Nephroblastoma Overexpressed/Cysteine-Rich Protein 61/Connective Tissue Growth Factor/Nephroblastoma Overexpressed Gene-3 (NOV/CCN3), a Selective Adrenocortical Cell Proapoptotic Factor, Is Down- Regulated in Childhood Adrenocortical Tumors

Mabrouka Doghman, Malika Arhatte, Hélène Thibout, Giovanna Rodrigues, Juliana De Moura, Sébastien Grosso, Alina Nico West, Maryvonne Laurent, Jean-Christophe Mas, André Bongain, Gerard P. Zambetti, Bonald C. Figueiredo, Patrick Auberger, Cécile Martinerie, and Enzo Lalli

Institut de Pharmacologie Moléculaire et Cellulaire Centre National de la Recherche Scientifique Unité Mixte de Recherche 6097 and Université de Nice Sophia Antipolis (M.D., M.A., G.R., J.D.M., E.L.), 06560 Valbonne, France; Institut National de la Santé et de la Recherche Médicale U515 (H.T., M.L., C.M.), Hôpital St. Antoine, UPMC, Paris 6, 75012 Paris, France; Institut National de la Santé et de la Recherche Médicale U526 (S.G., P.A.), Faculté de Médecine, 06107 Nice, France; Department of Biochemistry (A.N.W., G.P.Z.), St. Jude Children’s Research Hospital, Memphis, Tennessee 38105; Service de Pédiatrie (J .- C.M.) and Service de Gynécologie Obstétrique (A.B.), Hôpital de l’Archet, Centre Hospitalier Universitaire, 06200 Nice, France; and Instituto de Pesquisa Pelé Pequeno Principe and Centro de Genética Molecular e Pesquisa do Câncer em Crianças (B.C.F.), Curitiba, 80.060-110 Paraná, Brazil

Context: Childhood adrenocortical tumors (ACTs) have a fetal ad- renal phenotype and overexpress steroidogenic factor-1 (SF-1). Nephroblastoma overexpressed (NOV)/cysteine-rich protein 61/con- nective tissue growth factor/nephroblastoma overexpressed gene-3 mRNA is significantly down-regulated in childhood ACTs.

Objective: The objective of the study was to measure NOV protein levels in childhood ACTs and characterize NOV expression regulation and biological function in human adrenocortical cells.

Design and Setting: Protein extracts from ACT and normal adrenal cortex samples, human adrenocortical carcinoma H295R, primary adrenocortical tumors and fetal adrenal cultures, tissue culture su- pernatants, and cell lysates from H295R cells overexpressing SF-1 in an inducible fashion were used.

Main Outcome Measures: NOV protein levels were measured by enzyme-linked immunoassay and immunoblot. Transient transfec- tion assays were used to study the activity of NOV promoter. Terminal

deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling, caspase assays, and flow cytometry were used to assess the proapoptotic activity of NOV on cells in culture.

Results: NOV mRNA and protein expression is lower in childhood ACTs than in normal adrenal cortex. No significant difference was observed between adenomas and carcinomas. SF-1 overexpression down-regulates NOV at the transcriptional level. NOV has a selective proapoptotic activity toward human adrenocortical cells. The C-ter- minal domain of NOV is responsible for its proapoptotic effect. NOV protein is expressed in DAX-1-positive human fetal adrenal cells.

Conclusions: NOV is a selective proapoptotic factor for human ad- renocortical cells. Reduced expression of NOV in ACTs may play an important role in the process of childhood ACT tumorigenesis, ac- counting at least in part for the defect of apoptotic regression of the fetal adrenal that has been proposed to be responsible for tumor formation. (J Clin Endocrinol Metab 92: 3253-3260, 2007)

T HE STRUCTURE OF the human adrenal cortex is dif- ferent during fetal life and after birth. In fact, during intrauterine life, most of the inner part of the adrenal cortex (fetal zone) is occupied by large, steroid-secreting cells,

First Published Online June 19, 2007

Abbreviations: ACT, Adrenocortical tumors; CCN, connective tissue growth factor; DAPI, 4’,6’-diamidino-2-phenylindole; DAX-1, dosage- sensitive sex reversal-adrenal hypoplasia congenita critical region on X chromosome-gene 1; FITC, fluorescein isothiocyanate; IGFBP, IGF bind- ing protein; NOV, nephroblastoma overexpressed; P450scc, P450 side- chain cleavage; SF-1, steroidogenic factor-1; TUNEL, terminal deoxy- nucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling.

JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the en- docrine community.

which are differentiated to synthesize large amounts of de- hydroepiandrosterone and its sulfate derivative. Conversely, a few layers of cells lying in an outer, subcapsular position (definitive zone) are nonsteroidogenic and constitute a pro- liferative compartment in which most probably the precur- sors of fetal zone cells are generated (1-3). An additional zone (transitional zone) develops between the definitive and the fetal zone late in pregnancy and synthesizes cortisol. The fetal adrenal undergoes a process of rapid involution and remodeling after birth (1, 4). The fetal zone regresses by apoptosis, whereas cells of the definitive zone proliferate and differentiate to form the glomerulosa, fasciculata, and re- ticularis zones of the adult adrenal cortex. Childhood adre- nocortical tumors (ACTs) are thought to be derived from the fetal adrenal, probably through defective apoptosis because of their age distribution, their pattern of hormone secretion,

