A proteomic approach to analysing spheroid formation of two human thyroid cell lines cultured on a random positioning machine

RESEARCH ARTICLE

A proteomic approach to analysing spheroid formation

of two human thyroid cell lines cultured on a random

positioning machine

Jessica Pietsch1,2, Albert Sickmann3, Gerhard Weber 4, Johann Bauer 5, Marcel Egli 6,Robert Wildgruber 7, Manfred Infanger 8 and Daniela Grimm2,9

1 Department of Biology, Chemistry, Pharmacy, FU Berlin, Berlin, Germany2 Institute of Clinical Pharmacology and Toxicology, Charite-Unversit .atsmedizin Berlin, Berlin, Germany3 ISAS, Institute for Analytical Sciences, Dortmund, Germany4 FFE-Service GmbH, Kirchheim, Germany5 Max-Planck Institute of Biochemistry, Martinsried, Germany6 Space Biology Group, ETH Zurich, Zurich, Switzerland7 Becton and Dickinson, Martinsried, Germany8 Centre hospitalier de Luxembourg, Luxembourg9 Department of Pharmacology, Aarhus University, Aarhus, Denmark

Received: December 23, 2010

Revised: January 26, 2011

Accepted: February 17, 2011

The human cell lines FTC-133 and CGTH W-1, both derived from patients with thyroid

cancer, assemble to form different types of spheroids when cultured on a random positioning

machine. In order to obtain a possible explanation for their distinguishable aggregation

behaviour under equal culturing conditions, we evaluated a proteomic analysis emphasising

cytoskeletal and membrane-associated proteins. For this analysis, we treated the cells by

ultrasound, which freed up some of the proteins into the supernatant but left some attached

to the cell fragments. Both types of proteins were further separated by free-flow IEF and SDS

gel electrophoresis until their identity was determined by MS. The MS data revealed differ-

ences between the two cell lines with regard to various structural proteins such as vimentin,

tubulins and actin. Interestingly, integrin a-5 chains, myosin-10 and filamin B were only

found in FTC-133 cells, while collagen was only detected in CGTH W-1 cells. These analyses

suggest that FTC-133 cells express surface proteins that bind fibronectin, strengthening the

three-dimensional cell cohesion.

Keywords:

Cell biology / Cytoskeleton / Free-flow electrophoresis / Random positioning machine /

Spheroids / Three-dimensional cell growth

1 Introduction

Most tissue cells need adhesion to a solid surface in order to

stay viable upon dissociation and suspension in fluid.

Hence, for in vitro studies, tissue cells are suspended in

plastic dishes, where they are allowed to sediment and to

anchor to the inner polystyrene surface with the help of

extracellular matrix released into the nearer surrounding

and of contractile forces generated within the cells [1]. Some

decades ago, it had been observed that tissue-derived cancer

cells can survive, when their anchorage to a suitable surface

is prevented by agar (liquid overlay method [2]) or by stirring

Abbreviations: 1 g, normal gravity; emPAI, exponentially modi-

fied protein abundance index; FFE, free-flow electrophoresis;

FF-IEF, free-flow isoelectric focusing; M-scores, MASCOT scores;

RPM, random positioning machine

Correspondence: Dr. Johann Bauer, Max-Planck-Institut f .ur

Biochemie, Am Klopferspitz 18a, D-82152 Martinsried, Germany

E-mail: [email protected]

Fax: 149-89-1417931

& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com

Proteomics 2011, 11, 2095–2104 2095DOI 10.1002/pmic.201000817

(spinnerflask technique [3]). In this case they form three-

dimensional multicellular spheroids resembling small

microtumours in vivo rather than cells grown in two-

dimensional monolayers, because they form three-dimen-

sional intercellular contacts. Since they lack the interaction

with other healthy cells (e.g. fibroblasts and endothelial

cells), they represent an experimental model that is inter-

mediate in its complexity between monolayer cultures and

tumours grown in vivo.

Humans developed several health problems during and

after a space mission [4]. For this reason, interest arose to

study cells cultured in space. Since such experiments are

expensive and multifaceted, their careful preparation

appeared necessary. Therefore, machines were developed

which allow culturing cells under conditions of lacking a

defined gravity vector. Among others, rotating wall vessel

devices and the random positioning machine (RPM) proved

rather useful for preparing space flight experiments [5, 6].

The RPM can continuously and randomly change the

orientation of cell monolayers growing at the bottom of

culture flasks relative to the gravity vector so that sedi-

mentation of the cells is omitted. When we cultured human

tissue cells on an RPM, we observed repeatedly that rando-

misation of the gravity vector by the RPM induced not only

malignant human thyroid cells to leave the monolayer and to

continue growth within a three-dimensional spheroid [7], but

triggered also adherent endothelial cells to form tubes [8, 9] or

chondrocytes to form little pieces of cartilage [10]. This cell

type specific formation of three-dimensional aggregates on

the RPM [7–10] suggested that forces due to biochemical

components actually expressed on the cell surfaces predo-

minantly trigger initial interactions of cells floating in a fluid

in the absence of significant sheer [11] or vibration forces

[manuscript in preparation]. Thus, spheroid formation on the

RPM was considered a model to identify molecules, which

induce three-dimensional cell aggregation in vitro. It is

accompanied by an increase in apoptosis and an elevation of

extracellular matrix proteins in thyroid cells and endothelial

cells [7, 8]. Therefore, we started to characterise the cells by

determining several markers of apoptosis, cell adhesion

and extracellular matrix proteins with the help of flow cyto-

metry and Western blotting. However, both techniques are

limited by the need for appropriate antibodies. Therefore,

a further exploration of proteins was performed using MS.

The aim was to detect alterations in as many proteins as

possible, regardless of whether they were strongly or weakly

expressed.

