Proteomic profiling identifies specific histone species associated with leukemic and cancer cells
© Singh et al. 2015
Received: 7 January 2015
Accepted: 12 August 2015
Published: 27 August 2015
Chromatin is an extraordinarily complex structure. Much of this complexity results from the presence of numerous histone post-translational modifications and histone variants. Alterations in the patterns of histone post-translational modifications are emerging as a feature of many types of cancer and have been shown to have prognostic value.
We have applied a liquid chromatography/mass spectrometry-based approach to comprehensively characterize the histone proteome in primary samples from chronic lymphocytic leukemia (CLL) patients, as well as bladder and breast cancer cell culture models. When compared to non-malignant CD19+ B cells from healthy donors, the CLL histone proteome showed a distinct signature of differentially expressed species, spanning all the histones studied and including both post-translationally modified species and unmodified, non-allelic replication-dependent histone isoforms. However, the large changes in histone H3 and H4 that are characteristic of many cancer types were not observed. One of species of H2A (mass = 14,063 Da) was the most strongly associated with time to treatment in CLL patients. CLL patient samples also demonstrated histone profiles that were distinct from those of the bladder and breast cancer cells.
Signatures of histone profiles are complex and can distinguish between healthy individuals and CLL patients and may provide prognostic markers. In addition, histone profiles may define tissue specific malignancies.
KeywordsHistone Post-translational modification Chromatin Chronic lymphocytic leukemia
The packaging of eukaryotic genomes with histones to form chromatin is essential for the necessary condensation and protection of DNA. Given the central role of histones in chromatin, there have been a number of studies that have examined histone proteins to identify alterations, particularly post-translational modifications, associated with cancer. Indeed, several of the post-translational modifications may have clinical utility as prognostic markers. For example, decreased levels of histone H4 lysine 16 acetylation (H4 K16Ac) and H4 lysine 20 trimethylation (H4 K20me3) are found in a number of human cancers, while low global levels of histone H3 lysine 9 dimethylation (H3 K9me2) and H3 lysine 18 acetylation (H3 K18Ac) predict poorer outcome in prostate, lung, kidney and pancreatic cancer [1–9].
The intricate mechanisms of regulation that involve chromatin are largely based on the enormous complexity built into this structure. While histone post-translational modifications are clearly an important source of complexity, variations in the primary sequences of specific histone proteins are also a prime contributor. There are two basic groups of histone proteins, the replication-dependent histones and the replication-independent histone variants. The replication-dependent histones are encoded by multi-gene families that are found in several distinct clusters in the human genome. The expression of these histone genes is tightly coupled to DNA replication and they serve to package DNA into chromatin during DNA replication. The replication-independent histone variants, such as histone H3.3 and histone H2AZ, are typically expressed constitutively throughout the cell cycle and are found as single genes dispersed throughout the genome. Histone variants can play an important role in carcinogenesis, where driver mutations in histone H3.3 have been identified in pediatric glioblastoma [10–12].
While primary sequence diversity in histones is often thought to be the result of the incorporation of histone variants, the multiple replication-dependent histone genes do not all encode identical proteins [13, 14]. For example, there are 16 genes that code for replication-dependent histone H2A and they produce 11 distinct polypeptides. These histone species will be referred to as histone isoforms to distinguish them from the more familiar histone variants. The replication-dependent histone isoforms typically vary from each other by a small number of amino acids. This high degree of identity makes them challenging to study using typical techniques that would require the generation of isoform-specific antibodies. However, they can be readily resolved by LC/MS due to the change in mass (see Additional file 1: Tables S1–S4). The clinical importance of histone isoforms is highlighted by recent reports that identified alterations in specific histone H2A isoforms in B cells isolated from patients with chronic lymphocytic leukemia (CLL) and in estrogen receptor positive breast cancer tissues [15–17].