and their molecular phenotype (5, 6). These tumors can arise isolated or in the context of genetic syndromes (for a review see Ref. 7) and most commonly present with virilization due to production of androgenic steroids, which can be associ- ated with Cushing syndrome. A peculiar epidemiological feature of childhood ACTs is their high incidence in the states of Paraná and São Paulo in southern Brazil, where they are almost invariably found associated with a specific germline TP53 mutation (R337H) (8). In addition, a role for the tran- scription factor steroidogenic factor-1 (SF-1) in childhood ACT pathogenesis has been suggested based on its ampli- fication and overexpression in this type of cancer (9, 10). It is remarkable that an increase in SF-1 dosage is by itself sufficient to increase proliferation of human tumor adreno- cortical cells and to cause adrenocortical tumors in mice (Doghman, M., T. Karpova, G. Rodrigues, M. Arhatte, J. De Moura, L. R. Cavalli,, V. Virolle, P. Barby, G. P. Zambetti, B. C. Figueiredo, L. L. Heckert, E. Lalli, submitted for publication).

A gene expression profiling study performed in childhood ACTs has recently shown that distinct patterns of gene ex- pression distinguish tumors from normal adrenal cortex. Nephroblastoma overexpressed (NOV)/cysteine-rich pro- tein 61/connective tissue growth factor/nephroblastoma overexpressed gene (CCN)-3 is one of the transcripts that is most powerfully down-regulated in childhood ACTs, inde- pendently from their degree of malignancy (11). Conversely, it was previously shown that loss of NOV expression is associated with a poor prognosis in adult adrenocortical tumors (12). NOV is a multidomain-secreted protein, mem- ber of the CCN family, whose expression was described to be mostly restricted to the definitive zone of the fetal adrenal cortex (13). Its known functions are related to cell adhesion and angiogenesis, and its expression is negatively correlated to proliferation and associated with differentiation in many cellular systems (for review see Ref. 14). Here we show that NOV is strikingly down-regulated at the protein level in childhood ACTs and in human adrenocortical cells in a man- ner dependent on SF-1 dosage and characterize its selective proapoptotic activity for human adrenocortical cells.

Subjects and Methods

Tumor samples

Twenty-four childhood ACTs (five adenomas, 18 carcinomas, and one of undetermined histological type) and seven normal adrenal cortex samples were used for mRNA studies. These were the same samples that were used for gene expression profiling using Affymetrix microarrays (11). For protein expression studies, tumor samples were obtained at the time of surgery from eight children with ACTs (three boys and five girls). Five patients presented with virilization and three with virilization associated with Cushing’s syndrome. Histologically, four tumors were classified as adenomas and four as carcinomas. Eight normal adrenal glands resected from age-matched children undergoing surgery for Wilms’ tumor were used as controls. Samples were immediately snap frozen in liquid nitrogen before processing for protein extraction. In one case of adrenocortical adenoma in a 5-yr-old child, tumor cells were dissociated with collagenase and cultured in OptiMEM (Invitrogen, Carlsbad, CA) containing 2% fetal calf serum, following an established protocol (15). All patients and control subjects were included in the study after one of the parents or legal representatives signed an informed consent form approved by the Ethics Committee of the Hospital de Clínicas of the Federal University of Paraná.

Human fetal adrenal tissue preparation

Adrenal glands were obtained from human fetuses aged 10-14 wk after elective termination of pregnancy. The research protocol was pre- viously submitted to the French national Agence de la Biomédecine and local Ethical Committee of the Nice University Hospital. Fetal adrenal cells were isolated and cultured as described (15).

Immunofluorescence

It was performed as described (16) using the following antibodies: rabbit anti-NOV (K19M); mouse anti-dosage-sensitive sex reversal- adrenal hypoplasia congenita critical region on X chromosome-gene 1 (DAX-1) (2F4); rabbit anti-P450 side-chain cleavage (P450scc; gift of Dr. W. L. Miller) (17); rabbit anti-SF-1 (Upstate, Lake Placid, NY). As negative control for primary antibodies, species-matched IgG was used. Cell nuclei were counterstained with 4’,6’-diamidino-2-phe- nylindole (DAPI).

Cell culture

H295R cells (American Type Culture Collection, Manassas, VA) were cultured in DMEM/F-12 supplemented with 2% NuSerum (Becton Dick- inson, Lincoln Park, NJ), 1% ITS Plus (Becton Dickinson), and antibiotics. H295R cell clones overexpressing SF-1 in a doxycycline-inducible fash- ion (H295R TR/SF-1) were produced and maintained as described (Doghman, M., T. Karpova, G. Rodrigues, M. Arhatte, J. De Moura, L. R. Cavalli, V. Virolle, P. Barby, G. P. Zambetti, B. C. Figueiredo, L. L. Heckert, E. Lalli, submitted for publication). Proliferation assays were performed in the absence of NuSerum. Hela cells were cultured in DMEM (1 g/liter glucose) supplemented with 10% fetal calf serum (Invitrogen) and antibiotics. COS cells were cultured in DMEM (4.5 g/liter glucose) supplemented with 5% fetal calf serum (Invitrogen) and antibiotics. Jurkat cells were cultured in RPMI 1640 supplemented with 5% fetal calf serum (Invitrogen) and antibiotics.

cDNA amplification and real-time PCR analysis for NOV mRNA expression

This was performed as described (11). Results were normalized to ubiquitin expression levels using the 44C, method (18).

