SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

Microvascular invasion (MiVI) is a major risk factor in postoperative tumor recurrence and mortality in hepatocellular carcinoma (HCC). Unfortunately, this histological feature is usually missed by liver biopsy because of limited sampling, and MiVI is commonly detected only after surgery and examination of the full resected specimen. To date, there exists no reliable tool for identifying MiVI prior to surgical procedures. This study aimed to compare the proteome of HCC with and without MiVI in order to identify surrogate biomarkers of MiVI. A training cohort comprising surgically resected primary HCC with MiVI (n = 30) and without MiVI (n = 26) was subjected to matrix-assisted laser desorption ionization imaging mass spectrometry (MALDI IMS). Comparative analysis of acquired mass spectra of the two groups yielded 30 differential protein peaks, among which 28 were more strongly expressed in HCC with MiVI. Among these, two peaks were identified as N-term acetylated histone H4 dimethylated at lysine (K) 20, and N-term acetylated histone H4 dimethylated at K20 and acetylated at K16. Both peaks were validated in the training cohort and in an independent validation cohort (n = 23) by immunohistochemistry and western blot. Conclusion: These results demonstrate the potential of MALDI IMS for uncovering new relevant biomarkers of MiVI in HCC, and highlight the role of epigenetic modifications in the prognosis of HCC. Preoperative detection of modified forms of histone H4 expression in tumor biopsies would be helpful in management of patients with HCC. (Hepatology 2013;53:983–994)

Abbreviations
ACN

acetonitrile

H4K16ac

histone H4 acetylated at lysine 16

H4K20me2

histone H4 dimethylated at lysine 20

HCC

hepatocellular carcinoma

HCC/MiVI+

hepatocellular carcinoma with microvascular invasion

HCC/MiVI−

hepatocellular carcinoma without microvascular invasion

LT

liver transplantation

LTQ

linear trap quadrupole

K

lysine

MALDI IMS

matrix-assisted laser desorption ionization imaging mass spectrometry

MiVI

microvascular invasion

MS

mass spectrometry

MS/MS

tandem mass spectrometry

nano-LC-ESI-MS/MS

nano-liquid chromatography, electrospray ionization, tandem mass spectrometry

PTMs

posttranslational modifications

RP-micro-LC

reverse-phase micro-liquid chromatography

SA

sinapinic acid

TFA

trifluoroacetic acid

WB

western blot

Hepatocellular carcinoma (HCC) is the main primary malignancy of the liver, and ranks as the third most common cause of cancer-related death worldwide, despite significant efforts made in follow-up of cirrhosis patients and screening of HCC.[1] HCC prognosis remains poor, mainly related to a high rate of tumor recurrence following surgical treatment that attains 70% at 5 years after liver resection and 15%-30% after liver transplantation (LT).[2] Microvascular invasion (MiVI) is a major risk factor for recurrence and mortality in HCC.[3-6] Importantly, MiVI, defined as malignant cells in peritumoral vessels, can only be assessed after careful histological assessment of the whole surgical specimen. Thus, there is a critical need for blood or tissue surrogate markers able to predict MiVI before resection or LT. Such a marker might be helpful in assessing optimal therapeutic strategy. As MiVI reflects the aggressive potential of HCC, previous studies sought to identify molecular markers of MiVI in tumor tissue. Indeed, several gene expression signatures that correlate with vascular invasion have already been found.[7, 8] However, none of them are currently used in clinical practice. Indeed, applicability of this molecular technique to small biopsy specimens is one of the limits of this approach. In contrast, protein biomarkers seem to be more suitable candidates, as they can be more easily detected on routine biopsy samples by immunohistochemistry.

Proteomics, concerned with analysis of the entire set of proteins expressed in a biological sample (e.g., tissue or serum), fills the gap between the information encoded by the genome and cell function. Among analytical techniques of proteomic analysis, matrix-assisted laser desorption ionization imaging mass spectrometry (MALDI IMS) has emerged in the last decade as an attractive approach for combined morphologic and direct molecular tissue analysis.[9] Indeed, MALDI IMS can be used in combination with microscopy, enabling spatially resolved, unlabeled imaging of various peptides/proteins in their histologic context and allocation of molecular profiles to specific cell types, such as tumoral, preneoplastic, stromal and inflammatory cells.[10, 11] Interestingly, several studies have already shown the potential of MALDI IMS for identifying molecular markers associated with diagnosis or prognosis of various tumors, including HCC.[12-15] Using a MALDI IMS-based proteomic approach, we compared the tissue proteome of HCC samples with and without MiVI. We identified two modified forms of histone H4 as tissue biomarkers associated with MiVI and validated results in an independent series.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