In order to make scarce membrane-associated proteins

accessible without needing to discard the more abundant

ones [12], we used free-flow-isoelectric focusing (FF-IEF) for

efficient preparation of the proteins for highly developed

LC-MS/MS [13, 14]. In the first step, the cells were sonicated

and the proteins freed into the cell suspension fluid were

separated by centrifugation from the proteins that remained

attached to the membrane fragments. Both protein frac-

tions were subsequently prepared for matrix-free IEF using

free-flow electrophoresis (FFE). After the proteins had been

fractionated according to their pI, they were subjected to

SDS gel electrophoresis. Then, their identities were deter-

mined by MS [15].

Recently, we applied this method to investigate cells of

four different human thyroid cell lines after they had been

cultured under normal gravity (1 g) and RPM exposure.

Again, the cells were sonicated, and proteins released into

the supernatant and those remaining attached to the cell

fragments were fractionated by the FF-IEF technique [16].

The fractions obtained were further separated by SDS gel

electrophoresis. Selected gel pieces were excised and their

proteins were determined by MS. After a first evaluation of

the data, we published an overview on these studies

describing the method and reporting the proteins, which

appeared first identifications in human thyroid cells and

such which seemed to be differently expressed under 1 gand RPM exposure [16].

Meanwhile, we observed that the human cancer cell lines

FTC-133 [17] and CGTH W-1 [18], which were both derived

from patients suffering with thyroid carcinomas, assembled

to form three-dimensional spheroids of diverging size when

they were cultivated on an RPM. When sorting through

more than 1200 proteins determined in the study described

above [16], we recognised a number of proteins that may

explain the different aggregation behaviours found in these

two types of cells. Therefore, we proceeded to evaluate the

data, emphasising structural and membrane-associated

proteins of the FTC-133 and CGTH W-1 cells because these

proteins are most strongly affected, when gravity forces are

annulled [19] and simultaneously play an important role

when cells transit from two- to three-dimensional growth on

the RPM in vitro [10]. We were interested whether we would

find differences of proteins which connect cells by direct

binding to the surroundings or of proteins which generate

contractile forces required for cell–cell or for cell–surface

interactions [1]. The evaluation revealed differences between

the two cell lines with regard to abundant and less frequent

structural proteins.

2 Materials and methods

2.1 Culturing of the FTC-133 and CGTH W-1 cell lines

The human carcinoma cell lines FTC-133 [17] and CGTH

W-1 [18] were cultured in RPMI-1640 medium, which was

supplemented as indicated previously [16]. Each cell line was

seeded in 12 T-25 culture flasks and incubated at 371C and

5% CO2 in a normal laboratory incubator until subconfluent

monolayers had been grown. Then, all flasks were filled

completely with culture medium and six flasks of each cell

line were kept incubated under 1 g for further three days,

while the other six were mounted onto an RPM (Dutch

Space, Leiden, The Netherlands) for the same amount of

time. The RPM was operated in a random walk (basic)

2096 J. Pietsch et al. Proteomics 2011, 11, 2095–2104


which means that the experiment platform connected to two

perpendicular frames was moved according to a ‘‘random

walk scenario’’ continuously generated by computer soft-

ware and randomly changing the orientation of objects

relative to the gravity vector [6, 7]. Adjusted at a speed of

601/s, a residual sedimentation force of 10�2 g remained

within a distance of about 10 cm to the centre of rotation

[20]. Therefore, the flasks were mounted on the RPM inside

a radius of about 8 cm to the centre of rotation. Immediately

after stopping the RPM at the end of the 3 days incubation

period, pictures of the native spheroids were taken using

phase contrast microscopy (Olympus, Hamburg, Germany)

[9]. Then the cells were harvested, shock-frozen with liquid

nitrogen and stored at �801C until use [16].

2.2 Western blot analysis

SDS-PAGE, immunoblotting and densitometry were carried

out following routine protocols [21]. Antibodies were applied

at appropriate dilutions to quantify vimentin (dilution:

1:200; Sigma-Aldrich, St. Louis, MO, USA), tubulin b(dilution: 1:100; Sigma-Aldrich) and actin (dilution: 1:500;

Cell Signalling Technology, Danvers, MA, USA).

2.3 Cell preparation

Immediately prior to the FFE experiments, the cells were

thawed and suspended in 0.5 mL HEPES buffer (10 mM

HEPES, 15 mM MgCl2, 10 mM KCl and 0.2% DTT)

containing one tablet of protease inhibitor ‘‘Complete Mini’’

(Roche, Basel, Switzerland) per 10 mL fluid. Subsequently,

they were sonicated for 30 s on ice with a Soniprep 150

setting. The sonicated sample was centrifuged at 25 000� gat 41C for 30 min. The supernatant containing the sonica-

tion-freed proteins was collected and the HEPES buffer was

exchanged for a lysis medium containing 7 M urea, 2 M

thiourea, 4% CHAPS, 1% ASB-14 and 10 mM DTT by

ultrafiltration, until the concentration of the original buffer

was decreased by more than 100 times to a protein

concentration of 2.5 mg/mL. The pellet with the cell frag-

ment proteins was resuspended in 1 mL HEPES buffer and

washed once in this buffer. Subsequently, the pellet was

resuspended in lysis medium as described above and

centrifuged at 25 000� g at 41C for 60 min. The proteins

were dissolved in lysis medium at a concentration of

approximately 1 mg/mL before being applied to FF-IEF [16].