We have used LC/MS to quantify the relative abundance of every detectable histone species in a large set of primary tumor samples. By comparing the histone profiles between normal B cells and B cells isolated form CLL patients, we have identified a number of specific histone species with levels significantly altered in the CLL cells. In addition, histone profiles were correlated with clinical data from CLL patients to identify species that may warrant further study. The most promising was a specific histone H2A species (14,063 MW) where expression was indicative of a shortened time to treatment. We have also comprehensively analyzed histone profiles from cell culture models of bladder and breast cancer to determine whether histone alterations are specific to CLL or whether they are a more general feature of cancer cells. Interestingly, our analyses indicated that significantly decreased levels of histone H4 lysine 16 acetylation are not a ubiquitous property of cancer cells. Our studies highlight the enormous heterogeneity of histone isoforms present in human cells and indicate the importance of quantifying the histone proteome to understand the impact of chromatin structure on carcinogenesis.
Histones were isolated from CLL patient samples (n = 87) or CD19+ cells from healthy volunteers (n = 5) by acid extraction. The extracted histones were analyzed by LC–MS as described in the methods section. For each core histone peak eluting from the LC, we selected a specific mass range that encompassed all of its isoforms in an unmodified state as well as the possible spectrum of modifications that could occur. We then tested for differential expression by determining the relative abundance of each molecular species in that mass range between CLL patient samples and normal donors as well as among varying states of disease aggression in bladder and breast cell lines. An overview of our methodology has been shown in Additional file 2: Figure S1.
Global changes in histone H2A species in leukemic and cancer cells
Chronic lymphocytic leukemia
The most significantly altered isoform of H2A had a molecular mass of 14,063 Da (Fig. 1c). This mass is compatible with post-translationally modified versions of several H2A isoforms. For example, this mass is consistent with the H2A 1H Me3Ac3Phos1, H2A 2C Me3Ac1Phos1 or H2A 2A Me1Ac1. This species was easily detectable in all of the healthy control samples (Fig. 1c), but the CLL patient samples appeared to represent two distinct groups: one with the majority of patients who had completely undetectable levels and a second, small group with detectable levels. Intriguingly, presence of this species was associated with shorter time to treatment and was the only species with FDR adjusted p-value <0.10 (Fig. 1d). It is unclear why higher expression that is closer to normal levels would be deleterious and needs further study.
Bladder and breast cancer models
To determine which changes in histone profiles observed in CLL patients are disease- or tissue-specific and which might be more generally indicative of a cancerous state, we characterized the complement of histone proteins in two solid tumor model systems. The first is a set of cell lines that represent a range of histopathologic subtypes of bladder cancer. The cells that we have examined are normal bladder epithelium, normal bladder epithelial cells that have been immortalized with hTERT, a cell line derived from a non-invasive bladder cancer (RT4) and cell lines derived from invasive bladder cancers (T24 and UMUC3) . Even though none of the histone species met our criteria for statistical significance, we did observe that levels of H2A 1C dropped when the normal bladder epithelial cells became immortalized with hTERT, consistent with the recent report that decreased levels of H2A 1C expression resulted in increased proliferation of cells in vitro .
Global changes in histone H2B species in leukemic and cancer cells
Chronic lymphocytic leukemia
Bladder and breast cancer models
Global changes in histone H3 species in leukemic and cancer cells
Histone H3 Peak 1
Chronic lymphocytic leukemia
A representative spectrum of peak 1 histone H3 is shown in Fig. 5a. We observed down-regulation of several peak 1 histone H3 species in the CLL tissue samples (Table 1). Intriguingly, while changes in H3 acetylation and methylation have been linked to a number of cancers, the most significantly down-regulated species in peak 1 were also phosphorylated (Fig. 5b, c) [1–3, 5–7, 21].
Bladder and breast cancer models
Comparison of the peak 1 histone H3 profiles in the series of bladder cancer cell lines indicated that these histones were largely unchanged. The only species that increased in abundance with cancer aggression were unmodified and monomethylated versions of H3.2. However their adjusted p-values did not reach statistical significance. A few of the peak 1 histone H3 species changed in the breast cancer cell lines. The most significantly altered species was a decreased level of the mono-methylated, mono-phosphorylated form of the lone histone H3 isoform found in histone cluster 3 (H3/3 or H3.1t, Table 2). Intriguingly, the abundance of this species, which was originally thought to be testis-specific, was also significantly down-regulated in the CLL patient samples (Table 1) [22, 23]. Some of the species that are altered in bladder and breast cancer are shown in Fig. 5d–f.
