mesothelial cells can detach from the mesentery and differentiate into macrophage-like cells

Mesothelial cells can detach from the mesentery

and differentiate into macrophage-like cells

SANDOR KATZ, PETRA BALOGH and ANNA L. KISS

Department of Human Morphology and Developmental Biology, Semmelweis University, Budapest, Hungary

Katz S, Balogh P, Kiss AL. Mesothelial cells can detach from the mesentery and differentiate intomacrophage-like cells. APMIS 2011; 119: 782–93.

Peritoneal cell suspension is composed of heterogeneous cell population. Macrophages are the mostnumerous cells among them. They can originate from different sources and can be resident, exudateand elicited. When we used Freund’s adjuvant to elicit peritoneal macrophages, cells having largeamount of caveolae on their plasma membrane appeared in the peritoneal wash. The number of thesecaveolae-rich cells increased by the time of the Freund’s adjuvant treatment. Although their morphol-ogy was different form from the common macrophages, they were labelled with pan-macrophage anti-bodies. As the origin of these cells is unknown in this work, we tried to find out where they canoriginate from. Our interest turned towards the mesothelial cells. We found that the adjuvant treatmentresulted in significant morphological changes in these cells and stimulate them to leave the surface ofthe mesentery. By the time of the adjuvant treatment, the macrophage markers expression increased inthe mesothelial cells and more cells were found to detach from the mesentery. These results stronglysuggest that under special stimuli mesothelial cells can leave the mesentery and differentiate into phago-cytotic (macrophage-like) cells. These data raises the idea that mesothelial cells might not entirely dif-ferentiated and represent a multipotential cell lineage. To study whether this is the case we used anti-nestin antibody, which is a specific marker for multifunctional, multi-lineage progenitor cells. Mesothe-lial cells showed strong labelling with this antibody indicating that these cells really represent a ‘young’,not entirely differentiated cell population.

Key words: Caveolae; macrophages; mesothelial cells; multi-lineage progenitor cells.

Anna L. Kiss, Department of Human Morphology and Developmental Biology, SemmelweisUniversity, Budapest, T}uzolto u. 58. Hungary. e-mail: [email protected]

Macrophages are the major differentiated ele-ments of the mononuclear phagocytotic system.Macrophages are widely distributed in many tis-sues and organs of the body and display diversemorphological and functional characteristics.These highly active phagocytes are found inlarge number in the peritoneal cavity, wherethree groups of phagocytes can be distinguished:(i) resident (or tissue) and (ii) exudate macro-phages (1). The resident macrophages reside inthe peritoneal cavity under normal (steady state)conditions; they are self-sustaining by local

proliferation in the ‘milky spots’ of the mesen-tery (2). Resident macrophages can also origi-nate from precursors migrating into theperitoneal cavity. Exudatemacrophages are spe-cifically derived from monocytes and representthe population of phagocytes which enters theperitoneal cavity (3). A particular stimulusresults in an accumulation of a heterogeneouspopulation of mononuclear phagocytes, called(iii) elicitedmacrophages (4).In our previous work, we used complete Fre-

und’s adjuvant injection to elicit macrophagesand found that the number of peritoneal macro-phages has significantly increased. It was alsoReceived 28 February 2011. Accepted 25 July 2011

APMIS 119: 782–793 � 2011 The Authors

APMIS � 2011 APMIS

DOI 10.1111/j.1600-0463.2011.02803.x

782

surprising that cells having large amount ofcaveolae on their plasma membrane appeared inthis elicited cell population (5). As these cellswere labelled with a macrophage marker (ED1),we called them caveolae-rich macrophages (6).We were very much interested in where thesecells could originate from. As the mesothelialcells are present all over the peritoneal cavity,covering the internal organs of the abdomen, weasked whether these cells can contribute to theincreased number of phagocytotic cells. It is wellknown that mesothelial cells are multifunction-al, can take part in antigen presentation bysecreting interleukin (IL)-15 (7) and play impor-tant roles in tumour cell adhesion and growth(8), coagulation and fibrinolysis (9). Theysecrete various immunomodulators by whichthey can also take part in inflammatory reac-tions and tissue repair (10). Mesothelial cells areactively involved in transport and movement offluid and cells across the serosal cavities (11);they represent a protective barrier against physi-cal damage and invading organisms.Taking together all these facts, our question