A

FIG. 1. NOV is down-regulated in childhood ACTs at the transcript and protein level. A, NOV mRNA levels were measured in seven normal adrenals and 24 childhood ACT samples by quantitative RT-PCR. NOV mRNA levels are significantly down-regulated in tumors, compared with normal adrenal (Mann-Whitney U test; P = 0.002). ** , P < 0.05. B, Immunoblot showing NOV protein expression in eight childhood ACTs (T1-T4 adenomas; T5-T8 carcinomas) and eight normal adrenals. B-Tu- bulin levels show equivalent protein loading of gel lanes.

150

Relative expression

125

100

75

**

50

25

0

Normal

ACT

B

kDa

N1 N2 N3

N4

T1

T2

T3

T4

N5 N

N7 N8 18

T5 T6 T7 T8

66 -

45 -

NOV

35 -

25

₿-tubulin

FIG. 2. NOV production is negatively regulated by SF-1 overexpression in H295R cells. A, NOV protein levels are decreased in the culture su- pernatant of H295R TR/SF-1 cells treated with doxycycline, compared with untreated cells. Cells were incubated for 48 h in the absence or presence of Dox, medium was collected, and NOV levels were assayed by a specific enzyme immunoassay and Western blot. Culture me- dium was replaced, and measurements were re- peated after 24 h. Significant differences existed between NOV levels in the absence (white his- tograms) and presence (black histograms) of Dox at both 48 and 72 h (P = 0.0055 and P = 0.0378, respectively, t test). Conversely, Dox treatment did not modify the levels of the secreted protein IGFBP2 in the supernatant of H295R TR/SF-1 cells at both 48 and 72h .* , P<0.01 ;** , P<0.05. B, SF-1 overexpression reduces NOV promoter activity. NOV promoter -2540, - 625, and -492 luciferase reporters were transfected in H295R TR/SF-1 or parental H295R TR cells. Cells were subsequently cultured for 48 h in the presence or absence of Dox. For each promoter fragment, luciferase activity in Dox-treated cells (white histograms) was compared with the activity of untreated cells (black histograms). No effect of Dox treatment was detected in transfected parental cells. Data shown are the average of three independent transfection ex- periments, each one performed in duplicate.

A

30

NOV (ng/ml)

20

**

*

10

0

NOV

IGFBP2

+

+

Dox

48 h

72 h

B

Dox

Parental clone

+

luc

-492 NOV promoter

H295R TR/SF-1

+

Parental clone

+

luc

-625 NOV promoter

H295R TR/SF-1

+

Parental clone

+

luc

-2540 NOV promoter

H295R TR/SF-1

+

0

5000

10000

15000

RLU

NOV assays

Secreted NOV protein was quantified by an enzyme linked immu- noassay, as described (19) or extracted from culture supernatants by heparin-Sepharose chromatography, as described (20), and detected by immunoblot.

Immunoblots

The immunoblots were performed as described (10). Normal and tumor adrenal samples were homogenized in Laemmli buffer, and pro-

FIG. 3. NOV reduces H295R cell number in a dose-de- pendent manner. A, H295R cells were cultured in me- dium without NuSerum in the presence of protein buffer only (Ctrl) or with increasing doses [0.01 (0.25 nM), 0.05 (1.25 nM), 0.1 (2.5 nM), and 0.5 (12.5 nM) µg/ml] of re- combinant NOV protein added. Cell numbers were counted after 4 d. Data are normalized by the number of cells present in buffer-treated cultures. Data shown are the average of two experiments performed in duplicate. B, Cell cycle distribution of untreated H295R cells (left) and cells treated with 0.5 µg/ml NOV protein for 4 d (right).

teins were separated on SDS-PAGE and blotted on a polyvinyl difluoride membrane (GE Healthcare, Indianapolis, IN). Antibodies used for im- munoblot were: rabbit anti-NOV (K19M), mouse anti-ß-tubulin (Sigma, St. Louis, MO), and goat anti-IGF binding protein (IGFBP)-2 (Santa Cruz Biotechnology, Santa Cruz, CA).

Transient transfection and luciferase assays

H295R cells were transfected with NOV promoter reporters (20) and pCH110 as a ß-galactosidase expression vector using the jetPEI reagent

A

% of untreated cells

100

T

75

50

T

25

0

Ctrl

NOV

B

0 NOV

0.5 NOV

#

8

2

2

8

B

-

2

&

.

.

-

0

Q

0

200

400

600

1000

FLDE

FLZH

at a nitrogen to phosphate ratio of 5, following the manufacturer’s (Polyplus Transfection, Illkirch, France) instructions. Luciferase assays were realized using a Luminoskan Ascent (Thermo Labsystems, San Jose, CA) luminometer. Results were normalized by ß-galactosidase activities.

Recombinant NOV proteins

The recombinant NOV proteins were produced and purified as de- scribed (19).

Flow cytometry

H295R cells were fixed in 70% ethanol and then treated with RNase A (1 µg/ml) for 30 min at 37 C. DNA was stained with propidium iodide (50 µg/ml), and cells were analyzed for cell-cycle distribution with a FACScan instrument (Becton Dickinson). Results are expressed as pro- liferating index (sum of percentages of cells in S + G2/M phases).