Histone H3 Peak 2
Chronic lymphocytic leukemia
The peak 2 population of histone H3 contained an entirely different complement of species and, based solely on the masses observed, there is no obvious explanation of why this pool of molecules has a distinct LC elution profile (Fig. 5g). When comparing peak 2 H3 from healthy and CLL B cells, many species showed a statistically significant change in abundance (Table 1). Notably, many phosphorylated forms of H3.1t with differing degrees of methylation were altered.
Bladder and breast cancer models
As observed with peak 1, the histone H3 pool from peak 2 was also not largely different in either bladder or breast cancer cells. The only species that showed a statistically significant change in breast cancer cells were the highly modified forms of H3.1 (Table 2). In the bladder cancer cell lines, only the ummodifed form of histone H3.2 showed an increased level in aggressive cell lines, though it did not reach statistical significance (Additional file 1: Table S6a).
Global changes in histone H4 species in leukemic and cancer cells
Chronic lymphocytic leukemia
Among the core histones, histone H4 contains the least variation. There are no replication-independent forms of histone H4 and of the 14 replication-dependent H4 genes, 12 encode identical protein products. The other two genes encode minor, but detectable, forms of H4 (Additional file 1: Table S4). Similar to the situation with H3, comparison of healthy and CLL B cells showed that H4 levels were mostly unchanged, with 11 of the low abundance H4 species showing a statistically significant change (Table 1).
Bladder and breast cancer models
Previous global analyses of histones in cancer have focused on the post-translational modification patterns of histones H3 and H4. These studies have identified several specific modifications that strongly correlate with tumorigenesis. In particular, decreases in the levels of H4 lysine 16 acetylation and lysine 20 methylation and H3 lysine 18 acetylation, lysine 4 methylation and lysine 9 methylation have been seen in a large number of tumor types and tumor cell models [1, 5, 6, 8, 21, 24–27]. However, we have found that the abundance of the post-translationally modified forms of H3 and H4 are quite stable across a large number of CLL patient samples with most of the changes limited to relatively low abundance species. This suggests that alterations of H3 and H4 are not universally hallmarks of malignant transformation.
The paucity of changes in high abundance species observed in CLL patient samples may have a number of explanations. From a biological perspective, CLL may be distinct from the limited range of other malignancies that have been explored, such as prostate and lung cancer. There may also be technical reasons that could prevent our analyses from detecting such changes. Our study has relied on mass spectrometry rather than immunological methods. This allows us to measure the relative abundance of a specifically modified histone species but does not allow us to localize the modifications to a specific residue. Hence, if there were a decrease in the level of H4 lysine 16 acetylation in the CLL cells but a concomitant increase in the acetylation of another lysine residue on H4, there would be no net change in acetylation detected.
There was a more dramatic change in the profile of histone H4 in the breast cancer cell lines. There was a much higher level of the 11,349 Da species in the M4 cells relative to the M1 cells. Based on this molecular weight, this species is likely to be the di-methylated/mono-acetylated form of H4. Based on previous analyses, the majority of this species typically contains H4 lysine 20 dimethylation and lysine 16 acetylation . Progressive increases in this methylated/acetylated species of H4 across the spectrum of aggressiveness of this series of breast cancer cell lines are contrary to immunohistochemical analysis of breast tumor samples . However, while we appreciate that no cell culture system fully recapitulates the complexity of human disease and that there are differences in the information obtained between mass spectrometry and immunohistochemical techniques, these differences might give us insight into the behavior of different subtypes of breast cancer.
One of the most important reasons for the use of mass spectrometry to characterize the histone proteome is that it is an unbiased approach that does not rely on the existence of modification-specific antibodies. As there are relatively few immunological reagents specific for the characterization of post-translational modifications on histones H2A and H2B, our analysis is the most comprehensive study of these histones in cancer, to date. Our results indicate that the patterns of histones H2A and H2B found in both primary tumor tissue samples and cancer cell line models are highly dynamic. These changes were observed both between healthy and cancer cells and between different types of cancer cells. For example, the pattern of H2B observed in B cells was very different from that observed in the solid tumor cell lines. This suggests the possibility that there may be tissue specific variations in the patterns of histone H2A and H2B species.