was whether these multifunctional cells coulddetach from the mesentery and differentiate intophagocytotic cells. To answer this question, wecarried out detailed morphological and immu-nocytochemical studies on mesothelial cells ofcontrol and complete Freund’s adjuvantinjected rats’ mesentery. Our light and electronmicroscopic results revealed that Freund’s adju-vant injection (elicitation) has significantlychanged the morphology of these simple squa-mous epithelial cells. They became rounded,rather cuboidal, many of them seem to detachfrom the basement membrane. By the time ofthe elicitation, numerous free, macrophage-likecells were found close to the surface of the mes-entery. These free, macrophage-like cells andmany mesothelial cells as well were labelled withmacrophage-specific markers (ED1 and OX43).In mesothelial cells, both macrophage markers’expression has significantly increased by thetime of the adjuvant treatment. When we useddouble labelling with mesothelin (a specificmesothelial cell marker) and ED1, mesothelialcells as well as free cells close to the mesenterywere found to be labelled with both mesothelinand ED1. The number of double labelled cellshas significantly increased by the time of elicita-tion. These data strongly support the idea that

under special stimuli, mesothelial cells cantransdifferentiate into phagocytotic (macro-phage-like) cells, indicating that they are notentirely differentiated or they still keep theirmultipotential character. To study whethermesothelial cells represent a ‘young’, multipo-tential cell population, we used anti-nestin anti-body. Nestin is a well-known and characteristicmarker of multi-lineage progenitor cells; itspresence indicates multipotentiality and regen-erative capacity (12). We found that mesothelialcells were strongly labelled with anti-nestin anti-body, indicating that these cells are really notentirely differentiated cells.Intermediate filaments are important indica-

tors of epithelial differentiation (13). Rat meso-thelial cells co-express vimentin (mesenchymal)and the simple epithelial cytokeratins (14).While cytokeratins predominate in situ, the lossof cytokeratin production can be a sign ofchanging the phenotype. The Freund’s adjuvanttreatment resulted in a loss of cytokeratinexpression in rat mesentheric mesothelial cells.These data support the idea that mesothelialcells have the ability to change their phenotypeand can undergo an epithelial to mesenchymaltransition.

MATERIALS AND METHODS

Materials

Mesentery isolated from control and elicited Spra-gue–Dawley male rats (200–400 g). For elicitation,1 mL complete Freund’s adjuvant (Sigma, Steinheim,Germany) was injected into the peritoneal cavity (24and 48 h). The Freund’s adjuvant is a solution ofantigen emulsified in mineral oil and used as animmuno potentiator. The complete form is composedof inactivated and dried mycobacteria (Mycobacte-rium tuberculosis), has a very low toxicity. Injectedintraperitonally, it can cause peritonitis.

Antibodies

Polyclonal anti-caveolin-1 antibody (1:100) was pur-chased from Transduction Laboratories. Macrophagespecific antibody (OX43) was obtained from Dr PeterBalogh Department of Immunology and Biotechnol-ogy, University of Pecs (Hungary). Monoclonal anti-body ED1 (1:400), a rat macrophage marker, was agenerous gift of Prof. Dr C.D. Dijkstra (Departmentof Cell Biology, Medical Faculty, Vrije University,

MESOTHELIAL CELL CAN BECOMEMACROPHAGE

� 2011 The Authors APMIS � 2011 APMIS 783

Amsterdam, the Netherlands). Anti-rat mesothelin(1:200) was purchased from Immuno-Biological Lab-oratories Co. Ltd (Takashaki, Gumma, Japan). Anti-nestin (1:45) antibody was purchased from MilliporeChemicon, Atlanta, GA, USA. Anti-cytokeratin anti-body (Lu5, 1:200) was purchased from Biomedicals,Augst, Switzerland.

For confocal microscopy, Alexa Fluor 488 (Molec-ular Probes, Leiden, the Netherlands)-conjugated sec-ond antibodies (1:100) were used. The nuclei werelabelled by DAPI (Vectashield DAPI mounting med-ium, Burlingame, Canada).

Experimental procedures

Mesentery was isolated from both control and Fre-und’s adjuvant-injected animals. The samples werefixed either in 2% glutaraldehyde (GA) in Millonig’s(0.2 M NaH2PO4 and 0.5 M NaOH) phosphate buf-fer (1 h, room temperature) or 4% formaldehyde(FA) in 0.1 M phosphate buffer (PBS, containing13 mM NaCl, 2 mM KCl, 8 mM Na2HPO4, 1.5 mMKH2PO4, pH 7.4). The fixation was followed bywashing in PBS and the adipose tissue was removedfrom the surface of the mesentery. The GA-fixedmaterial was proceeded to electron microscopicembedding, while the FA-fixed samples were used forimmunocytochemistry.