Apoptosis and caspase activity assays

Apoptosis was assayed using the TdT-based in situ cell death detec- tion kit (Roche, Indianapolis, IN). For caspase activity assays, H295R cells were lysed for 30 min at 4 C in lysis buffer [50 mM HEPES (pH 8), 150 mM NaCl, 20 mM EDTA, 1 mm phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, 10 µg/ml aprotinin, and 0.2% Triton X-100], and lysates were cleared at 10,000 × g for 15 min at 4 C. Each assay (in triplicate) was performed using 20 µg of protein lysate prepared from control or NOV-stimulated cells and complemented with 5 mm dithio- threitol. Briefly, cellular extracts were incubated in a 96-well plate with 0.2 mM of Ac-DEVD-AMC (caspase-3), Ac-LEHD-AMC (caspase-9), or

Ac-IETD-AMC (caspase-8) as substrates at 37 C. Caspase activity was measured after absorbance at 460 nm (excitation at 390 nm) in the absence or presence of 1 µM of Ac-DEVD-CHO, Ac-LEHD-CHO, and Ac-IETD-CHO, for caspase-3,-9, and -8, respectively. To assay for active caspase-3 levels in Jurkat T cells, cells were treated for 8 or 24 h in the presence of either NOV (2.5 µg/ml) or CH11 anti-Fas antibody (50 ng/ml), as a control of induction of apoptosis. Cells were then washed with PBS, fixed, and permeabilized. Subsequently cells were incubated for 30 min with antiactive-caspase-3-fluorescein isothiocyanate (FITC) monoclonal antibody (BD Biosciences, Franklin Lakes, NJ). After wash- ing, samples were immediately analyzed with a FACScan (Becton Dickinson).

Results

NOV protein is down-regulated in childhood ACT

By microarray analysis we have shown that NOV is one of the transcripts that is more significantly down-regulated in a series of 24 childhood ACTs, compared with normal ad- renal cortex (11). NOV mRNA levels measured by quanti- tative RT-PCR are significantly lower in childhood ACTs than normal adrenal cortex (Fig. 1A). To verify whether NOV protein levels correlate with mRNA levels, we compared NOV protein expression in a series of eight tumors (four adenomas and four carcinomas) and eight samples of age- matched normal adrenal cortex. NOV protein levels are in- variably reduced in all tumor samples, compared with nor- mal tissue, for both adenomas and carcinomas (Fig. 1B).

FIG. 4. NOV is a selective proapoptotic factor for hu- man tumor adrenocortical cells. A, Structure of the NOV protein. The different protein domains are shown with different colors. SP, Signal peptide; VWC, von Willebrand factor type C domain; TSP1, thrombospondin type I repeat; CT, cysteine-knot do- main. A variable amino acid stretch connects do- mains von Willebrand factor type C domain and thrombospondin type I repeat. The extensions of the N-terminal (N-ter) and C-terminal (C-ter) domains expressed in recombinant form are indicated. B, H295R cells were treated with buffer (Ctrl) or NOV full-length (FL), N- (N-ter) and C-terminal (C-ter) fragments (2.5 µg/ml each) for 24 h, and the percent- age of apoptotic cells was measured by TUNEL stain- ing. Apoptosis was significantly increased in cells treated with full-length NOV and NOV C-terminal fragment, as compared with buffer-treated cells (P = 0.0026 and P = 0.0398, Fisher's exact test). On the right side, fluorescence microscopy view of DAPI staining (top panels) and TUNEL staining (bottom panels) of H295R cells treated with buffer (left pan- els) or full-length NOV (2.5 µg/ml; right panels). Size bar, 50 pm. ** , P < 0.05. C, H295R, HeLa, and COS cells were treated for 24 h with buffer (white histo- grams) or 2.5 µg/ml NOV (black histograms), and percentage of apoptotic cells was measured by TUNEL staining. Data shown represent the average of two experiments performed in duplicate.

A

SP IGFBP

VWC

TSP1

CT

N-ter

C-ter

B

30

**

% apoptotic cells

DAPI

**

20

10

TUNEL

0

Ctrl

NOV FL

N-ter

C-ter

Ctrl

NOV

C

20

% apoptotic cells

15

10

5

0

Ctrl

NOV

NOV

Ctrl

NOV

H295R

HeLa

COS

B-Tubulin staining shows similar protein loading in each sample (Fig. 1B).

Increase of SF-1 dosage in H295R cells represses NOV expression at the transcriptional level

H295R cells are the only differentiated human adrenocor- tical cell line with a fetal adrenal phenotype currently avail- able (21). They express SF-1 endogenously (16). We estab- lished cellular clones derived from H295R cells that overexpress SF-1 in a doxycycline-inducible fashion (H295R TR/SF-1) to produce a cellular model for childhood ACT (Doghman, M., T. Karpova, G. Rodrigues, M. Arhatte, J. De Moura, L. R. Cavalli, V. Virolle, P. Barby, G. P. Zambetti, B. C. Figueiredo, L. L. Heckert, E. Lalli, submitted for publica- tion). Gene expression profiling of H295R TR/SF-1 cells re- vealed that NOV is one of the transcripts that is more pow- erfully down-regulated when SF-1 dosage is increased. Microarray results were confirmed by quantitative RT-PCR in two different H295R TR/SF-1 clones (Doghman, M., T. Karpova, G. Rodrigues, M. Arhatte, J. De Moura, L. R. Cav- alli, V. Virolle, P. Barby, G. P. Zambetti, B. C. Figueiredo, L. L. Heckert, E. Lalli, submitted for publication). NOV protein levels are significantly reduced in the supernatant of H295R TR/SF-1 cells treated with doxycycline for either 48 or 72 h to induce SF-1 overexpression (Fig. 2A). The IGFBP2-se- creted protein harbors an IGF-binding domain similar to the one present in the N-terminal half of NOV and is readily detectable in the tissue culture supernatant of H295R cells (22). Remarkably, doxycycline treatment of H295R TR/SF-1 cells does not appreciably modulate IGFBP2 levels in their culture supernatant (Fig. 2A), showing the specificity of the effect of increased SF-1 levels to negatively regulate NOV expression in H295R human adrenocortical cells.