Patients and Tissue Specimens

HCC developing in advanced chronic liver disease (fibrosis stage F3 or F4 according to the METAVIR classification) obtained from liver resections or LT were retrospectively retrieved from the pathological files of Beaujon Hospital (1994-2010). Among them, 56 cases of HCC without preoperative treatment and for which frozen samples were available were selected for MALDI IMS analysis (training cohort). Tissues were snap-frozen after surgery and stored at −80°C. Main clinical data were collected and pathological features were reevaluated (V.P., N.P.), especially for the presence of MiVI defined as the presence of tumor cells in a vessel located in the tumor capsule and/or in surrounding liver parenchyma. A mean of one paraffin block/tumor/cm and five paraffin blocks of peritumoral liver were reviewed for each case. Based on the presence of MiVI, two groups were defined: HCC without (HCC/MiVI−, n = 26) and HCC with MiVI (HCC/MiVI+, n = 30). HCC/MiVI− that showed satellite nodules were excluded. An independent cohort comprising 23 HCC samples was used for validation. Additionally, immunohistochemistry was performed in 13 HCC preoperative biopsy specimen (HCC/MiVI−, n = 8; HCC/MiVI+, n = 5). Ten cases were from the train/validation cohorts, and three additional cases fulfilling our initial inclusion criteria were selected. Study protocols were in conformity with the ethical guidelines of the 1975 Declaration of Helsinki and approved by the local Institutional Review Board and Ethical Committee. All subjects gave informed consent.

Sample Preparation for MALDI IMS

For each case, one representative block of frozen tumor sample (mean area 1 cm2) was selected for MALDI IMS. The block was cryosectioned into 10-μm thick sections. One section was stained with hematoxylin and eosin for histological examination to check the adequacy of the frozen material and the subsequent section was placed on conductive indium-tin-oxide-coated glass slides (Bruker Daltonics, Bremen, Germany), dried under a vacuum, briefly washed in 70% and 100% ethanol, and dried again. The section was then coated with the MALDI matrix (sinapinic acid at 10 mg/mL in water/acetonitrile 40:60 [v/v] with 0.2% trifluoroacetic acid) using the ImagePrep matrix application device (Bruker Daltonics) following the standard protocol. Acetonitrile (ACN), ethanol, and the MALDI matrix (sinapinic acid, SA) were purchased from Sigma-Aldrich (Saint-Quentin Fallavier, France). Trifluoroacetic acid (TFA) was purchased from Fisher Scientific (Illkirch, France).

MALDI IMS Experiments

MALDI IMS measurement was performed as previously described.[15] A comprehensive description of MALDI IMS workflow, including sample preparation, measurement, and data interpretation, can be found in the review by Balluff et al.[11]

Molecular Identification of Protein Peaks

For further molecular identification, protein extracts obtained from frozen HCC/MiVI+ samples were subjected to mass spectrometry (MS) and MS/MS analysis. Tissue extracts were prepared as previously described.[15]

Fractionation on Reverse-Phase Micro-Liquid Chromatography (RP-micro-LC)

Tissue extracts were further fractionated by RP-micro-LC using a microchromatography system (Ultimate 3000, Dionex, Thermo Fisher Scientific, Courtaboeuf, France). Fifteen microliters of the extract was loaded onto a C18 RP-microLC column (Acclaim PepMap300, 1.0 mm × 15 cm, particle size 5 μm, 300 Å; Dionex) at a flow rate of 50 μL/min. Bound proteins were eluted with a linear gradient of 4%-90% solvent B (B = ACN containing 0.1% TFA, A = 98% H2O containing 0.1% TFA, 2% ACN, v/v) over a period of 90 minutes. Eluate was collected with Probot (Dionex) on a 96-well microplate. Collection was performed on the whole plate every 30 seconds with a mean delay of 10 minutes. The fractions were evaporated in a vacuum centrifuge and resuspended in 10 μL of water. An aliquot of each fraction (1 μL) was withdrawn, mixed with 5 mg/mL α-cyano-4-hydroxycinnamic acid (LaserBio Labs, Sophia Antipolis, France) in 50:50:0.1 (ACN/water/TFA) and analyzed by MALDI-MS on an AutoFlex III MALDI-TOF/TOF (Bruker Daltonics). Fractions containing m/z of interest were further analyzed by nano-liquid chromatography, electrospray ionization, tandem mass spectrometry (nano-LC-ESI-MS/MS) using the bottom-up approach.

Enzymatic Digestion

For enzymatic digestion, the fraction of interest was transferred to a tube. Tryptic digestion (10 ng/μL trypsin [Sigma Aldrich], 50 mM NH4HCO3) and digestion with Asp-N (10 ng/μL Asp-N [Roche Applied Science, Meylan, France], 100 mM NH4HCO3) were performed overnight at 37°C. The digests were evaporated in a vacuum centrifuge and resuspended in 12 μL of water with 0.1% formic acid prior to nano-LC/ESI-MS/MS experiments.

MS and MS/MS Analyses of Digests

The analysis was performed using a nano-chromatography system (Easy nLC, Proxeon, Thermo Fisher Scientific, Courtaboeuf, France) connected to a linear trap quadrupole (LTQ) Velos Orbitrap (Thermo Fisher Scientific) mass spectrometer under the same chromatographic conditions as previously described.[15] Peptides were analyzed in the Orbitrap in full ion scan mode at a resolution of 30,000 and a mass range of 400-1,800 m/z. Fragments were obtained by collision-induced dissociation activation with collisional energy of 40% and an activation Q of 0.250 for 10 ms, and were analyzed in the LTQ. MS/MS data were acquired in a data-dependent mode in which the 20 most intense precursor ions were isolated, with dynamic exclusion of 20 seconds and exclusion mass width of 10 ppm. To improve sensitivity of the modified peptides, inclusion list-dependent acquisition on the Orbitrap mass spectrometer was performed in a second step. The same chromatographic and MS conditions as described above were used. The ion trap MS/MS max ion time was increased to 200 ms. Four scan events were acquired continuously on the four selected precursors targeted for MS/MS acquisition.