Immunocytochemistry

The FA-fixed samples were stored until further pro-cessing in 1% formaldehyde at 4 �C. For semithincryosectioning and immunolabelling, the fixed sam-ples were washed with 0.05 M glycin in PBS, infil-trated gradually with gelatine at 37 �C (1%, 2% 5%and 10% in PBS). The 10% gelatine (containing mes-entery) was solidified on ice and cut into small blocks.For cryoprotection, the blocks were infiltrated over-night with 2.3 M sucrose at 4 �C and afterwardsmounted on aluminium pins and frozen in liquidnitrogen. The 0.8-lm thick frozen sections were cutusing Leica Ultracut S ultramicrotome (Vienna ⁄Aus-tria). To pick up the sections 1:1 mixture of 2.3 Msucrose and 1.8% methylcellulose was used (15). Thesections were incubated with the first antibodies at4 �C (overnight), after washing this was followed by1 h incubation (at room temperature) with the secondantibody. In some cases, we used double labellingimmunocytochemistry to detect mesothelin and ED1.Bio-Rad (Ontario, Canada) Radiance 2100 Rainbowconfocal microscope was used for confocal images.

Electron microscopy

The GA-fixed samples were washed in Millonig’s buf-fer, which was followed by washing in 0.1%

cacodylate buffer and postfixation in 1% OsO4 (1 h at4 �C in 0.1% cacodylate buffer, pH 7.4). A pre-embedding staining with 1% uranyl acetate in dis-tilled water was performed for 1 h at 4 �C. The sam-ples were dehydrated and embedded in araldite. Thesemithin sections were stained with toluidine blue. Ul-trathin sections were contrast stained with lead cit-rate.

Quantitation methods

To determine the ratio of mesothelin-positive cells ineach experimental group, 150 cells were photo-graphed at random. The mesothelin-labelled cellswere counted and expressed as percentage of the totalnumber of cells present in the peritoneal washes.To calculate the ratio of the caveolae-containing

cells, 50 electron micrographs were randomly takenfrom each experimental group with a basic magnifica-tion of ·700. Caveolae-containing cells as well as thetotal number of the cells were counted. The resultswere expressed as percentage of the total number ofthe cells.

RESULTS

Morphology of the mesothelial cells: effect of Freund’s

adjuvant injection

Mesothelial cells present on the surface of thecontrol (non-treated) rats’ mesentery are elon-gated, flat cells. They rest on a thin basementmembrane. In toluidine blue-stained semithinsections, they can be identified according theirnuclei, their cytoplasm is hardly seen andappears as wavy lines (Fig. 1A). The connectivetissue between the two layers of mesothelial cellscontains blood vessels (arterioles and capillar-ies) and collagen fibres. There are only few cells(mainly fibrocytes, mast cells and some granulo-cytes) present in this layer (Fig. 1A). On elec-tron microscopic pictures, many ‘holes’, emptycircles are seen on both the basal and the apicalplasma membrane (Fig. 1B). Using higher mag-nification, it appeared that these ‘holes’ areflask- or omega-shaped, caveola-like plasmamembrane invaginations (Fig. 1C). Althoughthese membrane invaginations are present onboth the apical and basal plasma membrane,they are more numerous on the basal surface.The basement membrane is well seen beneaththe mesothelial cells, sometimes the long ancient(primary) cilium was found on the apical part of

KATZ et al.

784 � 2011 The Authors APMIS � 2011 APMIS

these cells (Fig. 1C). After 24 and 48 h of theFreund’s adjuvant injection, the morphology ofthe mesothelial cells significantly changed. Ourtoluidine blue-stained semithin sections clearlyshowed that the flat squamous cells becamerounded, rather cuboidal-shaped, some of themseem to leave the basement membrane

(Fig. 2A,C). Phagocytosed materials (dark par-ticles) and endocytosed lipid droplets from theFreund’s adjuvant (white vacuoles) are alsopresent in the cytoplasm of these cells. In somearea, the connective tissue became significantlythicker. After 48 h treatment, many free cellswere found close to the surface of the mesentery

A

C

B

Fig. 1. Light (A) and electron microscopic (B) pictures of mesentery isolate from control (non-treated) rats. Onthe toluidine-blue stained semithin sections (A) the mesothelial cells are seen as elongated, flat cells; they can beidentified according their nuclei (arrows), their cytoplasm appears as wavy lines. The connective tissue between thetwo layers of mesothelial cells contains blood vessels (arterioles and capillaries) and collagen fibres. There are onlyfew cells present in this layer. On electron microscopic pictures, (B) many ‘empty holes’ are seen on both the basaland apical surfaces. With higher magnification (C), the basement membrane is well seen beneath the control cell.Large amount of omega- or flask-shaped invaginations are present on both the apical and basal plasma mem-branes, but they are more numerous on the basal surface [This cell has long ancient cilium (white c) on its apicalpart]. C, capillaries; A, arteriole; M, mast cells; cf, collagen fibres; bm, basement membrane; n, nucleus. Bars: (A)20 lm; (B) 6 lm; (C) 1 lm.