Increased SF-1 dosage regulates NOV expression at the transcriptional level because doxycycline treatment reduced activity of NOV promoter in H295R TR/SF-1 cells, whereas it had no effect on NOV promoter activity in the parental H295R TR clone (Fig. 2B). Deletion analysis showed that the activity of the shortest NOV promoter fragment analyzed (-492) is still repressed by elevated SF-1 levels in H295R TR/SF-1 cells. In the minimal NOV promoter sequences sen- sitive to inhibition by increased SF-1 levels, no putative SF-1 response element can be found by MatInspector (Genomatix, Munich, Germany) analysis. Furthermore, NOV promoter reporters are not sensitive to inhibition by SF-1 after transient transfection in COS cells (data not shown). This data suggest that the specific effect of increased SF-1 levels on NOV pro- moter activity in adrenocortical cells is mediated by other transcription factors whose levels, or activities, are regulated by SF-1 dosage.

NOV induces apoptosis of human tumor adrenocortical cells through caspase activation

To study the biological effects of NOV on human adre- nocortical cells, we treated H295R cells with increasing doses of recombinant full-length NOV protein and monitored their growth over a 4-d period. NOV decreases cell number in a dose-dependent manner (Fig. 3A). This effect of NOV cannot be ascribed to block of cell-cycle progression because the

proliferating index of H295R cells treated with the highest dose of NOV (0.5 µg/ml) was not decreased, compared with untreated cells (36.7 vs. 34.8, respectively; Fig. 3B). On the other hand, full-length NOV in nanomolar concentrations induces a significant increase in the percentage of apoptotic H295R cells, as detected by terminal deoxynucleotidyl trans- ferase-mediated deoxyuridine triphosphate nick end label- ing (TUNEL) staining, after a 24-h treatment (Fig. 4B). The C-terminal NOV fragment, encompassing the sequence con-

FIG. 5. NOV induces apoptosis through caspase activation. A, Caspase inhibitors inhibit apoptosis induced by NOV. H295R cells were treated for 24 h with buffer (Ctrl) or recombinant full-length NOV protein (2.5 µg/ml), alone or in the presence of caspase inhibitors Z-VAD (pan-caspase), Z-DEVD (caspase-3), Z-IETD (caspase-8), and Z-LEHD (caspase-9) (50 µM each). Percentage of apoptotic cells was measured by TUNEL staining. Data shown represent the average of two experiments performed in duplicate. B, Assay of caspase-3 (left), caspase-8 (center), and caspase-9 activities (right) in H295R cells treated with buffer (white histograms) or 2.5 µg/ml full-length NOV protein (black histograms) for 24 h. C, Jurkat T cells were treated for 8 or 24 h with recombinant NOV protein or CH11 anti-Fas antibody. Cells were stained with antiactive caspase-3 FITC antibody, washed, and the percentage of FITC-positive cells measured by flow cytom- etry.

A

30-

% apoptotic cells

20

10

0

Ctrl

Z-VAD

Z-DEVD

Z-IETD

Z-LEHD

NOV

B

Caspase 3

Caspase 8

Caspase 9

12.5

A Fluo/min/mg protein

10.0

A Fluo/min/mg protein

1.5

A Fluo/min/mg protein

1.5

7.5

1.0

1.0

5.0

0.5

0.5

2.5

0

0

0

Ctrl

NOV

Ctrl

NOV

Ctrl

NOV

C

100

1.88%

Ctrl

0

100

102

182

30.20%

CH11 24h

Number of events

Number of events

100

0-

103

27.44%

CH11 6h

100

1000

101

102

103

104

2.17%

NOV 24h

100

0

191

182

a

.

1.80%

NOV 6h

Active caspase 3

0

I 4

Active caspase 3

taining the thrombospondin-type I repeat and the CT (Fig. 4A), induces apoptosis of H295R cells nearly as efficiently as the full-length protein, whereas the N-terminal fragment, encompassing the IGFBP and von Willebrand factor type C domain domains, is inactive (Fig. 4B). Remarkably, NOV effect appears to be relatively selective for adrenocortical cells because it has no proapoptotic activity on other cell lines tested (Hela, COS; Fig. 4C). NOV-induced apoptosis is sen- sitive to the pan-caspase inhibitor Z-VAD and is also reduced by treatment with the selective caspase-3 (Z-DEVD), caspase-8 (Z-IETD), and caspase-9 (Z-LEHD) inhibitors (Fig. 5A). Consistent with these results, NOV induces a significant increase of caspase-3, caspase-8, and caspase-9 activities in H295R cells (Fig. 5B). Remarkably, NOV was unable to ac- tivate caspase-3 in Jurkat T leukemic cells (Fig. 5C), whereas this caspase was activated by anti-Fas treatment, as previ- ously described (23). Furthermore, NOV effect on apoptosis of primary tumor adrenocortical cells was measured. Adre- nocortical tumor cells in culture keep their differentiated phenotype, with the majority of cells expressing transcrip- tion factors SF-1 and DAX-1 (Fig. 6A) and steroidogenic enzymes P450scc and P450C17 (not shown). Similarly to the H295R cell line, apoptosis of primary adrenocortical tumor cells could be significantly elicited by recombinant NOV protein in SF-1-positive cells (Fig. 6B and Table 1).