Data Processing

Data were processed with Proteome Discoverer 1.3 software (Thermo Fisher Scientific) coupled to an in-house Mascot search server (Matrix Science, Boston, MA; v. 2.3.02). The mass tolerance of fragment ions was set at 10 ppm for precursor ions and 0.6 Da for fragments. The maximum number of missed cleavages was limited to 2. MS-MS data were searched against the SwissProt database with human taxonomy and the appropriate enzyme. The following modifications were used as variable modifications: oxidation (M), phosphorylation (STY), acetylation (N-term, K), and dimethylation (K). A reversed database approach was used for false-discovery rate estimation. A threshold of 1% was chosen for this rate.

Western Blot and Immunohistochemistry

Standard methods for western blotting (WB) and immunohistochemical stainings are shown in the supporting Materials and Methods. For each case, an immunohistochemical semiquantitative score (0-300) was calculated (% of stained cells [0-100] × staining intensity [1 weak; 2 intermediate; 3 strong]), blinded to the presence/absence of MiVI.

Statistical Analyses

MALDI IMS Datasets

A Comparative proteomic analysis using cross-classification design was performed. We considered the two groups of MALDI IMS datasets representing HCC/MiVI− and HCC/MiVI+ and performed statistical comparison with the aim of establishing mass to charge (m/z)-values discriminative for HCC/MiVI+. For comparison, we developed the cross-classification design,[16] wherein each dataset from the first group (HCC/MiVI−) was first compared with all datasets from the second group (HCC/MiVI+); next, each dataset from the second group was compared with all datasets from the first group. In our study, 56 comparisons or cross-classifications of MALDI IMS datasets were performed. Each cross-classification was performed using the classification algorithm developed for discovery of biomarkers in serum protein profiles, which combines discrete wavelet transformation with support vector machines, and exploits statistically sound evaluation using double cross-validation.[17] For each cross-classification, the algorithm provides a list of 20 m/z-values most discriminative for HCC/MiVI+. After 56 cross-classifications, we selected those m/z-values found to be discriminative in at least 20 cross-classifications. The classification was performed using custom-made scripts in MatLab (MathWorks, Natick, MA). MALDI images showing spatial distribution of selected ions in the tissue sections were plotted and overlaid with optical images in the SCiLS Lab software (SCiLS, Bremen, Germany) after automatic hotspot removal and edge-preserving denoising applied for improved visualization.

Clinicopathological Characteristics, Immunohistochemistry, and WB Analysis

Continuous variables are expressed as mean ± standard deviation and were compared using the Student's or Mann-Whitney test when appropriate. Categorical variables are expressed in absolute values and percentages and were compared with Fisher's exact test. P < 0.05 was considered significant. Statistical analysis was carried out with BiostaTVG software.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information
Patients

Detailed clinicopathological characteristics of the training and validation cohorts are shown in Table 1. The two cohorts were comparable for all criteria except tumor size. HCC of the validation cohort were significantly smaller than in the training cohort (2.9 ± 1.5 cm versus 4.7 ± 3.1 cm, P = 0.002). In addition, for each cohort both groups (HCC/MiVI+ and HCC/MiVI−) were comparable for all criteria (age, gender, etiology, tumor size, and grade of differentiation), except for the number of tumors nodules in the validation cohort (Supporting Table 1). Satellite nodules were present in 36% (15/42) cases of HCC/MiVI+.

Table 1. Patients' Clinicopathological Characteristics
 Training Cohort (n = 56)Validation Cohort (n = 23)P Value
  1. SD, standard deviation; ns, not significant.

Age (years)   
Mean (± SD)58 (±12.5)61.5 (±10)ns
Gender   
Males, n (%)46 (82)19 (83)ns
Etiology of chronic liver disease, n (%)   
HCV22 (39)8 (35) 
HBV15 (27)5 (22)ns
Alcohol6 (11)2 (8.5) 
NASH3 (5)4 (17.5) 
Hemochromatosis1 (2) 
Wilson's disease1 (2) 
Mixed3 (5)2 (8.5) 
Undetermined5 (9)2 (8.5) 
Tumor size (cm)   
Mean(± SD*)4.7 (±3.1)2.9 (±1.5)0.002
Number of tumors, n (%)   
Single49 (87)19 (83)ns
Multiple7 (13)4 (17) 
Tumor grade, n (%)   
Well differentiated26 (46)8 (35)ns
Moderately differentiated25 (45)14 (61) 
Poorly differentiated5 (9)1 (4) 
Microvascular invasion (MiVI), n (%)   
MiVI+30 (53)12 (52)ns
MiVI−26 (47)11 (48) 
MALDI IMS

The cross-classification design for comparison of peak intensities between spectra of HCC/MiVI+ and those of HCC/MiVI− identified 30 protein peaks which were discriminative in at least 20 cross-classifications (Table 2). Twenty-eight peaks were higher in HCC/MiVI+, whereas two were higher in HCC/MiVI− (m/z 8060 and 6012, respectively). Mean spectra of HCC with and without MiVI are shown in Fig. 1.