(Fig. 2C). The morphology of these free cellswas very much similar to that of the mesothelialcells. More cells (mast cells, plasma cells, granu-locytes) appeared in the mesenteric connectivetissue. Our electron microscopic studies revealedthat long surfaces of the mesentery have nomesothelial cells on the basement membrane(Fig. 2D). Cells that were partly detached fromthe basement membrane were often seen.Although the omega-shaped invaginations werepresent on the plasma membrane of the Fre-und’s adjuvant-treated cells as well, they wereless numerous on the basal plasma membrane(Fig. 2B). When we used anti-caveolin-1 anti-body to identify caveolae, plasma membrane ofthe mesothelial cells present on both surface ofthe mesentery showed strong immunoreactivity(Fig. 3A–C), suggesting that these membraneinvaginations were really caveolae. As the mor-phology of the mesothelial cells has changed, we

could detect more delicate caveolin-1 labelling.Interestingly in the 48 h treated samples, thelabelling appeared also in the cytoplasm(Fig. 3C).Our quantitative data (Table 1) showed that

after Freund’s adjuvant treatment the numberof macrophages present in the peritoneal exu-dates was significantly (�10 or 100 times) higherindicating that the treatment induced a promi-nent cell migration into the peritoneal cavity.

Macrophage markers (ED1 and OX43) are expressed

in mesothelial cells

Peritoneal macrophages are known to expressED1 antigen – a specific pan-macrophage mar-ker in rat – on their plasma membrane as wellas on the membranes of the phagolysosomalapparatus (16). To study whether mesothelialcell can be transformed into macrophage-like

A C

D

Fig. 2. After 24 h adjuvant injection, the flat squamous cells become cuboidal-shaped (A). Phagocytized materi-als (dark particles) and endocytosed lipid droplets from the Freund’s adjuvant (white vacuoles) are also presentin their cytoplasm. Free cell (*) is also found close to the surface of the mesentery. The connective tissue betweenthe two-layered mesothelium contains larger amount of cells. Adjuvant treatment for 24 h changes the distribu-tion of the flask- or omega-shaped plasma membrane invaginations: fewer of them are present on the basal sur-face of the cell (B). After 48 h treatment with Freund’s adjuvant (C) round-shaped cells seem to detach from thebasement membrane. Free cells (*) are also present close to the surface of the mesentery. On our electron micro-scopic picture (D), it is well seen that larger areas of the mesentery have no mesothelial covering (arrows). Many‘free’ cells (*) can be found close to the surface. bm, basement membrane; G, Golgi apparatus; n, nucleus; M,mast cell. Bars: (A) 40 lm; (B) 1 lm; (C) 30 lm; (D) 10 lm.

KATZ et al.


cells, we used this pan-macrophage marker.Mesothelial cells of control non-treated rats’mesentery were not labelled with ED1 anti-body (Fig. 4A). Freund’s adjuvant treatmentresulted in ED1 expression: our double label-ling experiments clearly showed that mesoth-elin-labelled mesothelial cells started to expressED1. After 24 h adjuvant treatment cells withdelicate ED1 labelling on their plasma mem-brane as well as strongly double labelled cellswere present on the surface of the mesentery(Fig. 4B). Forty-eight-hour adjuvant treatmentincreased the number of double labelled cells(Fig. 4C). The cells that were already free, not

connected to the mesentery were often foundto be double labelled (Fig. 4D). In some areaof the mesentery, where mesothelial cellsseemed to move apart (to open like a ‘gate’),ED1 positive (but mesothelin negative) cellsseem to migrate out from the connective tissue(Fig. 4B). Using another macrophage marker(OX43), we found similar results (not showed).These data clearly showed that special stimulican induce mesothelial cells to express macro-phage markers indicating that they mighttransdifferentiate into macrophages (or macro-phage-like cells). These transdifferentiated cellscan detach from the basement membrane and

A B

C

Fig. 3. Caveolin-1 detection on semithin frozen section. Caveolin-1 labelling (green) indicates the presence ofcaveolae on the plasma membrane. (A) As the control mesothelial cells are flat and the basal and apical cell mem-branes are difficult to discriminate, caveolin-1 labelling appears as a thick line. In some cells, where the nuclei arelocated in the section plane, strong labelling is seen on both the apical and the basal plasma membranes. Cellspresent in the connective tissue are slightly labelled with anti-caveolin-1 antibody. (B) After 24 h Freund’s adju-vant treatment both the basal and the apical plasma membranes are strongly labelled with anti-caveolin-1. (C) Inthe 48 h treated sample, a delicate but definitive caveolin-1 labelling can be detected on the plasma membrane aswell as in the cytoplasm of the mesothelial cells. Some cells (mainly fibroblasts) present in the connective tissueare also labelled with anti-caveolin-1 (B and C). Nuclei are labelled with DAPI (blue). Bars: (A) 36 lm; (B)34 lm; (C) 38 lm.



contribute to the increased number of phago-cytotic cells.