NOV protein is expressed in human fetal adrenal cells

Previous studies have shown that NOV mRNA is detect- able in the human fetal adrenal, mostly in the definitive zone (13). Adrenocortical fetal adrenal cells in culture isolated from fetuses aged 11-14 wk after conception express the DAX-1 transcription factor (24) and the first enzyme in the steroidogenic cascade, P450scc (Fig. 7A). NOV protein is detected in the cytoplasm of the same cells that express

TABLE 1. Recombinant NOV protein triggers apoptosis of primary adrenocortical tumor cells
SF-1-positive apoptotic cellsNonapoptotic SF-1- positive cells
14468Control cells
46419NOV-treated cells

Primary tumor adrenocortical cells isolated from one childhood ACT were isolated by collagenase treatment and cultured in eight- well chamber slides. Equivalent volumes of protein buffer (control) or recombinant full-length NOV protein (10 µg/ml) were added. Num- bers of apoptotic (TUNEL staining) SF-1-positive cells in the cultures were counted after an 18-h incubation. The difference in the number of apoptotic cells between control and NOV-treated samples is sig- nificant (P < 0.0001, Fisher’s exact test).

DAX-1 (Fig. 7B). Remarkably, DAX-1-negative cells, which are not steroidogenic (Fig. 7A), do not express NOV.

Discussion

Members of the CCN family of proteins are multimodular factors associated with the extracellular matrix endowed with diverse physiological functions. Their biological activ- ities are often dependent on the specific cellular context, and they may even have opposing effect in different cell types (reviewed in Ref. 25). Here we have shown that NOV/CCN3 has a potent proapoptotic effect on human adrenocortical cells. Remarkably, NOV proapoptotic activity appears to be relatively selective for adrenocortical cells because other cell types are not sensitive to NOV. This is the first time that direct evidence is presented showing a proapoptotic effect of NOV through caspase activation. However, NOV has been associated with the inhibition of proliferation of several other cancer cell types (26-28). Together with data shown here, those reports indicate that the proapoptotic effect of NOV may target a restricted spectrum of tissues. Other members

FIG. 6. Primary adrenocortical tumor cells express transcription factors SF-1 and DAX-1 and are sensitive to NOV pro- apoptotic effect. A, Immunofluorescence shows nuclear SF-1 staining (green) and predominantly nuclear DAX-1 staining (red) in tumor cells. Nuclei were stained with DAPI dye. Size bar, 20 um. B, Cul- tured primary adrenocortical tumor cells were treated with either protein buffer (Ctrl) or recombinant full-length NOV protein (10 µg/ml) for 18 h. Apoptotic cells were detected by TUNEL staining (bot- tom panels). Nuclei were stained with DAPI dye (top panels). Size bar, 50 um.

A

O

DAPI

SF-1

DAX-1

SF-1/DAX-1

B

DAPI

TUNEL

Ctrl

NOV

A

FIG. 7. NOV protein is expressed in human fetal adrenal cells. Cells isolated from the adrenal gland of a fetus 11 wk after conception were stained with anti-P450scc, anti-NOV, and anti-DAX-1 antibodies. Nuclei were stained with DAPI dye. A, P450scc and DAX-1 are co- expressed by steroidogenic cells, whereas P450scc-negative cells in the culture (arrowheads) do not express DAX-1. B, NOV protein is exclusively detected in the cytoplasm of cells that also express nuclear DAX-1. DAX-1-negatiwe cells do not express NOV (arrowheads). Size bar, 20 um.

DAPI

P450scc/DAX-1

B

DAPI

NOV/DAX-1

of the CCN family may function as growth factors or induce apoptosis, depending on the cell type (29-31). The molecular bases of those differential effects are still unknown, but they might be accounted for by differential expression of NOV receptors and/ or coupling to different signaling pathways in different cell types.

It is remarkable that in childhood ACTs, NOV expression is invariably reduced at the transcript and protein level (Fig. 1). Reduced expression of an endogenous proapoptotic factor may play an important role in the process of childhood ACT tumorigenesis, accounting at least in part for the defect of apoptotic regression of the fetal adrenal that has been pro- posed to be responsible for tumor formation (6). Signifi- cantly, in childhood ACTs, NOV levels are down-regulated in both the adenoma and carcinoma groups (Ref. 11 and our data). Importantly, increase of SF-1 dosage in human tumor adrenocortical cells is sufficient to repress NOV expression at the transcriptional level (Fig. 2). SF-1 overexpression in childhood ACTs is also a widespread finding, independent from clinical outcome (10). This is at variance with data reported in adult ACT, in which NOV expression is lower in carcinomas, compared with adenomas, and associated with a more severe prognosis (12). Our results suggest that in childhood ACTs, SF-1 overexpression and NOV down-reg- ulation may be essential for tumor formation but not for

progression to a malignant phenotype. Conversely, loss of NOV expression in adult adrenocortical carcinomas may reflect their loss of the differentiated adrenal cortex pheno- type, similarly to what has been shown for other transcripts that are expressed at high levels in normal adrenal cortex and down-regulated in carcinomas (32, 33).