2.4 FF-IEF

The FFE separations were conducted in the FF-IEF mode

using a BDTM FFE System (BD Diagnostics, Munich,

Germany), exactly as described elsewhere [16, 22]. Briefly, a

stable pH gradient was created between the anodal (pH 3)

and cathodal (pH 10) edges of the separation medium,

which contained 7 M urea, 2 M thiourea and 250 mM

mannitol in aqueous solution, in addition to the ampholytes

used. Each sample was infused into the chamber at a rate of

1 mL/h through inlet S2, which entered the separation

chamber opposite the 48th sample collection fraction out of

96, where the chamber fluid pH was 7. Electrophoresis was

performed at a separation buffer flow speed of 60 mL/h,

101C, and a voltage of 520 V was applied perpendicular to

the flow of the 10-cm-wide separation medium film. The

separated samples were collected in 96-well plates. Collec-

tion began 25 min after starting a run and lasted approxi-

mately 6 min. The fractions were used for further analyses

and had fluid pH levels of 5.45–5.86 (A1), 6.5–6.7 (A2),

7.3–7.5 (C2) and 8.8–9.2 (C1) (see Fig. 1 of ref [16]).

2.5 Sample preparation for MS

The FF-IEF fractions of interest were first concentrated

30- to 40-fold and thereafter subjected to SDS-PAGE

as described by Pietsch et al. [16]. The proteins were stained

by Coomassie Brilliant Blue G-250 (Bio-Rad, Munich,

Germany). The gel bands of interest were cut out and

treated as described earlier [23].

2.6 Protein identification by MS

The proteins were analysed using nano-LC-MS/MS accord-

ing to a protocol already described by Pietsch et al. [16].

Briefly, an UltiMate 3000 nano-LC system (Dionex, Idstein,

Germany) coupled to an ESI-linear ion trap (LTQ XL,

Thermo Electron, Karlsruhe, Germany) was employed. The

LC setup consisted of an autosampler (WPS, Dionex) and a

column compartment (FLM, Dionex) before nano-

LC separation (UltiMate 3000, Dionex). Gradient elution

was performed using a linear gradient from 95% solvent A

(0.1% formic acid) to 50% solvent B (84% ACN and 0.1%

formic acid) for 33 min. The peptides were directly eluted

into the ESI-linear ion trap (LTQ XL) using distal-coated

fused-silica tips (New Objectives, Woburn, MA, USA) with

the spray voltage set to 1800 V. A survey scan (m/z400–2000) was followed by five MS/MS scans that frag-

mented the five most intensive peptide signals (1000 cps,

30 ms). Duplicate detection of one mass within 30 s led to

dynamic exclusion for 180 s. The mass spectra obtained

from the LC-MS/MS analysis were used to identify the

corresponding peptides using the MASCOTTM algorithm

(version 2.1.6) [24]. The raw data were converted by the LCQ-

DTA.EXE plug-in to the MASCOT Daemon software

using the following parameters: (i) minimum mass: 400 Da,

(ii) maximum mass 3000 Da, (iii) grouping tolerance 1.4,

(iv) minimum scans/group: 1, (v) intermediate scans: 1 and

(vi) precursor charge: auto. Searches were conducted against

the current FASTA database of Homo sapiens using the

Proteomics 2011, 11, 2095–2104 2097


following parameter set: (i) fixed modification: carbamido-

methyl (C); (ii) variable modification: oxidation (M);

(iii) peptide and MS/MS tolerance: 70.5 Da; (iv) ion score

cut-off point: 35; (v) significance threshold of 0.05 and

(vi) enzyme trypsin with miss cleavage: max. 1. After

manual validation, a protein was identified when at least two

different peptides with a score of 435 were found and the

cumulative score was 4100. The exponentially modified

protein abundance index (emPAI) was calculated according

to Ishihama et al. [25].

2.7 Statistical analysis

Statistical analysis was performed using SPSS 16.0

(SPSS, Chicago, IL, USA). All data are expressed as

mean7standard deviation (SD). Differences were consid-

ered significant at the level of po0.05.

3 Results and discussion

3.1 Spheroids of FTC-133 and CGTH W-1 cells

In this study we compared two human thyroid carcinoma cell

lines. The cell line FTC-133 was derived of a lymph node

metastasis, observed in a male suffering of a follicular thyroid

carcinoma. The cells are tumourigenous in nude mice,

express thyroglobulin and react sensitive to TSH [17]. The cell

line CGTH W-1 was established from a sternal metastasis of

a follicular thyroid tumour. When injected subcutaneously

into SCID mice the cells formed subcutaneous tumours and

distant metastases [18]. Both lines grew as monolayers when

cultured under normal laboratory conditions (1 g; Fig. 1A and

C) and showed a three-dimensional growth in vitro when

incubated on an RPM (1 g; Fig. 1B and D). However, the

spheroids which all formed within three days of RPM culti-

vation diverged in size. We found large structures with

diameters of up to 1 mm only in those six culture flasks,

which contained FTC-133 cells (Fig. 1B). Although the six

culture flasks containing CGTH W-1 were carefully examined

after incubation on the RPM, only rather small spheroids

(o0.2 mm in diameter) were observed in these cultures

(Fig. 1D). In order to obtain a possible explanation for the

formation of these distinguishable spheroids by the related

human carcinoma cell lines CGTH W-1 and FTC-133, we

evaluated the proteomic analysis data mentioned above [16].

The proteins involved in maintaining cell shape and motility

were of special interest since the transition from growth

within a two-dimensional monolayer to growth within a

three-dimensional aggregate is thought to be accompanied by

a re-organisation of such proteins [26].

1g RPM

FTC-133

A B

CGTH W-1

C D

50 μm 50 μm

50 μm 50 μm

Figure 1. Cells of the FTC-133 (A and B) and

CGTH W-1 (C and D) cell lines after 3 days of

incubation under normal laboratory condi-

tions (A and C) and on an RPM (B and D).