Cytokeratin expression in mesothelial cells

By the time of the Freund’s adjuvant treatment,the cytokeratin expression has significantlydecreased in rat mesenteric mesothelial cells.Although the non-treated mesothelial cells areflat, their cytoplasm is hardly seen, all of themwere found to be labelled by anti-cytokeratinantibody (not showed). After 24 h Freund’sadjuvant treatment, the morphology of the cellsstarted to change, the cytokeratin labelling isbetter seen in the cytoplasm (Fig. 5A). All themesothelial cells were labelled with anti-cytoker-atin antibody; 48 h treatment resulted in a sig-nificant decrease in cytokeratin expression. Wecould detect only a few cells with a delicate cyto-keratin network in their cytoplasm (Fig. 5B).The majority of the mesothelial cells, however,did not show cytokeratin labelling. Sometimeswe could find cytokeratin-positive mesothelialcells detaching from the surface of the mesen-tery (Fig. 5B). When we used double labellingwith anti-mesothelin and anti-cytokeratin anti-bodies, we found that after 48 h of the treat-ment, the majority of the mesothelial cells havealready lost their cytokeratin (Fig. 5C).

Mesothelial cells are nestin positive

To check whether mesothelial cells are young,not entirely differentiated cells, we used anti-nestin antibody. On the surface of the controlmesentery, mesothelial cells were labelled withanti-nestin antibody (Fig. 6A). Cells present inthe mesenteric connective tissue did not showlabelling with this antibody. In the 24 h adju-vant-treated mesothelium, the nestin labellingwas similar. Cells unambiguously detachingfrom the basement membrane and those thathave already lost their connection with the mes-entery were nestin positive (Fig. 6B). In the 48 h

sample, all mesothelial cells on the surface ofthe mesentery showed a delicate, but definitivenestin labelling (Fig. 6C). It has to be pointedout that Freund’s adjuvant treatment increasedthe number of the nestin-positive cells presentin the mesenteric connective tissue as well(Fig. 6C).

DISCUSSION

In earlier studies, we injected complete Freund’sadjuvant into rats’ peritoneal cavity to elicitmacrophages (for 24 and 48 h). We found thatby the time of the elicitation, the number ofperitoneal cells has significantly increased.Although the peritoneal cells are heterogeneous,macrophages are the most dominant cells in thiscell population. Our electron microscopicstudies revealed that after Freund’s adjuvantinjection, cells having numerous omega- orflask-shaped membrane invaginations on theirplasma membrane appeared in the peritonealwash. These membrane invaginations werestrongly labelled with anti-caveolin antibodyindicating that they were caveolae (5). We calledthem caveolae-rich cells (6). By the time of theelicitation, the number of macrophages as wellas these caveolae-rich cells has significantlyincreased in the peritoneal wash. Although themorphology of these cells was different from thetraditional macrophages, their labelling with amonoclonal antibody (ED1) suggests that theywere macrophages. (ED1 exclusively recognizescells of the mononuclear phagocyte system andwidely used as a pan-macrophage marker in therat; 16–18). The question arose as to what canbe the source of these phagocytotic cell invadingto the peritoneal cavity and where these caveo-lae-rich cells (labelled with macrophage marker)could originate from. Although we often foundED1-positive cells migrating out from the meso-thelium-free parts of the mesentery, we do thinkthat the two well-known sources of peritoneal

Table 1. Quantitative data about the number of macrophages, the ratio of caveolae-rich macrophages andmesothelin positive cells present in the peritoneal wash

Control 24 h elicitation 48 h elicitation

Number of macrophages ⁄mL peritoneal wash 8 · 106 8 · 107 2 · 108

Ratio of caveolae-rich macrophages (%) 0.5 ± 0.01 3–6 ± 0.02 18–20 ± 0.005Ratio of mesothelin-positive cells (%) 0.7 ± 0.01 5 ± 0.01 15 ± 0.02

KATZ et al.


macrophages – monocyte-derived exudate mac-rophages and locally proliferating resident mac-rophages – (1) would not provide such a largeamount of cells. Mesothelial cells are present allover the body taking part in formation of serousmembrane. As electron microscopic studiesshowed that both the apical and the basalplasma membrane of these cells had numerousomega- or flask-shaped invaginations, calledcaveolae (19), our interest turned towards them.Mesothelial cells are known to be multifunc-tional: they can take part in antigen presenta-tion by secreting interleukin IL-15 (7), they playimportant roles in tumour cell adhesion andgrowth (8), coagulation and fibrinolysis (9).