In cultured human fetal adrenal cells, NOV is exclusively expressed by cells that are also positive for DAX-1 staining. Previous studies have shown that in the human fetal adrenal NOV mRNA is mainly expressed in the outer definitive zone (13). This zone represents the adrenal cortex proliferative compartment, in which the precursors of steroidogenic cells are generated. Because of its proapoptotic effect on adreno- cortical cells, it is tempting to speculate that in the physio- logical setting, NOV may play an important role in the dy- namic interplay of diffusible factors that regulate (1, 2, 34, 35) the proliferation and differentiation of fetal adrenal cells.

Acknowledgments

We thank Dr. Nicole Gallo-Payet for advice about fetal adrenal cell culture, Dr. Walter L. Miller for gift of the anti-P450scc antibody, and the personnel of the Service d’Orthogénie of the Archet Hospital for collaboration.

Received February 14, 2007. Accepted May 31, 2007.

Address all correspondence and requests for reprints to: Enzo Lalli, M.D., Institut de Pharmacologie Moléculaire et Cellulaire, Centre Na- tional de la Recherche Scientifique Unité Mixte de Recherche 6097, 660 Route des Lucioles, Sophia Antipolis, 06560 Valbonne, France. E-mail: ninino@ipmc.cnrs.fr.

This work was supported by a Centre National de la Recherche Scientifique ATIP grant, contrat d’interface Institut National de la Santé et de la Recherche Médicale-Centre Hospitalier Universitaire de Nice, CAPES-COFECUB (419/03), Fondation Recherche Médicale, Associa- tion Recherche sur le Cancer (3142), and Princess Grace Foundation (to E.L.); and National Institutes of Health Grant CA63230 (to G.P.Z.). M.D. is a recipient of a Fondation Recherche Médicale fellowship.

Disclosure Information: All authors have nothing to declare.

References

1. Mesiano S, Jaffe RB 1997 Developmental and functional biology of the pri- mate fetal adrenal cortex. Endocr Rev 18:378-403

2. Spencer SJ, Mesiano S, Lee JY, Jaffe RB 1999 Proliferation and apoptosis in the human adrenal cortex during the fetal and perinatal periods: implications for growth and remodeling. J Clin Endocrinol Metab 84:1110-1115

3. Folligan K, Bouvier R, Targe F, Morel Y, Trouillas J 2005 Etude histologique et moléculaire de la corticosurrénale fœtale humaine. Ann Endocrinol 66:519- 526

4. Ben-David S, Zuckerman-Levin N, Epelman M, Shen-Orr Z, Levin M, Sujov P, Hochberg Z 2007 Parturition itself is the basis for fetal adrenal involution. J Clin Endocrinol Metab 92:93-97

5. Michalkiewicz E, Sandrini R, Figueiredo B, Miranda EC, Caran E, Oliveira- Filho AG, Marques R, Pianovski MA, Lacerda L, Cristofani LM, Jenkins J, Rodriguez-Galindo C, Ribeiro RC 2004 Clinical and outcome characteristics of children with adrenocortical tumors. An analysis of 254 cases from the International Pediatric Adrenocortical Tumor Registry. J Clin Oncol 22:838- 845

6. Wilkin F, Gagné N, Paquette J, Oligny LL, Deal C 2000 Pediatric adreno- cortical tumors: molecular events leading to insulin-like growth factor II gene overexpression. J Clin Endocrinol Metab 85:2048-2056

7. Koch CA, Pacak K, Chrousos GP 2004 The molecular pathogenesis of hered- itary and sporadic adrenocortical and adrenomedullary tumors. J Clin Endo- crinol Metab 87:5367-5384

8. Ribeiro RC, Sandrini F, Figueiredo B, Zambetti GP, Michalkiewicz E, Laf- ferty AR, DeLacerda L, Rabin M, Cadwell C, Sampaio G, Cat I, Stratakis CA, Sandrini R 2001 An inherited p53 mutation that contributes in a tissue-specific manner to pediatric adrenal cortical carcinoma. Proc Natl Acad Sci USA 98:9330-9335

9. Figueiredo BC, Cavalli LR, Pianovski MA, Lalli E, Sandrini R, Ribeiro RC, Zambetti G, DeLacerda L, Rodrigues GA, Haddad BR 2005 Amplification of

the steroidogenic factor 1 gene in childhood adrenocortical tumors. J Clin Endocrinol Metab 90:615-619

10. Pianovski MA, Cavalli LR, Figueiredo BC, Santos SC, Doghman M, Ribeiro RC, Oliveira AG, Michalkiewicz E, Rodrigues GA, Zambetti G, Haddad BR, Lalli E 2006 SF-1 overexpression in childhood adrenocortical tumours. Eur J Cancer 42:1040-1043

11. West AN, Neale GA, Pounds S, Figueiredo BC, Galindo CR, Pianovski MAD, Oliveira Filho AG, Malkin D, Lalli E, Ribeiro R, Zambetti GP 2007 Gene expression profiling of childhood adrenocortical tumors. Cancer Res 67:600-608