Table 2. Peaks Discriminative Between HCC With (MiVI+) and Without (MiVI−) Microvascular Invasion (MiVI) Which Were Found by Means of the Cross-Classification Design Comparing Mass Spectra from Corresponding Tissue Sections
Peak (m/z)Number of Cross Classifications in Which the Peak is DiscriminativeP ValueTrends HCC MiVI+/ HCC MiVI−Mean Intensity in HCC/MiVI+ (a.u.)Mean Intensity in HCC/MiVI− (a.u.)
  1. a.u., arbitrary unit.

5654271.00[UPWARDS ARROW]6.96.6
3790261.00[UPWARDS ARROW]6.14.5
5822301.00[UPWARDS ARROW]4.93.4
5701251.00[UPWARDS ARROW]4.54
7763291.00[UPWARDS ARROW]4.13
5471531.00[UPWARDS ARROW]3.73
5507241.00[UPWARDS ARROW]3.72.8
7049261.00[UPWARDS ARROW]3.52.7
6469291.00[UPWARDS ARROW]32.3
5905201.00[UPWARDS ARROW]32.3
7263261.00[UPWARDS ARROW]2.82.1
5593301.00[UPWARDS ARROW]2.82.2
7276251.00[UPWARDS ARROW]2.61.9
7433301.00[UPWARDS ARROW]2.62
3644291.00[UPWARDS ARROW]2.61.9
7301261.00[UPWARDS ARROW]2.41.9
7481341.00[UPWARDS ARROW]2.11.7
8968261.00[UPWARDS ARROW]21.4
6012271.00[DOWNWARDS ARROW]22.1
7545241.00[UPWARDS ARROW]21.3
7801261.00[UPWARDS ARROW]1.91.2
8053261.00[UPWARDS ARROW]1.91.3
7522261.00[UPWARDS ARROW]1.71.4
10042561.00[UPWARDS ARROW]1.61.1
8904541.00[UPWARDS ARROW]1.60.9
8031301.00[UPWARDS ARROW]1.41.1
8993301.00[UPWARDS ARROW]1.41.1
8848251.00[UPWARDS ARROW]1.20.8
8060301.00[DOWNWARDS ARROW]1.161.22
9553301.00[UPWARDS ARROW]1.10.7
image

Figure 1. Mean spectra of HCC with (red) and without (blue) microvascular invasion (MiVI) representative of the training cohort. Higher magnification spectra highlight two differential peaks (m/z 3790 and 5654), overexpressed in HCC/MiVI+. a.u., arbitrary unit.

Download figure to PowerPoint

Protein Peaks Identification

m/z 5654 and 3790, the two protein peaks showing the highest intensities in HCC/MiVI+ (Table 2, Fig. 1), were detected by MALDI MS from tissue extracts after fractionation on RP-micro LC (Fig. 2). Despite MALDI ionization source normally generates +1 charged ions, m/z 5654 turned out to correspond to the doubly charged (MH22+) of m/z 11304 (Fig. 2A,C,D), and m/z 3790 to the triply charged (MH33+) of m/z 11346 (Fig. 2A,B,D). Therefore, m/z 11304 and 11346 were selected for molecular identification.

image

Figure 2. Detection of differential peaks in tissue extracts of HCC with microvascular invasion. MALDI MS spectra obtained after fractionation on reverse phase micro-liquid chromatography revealed m/z 5654 (A,C) and m/z 3782 (corresponding to the differential peak m/z 3790) (B), which corresponded to the doubly charged (MH22+) of m/z 11304 (D) and the triply charged (MH33+) of m/z 11346 (D), respectively. a.u., arbitrary unit.

Download figure to PowerPoint

Tryptic digestion of the fraction containing m/z 11304 and 11346, followed by nano-LC-ESI-MS/MS analysis, resulted in the characterization of human histone H4 (P62805, nominal mass 11.4 KDa) with a Mascot score of 287, sequence coverage of 66% and seven peptides assigned (Supporting Table 2). Since the average mass (MavH+) of human histone H4 is 11237 Da, we expected that m/z 11304 and 11346 Da peaks would correspond to histone H4 with posttranslational modifications (PTMs). Since N-term regions of histones are privileged sites of PTMs, digestion of histone H4 by ASP-N was performed in order to characterize the N-term fragment [1-23]. MALDI MS analysis of the digest resulted in detection of the fragments [1-23] at MH+ = 2430.28 and 2472.31 Da, corresponding to the expected mono-isotopic mass of the fragment [1-23] (MH+ = 2360.43 Da) with mass change of +70 and +112 Da, respectively (Fig. 3A). Nano-LC-ESI-MS/MS analysis of this digest led to characterization of human histone H4 with a Mascot score of 139, sequence coverage of 56%, and three peptides assigned. Peptide [1-23] + 70 corresponded to the N-term acetylated peptide dimethylated at lysine (K) 20 (Supporting Table 3). Inclusion list-dependent acquisition confirmed identification of peptide [1-23] + 70 and enabled characterization of peptide [1-23] + 112 as the N-term acetylated peptide acetylated at K16 and dimethylated at K20 (Fig. 3B,C; Supporting Fig. 1). Demonstrative MALDI images of m/z 5654 and 3790 are shown in Fig. 4. These 2D ion-intensity maps clearly showed that both peaks were expressed by tumoral cells, but not by stromal tissue.