The dynamic nature of histone H2A is not limited to its pattern of post-translational modifications. We also detect a wide range of variation in the abundance of specific replication-dependent isoforms of histone H2A. While the functional analysis of these distinct replication-dependent H2A isoforms is only beginning, the observation that their abundances are altered in primary tumor tissue supports the intriguing possibility that they may encode functionally distinct molecules [16, 17].
The most clinically relevant species in the histone complement of CLL cells may be that with a molecular mass of 14,063 Da. The CLL patient samples could be divided into two groups based on the presence or absence of this species. Those patients with a detectable level of this species had a significantly shorter time to treatment; its prognostic relevance will need to be verified in an independent dataset. Additionally, identification of the molecular make-up of this species will be needed to determine whether this form of histone H2A plays a functional role in disease progression.
We have performed a comprehensive proteomic analysis of the core histones from CD19+ B cells from healthy individuals and the malignant counterpart in CLL patient samples as well as from bladder and breast cancer cell line models. Using LC/MS, we have quantified the relative abundance of every detectable histone peak to identify histone species whose abundance correlates with the presence or severity of disease. Our results identified interesting candidates whose levels show a significant correlation to disease aggression and can be explored further. In addition, we demonstrated that the abundance of the vast majority of histone species do not significantly change in primary tumor samples and cell model systems suggesting that at least some malignancies are not accompanied by large-scale alterations in the global histone proteome.
Cell lines and culture conditions
Peripheral blood was obtained from patients diagnosed with CLL by NCI 1996 criteria (Ref: Cheson et al., Blood. 1996; 87(12):4990–7) or from healthy individuals under protocols approved by the Ohio State University Institutional Review Board, according to the Declaration of Helsinki. Clinical status of CLL patients varied, but all CLL samples contained elevated lymphocyte count (>30,000 CD19+ cells/uL). CD19+ cells were obtained by negative selection using reagents from StemCell Technologies (Vancouver BC). M1–M4 cells were a gift from Dr. Tsonwin Hai. The cells were grown in DMEM/F-12 with 5 % horse serum (Invitrogen), 0.029 M sodium bicarbonate (Sigma), 10 mM HEPES (Sigma), 10 µg/ml insulin (Sigma), 10 ng/ml EGF (Millipore), 0.5 µg/ml hydrocortisone (Sigma), 100 ng/ml cholera toxin (Calbiochem) and 1 % penicillin/streptomycin (Sigma). hTERT cells are immortalized breast epithelial cells, RT4 cells are transformed but non-malignant, T24 cells are malignant but non-metastatic and UM-UC-3 cells are malignant and highly metastatic. The cells were grown in DMEM, supplemented with 10 % FBS and 1 % penicillin/streptomycin (Sigma). All the cell types as described above were incubated at 37 °C in a humidified atmosphere with 5 % CO2 and 95 % air.
Histones were prepared with standard acid extraction procedure as described . Extracted histones were subjected to LC–MS analysis. Characterization was performed by HPLC separation (Dionex, Waltham, MA) in line with either a MicroMass Q-TOF (MicroMass, Milford, MA) or AmaZon ETD (Burker, Billerica, MA) mass spectrometers.
For the Q-TOF MS analysis, approximately 20 μg of extracted histones were HPLC separated on a Dionex Ultimate 3000 capillary/nano HPLC (Dionex, Waltham, MA) fitted with a 1.0 × 150 mm C18 column (Discovery Bio wide pore C18 column, 5 μm, 300 Å, Supelco, USA) as described previously by Wang et al. . Mobile phase A contained 0.05 % TFA (aq) (Pierce, Rockford, IL). Mobile phase B contained 0.05 % TFA in acetonitrile (EMD Millipore, Billerica, MA). Initial conditions were 20 % B, with linear increases to 30 % at 2 min, 35 % at 10 min, 50 % B at 30 min, 60 % at 35 min and 95 % at 36 min. The column was washed at 95 % B for 4 min. Column equilibration was conducted at the initial conditions for 15 min. Total run time was 55 min. Blank injections of the same gradient were made between each sample injection. The mass range (500-2500 m/z) was scanned every 0.6 s in positive ion mode. A representative raw spectrum for each histone variant has been provided in Additional file 2: Fig. S2. Savitzky Golay smoothing (smooth window ± 3 channels, 2 cycles), mass deconvolution (MaxEnt algorithm, 1.0 Da/channel resolution, 0.750 Da uniform Gaussian width at half height damage model, 33 % left right minimum intensity ratios, iterate to convergence) and mass analysis were conducted using the MassLynx software 4.0 (Waters Corp., Milford, MA).
