Autoantibody signatures defined by serological proteome analysis in sera from patients with cholangiocarcinoma

The design of the study is summarized in Fig. 1. Briefly, sera from a pool of ten normal subjects, and from eight patients with CC,
were tested on two-dimensional immunoblots performed using two CC cell lines and human
liver from healthy subjects. In addition, five CC sera were tested on these three
substrates and also on immunoblots performed using tumour and non-tumour tissue from
the same CC patients. Spots of interest were identified by mass spectrometry (MS),
and autoantigens categorised according to the Gene Ontology project before further
classification.

Fig. 1. Design of the experiment. CC cholangiocarcinoma

Serum samples and human tissue specimens

All patients gave their informed consent for the collection of blood and tissue samples.
Specimens were conserved at ?80 °C, with approval of the “Committee of the Biobanque
of Centre Hépato-Biliaire”, managed by the “Biological Resource Centre CRB Paris-Sud”.
All subjects signed a written informed consent form regarding this analytical study.

Thirteen serum samples from CC patients followed by the Centre Hépato-Biliaire at
Hôpital Paul-Brousse, were analysed. All the patients fulfilled the international
criteria for the diagnosis of CC. Ten pooled sera from healthy volunteers were used
as controls.

The CC tissues and adjacent non-tumour liver tissues used for this study were collected
from five CC patients who were being treated surgically in our centre. After resection,
the specimens were rinsed thoroughly in ice-cold normal saline and stored at ?80 °C.
Necrotic tissues were excluded, and pathological examination of the non-tumour liver
tissues by an expert (CG) confirmed that they contained no tumour. Normal liver tissue
specimens were obtained from patient who had been transplanted for amyloid neuropathy.

All liver tissues were homogenized using a Potter-Elvejhem apparatus, with 10 mM Tris,
50 mM sucrose, 1 mM EDTA and 1 mM phenylmethyl sulphonide fluoride (PMSF). Homogenates
were lysed in buffer with 50 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 % triton
(v/v), 0.2 % SDS (w/v) and 1 % (v/v) nuclease mix (GE Healthcare).

Cell lines

Two human cholangiocarcinoma cell lines, CCSW1 and CCLP1, were obtained from the European
Cell Culture Bank, and cells were grown in Dulbecco’s modified Eagle’s medium (DMEM)
supplemented with 10 % (v/v) heat inactivated bovine fœtal serum (BFS), 1 % (v/v)
minimal essential medium of non-essential amino acids, 1 mmol/L sodium 2-oxopropanoate,
and standard concentrations of penicillin plus streptomycin. Whole cell proteins were
extracted from the cell lines. Cell lysis was performed with 20 mM Tris (pH 7.5),
150 mM NaCl, 1 % NP40 (Sigma) (v/v), 1× protease inhibitor (Roche, Germany) and 1×
phosphatase inhibitor.

Two-dimensional gel electrophoresis (2-DE) and immunoblotting

Proteins from the lysed homogenates and cell lines were precipitated using the 2-D
Clean up kit (GE Healthcare) and the final protein concentration was measured with
the 2-D Quant kit (GE Healthcare). Protein samples of 250 ?g for future immunotransfer,
or 1 mg for future Coomassie blue staining, were mixed with IEF buffer (7.5 M urea,
2.2 M thiourea, 4 % (w/v) CHAPS, 0.6 % (v/v) immobilised pH gradient (IPG) buffer
at pH 3–10, 0.8 % (v/v) Destreak
^®
solution (GE Healthcare) and orange G. For each sample, the proteins were applied
to an immobiline Dry Strip
^®
(pH range 3–10, 13 cm; GE Healthcare). After overnight rehydration at room temperature,
the IEF procedure was performed by applying voltage that was gradually increased to
a maximum of 23,000 V/h.

Each IPG strip was then equilibrated with a solution containing 6 M urea, 0.075 M
Tris (pH 8.8), 30 % (v/v) glycerol, 2 % (w/v) SDS, 2 % (w/v) DTT and pyronine for
15 min. The strips were equilibrated again by replacing DTT with 5 % (w/v) idoacetamide,
for a further 15 min. The IPG strips were applied to 10 % SDS-PAGE for a second dimension
protein separation. For subsequent immunoblotting, the proteins were transferred to
nitrocellulose membranes and then blocked for 1 h with 50 mL blocking buffer: 5 %
(w/v) non-fat powdered milk in TBS-T (Tris Buffer Saline-Tween 20) PH 7.4, (Tris 1 M
2 % (w/v), NaCl 0.8 % (w/v), Tween 0.1 % (v/v). The filters were then probed with
sera diluted 1:2000 in TBS-T, and finally incubated with 1:3000 diluted horseradish
peroxidase-conjugated anti-human immunoglobulin (BioRad). The proteins were detected
by chemiluminescence according to the manufacturer’s instructions (ECL Plus™ Western
Blotting Detection kit, GE Healthcare). After transfer, the resulting gels were silver-stained.

The analysis was performed in triplicate. After standard immunoblotting, the patterns
produced by CC sera were compared with those given by normal sera using scanning and
superimposition by means of Adobe Photoshop
^®
Software. Spots of interest were defined as those which were only stained by CC sera.