image

Figure 3. Identification of histone H4 posttranslational modifications. (A) MALDI MS analysis of the Asp-N digest of histone H4 resulted in detection of fragments at MH+ = 2430.28 and 2472.31 Da, corresponding to the expected monoisotopic mass of the N-term fragment (MH+ = 2360.43 Da) with mass change of +70 and +112 Da, respectively. Collision-induced dissociation fragmentation during nano-LC-ESI-MS/MS analysis of N-term fragments +70 Da and +112 Da enabled characterization of the N-term acetylated peptide, dimethylated at lysine (K) 20 (B), and the N-term acetylated peptide, acetylated at K16 and dimethylated at K20 (C), respectively. 28: mass (Da) of the dimethyl group; 42: mass (Da) of the acetyl group.

Download figure to PowerPoint

image

Figure 4. MALDI IMS of a representative tissue section of HCC with microvascular invasion. (A) Hematoxylin and eosin staining showing areas of tumor cells with distinct morphologies (T1/T2) and stroma (S). MALDI images showing spatial distribution of ions with m/z 3790 (B) and 5654 (C), highlighting heterogeneous expression of proteins in tumor cells. a.u., arbitrary unit.

Download figure to PowerPoint

Modified Forms of Histone H4 Tissue Expression

Tissue expression of the two modified forms of histone H4 was then evaluated in the training and validation cohorts using specific antibodies targeting histone H4 dimethylated at K20 (H4K20me2) and acetylated at K16 (H4K16ac). H4K16ac and H4K20me2 protein expression was found at significantly higher levels in HCC/MiVI+ in the training cohort by WB (P = 0.02) (Fig. 5B). Immunostaining showed strong specific nuclear expression of both H4K20me2 and H4K16ac by tumor cells (Fig. 5). Semiquantitative counting of immunostained tumor nuclei showed significant overexpression, in HCC/MiVI+, of both H4K16ac (mean 107; 95% confidence interval [CI] 78-137 in HCC/MiVI+ versus 53; 95% CI 26-79 in HCC/MiVI−, P = 0.01) and H4K20me2 (mean 242; 95% CI 220-264 in HCC/MiVI+ versus 187; 95% CI 155-219 in HCC/MiVI−, P = 0.01) A similar significant difference was also observed in the validation cohort for both H4K16ac (mean 47; 95% CI 14-80 in HCC/MiVI+ versus 10; 95% CI 5-26 in HCC/MiVI−, P = 0.01) and H4K20me2 (mean 182; 95% CI 163-200 in HCC/MiVI+ versus 134; 95% CI 86-183 in HCC/MiVI−, P = 0.05) (Fig. 6A). H4K16ac and H4K20me2 expression was higher in tumor cells than in hepatocytes in both the training and validation cohorts (data not shown). H4K20me2 and H4K16ac immunostainings, performed in satellite nodules from five cases of HCC/MiVI+, showed a similar pattern of expression compared to the main nodule. Increased expression of both H4K20me2 and H4K16ac was observed in biopsy specimens of HCC/MiVI+ compared to HCC/MiVI−, but the difference was not significant (H4K20me2: mean 260 versus 200, P = 0.25; H4K16ac: mean 100 versus 75, P = 0.55) (Fig. 6B).

image

Figure 5. Tissue expression of the two modified forms of histone H4 in the training cohort. (A) Nuclear overexpression of histone H4 dimethylated at lysine 20 (H4K20me2) and histone H4 acetylated at lysine 16 (H4K16ac) in HCC with microvascular invasion (MiVI+) (right panel) compared to HCC without MiVI (MiVI−) (left panel) (×400). Lymphocyte expression was used as an internal control. Quantification of immunostaining showed significant overexpression in HCC/MiVI+ (**P < 0.01; Mann-Whitney test). (B) Significant overexpression of H4K20me2 and H4K16ac in HCC/MiVI+ (four cases) compared to HCC/MiVI− (three cases) by WB (*P < 0.05; Student's test).

Download figure to PowerPoint

image

Figure 6. Immunostaining of histone H4 dimethylated at lysine 20 (H4K20me2) and histone H4 acetylated at lysine 16 (H4K16ac) in the validation cohort and in biopsy specimen. (A) Validation cohort: H4K20me2 and H4K16ac overexpression in HCC with microvascular invasion (MiVI+) compared to HCC without MiVI (MiVI−) (×400) (*P < 0.05; **P < 0.01; Mann-Whitney test). (B) Representative immunostaining of H4K20me2 and H4K16ac in biopsy specimen of HCC with and without MiVI.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

MiVI is a major prognostic factor for HCC and was recently identified as a better predictor of tumor recurrence and overall survival following surgical resection for HCC than the Milan criteria, which are based on the number/size of tumors.[18, 19] In addition, Mazzafero et al.[20] demonstrated that patients fulfilling the up-to-seven criteria (HCC with 7 as the sum of the size of the largest tumor [in cm] and the number of tumors) but without MiVI, achieved 5-year overall survival similar to that of patients who met the Milan criteria, suggesting that MiVI could be a new selection criterion for LT. Since MiVI is not detectable by imaging techniques or even tumor biopsy, MiVI surrogate molecular markers are clearly needed for selection of patients for surgical resection or LT, and would be helpful in assessing HCC prognosis.