Spheroids formed by the different cell lines

diverged in size.



3.2 Vimentin

In our study, vimentin was the most abundant protein

found in both cell lines. Vimentin is an intermediate fila-

ment that contributes to the maintenance of the structural

and mechanical integrity of cells and tissues [27]. This

protein was firmly anchored within the cell membranes, and

sonication did not cause a significant release into the

supernatant. Although 78 pieces were excised from two gels

loaded with sonication-freed proteins and analysed accord-

ing to the methods described above, only one gel piece was

found to contain vimentin of the CGTH W-1 cell line, with a

very low MASCOT score (M-score) of 134. In contrast, when

the cell-fragment proteins of FTC-133 and CGTH W-1 cells

were loaded onto the gels, all of the 31 gel pieces excised

after electrophoresis showed vimentin. According to the

emPAI values, the highest quantities of vimentin were

found when gel pieces were investigated that had been

loaded with FFE fractions of CGTH W-1 cells collected at

pH 5.45–5.86 and then excised, where the protein marker

indicated an apparent molecular weight (MW) around

55 kDa (Table 1A). This result indicates that most of the

vimentin was processed without proteolytic degradation, but

it had not completely reached its pI of around 5.0 (Swiss-

Prot, http://expasy.org/sprot/, P08670) when it left the FFE

chamber. Hence, a minimal level of protein precipitation

can be assumed [28]. High quantities of vimentin were also

found in pieces of the gels that had been loaded with FFE

fractions collected at pH 5.45–5.86 and then excised, where

the protein marker indicated an apparent MW around

50 kDa. Vimentin in these gel pieces was determined with

higher M-scores and emPAI values for the CGTH W-1 cells

than the FTC-133 cells. These M-scores and emPAI values

(Table 1A) appeared to represent the quantities of vimentin

in both cell lines. This was confirmed by Western blot

analyses of whole cell lysates, which revealed a higher

concentration of vimentin in the CGTH W-1 cells cultured

under 1 g condition than in the RPM-exposed CGTH W-1

cells and in the FTC-133 cells cultured at 1 g (Fig. 2A).

Interestingly, vimentin in FTC-133 cells was enhanced

during the incubation on the RPM (Fig. 2A).

3.3 The tubulin system

Tubulin isoforms are building stones of microtubuli, which

are physically robust polymers with an intrinsic resistance to

bending and compression [29]. Microtubuli are gravity-

sensitive in human lymphocytes [30]. Isolated from bovine

brain cells, their self-organisation was inhibited during

space flight [31]. Both investigated thyroid cell lines

contained tubulin a- and tubulin b-chains. Table 1B shows

the tubulin content of the gel pieces that had been loaded

with FFE fractions collected at pH 5.45–5.86 and then cut at

a gel site where an apparent MW around 50 kDa was indi-

cated. Of these gel pieces, the ones containing sonication-

freed CGTH W-1 proteins showed tubulin a-1B chains with

higher M-scores than the tubulin b-chains. But the M-scores

and emPAI values determined for the tubulin b-chains were

two- to fourfold higher than for the tubulin a-1B-chains in

both cell lines when the cell-fragment proteins were

analysed. These emPAI values suggest that many b-chains,

firmly attached to cell fragments, interact with more tubulin

a-than a-1B-chains. Although emPAI values decreased

when tubulin b-chains of both types of cells cultured under

RPM exposure were determined by MS (Table 1B), parallel

Western blots of whole cell lysates indicated that RPM

exposure had reduced the content of tubulin b-chains in

CGTH W-1 cells but not in FTC-133 cells (Fig. 2B). These

results suggest that rather similar amounts of tubulin

b-chains remained in cell fragments during sonication,

while different quantities were present in whole cells.

Table 1. M-scores and emPAI values of vimentin and tubulin chains

Sonication-freed Cell-fragment

A: Vimentin 50 kDa 55 kDa

CGTH W-1 1 g 21480/147.08 (A1) 19542/298.87 (A1)CGTH W-1 RPM 18089/101.39 (A1) 18651/279.41 (A1)FTC-133 1 g 9103/77.36 (A1) n.a.FTC-133 RPM 10578/59.66 (A1) n.a.

B: Tubulin chains 50 kDa 50 kDa

a-1B b a-1B b

CGTH W-1 1 g 5966 (A1) 5388 (A1) 3020/9.22 (A1) 5849/23.09 (A1)CGTH W-1 RPM 6569 (A1) 2348 (A1) 1146/4.48 (A1) 4625/20.89 (A1)FTC-133 1 g n.a. n.a. 2698/10.12 (A1) 7087/35.18 (A1)FTC-133 RPM n.a. n.a. 1739/6.01 (A1) 6133/17.67 (A1)

The kDa specification refers to the excised and analysed SDS-gel pieces. For all proteins the M-scores and behind the ‘/’ the emPAI values,if available, are given; n.a., not analysed; RPM, cells had been cultured on the RPM prior to analysis; A1, FFE fractions with pH 5.45–5.86.

Proteomics 2011, 11, 2095–2104 2099


3.4 Proteins of the actin system

Actin is a vital component of the cellular cytoskeleton and it

essentially contributes to cell motility and tissue remodel-

ling [32, 33]. In our study, actin was found among the

sonication-freed proteins as well as among the cell-fragment

proteins. Within the FFE, actin migrated towards the anode

and was collected in fractions with a pH between 5.45 and

5.86 [16]. These FFE fractions were electrophoresed by SDS

gel electrophoresis. Subsequently, gel pieces containing

proteins with a MW between 40 and 45 kDa were excised

and analysed. The data showed higher values of M-scores for

actin when the sonication-freed protein fractions of CGTH

W-1 and FTC-133 cells were investigated than when the cell-

fragment proteins were studied (Table 2A). It was very

interesting that in comparison to the cells grown at 1 g, the

M-scores determined for actin were enhanced when CGTH

W-1 cells had been cultured on the RPM, but they decreased

when the FTC-133 cells had been incubated on the RPM.