For AmaZon ETD analysis, approximately 100 ng of extracted histones were subjected to a modified LC–MS method described by You et al. . HPLC separation was performed on a Dionex Ultimate 3000 capillary/nano HPLC (Dionex, Waltham, MA) fitted with a 0.3 × 150 mm Magic C8 column (5 μm, 300 Å, Michrom Bioresources, Auburn, CA). Mobile phases were HPLC water (J.T. Baker, Center Valley, PA) and acetonitrile (EMD Millipore, Billerica, MA) each supplemented with 0.5 % formic acid (v/v, Acros Organics, Waltham, MA). Initial conditions were 20 % B held for 5 min, the gradient increased to 30 % B at 25 min and 35 % B at 65 min each with a convex curvature of 2. From 65 to 78 min, the gradient increased from 35 to 48 % B with a convex curvature of 4. Column equilibration at the initial conditions was conducted 12 min. Total run time was 90 min. Blank injections of the same gradient were made between each sample injection. A 500–2500 m/z mass range was scanned every 0.2 s in positive ion mode. A representative raw spectrum for each histone variant has been provided in Additional file 2: Fig. S3. Savitzky Golay smoothing (3 m/z width, 2 smooth cycles), Maximum Entropy deconvolution (data point spacing auto m/z, instrument peak width 0.3, resolution normal) and data analysis was conducted using the Bruker ESI Compass Data Analysis 1.3 software for AmaZon (ver 4.0 SP4, Bruker, Billerica, MA).
For quantitation of an isoform, the peak intensities of all the isoforms corresponding to a particular histone were pooled. The abundance of a particular isoform was then calculated relative to the pooled intensity of all detectable forms of that histone. A given peak was therefore quantitated relative to the total amount of histone present in the sample. The same approach was followed for the results derived from both the Q-Tof and Amazon mass spectrometers.
For each histone, the expression for a particular species was measured relative to the total expression across all species. The nonparametric Wilcoxon rank sum test was used to compare (1) the proportion expressed between normal donor and CLL patient samples and (2) the proportion expressed between patients samples with CLL and Zap70 positive (>20 %) versus negative protein. In these patients, time to treatment was measured from the date of diagnosis until the date of first treatment, censoring those who had not yet started treatment at the date of last follow-up. Associations between expression of a species (using median expression to group patients as high and low expressers) and time to treatment were evaluated using the score test from proportional hazards models; estimates of hazard ratios with confidence intervals were estimated from the model. Difference in time to treatment between high and low expressers for a particular species is shown graphically in a Kaplan–Meier plot. All tests are 2-sided and statistical significance was declared when the false discovery rate (FDR) adjusted p-value was less than 0.05, allowing for 5 % of the significant tests to be false positives. If none of the variants are statistically significant when applying the multiple testing corrections then it may be possible that the findings are false positives. However, to provide complete information, all the isoforms have been ordered for each analysis from smallest to largest p-value without any correction, with results corresponding to the smallest p-value as most reliable. The results have been sorted as separate tables based on if the isoforms were significant at least at the raw p-value levels (Additional file 1: Tables S5a, S6a and S7a) or non-significant even at raw p-value levels (Additional file 1: Tables S5b, S6b and S7b). A few of the isoforms had an FDR < 0.05, but were unidentified and therefore omitted from the study. The differences in expression across breast and bladder cell lines with varying degrees of aggressiveness were screened using one-way ANOVA. The cell lines were analyzed in triplicates and error bars represent the standard error of mean between the samples. All statistical analyses were conducted using the SAS (University Edition) statistical software package (SAS Institute Inc., Cary, NC).