In the present study we used an on-tissue proteomic approach, MALDI IMS, to compare the tissue proteome of HCC with and without MiVI in order to identify new tissue surrogate biomarkers of MiVI. Our study shows that biomarkers of MiVI can be detected in tumor cells remote from the vascular invasion site, indirectly confirming results of previous genomic studies.[7, 8] Indeed, 30 protein peaks were differentially expressed with respect to the presence of MiVI. Among these, two peaks (m/z 3790 and 5654) were identified as modified forms of histone H4, providing the first evidence of an association between aberrant PTMs of histone H4 and MiVI in HCC.

MALDI IMS is a powerful tool for characterizing the proteomes of various tissues.[21] In this approach, the tissue section remains intact, protecting the spatial distribution of proteins and restricting preanalytical protein alterations induced by homogenization and extraction.[22] Furthermore, this technology allows detection of PTMs, which play an essential role in regulation of protein functions. Indeed, biomarkers detected in the present study would not have been detected by genomic profiling. The ability to visualize cellular distribution of potential markers is another attractive feature of MALDI IMS. Overlaying the histological image of m/z 3790 and 5654 confirms that these markers are specifically expressed by tumor cells and not by stromal tissue. In addition, 2D-ion intensity images, as expected, showed significant intratumoral heterogeneity.

Tryptic and Asp-N peptide fingerprinting associated with accurate mass determination by nano-LC/ESI-MS/MS indicated, with a high level of confidence, that m/z 3790 and 5654 corresponded to modified forms of histone H4, results that were further confirmed by immunohistochemistry and WB analyses. Interestingly, previous studies reported an association between modified forms of histone H3 or histone H3 modifying enzymes and both vascular invasion[23, 24] and poor prognosis in HCC,[25] supporting evidence for a major role of histone PTMs in HCC progression.

Histones are key molecules in epigenetic regulation of many DNA-based processes, including transcription, replication, and DNA repair, through PTMs of their N-terminal “tails.”[26] These modifications, directed by specific histone-modifying enzymes, function either by disrupting chromatin contacts or by affecting recruitment of nonhistone proteins to chromatin.[26] Yet the underlying mechanism of these epigenetic alterations remains largely unknown. Nevertheless, emerging evidence suggests that misregulation of covalent histone modifications contributes to the initiation and progression of human cancers,[27] and that changes in global levels of individual histone modifications are predictive of clinical outcome in various tumors.[28, 29] Acetylation at lysine 16 of histone H4 (H4K16ac) is a prevalent reversible PTM in eukaryotes, and recent studies have highlighted its significance. Indeed, H4K16ac, the level of which is regulated by histones acetyltranferase hMOF and deacetylase SIRT1, has been shown to play a central role in active transcription, presumably by facilitating chromatin decondensation.[30] However, the precise role of this PTM in carcinogenesis remains unknown, with discordant results across publications.[31, 32] In our study, HCC/MiVI+ had a higher level of H4K16ac than HCC/MiVI−, suggesting that increased H4K16ac in HCC/MiVI+ could be related to deregulated expression of hMOF and/or SIRT1, favoring transcription of genes implicated in tumor progression/invasion. However, the underlying mechanisms need to be investigated and genes targeted by these modifications must be identified. Dimethylation at lysine 20 of histone H4 (H4K20me2) is frequently observed[33] and several studies have underlined its importance in various DNA-templated processes such as DNA double-strand break repair.[34] More recently, it has been shown that ORC1—a component of ORC (origin of replication complex) which mediates pre-DNA replication licensing—contains a bromo adjacent homology domain that specifically recognizes H4K20me2, suggesting a critical role for H4K20me2 in cell cycle progression.[35] Indeed, increased expression of H4K20me2 in HCC/MiVI+, as observed in our study, could be involved in increased cell proliferation, contributing to the aggressive behavior of tumors with MiVI.

From a diagnostic perspective, immunodetection of surrogate MiVI biomarkers on tumor biopsies could have a significant impact on clinical practice. Although MiVI closely correlates with tumor size,[4] a significant subset of small HCC (≤3 cm) can show MiVI.[36] Indeed, these markers would be valuable for selection of patients at an early stage of HCC but with more aggressive disease for clinical trials of new adjuvant agents, thus avoiding unnecessary surgery or LT with implications for the restricted donor pool. Interestingly, overexpression of the two modified forms of histone H4 in HCC/MiVI+ was confirmed in the validation cohort comprising significantly smaller tumors (2.9 cm ± 1.5), underlying the potential diagnostic value of these biomarkers independently of tumor size. In order to assess whether H4K20me2 and H4K16ac may also detect macrovascular invasion that should have been missed by imaging, we also studied a few cases of HCC with macrovascular invasion (data not shown). Both markers were also expressed in tumoral cells, suggesting that these markers globally predict vascular invasion. Increased expression of both H4K20me2 and H4K16ac in biopsy specimens of HCC/MiVI+ compared to HCC/MiVI− was observed by immunohistochemistry. Although the difference was not significant, these preliminary results are encouraging and deserve further investigation.