The transferred and silver-stained gels and their corresponding immunoblots were also
scanned, and after Adobe Photoshop
^®
software analysis, the spots of interest were localised on the silver-stained gels.
These spots were then localised together on the corresponding scans of Coomassie blue-stained
gels. Immunoreactive spots obtained with at least 30 % of CC sera were then identified
using MS.

Procedures for protein and peptide preparation

The spots of interest were excised manually. Cysteine reduction was performed with
10 mmol/L DTT-100 mmol/L NH
₄
HCO
₃
for 45 min at 56 °C, and protein alkylation was carried out with 55 mmol/L iodoacetamide-100 mmol/L
NH
₄
HCO
₃
for 30 min in the dark at room temperature, the gel pieces being washed successively
with 100 mmol/L NH
₄
HCO
₃
, a 1:1 (by volume) mixture of 100 mmol/L NH
₄
HCO
₃
and acetonitrile, and acetonitrile, before being dried again. The gel pieces were
then rehydrated for 45 min at 4 °C in a digestion buffer containing 50 mmol/L NH
₄
HCO
₃
, 5 mmol/L CaCl
₂
, and 12.5 mg/L trypsin. Peptides generated through proteolytic digestion were extracted
by incubation in 10 g/L formic acid for 15 min, which was followed successively by
two extractions with 10 g/L formic acid-acetonitrile (1:1 by volume) and acetonitrile.
The extracted peptides were pooled and dried out in a SpeedVac centrifuge before mass
spectrometry (MS) analysis.

Mass spectrometry analysis

LC–MS measurements were obtained using a nano LC system (Ultimate 3000; Dionex) coupled
online to a hybrid linear ion trap/Orbitrap
^®
MS (LTQ OrbitrapVelos; Thermo Fisher Scientific, Bremen, Germany). One microlitre
of protein digest was injected onto the nano LC system, which contained a C18 trap
column (PepMap C18, 300 ?mID × 5 mm, 5 ?m particle size and 100 Å pore size; Dionex)
and a 15 cm long analytical column (Acclaim pepmap RSLC 75 µm × 15 cm, nanoViper C18,
2 µm, 100 Å). The peptides were separated according to the following gradient: 100 %
solvent A (0.1 % formic acid in water) for 3 min., 0–55 % solvent B (80 % acetonitrile
in water with 0.1 % formic acid) for 25 min., 50–90 % solvent B for 1 min and 90 %
solvent B for 5 min. A high resolution full scan MS was obtained from the Orbitrap
^®
(resolution 30,000; AGC 1,000,000), and MS/MS spectra were obtained by CID (collision-induced
dissociation) fragmentation, with an isolation window of 3 Da. A data-dependent top
5 (one full MS and 5 MS/MS) was obtained with the dynamic exclusion option switched
on. Spots that were reactive with fewer than 30 % of sera were not identified by MS.

Data analysis

The data were analysed using Discoverer Proteome 1.4 software. The database is a human
Swiss-Prot, the mass error for the precursor ions (full MS) being less than 10 ppm.
The mass error for ions from the MS/MS spectra is reported to be less than 0.6 Da.
Searches for peptide mass are made between 350 and 5000 Da with a time retention ranging
from 10 to 50 min. A miss cleavage site is tolerated. Dynamic modification was enabled
for N
_ter
acetylation, the oxidation of methionine and histidine, and the carbamidomethylation
of amino acids, aspartic acid and glutamic acid. The static carbamidomethyl modification
of cysteine was enabled. Peptide identifications are validated by determining false
positives using the Target decoy PSM validator. This is high if the false positive
rate (FDR or false Discovery Rate) is less than 1 %, low if the FDR is greater than
5 % or average (between 1 and 5 %). Peptide identification Xcorr were calculated by
correlating the MS/MS experimental spectrum with the theoretical MS/MS spectrum generated
by Proteome Discoverer 1.4 software.

Detection of anti vimentin and anti actin antibodies by immunofluorescence as a validation
technique

The presence of anti-vimentin and anti-actin antibodies was determined using indirect
immunofluorescence on monolayers of colchicine-treated Hep2 cells, as described elsewhere
11], 12]. Briefly, Hep2 monolayer cells culture was home-prepared. This culture performed
on slide was incubated with colchicine 0.0014 % (w:v) (Sigma) diluted in minimum essential
medium Eagle (Eurobio), glutamine supplemented, during 20 h at 37 °C. After three
washes with phosphate-buffered saline (PBS), and acetone-fixed, the monolayer cells
was incubated with sera at the dilution of 1/40, in PBS, during 30 min. After PBS
washing (3×), monolayer cells were revealed using a fluorochrome–labeled, polyclonal
antihuman IgG, IgA, IgM antiserum (BioRad Laboratory), 30 min incubated. Vimentin
appear to be collapsed into thick perinuclear coils if the tested serum was positive
for anti-vimentin antibodies, and a typical pattern of actin cables was strongly stained
if serum was positive for anti-actin antibodies.