These M-scores indicated a tendency of RPM exposure-

regulated actin expression because the Western blot

analyses revealed a more significant reduction of actin in

RPM-exposed FTC-133 cells than in the RPM-exposed

CGTH W-1 cells compared to the 1 g control cells, respec-

tively (Fig. 2C).

In addition to actin, we detected proteins that are known

to bind to actin [34–36]. High MW actin-binding proteins

such as myosin-9, plectin and filamin A were mainly found

in the cell-fragment fractions of both cell lines (Table 2B).

The sonication-freed fractions of both CGTH W-1 and FTC-

133 cells contained filamin A, while myosin-9 and plectin

were only detected in the sonication-freed fractions of the

CGTH W-1 cells. Furthermore, actinin a 4 was exclusively

found among the cell-fragment proteins of FTC-133 cells,

whereas tropomodulin 3, which interrupts actin poly-

merisation by capping the ends of actin molecules [37], was

found among sonication-freed and cell-fragment proteins of

CGTH W-1 cells only (Table 2B). Like actin, all of these

proteins migrated towards the anode within the FFE and

reached fractions with pH 5 5.45–5.86.

Various proteins that regulate cellular actin activities

were only found within the sonication-freed fractions

[38, 39]. Interestingly, these types of proteins showed vary-

ing electrophoretic mobilities. During FF-IEF, tropomyosin

a-4 chains and moesin of FTC-133 and CGTH W-1 cells

migrated towards the anode, whereas cofilin 1 of both cell

lines moved towards the cathode. In addition, cofilin 2 and

ezrin of the FTC-133 cells were found lying at the cathodal

side of the injection point in the FFE fractions. However,

profilin 2 showed up in the FFE-fractions with pH 5

5.45–5.86 when the CGTH W-1 proteins were studied but in

fractions with pH 5 7.3–7.5 when the FTC-133 cells were

investigated (Table 2C). Whether or not different electro-

phoretic mobilities of a protein indicate different phos-

phorylation remains to be determined.

3.5 High MW proteins

Considerable differences between the cell lines were recog-

nised when cell-fragment proteins were examined that had

migrated to fractions with pH 5.45–5.86 during FF-IEF and

to sites indicating apparent MWs between 110 and 160 kDa

during the subsequent SDS gel electrophoresis. Relevant gel

pieces contained a number of proteins involved in main-

taining cell structure, cell motility and tissue organisation.

The CAP-Gly domain-containing linker protein 1, integrin

a-5, integrin a-2, kinectin, desmuslin and dynactin subunit

1 were all detected, which have MWs as expected based on

the sites of excision of the gel pieces [40–42]. Of these

100# *

GAPDH100

GAPDH100

* #

GAPDH

60

80

60

80 # #

60

80

pro

tein

sito

met

ric u

nit * #

20

40

pro

tein

sito

met

ric u

nit

Act

inp

Rel

ativ

e de

ns

20

40

n p

rote

inV

imen

tin si

tom

etric

uni

tR

elat

ive

dens

20

40

Tu

bu

linβ

Rel

ativ

e de

ns1g RPM

79 7161 630

FTC-133 CGTH W-1

26 6132 330

FTC-133 CGTH W-1

60 6965 260

FTC-133 CGTH W-1

1g RPM

A B C

Figure 2. Western blot analyses: Whole cell lysates of the cell lines FTC-133 and CGTH W-1 (n 5 6, each group) were analysed for vimentin

(A), tubulin b (B) and actin (C) after 3 days of culturing under the conditions of 1 g (black columns) and RPM-exposure (grey columns).

Western blot membranes of the respective proteins and of GAPDH are shown above the diagrams. Densitometric values of each protein

band were normalised to the densitometric values of the respective GADPH bands. Significant differences are shown either with a cross

(] po0.005) or with a star (�po0.05). RPM, cells had been cultured on the RPM prior to analysis.



proteins, the CAP-Gly domain-containing linker protein 1,

kinectin, desmuslin and dynactin subunit 1 were found

regardless of whether proteins of CGTH W-1 or FTC-133

cells had been loaded onto the gel, whereas integrin a-5

chains were only detected in FTC-133 cells, and integrin a-2

together with collagen were only detected in CGTH W-1 cell

proteins.

When analysing the gel pieces mentioned above, addi-

tional proteins such as myosin-9, myosin-10, spectrin

a-chain, spectrin b-chain, filamin A, filamin B, fibronectin

and plectin-1 were identified [43–46]. Their MWs were

higher than indicated by the sites where the gel pieces

were excised (http://expasy.org/sprot/). The migration of

these proteins to a gel site where lower MW proteins were

expected was obviously due to proteolytic cleavage, which

may have occurred during the preparation of the proteins.