In conclusion, our findings highlight the potential of MALDI IMS for uncovering relevant tissue molecular markers of MiVI in HCC that could be easily assessed in clinical practice by immunohistochemistry. We identified two modified forms of histone H4 as surrogate biomarkers of MiVI, supporting the significant role of epigenetic modifications in HCC prognosis. The relevance of the two markers in management of patients with HCC warrants further study evaluating their performance in biopsy specimens.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

We thank Nathalie Colnot, Sylvie Mosnier, Brigitte Boulangé, Olivier Boucher, Mohamed Achahboun, the “Réseau des Centres de Ressources Biologiques Foie-INCA,” and tumor biobank (Pathology department, Beaujon hospital) for technical support.

Author Contributions

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

V.P. and P.B. conceived the study. N.P., T.A., V.P., J.L.F., and M.M. wrote the article. N.P. collected clinical data. N.P., J.L.F., and S.L. assisted in the generation, analysis, and interpretation of MALDI IMS data. T.A. performed statistical analyses of MALDI IMS data. J.F.L., T.L., and J.M.C. were responsible for protein identification. J.B. participated in the writing of the article.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information
  • 1
    Ferlay J, Shin H, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893-2917.
  • 2
    Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology 2005;42:1208-1236.
  • 3
    Jonas S, Bechstein WO, Steinmüller T, Herrmann M, Radke C, Berg T, et al. Vascular invasion and histopathologic grading determine outcome after liver transplantation for hepatocellular carcinoma in cirrhosis. Hepatology 2001;33:1080-1086.
  • 4
    Sumie S, Kuromatsu R, Okuda K, Ando E, Takata A, Fukushima N, et al. Microvascular invasion in patients with hepatocellular carcinoma and its predictable clinicopathological factors. Ann Surg Oncol 2008;15:1375-1382.
  • 5
    Imamura H, Matsuyama Y, Tanaka E, Ohkubo T, Hasegawa K, Miyagawa S, et al. Risk factors contributing to early and late phase intrahepatic recurrence of hepatocellular carcinoma after hepatectomy. J Hepatol 2003;38:200-207.
  • 6
    Schlitt HJ, Neipp M, Weimann A, Oldhafer KJ, Schmoll E, Boeker K, et al. Recurrence patterns of hepatocellular and fibrolamellar carcinoma after liver transplantation. J Clin Oncol 1999;17:324-331.
  • 7
    Ho M-C, Lin J-J, Chen C-N, Chen C-C, Lee H, Yang C-Y, et al. A gene expression profile for vascular invasion can predict the recurrence after resection of hepatocellular carcinoma: a microarray approach. Ann Surg Oncol 2006;13:1474-1484.
  • 8
    Mínguez B, Hoshida Y, Villanueva A, Toffanin S, Cabellos L, Thung S, et al. Gene-expression signature of vascular invasion in hepatocellular carcinoma. J Hepatol 2011;55:1325-1331.
  • 9
    Caprioli RM, Farmer TB, Gile J. Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal Chem 1997;69:4751-4760.
  • 10
    Schwamborn K, Caprioli RM. Molecular imaging by mass spectrometry—looking beyond classical histology. Nat Rev Cancer 2010;10:639-646.
  • 11
    Balluff B, Rauser S, Ebert MP, Siveke JT, Höfler H, Walch A. Direct molecular tissue analysis by MALDI imaging mass spectrometry in the field of gastrointestinal disease. Gastroenterology 2012;143:544-549.e1-2.
  • 12
    Schwartz SA, Weil RJ, Thompson RC, Shyr Y, Moore JH, Toms SA, et al. Proteomic-based prognosis of brain tumor patients using direct-tissue matrix-assisted laser desorption ionization mass spectrometry. Cancer Res 2005;65:7674-7681.
  • 13
    Meding S, Balluff B, Elsner M, Schöne C, Rauser S, Nitsche U, et al. Tissue-based proteomics reveals FXYD3, S100A11 and GSTM3 as novel markers for regional lymph node metastasis in colon cancer. J Pathol 2012;228:459-470.
  • 14
    Cazares LH, Troyer D, Mendrinos S, Lance RA, Nyalwidhe JO, Beydoun HA, et al. Imaging mass spectrometry of a specific fragment of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase kinase 2 discriminates cancer from uninvolved prostate tissue. Clin Cancer Res 2009;15:5541-5551.
  • 15
    Le Faouder J, Laouirem S, Chapelle M, Albuquerque M, Belghiti J, Degos F, et al. Imaging mass spectrometry provides fingerprints for distinguishing hepatocellular carcinoma from cirrhosis. J Proteome Res 2011;10:3755-3765.
  • 16
    Lagarrigue M, Alexandrov T, Dieuset G, Perrin A, Lavigne R, Baulac S, et al. New analysis workflow for MALDI imaging mass spectrometry: application to the discovery and identification of potential markers of childhood absence epilepsy. J Proteome Res 2012;11:5453-5463.
  • 17