They secrete various immunomodulators bywhich they can also take part in inflammatoryreactions and tissue repair (10) and are activelyinvolved in transport and movement of fluidand cells across the serosal cavities (11). Meso-thelial cells represent a protective barrier againstphysical damage and invading organisms.Focusing on the mesentery, we found that the

Freund’s adjuvant treatment has significantlychanged both the morphology of the mesothe-lial cells and the cellular architecture of the mes-enteric connective tissue. Our light and electronmicroscopic results revealed that after Freund’sadjuvant injection, the simple squamous meso-thelial cells became rounded, rather cuboidal,

A C

B

D

Fig. 4. Mesothelin and ED1 double labelling on semithin frozen sections. (A) Control mesothelial cells arestrongly labelled with anti-mesothelin antibody (red). No ED1 (green) expression can be detected in these cells. (B)After 24 h adjuvant treatment most of the mesothelin-positive cells show ED1 labelling on their plasma mem-brane. The yellow colour indicates the co- localization of the mesothelin (red) and ED1 (green). Some cells arealready strongly labelled with both mesothelin and ED1 (arrowhead). We find areas (arrows) where mesothelialcells seemed to move apart (to open like a ‘gate’) and ED1-positive cells are migrating out from the connective tis-sue. One double labelled free cell (*) is also found close to the mesentery. (C) Forty-eight-hour treatment results ina strong cytoplasmic ED1 expression in many mesothelin-labelled cells (arrowheads) on the surface of the mesen-tery. Double labelled cells are also present among the free cells that have already lost their connection with the mes-entery (D). Nuclei are labelled with DAPI (blue). Bars: (A) 25 lm; (B) 28 lm; (C) 22 lm; (D) 20 lm.



many of them seem to detach from the basementmembrane. We often found areas of the mesen-tery where mesothelial cells were absent. By thetime of the adjuvant treatment more and more

macrophage-like cells were found freely, closeto the surface of the mesentery. These cells havealready lost their connection with the surface,but we often found partly detached cells as well.When we used double labelling with mesothelinand ED1, we found that in Freund’s adjuvant-treated rats the mesothelin-labelled cells startedto express macrophage markers. ED1- and mes-othelin-positive cells could be detected on thesurface of the mesentery but they were also pres-ent among the free cells close to the surface ofthe mesentery. The ED1 expression was moreand more emphasized by the time of the adju-vant treatment.Our electron microscopic pictures clearly

showed that Freund’s adjuvant treatment haschanged the distribution of caveolae as well.Comparing the number of caveolae present onthe apical and the basal membrane of controland adjuvant-treated cells, we found fewer onthe basal plasma membrane. It is well knownthat integrin-mediated adhesion regulates thecell detachment from the extracellular matrixand the internalization of cholesterol-enrichedmembrane micro domains (20). This internaliza-tion is mediated by dynamin-2 and requiresphosphorylation of caveolin-1 on tyrosine 14. Ashift in localization of phospho-caveolin-1 fromfocal adhesion to caveolae induces internaliza-tion of these membrane domain and results in adecrease in the number of caveolae upon celldetachment (21). We think that the decreasednumber of caveolae on the basal plasma mem-brane as well as the delicate cytoplasmic caveo-lin-1 labelling can be the first sign to indicatethat mesothelial cells have already started todetach from the basement membrane.The fact that mesothelial cells can express mac-

rophage marker(s) indicates that they can trans-differentiate into phagocytotic (macrophages-like) cells. We do not knowwhether these macro-phage markers are only temporally expressed inthese cells or the inflammatory stimulus caused ashift towards a final phagocyte differentiation. Ifmesothelial cells can really differentiate intomac-rophage-like cells, one can suppose thatmesothe-lial cells are ‘young’, not entirely differentiatedcells. To study whether this is the case, we usedantibody against nestin. Nestin is a characteristicmarker of multi-lineage progenitor cells and itspresence in cells indicates multi-potentiality andregenerative potential (12). Although nestin-

A

B

C

Fig. 5. Cytokeratin expression in mesothelial cells.(A) After 24 h of Freund’s adjuvant treatment cyto-keratin (green) expression is well seen in the cytoplasmof these cells. Cells that seem to detach from the sur-face of the mesentery are also labelled with anti-cyto-keratin antibody (arrows). (B) Forty-eight hours afterthe adjuvant injection, there are only few cells showingcytokeratin (as a network) in their cytoplasm (arrows).Some cells detaching from the surface of the mesenteryare still cytokeratin positive (*). (C) In the 48 h treatedsample, we found only few cells labelled with mesoth-elin (red) and the cytokeratin (arrow). The majority ofthe mesothelial cells, however, are already cytokeratinnegative. Nuclei are labelled with DAPI (blue). Bars:(A) 30 lm; (B) 36 lm; (C) 20 lm.