For example, myosin-9 was found in the gel pieces

mentioned above and in gel pieces where proteins could be

expected with an MW of around 230 kDa. When myosin-9 of

the latter gel pieces was analysed by LC-MS analysis, its

peptides were found to be distributed over the entire

sequence of the protein (Fig. 3). The peptides found for the

Table 2. M-scores of actin potein (A) and of proteins interacting with (B) or regulating actin (C)

kDa Sonication-freed Cell-fragment

CGTH W-1 FTC-133 CGTH W-1 FTC-133

1 g RPM 1 g RPM 1 g RPM 1 g RPM

A: Actin protein

Actin protein 43 5117 (A1) 6984 (A1) 5195 (A1) 3317 (A1) 2224 (A1) 4575 (A1) 516 (A1) 213 (A1)

B: Actin-binding proteins

Myosin-9 4110 482 (A1) 204 (A1) 3666 (A1) 3814 (A1) 4962 (A1) 3502 (A1)Plectin 4110 178 (A1) 627 (A1) 1263 (A1) 385 (A1) 260 (A1)Filamin A 4110 534 (A1) 400 (A1) 332 (A1) 264 (A1) 1720 (A1) 2342 (A1) 1481 (A1) 675 (A1)Actinin a 4 90 689 (A1) 692 (A1)Tropomodulin 3 42 399 (A1) 395 (A1) 479 (A1) 1036 (A1)

C: Actin-regulating proteins

Moesin 65 2205 (A1) 2637 (A1) 2026 (A2) 1717 (A2)Tropomyosin a 4 30 389 (A1) 826 (A1) 2205 (A1)Profilin 1 17 145 (C1) 383 (C1)Profilin 2 17 149 (A1) 142 (A1) 4151 (C2) 5022 (C2)Cofilin 1 20 323 (C1) 261 (C1) 739 (C2) 1305 (C2)Cofilin 2 17 332 (C2) 314 (C2)Ezrin 42 245 (C2)

The kDa specification refers to the excised and analysed SDS-gel pieces. RPM, cells had been cultured on the RPM prior to analysis; A1,FFE fractions with pH 5.45–5.86; A2, FFE fractions with pH 6.5–6.7; C2, FFE fractions with pH 7.3–7.5; C1, FFE fractions with pH 8.8–9.2.

30 kDa1g

140 kDa

14 kDaRPM

RPM

FTC-133

140 kDa1g

140 kDa

140 kDaRPM

1g

230 kDaRPM

CGTH W-1

230 kDa1g

N-termini C-termini1 1960

Amino acid sequence of Myosin-9

Figure 3. Sequence analysis of

myosin-9. The x-axis repre-

sents the amino acid sequence

of myosin-9 (1–1960 amino

acids). The columns above

reflect the consensus between

the whole myosin-9 sequence

(grey background) and the

peptides found (black bars).

The origin of the protein is

indicated on the left-hand side

of the columns. The kDa

specification on the right-hand

side of the columns refers to

the excised and analysed SDS-

gel pieces. 1 g, the cells had

been cultured at 1 g prior to

analysis; RPM, the cells had

been cultured on the RPM prior

to analysis.

Proteomics 2011, 11, 2095–2104 2101


apparent 140 kDa myosin-9 were spread over the c-terminal

60% of the protein’s sequence only, whereas peptides that

fitted into the head-like region of myosin-9 were detected at

sites in the gels where both 14 and 30 kDa proteins migra-

ted. This suggests that the myosin-9 head-like part is more

sensitive to proteases. Despite a possible partial digestion,

myosin-9, spectrin a-chain, spectrin b-chain, filamin A,

fibronectin and plectin-1 were found when comparable gel

pieces of both cell lines were studied. In contrast, myosin-

10, filamin B and chondroitin sulphate were only found in

gels loaded with cell-fragment proteins of the FTC-133 cell

line, whereas various types of collagen chains were only

determined in gels loaded with cell-fragment proteins of the

CGTH W-1 cells.

Taking the results altogether, both cell lines showed

different quantities of some structural and membrane-

associated proteins under normal 1 g conditions and of

others, only after RPM exposure (Fig. 2, Table 1). In addi-

tion, proteins were found, which showed altered migration

behaviour within the FFE depending on the original cell line

(Table 2). Major differences between the two cell lines were

observed with regard to vimentin, the tubulins, integrin,

myosin-10 and filamin B. At the moment, it cannot be

decided which of these proteins are responsible for spheroid

formation and it is still unclear which of them cause the

diverging sizes of the spheroids shown in Fig. 1, when

they are expressed at different basic levels. However, it

should be noted that the integrin a-5-chains, together with

the integrin b-1-chains, myosin-10 and filamin B, were

found in protein solutions of the FTC-133 cells only, where

the expression of b tubulin was nearly independent

of RPM exposure but that of actin was rather sensitive to

annulling gravity. It was recently reported that integrin

a-5-b1, together with fibronectin, contributes significantly

to the cohesion of tissue in vivo and of the three-

dimensional aggregates in vitro [47, 48]. Furthermore,

myosin 10 and filamin B can bind to actin and to integrin

b-chains [49, 50]. Thus, it might be that FTC-133 cells, but

not CGTH W-1 cells, use fibronectin for building the large

spheroids, even though the protein was produced by both

cell lines.

4 Concluding remarks

The application of FF-IEF permitted the characterisation of

a number of both abundant and less frequent membrane-

associated proteins that are involved in maintaining cell

structure, cell mobility and tissue organisation. After pre-

separation by FF-IEF, the MS analysis revealed differences

in protein expression between the two related human

carcinoma cell lines, which formed spheroids in different

ways when exposed to the RPM. This distinguishable

spheroid-forming behaviour was probably due to the inter-

action between several proteins, and surface proteins bind-

ing fibronectin could play a significant role in this context.

This study was supported by the European Space Agency(CORA-GBF-2004-003-II and CORA-GBF-2010-202), theGerman Space Agency DLR (BMWi grant 50WB0824), by theETH Zurich, Switzerland and by the ‘‘Ministerium f .ur Inno-vation, Wissenschaft, Forschung und Technologie des LandesNordrhein-Westfalen’’ and the ‘‘Bundesministerium f .ur Bildungund Forschung.’’

The authors have declared no conflict of interest.

5 References

[1] Discher, D. E., Janmey, P., Wang, Y. L., Tissue cells feel and

respond to the stiffness of their substrate. Science 2005,

310, 1139–1143.

