between H3K9me1 concentration and either nucleosome concentration (P = 0.17, Spearman rank test) or total DNA (P = 0.15), suggesting that the detection of H3K9me1 is independent of the concentration of circulating nucleosomes.