    Alexandrov T, Decker J, Mertens B, Deelder AM, Tollenaar RAEM, Maass P, et al. Biomarker discovery in MALDI-TOF serum protein profiles using discrete wavelet transformation. Bioinformatics 2009;25:643-649.
  • 18
    Lim K-C, Chow PK-H, Allen JC, Chia G-S, Lim M, Cheow P-C, et al. Microvascular invasion is a better predictor of tumor recurrence and overall survival following surgical resection for hepatocellular carcinoma compared to the Milan criteria. Ann Surg 2011;254:108-113.
  • 19
    Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Bozzetti F, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693-699.
  • 20
    Mazzaferro V, Llovet JM, Miceli R, Bhoori S, Schiavo M, Mariani L, et al. Predicting survival after liver transplantation in patients with hepatocellular carcinoma beyond the Milan criteria: a retrospective, exploratory analysis. Lancet Oncol 2009;10:35-43.
  • 21
    Stoeckli M, Chaurand P, Hallahan DE, Caprioli RM. Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nat Med 2001;7:493-496.
  • 22
    Seeley EH, Caprioli RM. MALDI imaging mass spectrometry of human tissue: method challenges and clinical perspectives. Trends Biotechnol 2011;29:136-143.
  • 23
    Cai M-Y, Hou J-H, Rao H-L, Luo R-Z, Li M, Pei X-Q, et al. High expression of H3K27me3 in human hepatocellular carcinomas correlates closely with vascular invasion and predicts worse prognosis in patients. Mol Med 2011;17:12-20.
  • 24
    Fan DN-Y, Tsang FH-C, Tam AH-K, Au SL-K, Wong CC-L, Wei L, et al. Histone lysine methyltransferase, SUV39H1, promotes HCC progression and is negatively regulated by microRNA-125b. Hepatology 2012;57:637-647.
  • 25
    He C, Xu J, Zhang J, Xie D, Ye H, Xiao Z, et al. High expression of trimethylated histone H3 lysine 4 is associated with poor prognosis in hepatocellular carcinoma. Hum Pathol 2012;43:1425-1435.
  • 26
    Kouzarides T. Chromatin modifications and their function. Cell 2007;128:693-705.
  • 27
    Wang GG, Allis CD, Chi P. Chromatin remodeling and cancer. Part I. Covalent histone modifications. Trends Mol Med 2007;13:363-372.
  • 28
    Seligson DB, Horvath S, Shi T, Yu H, Tze S, Grunstein M, et al. Global histone modification patterns predict risk of prostate cancer recurrence. Nature 2005;435:1262-1266.
  • 29
    Elsheikh SE, Green AR, Rakha EA, Powe DG, Ahmed RA, Collins HM, et al. Global histone modifications in breast cancer correlate with tumor phenotypes, prognostic factors, and patient outcome. Cancer Res 2009;69:3802-3809.
  • 30
    Shogren-Knaak M, Ishii H, Sun J-M, Pazin MJ, Davie JR, Peterson CL. Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 2006;311:844-847.
  • 31
    Fraga MF, Ballestar E, Villar-Garea A, Boix-Chornet M, Espada J, Schotta G, et al. Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human cancer. Nat Genet 2005;37:391-400.
  • 32
    Gupta A, Guerin-Peyrou TG, Sharma GG, Park C, Agarwal M, Ganju RK, et al. The mammalian ortholog of Drosophila MOF that acetylates histone H4 lysine 16 is essential for embryogenesis and oncogenesis. Mol Cell Biol 2008;28:397-409.
  • 33
    Schotta G, Sengupta R, Kubicek S, Malin S, Kauer M, Callén E, et al. A chromatin-wide transition to H4K20 monomethylation impairs genome integrity and programmed DNA rearrangements in the mouse. Genes Dev 2008;22:2048-2061.
  • 34
    Botuyan MV, Lee J, Ward IM, Kim J-E, Thompson JR, Chen J, et al. Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair. Cell 2006;127:1361-1373.
  • 35
    Kuo AJ, Song J, Cheung P, Ishibe-Murakami S, Yamazoe S, Chen JK, et al. The BAH domain of ORC1 links H4K20me2 to DNA replication licensing and Meier-Gorlin syndrome. Nature 2012;484:115-119.
  • 36
    Pawlik TM, Delman KA, Vauthey J-N, Nagorney DM, Ng IO-L, Ikai I, et al. Tumor size predicts vascular invasion and histologic grade: implications for selection of surgical treatment for hepatocellular carcinoma. Liver Transpl 2005;11:1086-1092.

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Author Contributions
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
hep26433-sup-0001-suppfig1.tif2028KSupporting Figure 1: Ion attribution of CID fragmentation spectra for peptides [1-23] + 70 (A) and [1-23] + 112 (B).
hep26433-sup-0002-supptab1.docx18KSupporting Table 1: Patients clinicopathological characteristics depending on the microvascular invasion (MiVI) status.
hep26433-sup-0003-supptab2.docx34KSupporting Table 2: Characterization of human histone H4 by Nano-LC-ESI-MS/MS analysis of the tryptic digest.
hep26433-sup-0004-supptab3.docx31KSupporting Table 3: Characterization of human histone H4 by Nano-LC-ESI-MS/MS analysis of the Asp-N digest.
hep26433-sup-0005-suppinfo.docx15KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.