[2] Yuhas, J. M., Li, A. P., Martinez, A. O., Ladman, A. J.,

A simplified method for production and growth of

muticellular tumor spheroids. Cancer Res. 1977, 37,

3639–3643.

[3] Sutherland, R. M., Sordat, B., Bamat, J., Gabbert, H. et al.,

Oxygenation and differentiation in multicellular spheroids

of human colon carcinoma. Cancer Res. 1986, 46,

5320–5329.

[4] White, R. J., Averner, M., Humans in space. Nature 2001,

409, 1115–1118.

[5] Barrila, J., Radtke, A. L., Crabbe, A., Sarker, S. F. et al.,

Organotypic 3D cell culture models: using the rotating wall

vessel to study host-pathogen interactions. Nat. Rev.

Microbiol. 2010, 8, 791–801.

[6] Grimm, D., Bauer, J., Infanger, M., Cogoli, A., The use of the

random positioning machine for the study of gravitational

effects on signal transduction in mammalian cells. Signal

Transduct. 2006, 6, 388–396.

[7] Grimm, D., Bauer, J., Kossmehl, P., Shakibaei, M. et al.,

Simulated microgravity alters differentiation and increases

apoptosis in human follicular thyroid carcinoma cells.

FASEB J. 2002, 16, 604–616.

[8] Infanger, M., Kossmehl, P., Shakibaei, M., Baatout, S. et al.,

Induction of three-dimensional assembly and increase in

apoptosis of human endothelial cells by simulated micro-

gravity: impact of vascular endothelial growth factor.

Apoptosis 2006, 11, 749–764.

[9] Grimm, D., Infanger, M., Westphal, K., Ulbrich, C. et al., A

delayed type of three-dimensional Growth of human

endothelial cells under simulated weightlessness. Tissue

Eng. Part A 2009, 15, 2267–2275.

[10] Ulbrich, C., Westphal, K., Pietsch, J., Winkler, H. D. F. et al.,

Characterization of human chondrocytes exposed to simu-

lated microgravity. Cell. Physiol. Biochem. 2010, 25,

551–560.

[11] Pardo, S. J., Patel, M. J., Sykes, M. C., Platt, M. O. et al.,

Simulated microgravity using the Random Positioning

Machine inhibits differentiation and alters gene expression

profiles of 2T3 preosteoblasts. Am. J. Physiol. Cell Physiol.

2005, 288, C1211–C1221.



[12] Moebius, J., Zahedi, R. P., Lewandrowski, U., Berger, C.

et al., The human platelet membrane proteome reveals

several new potential membrane proteins. Mol. Cell.

Proteomics 2005, 4, 1754–1761.

[13] Islinger, M., Eckerskorn, C., Volkl, A., Free flow electro-

phoresis in the proteomic era: a technique in flux. Electro-

phoresis 2010, 31, 1754–1763.

[14] Wisniewski, J. R., Zougman, A., Mann, M., Combination of

FASP and stage tip-based fractionation allows in-depth

analysis of the hippocampal membrane proteome.

J. Proteome Res. 2009, 8, 5674–5678.

[15] Obermaier, C., Jankowski, V., Schmutzler, C., Bauer, J.

et al., Free-flow isoelectric focusing of proteins remaining in

cell fragments following sonication of thyroid carcinoma

cells. Electrophoresis 2005, 26, 2109–2116.

[16] Pietsch, J., Kussian, R., Sickmann, A., Bauer, J. et al.,

Application of free-flow IEF to identify protein candidates

changing under microgravity conditions. Proteomics 2010,

10, 904–913.

[17] Goretzki, P. E., Frilling, A., Simon, D., Roeher, H. D., Growth

regulation of normal thyroids and thyroid tumors in man.

Recent Results Cancer Res. 1990, 118, 48–63.

[18] Lin, J. D., Chao, T. C., Weng, H. F., Huang, H. S., Ho, Y. S.,

Establishment of xenografts and cell lines from well-differ-

entiated human thyroid carcinoma. J. Surg. Oncol. 1996, 63,

112–118.

[19] Hughes-Fulford, M., Function of the cytoskeleton in grav-

isensing during spaceflight. Adv. Space Res. 2003, 32,

1585–1593.

[20] van Loon, J. J. W. A., Some history and use of the random

positioning machine, RPM, in gravity related research. Adv.

Space Res. 2007, 39, 1161–1165.

[21] Kossmehl, P., Schonberger, J., Shakibaei, M., Faramarzi, S.

et al., Increase of fibronectin and osteopontin in porcine

hearts following ischemia and reperfusion. J. Mol. Med.

2005, 83, 626–637.

[22] Nissum, M., Kuhfuss, S., Hauptmann, M., Obermaier, C.

et al., Two-dimensional separation of human plasma

proteins using iterative free-flow electrophoresis. Proteo-

mics 2007, 7, 4218–4227.

[23] Winkler, C., Denker, K., Wortelkamp, S., Sickmann, A.,

Silver- and coomassie-staining protocols: detection limits

and compatibility with ESI MS. Electrophoresis 2007, 28,

2095–2099.

[24] Perkins, D. N., Pappin, D. J., Creasy, D. M., Cottrell, J. S.,

Probability-based protein identification by searching

sequence databases using mass spectrometry data. Elec-

trophoresis 1999, 20, 3551–3567.

[25] Ishihama, Y., Oda, Y., Tabata, T., Sato, T. et al., Exponen-

tially modified protein abundance index (emPAI) for esti-

mation of absolute protein amount in proteomics by the

number of sequenced peptides per protein. Mol. Cell.

Proteomics 2005, 4, 1265–1272.