KATZ et al.


positive cells were present on the surface of thecontrol mesentery, after Freund’s adjuvant treat-ment the anti-nestin labelling was more

pronounced. Round shaped cells on the surfaceof the mesentery, cells that are unambiguouslydetaching from the basement membrane werestrongly labelledwith anti-nestin antibody.Intermediate filaments are important indica-

tors of epithelial differentiation (13). Mesothelialcells are unique as although they are derivedfrom mesoderm and express the mesenchymalintermediate filaments vimentin and desmin,they also express cytokeratins which are interme-diate filaments characteristic of epithelial cells(14). While cytokeratins predominate in situ, cul-tured mesothelial cells were found to expresssmall amount of cytokeratins (22). When westudied the cytokeratin expression in rat mesen-teric mesothelial cells, we found that Freund’sadjuvant treatment resulted in a loss of cytokera-tin expression. These results clearly support theidea that mesothelial cells have the ability tochange their phenotype and can undergo an epi-thelial-to-mesenchymal transition (11). Thereare data showing that mesothelial cells contrib-ute to the differentiation of the visceral smoothmuscle cells as well (23). Recently several obser-vations proved that even tissue-specific multipo-tent cells are capable of producing a widespectrum of cell types, regardless whether thesetissues are derived from the same germ layer ornot. This ability is frequently called plasticity ortransdifferentiation (24). Considering that iso-lated peritoneal macrophages can function asprimitive endothelial precursors (25), it wouldnot be surprising if this procedure is reversibleand mesothelial cells (that are very much similarto endothelial cells) can dedifferentiate into mac-rophages. There are data showing that mono-cytes can also be precursors of mesothelial cells(26, 27). Haematopoietic stem cell markers werefound to be expressed in peritoneal mesothelialcells supporting the idea that mesothelial cellsretain their multipotent or pluripotent capacityof differentiation (28). Our recent morphologicaland immunocytochemical results give furthersupport to the plasticity of the mesothelial cells.It has to be pointed out that the cellularcomposition and the cytoarchitecture of mesen-teric connective tissue isolated from Freund’sadjuvant-treated rats have also changed. Thisconnective tissue became thicker, more cells(mainly mast cells, granulocytes andmacrophag-es, fibroblast-like and not really identified cells)were found to be present in this layer. By the time

A B

C

Fig. 6. The flat mesothelial cells present on the sur-face of the control (A) mesentery are labelled withanti-nestin antibody (green). The labelling drawsaround the nuclei, the long, flat cytoplasm can be fol-lowed (arrow). Note that no labelling can be detectedin the mesenteric connective tissue cells. Freund’sadjuvant treatment results in an increase in the num-ber of the nestin-positive cells (B: 24 h and C: 48 hadjuvant treatment). Cells on the surface of the mesen-tery, those that are unambiguously detaching from thebasement membrane (*) and cells that are freely pres-ent close to the mesentery (arrowhead) are stronglylabelled with anti-nestin antibody. In the mesentericconnective tissue, the number of the nestin-positivecells also increased. Nuclei are labelled with DAPI(blue). Bars: (A) 18 lm; (B) 23 lm; (C) 25 lm.



of the Freund’s adjuvant treatment, the numberof the nestin-positive cells in the mesenteric con-nective tissue has also increased. It is not knownwhere these cells can originate from. We thinkthat they can derive either from resting cellsactivated by the Freund’s adjuvant stimulus ormulti-lineage progenitor cells migrate out fromthe blood to the connective tissue. Further exper-iments should be carried out to answer thesequestions.

The authors are very much grateful to Prof. Dr C.D.Dijkstra (Amsterdam, the Netherlands) to provideED1 antibody as a generous gift. We are gratefulto Margit Kutasi and Katalin L}ocsey-Szemere fortheir valuable technical work. We are gratefullyacknowledging Dr Elisabeth Fromm for the languagecorrection. This work was supported by TAMOP.4.2.1.B-09 ⁄1 ⁄KMR-2010-0001 grant.

REFERENCES

1. Ginsel LA, Rijfkogel LP, Deams WT. A dual ori-gin of macrophages? Review and hypothesis.Macrophage Biology 1985;621–49.

2. De Bakker JM, de Wit AW, Woelders H, GinselLA, Daems WT. On the origin of peritoneal resi-dent macrophages. II. Recovery of the residentmacrophage population in the peritoneal cavityand in the milky spots after peritoneal cell deple-tion. J Submicrosc Cytol 1985;7:41–51.

3. Van Furth R. Phagocytic cells: development anddistribution of mononuclear phagocytes in nor-mal steady state and inflammation. In: Gallin JI,Goldstein IM, Snyderman R, editors. Inflamma-tion. Basic Principles and Clinical Correlates.New York: Raven Press, 1988: 281–95.

4. Papadimitriou JM, Ashman RB. Macrophages:current view on their differentiation, structureand function. Ultrastruct Pathol 1989;13:343–72.

5. Kiss AL, Geuze HJ. Caveolae can be alternativeendocytotic structures in elicited macrophages.Eur J Cell Biol 1997;73:19–27.