Western Blotting: The M1–M4 cell lines were grown as described above. The cells were seeded in equal numbers, harvested after 48 h and histone preparation was carried out as described elsewhere (ref 29). The histones were quantitated using Bradford assay and about 20 μg of protein was used for western analysis. All the antibodies used in the analysis were purchased commercially (Abcam, Cambridge, MA). Briefly, the samples were loaded on 12.5 % polyacrylamide gels, transferred to PVDF membranes and Ponceau stained. The images were taken to confirm the equal loading followed by blocking in 5 % skimmed milk. The blots were the incubated in respective primary and HRP-conjugated secondary antibodies (G E Healthcare, UK) in 1:1000 and 1:5000 dilutions respectively. After each incubation, the blots were washed three times with TBST and finally developed using Pierce™ ECL Western Blotting Substrate (Life Technologies). The blots were then exposed to X-ray film for 60 s (Denville Scientific Inc, South Plainfield, NJ). The process was repeated three times using different passages of cell lines to ensure reproducibility.
RS isolated histones from patient and tissue culture samples, analyzed and quantitated mass spectra, designed experiments and drafted the manuscript. SWH performed the mass spectrometry. ASR performed statistical analyses. AM provided bladder cancer cells and helped design experiments. DML provided patient samples, purified healthy donor cells and helped design experiments. JMT-H prepared tissue culture cells for analysis. SKC provided bladder cancer cells, helped design experiments and helped draft the manuscript. JCB provided patient samples and helped design experiments. MAF supervised the mass spectrometry, analyzed data and helped design experiments. MRP helped design experiments, analyzed data and helped draft the manuscript. All authors read and approved the final manuscript.
We would like to thank Dr. Tsonwin Hai and the Barbara Ann Karmanos Cancer Institute (Detroit, MI, USA) for cell lines. This work was funded by grants from the National Institutes of Health [P01 CA101956 to M.R.P. and J.C.B., R01 062970 to M.R.P. and CA107106 to M.A.F.]; a Specialized Center of Research award from the Leukemia and Lymphoma Society (to J.C.B.); and The Ohio State University Comprehensive Cancer Center Molecular Carcinogenesis and Chemoprevention Program (to S.K.C. and A.M.). The authors declare that they have no competing financial interests.
Compliance with ethical guidelines
Competing interests The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Mosashvilli D, et al. Global histone acetylation levels: prognostic relevance in patients with renal cell carcinoma. Cancer Sci. 2010;101(12):2664–9.View ArticlePubMedGoogle Scholar
- Seligson DB, et al. Global levels of histone modifications predict prognosis in different cancers. Am J Pathol. 2009;174(5):1619–28.PubMed CentralView ArticlePubMedGoogle Scholar
- Seligson DB, et al. Global histone modification patterns predict risk of prostate cancer recurrence. Nature. 2005;435(7046):1262–6.View ArticlePubMedGoogle Scholar
- Fraga MF, et al. Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human cancer. Nat Genet. 2005;37(4):391–400.View ArticlePubMedGoogle Scholar
- Ellinger J, et al. Global levels of histone modifications predict prostate cancer recurrence. Prostate. 2010;70(1):61–9.View ArticlePubMedGoogle Scholar
- Elsheikh SE, et al. Global histone modifications in breast cancer correlate with tumor phenotypes, prognostic factors, and patient outcome. Cancer Res. 2009;69(9):3802–9.View ArticlePubMedGoogle Scholar
- Barlesi F, et al. Global histone modifications predict prognosis of resected non small-cell lung cancer. J Clin Oncol. 2007;25(28):4358–64.View ArticlePubMedGoogle Scholar
- Schneider AC, et al. Global histone H4K20 trimethylation predicts cancer-specific survival in patients with muscle-invasive bladder cancer. BJU Int. 2011;108(8 Pt 2):E290–6.View ArticlePubMedGoogle Scholar