[26] Chang, T. T., Hughes-Fulford, M., Monolayer and spheroid

culture of human liver hepatocellular carcinoma cell line

cells demonstrate distinct global gene expression patterns

and functional phenotypes. Tissue Eng. 2009, 15, 559–567.

[27] Ivaska, J., Pallari, H. M., Nevo, J., Eriksson, J. E., Novel

functions of vimentin in cell adhesion, migration, and

signalling. Exp. Cell Res. 2007, 313, 2050–2062.

[28] Weber, G., Bauer, J., Counterbalancing hydrodynamic

sample distortion effects increases resolution of free-flow

zone electrophoresis. Electrophoresis 1998, 19, 1104–1109.

[29] Desai, A., Mitchison, T. J., Microtubule polymerization

dynamics. Annu. Rev. Cell Dev. Biol. 1997, 13, 83–117.

[30] Lewis, M. L., Reynolds, J. L., Cubano, L. A., Hatton, J. P. et al.,

Spaceflight alters microtubules and increases apoptosis in

human lymphocytes (Jurkat). FASEB J. 1998, 12, 1007–1018.

[31] Papaseit, C., Pochon, N., Tabony, J., Microtubule self

organization is gravity dependent. Proc. Natl. Acad. Sci.

USA 2000, 97, 8364–8368.

[32] Pollard, T. D., Borisy, G. G., Cellular motility driven by

assembly and disassembly of actin filaments. Cell 2003,

112, 453–465.

[33] Watanabe, N., Inside view of cell locomotion through

single-molecule: fast F-/G-actin cycle and G-actin regulation

of polymer restoration. Proc. Natl. Acad. Sci. USA 2010, 86,

62–83.

[34] Nebl, T., Pestonjamasp, K. N., Leszyk, J. D., Crowley, J. L.

et al., Proteomic analysis of a detergent-resistant

membrane skeleton from neutrophil plasma membranes. J.

Biol. Chem. 2002, 277, 43399–43409.

[35] Fontao, L., Geerts, D., Kuikman, I., Koster, J. et al., The

interaction of plectin with actin: evidence for cross-linking

of actin filaments by dimerization of the actin-binding

domain of plectin. J. Cell Sci. 2001, 114, 2065–2076.

[36] Feng, Y. Y., Walsh, C. A., The many faces of filamin: a

versatile molecular scaffold for cell motility and signalling.

Nat. Cell Biol. 2004, 6, 1034–1038.

[37] Fischer, R. S., Yarmola, E. G., Weber, K. L., Speicher, K. D.

et al., Tropomodulin 3 binds to actin monomers. J. Biol.

Chem. 2006, 281, 36454–36465.

[38] Tsukita, S., Yonemura, S., Tsukita, S., ERM proteins: head-

to-tail regulation of actin-plasma membrane interaction.

Trends Biochem. Sci. 1997, 22, 53–58.

[39] Bugyi, B., Carlier, M. F., Control of actin filament treadmilling

in cell motility. Annu. Rev. Biophys. 2010, 39, 449–470.

[40] Mizuno, Y., Thompson, T. C., Guyon, J. R., Lidov, H. G. W.

et al., Desmuslin, an intermediate filament protein that

interacts with alpha-dystrobrevin and desmin. Proc. Natl.

Acad. Sci. USA 2001, 98, 6156–6161.

[41] Ross, J. L., Wallace, K., Shuman, H., Goldman, Y. E., Holz-

baur, E. L. F., Processive bidirectional motion of dynein-

dynactin complexes in vitro. Nat. Cell Biol. 2006, 8, 562–570.

[42] Tran, H., Pankov, R., Tran, S. D., Hampton, B. et al., Integrin

clustering induces kinectin accumulation. J. Cell Sci. 2002,

115, 2031–2040.

[43] Dahl, S. C., Geib, R. W., Fox, M. T., Edidin, M., Branton, D.,

Rapid capping in alpha-spectrin-deficient MEL cells from

mice afflicted with hereditary hemolytic anemia. J. Cell Biol.

1994, 125, 1057–1065.

[44] Wiche, G., Role of plectin in cytoskeleton organization and

dynamics. J. Cell Sci. 1998, 111, 2477–2486.

Proteomics 2011, 11, 2095–2104 2103


[45] Esue, O., Tseng, Y., Wirtz, D., Alpha-actinin and filamin

cooperatively enhance the stiffness of actin filament

networks. PLoS One 2009, 4, Article Number: e4411.

[46] Vincentr-Manzanares, M., My, X., Adelstein, R. S., Horwitz,

A. R., Non-muscle mysosin II takes centre stage in cell adhe-

sion and migration. Nat. Rev. Mol. Cell Biol. 2009, 10, 778–790.

[47] Caicedo-Carvajal, C. E., Shinbrot, T., Foty, R. A., Alpha 5

beta 1 integrin-fibronectin interactions specify liquid to

solid phase transition of 3D cellular aggregates. PLoS One

2010, 5, Article Number: e11830.

[48] Robinson, E. E., Foty, R. A., Corbett, S. A., Fibronectin

matrix assembly regulates alpha 5 beta 1-mediated cell

cohesion. Mol. Biol. Cell 2004, 15, 973–981.

[49] Zhang, H. Q., Berg, J. S., Li, Z. L., Wang, Y. L. et al., Myosin-

X provides a motor-based link between integrins and the

cytoskeleton. Nat. Cell Biol. 2004, 6, 523–531.

[50] van der Flier, A., Kuikman, I., Kramer, D., Geerts, D. et al.,

Different splice variants of filamin-B affect myogenesis,

subcellular distribution, and determine binding to integrin

beta subunits. J. Cell Biol. 2002, 156, 361–376.



Documents

A proteomic approach to analysing spheroid formation of two human thyroid cell lines cultured on a random positioning machine