6. Kiss AL, Turi A, Mullner N, Kantor O, Botos E.Caveolae and caveolin isoforms in rat peritonealmacrophages. Micron 2002;33:75–93.

7. Hausmann M, Rogachev B, Weiler M, Chaimo-vitz C, Douvdevani A. Accessory role of humanperitoneal mesothelial cells in antigen presenta-tion and T-cell growth. Kidney Int 2000;57:476–86.

8. Cunliffe WJ, Sugarbaker PH. Gastrointestinalmalignancy: rationale for adjuvant therapy usingearly postoperative intraperitoneal chemother-apy. Br J Surg 1989;76:1082–90.

9. Holmdahl L. The role of fibrolysis in adhesionformation. Eur J Surg Suppl 1997;577:24–31.

10. Mutsaers SE, Whitaker D, Papadimitriou JM.Stimulation of mesothelial cell proliferation byexudate macrophages enhances serosal woundhealing in a murine model. Am J Physiol 2002;160:681–92.

11. Mutsaers SE. Cells in focus: the mesothelial cells.Int J Biochem Cell Biol 2004;36:9–16.

12. Wiese C, Rolletschek A, Kania G, Blyszczuk P,Tarasova KV, Wersto RP, et al. Nestin expres-sion-a property of multi-lineage progenitor cells?Cell Mol Life Sci 2004;61:2510–22.

13. Tseng SC, Jarvinen MJ, Nelson WG, Huang JW,Woodcock-Mitchell JL, Sun TT. Correlation ofspecific keratins with different types of epithelialdifferentiation: monoclonal antibody studies. Cell1982;30:361–9.

14. Ferrandez-Izquierdo A, Navarro-Fos S, Gonz-alez-Devesa M, Gil-Benso R, Llombart-Bosch A.Immunocytochemical typification of mesothelialcells in effusions: in vivo and in vitro models.Diagnostic Cytopathol 1994;10:256–62.

15. Liou W, Geuze HJ, Slot JW. Improving structuralintegrity of cryosections for immunogold labeling.Histochem Cell Biol 1996;106:41–58.

16. Dijkstra CD, Dopp EA, Joling P, Kraal G. Theheterogeneity of mononuclear phagocytes in lym-phoid organs: distinct macrophages subpopula-tion in the rat recognized by monoclonalantibodies ED1, ED2 and ED3. Immunology1985;54:589–99.

17. Robinson AP, White TM, Mason DW. MRCOX-43: a monoclonal antibody which reacts withall vascular endothelium in the rat except thatbrain capillaries. Immunology 1986;57:231–7.

18. Damoiseaux JGMC, Dopp EA, Calame W, ChaoD, MacPherson GG, Dijkstra CD. Rat macro-phage lysosomal membrane antigen recognizedby monoclonal antibody ED1. Immunology1994;83:140–7.

19. Von Ruhland CJ, Campbell L, Gumbleton M,Jasani B, Newman GR. Immunolocalization ofcaveolin-1 in rat and human mesothelium. J His-tochem Cytochem 2004;52: 1415–25.

20. Del Pozo MA, Alderson NB, Klosses WB, ChiangHH, Anderson RGW. Integrins regulate Rac tar-geting by internalization of membrane domains.Science 2004;303:839–42.

21. Del Pozo MA, Alderson NB, Grande-Garcia A.Phospho-caveolin-1 mediates integrin-regulatedmembrane domain internalisation. Nat Cell Biol2005;7:901–8.

22. Craighead JE, Akley NJ, Gould LB, Libbus BL.Characteristics of tumors and tumor cells culturedfrom experimental asbestos-induced mesothelio-mas of rats. Am J Pathol 1988;129:448–62.

23. Nagy N, Magyar A, Olah I. A novel monoclonalantibody identifies all avian embryonic myogenic

KATZ et al.


cells and adult smooth muscle cells. Anat Embryol2001;204:123–34.

24. Filip S, English D, Mokry J. Issues in stem cellplasticity. J Cell Mol Med 2004;8:572–7.

25. AnghelinaM,MoldovanL, ZabuawalaT,Ostrow-skiMC,MoldovanNL.A subpopulation of perito-neal macrophages form capillary-like lumens andbranching pattern in vitro. J Cell Mol Med2006;10:708–15.

26. Johnson FR, Whitting HW. Repair of parietalperitoneum. Br J Surg 1962;49:653–60.

27. Ryan GB, Grobety J, Majno G. Mesothelialinjury and recovery. Am J Pathol 1973;71:93–102.

28. Foroutan T, Hosseini A, Pourfatholah AA, Solei-mani M, Alimoghadam K, Mosaffa N. Peritonealmesothelial progenitor or stem cell. J Biol Sci2010;10:460–4.



mesothelial cells can detach from the mesentery and differentiate into macrophage-like cells

Documents