- Waldmann T, Schneider R. Targeting histone modifications–epigenetics in cancer. Curr Opin Cell Biol. 2013;25(2):184–9.View ArticlePubMedGoogle Scholar
- Schwartzentruber J, et al. Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma. Nature. 2012;482(7384):226–31.View ArticlePubMedGoogle Scholar
- Khuong-Quang DA, et al. K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas. Acta Neuropathol. 2012;124(3):439–47.PubMed CentralView ArticlePubMedGoogle Scholar
- Lewis PW, et al. Inhibition of PRC2 activity by a gain-of-function H3 mutation found in pediatric glioblastoma. Science. 2013;340(6134):857–61.PubMed CentralView ArticlePubMedGoogle Scholar
- Albig W, et al. The human H2A and H2B histone gene complement. Biol Chem. 1999;380(1):7–18.View ArticlePubMedGoogle Scholar
- Marzluff WF, et al. The human and mouse replication-dependent histone genes. Genomics. 2002;80(5):487–98.View ArticlePubMedGoogle Scholar
- Su X, et al. Validation of an LC-MS based approach for profiling histones in chronic lymphocytic leukemia. Proteomics. 2009;9(5):1197–206.PubMed CentralView ArticlePubMedGoogle Scholar
- Su CH, et al. An H2A histone isotype regulates estrogen receptor target genes by mediating enhancer-promoter-3′-UTR interactions in breast cancer cells. Nucleic Acids Res. 2014;42(5):3073–88.PubMed CentralView ArticlePubMedGoogle Scholar
- Singh R, et al. Increasing the complexity of chromatin: functionally distinct roles for replication-dependent histone H2A isoforms in cell proliferation and carcinogenesis. Nucleic Acids Res. 2013;41(20):9284–95.PubMed CentralView ArticlePubMedGoogle Scholar
- Telu KH, et al. Alterations of histone H1 phosphorylation during bladder carcinogenesis. J Proteome Res. 2013;12(7):3317–26.PubMed CentralView ArticlePubMedGoogle Scholar
- Tang B, et al. TGF-beta switches from tumor suppressor to prometastatic factor in a model of breast cancer progression. J Clin Invest. 2003;112(7):1116–24.PubMed CentralView ArticlePubMedGoogle Scholar
- Santner SJ, et al. Malignant MCF10CA1 cell lines derived from premalignant human breast epithelial MCF10AT cells. Breast Cancer Res Treat. 2001;65(2):101–10.View ArticlePubMedGoogle Scholar
- Ellinger J, et al. Global histone H3K27 methylation levels are different in localized and metastatic prostate cancer. Cancer Invest. 2012;30(2):92–7.View ArticlePubMedGoogle Scholar
- Witt O, Albig W, Doenecke D. Testis-specific expression of a novel human H3 histone gene. Exp Cell Res. 1996;229(2):301–6.View ArticlePubMedGoogle Scholar
- Andersen JS, et al. Nucleolar proteome dynamics. Nature. 2005;433(7021):77–83.View ArticlePubMedGoogle Scholar
- Ellinger J, et al. Prognostic relevance of global histone H3 lysine 4 (H3K4) methylation in renal cell carcinoma. Int J Cancer. 2010;127(10):2360–6.View ArticlePubMedGoogle Scholar
- Rogenhofer S, et al. Global histone H3 lysine 27 (H3K27) methylation levels and their prognostic relevance in renal cell carcinoma. BJU Int. 2012;109(3):459–65.View ArticlePubMedGoogle Scholar
- Behbahani TE, et al. Alterations of global histone H4K20 methylation during prostate carcinogenesis. BMC Urol. 2012;12:5.PubMed CentralView ArticlePubMedGoogle Scholar
- Rogenhofer S, et al. Decreased levels of histone H3K9me1 indicate poor prognosis in patients with renal cell carcinoma. Anticancer Res. 2012;32(3):879–86.PubMedGoogle Scholar
- Knapp AR, et al. Quantitative profiling of histone post-translational modifications by stable isotope labeling. Methods. 2007;41(3):312–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Wang L, et al. Assaying pharmacodynamic endpoints with targeted therapy: flavopiridol and 17AAG induced dephosphorylation of histone H1.5 in acute myeloid leukemia. Proteomics. 2010;10(23):4281–92.PubMed CentralView ArticlePubMedGoogle Scholar
- You J, et al. High-sensitivity TFA-free LC-MS for profiling histones. Proteomics. 2011;11(16):3326–34.PubMed CentralView ArticlePubMedGoogle Scholar