12. Martinerie C, Gicquel C, Louvel A, Laurent M, Schofield PN, Le Bouc Y 2001 Altered expression of novH is associated with human adrenocortical tumor- igenesis. J Clin Endocrinol Metab. 86:3929-3940

13. Ratcliffe J, Nakanishi M, Jaffe RB 2003 Identification of definitive and fetal zone markers in the human fetal adrenal gland reveals putative developmental genes. J Clin Endocrinol Metab 88:3272-3277

14. Planque N, Perbal B 2003 A structural approach to the role of CCN (CYR61/ CTGF/NOV) proteins in tumourigenesis. Cancer Cell Intern 3:15

15. Breault L, Lehoux JG, Gallo-Payet N 1996 The angiotensin AT2 receptor is present in the human fetal adrenal gland throughout the second trimester of gestation. J Clin Endocrinol Metab 81:3914-3922

16. Lalli E, Ohe K, Hindelang C, Sassone-Corsi P 2000 Orphan receptor DAX-1 is a shuttling RNA binding protein associated to polyribosomes via messenger RNA. Mol Cell Biol 20:4910-4921

17. Black SM, Szklarz GD, Harikrishna JA, Lin D, Wolf CR, Miller WL 1993 Regulation of proteins in the cholesterol side-chain cleavage system in JEG-3 and Y-1 cells. Endocrinology 132:539-545

18. Livak KJ, Schmittgen TD 2001 Analysis of relative gene expression data using real-time quantitative PCR and the 2-AACt method. Methods 25:402-408

19. Thibout H, Martinerie C, Creminon C, Godeau F, Boudou P, Le Bouc Y, Laurent M 2003 Characterization of human NOV in biological fluids: an enzyme immunoassay for the quantification of human NOV in sera from patients with diseases of the adrenal gland and of the nervous system. J Clin Endocrinol Metab 88:327-336

20. Lafont J, Laurent M, Thibout H, Lallemand F, Le Bouc Y, Atfi A, Martinerie C 2002 The expression of novH in adrenocortical cells is down-regulated by TGFß1 through c-Jun in a Smad-independent manner. J Biol Chem 277:41220- 41229

21. Staels B, Hum DW, Miller WL 1993 Regulation of steroidogenesis in NCI- H295 cells: a cellular model of the human fetal adrenal. Mol Endocrinol 7:423- 433

22. Logié A, Boulle N, Gaston V, Perin L, Boudou P, Le Bouc Y, Gicquel C 1999 Autocrine role of IGF-II in proliferation of human adrenocortical carcinoma NCI H295R cell line. J Mol Endocrinol 23:23-32

23. Ricci JE, Maulon L, Luciano F, Guerin S, Livolsi A, Mari B, Breittmayer JP, Peyron JF, Auberger P 1999 Cleavage and relocation of the tyrosine kinase

P59FYN during Fas-mediated apoptosis in T lymphocytes. Oncogene 18:3963- 3969

24. Battista MC, Otis M, Cote M, Laforest A, Peter M, Lalli E, Gallo-Payet N 2005 Extracellular matrix and hormones modulate DAX-1 localization in the human fetal adrenal gland. J Clin Endocrinol Metab 90:5426-5431

25. Leask A, Abraham DJ 2006 All in the CCN family: essential matricellular signaling modulators emerge from the bunker. J Cell Sci 119:4803-4810

26. Gupta N, Wang H, McLeod TL, Naus CC, Kyurkchiev S, Advani S, Yu J, Perbal B, Weichselbaum RR 2001 Inhibition of glioma cell growth and tu- morigenic potential by CCN3 (NOV). Mol Pathol 54:293-299

27. Benini S, Perbal B, Zambelli D, Colombo MP, Manara MC Serra M, Parenza M, Martinez V, Picci P, Scotlandi K 2005 In Ewing’s sarcoma CCN3(NOV) inhibits proliferation while promoting migration and invasion of the same cell type. Oncogene 24:4349-4361

28. McCallum L, Price S, Planque N, Perbal B, Pierce A, Whetton AD, Irvine AE 2006 A novel mechanism for BCR-ABL action: stimulated secretion of CCN3 is involved in growth and differentiation regulation. Blood 108:1716-1723

29. Chien W, Kumagai T, Miller CW, Desmond JC, Frank JM, Said JW, Koeffler HP 2004 Cyr61 suppresses growth of human endometrial cancer cells. J Biol Chem 279:53087-53096

30. Todorovic V, Chen CC, Hay N, Lau LF 2005 The matrix protein CCN1 (CYR61) induces apoptosis in fibroblasts. J Cell Biol 171:559-568

31. Croci S, Landuzzi L, Astolfi A, Nicoletti G, Rosolen A, Sartori F, Follo MY, Oliver N, De Giovanni C, Nanni P, Lollini PL 2004 Inhibition of connective tissue growth factor (CTGF/CCN2) expression decreases the survival and myogenic differentiation of human rhabdomyosarcoma cells. Cancer Res 64: 1730-1736

32. Lefrançois-Martinez AM, Bertherat J, Val P, Tournaire C, Gallo-Payet N, Hyndman D, Veyssière G, Bertagna X, Jean C, Martinez A 2004 Decreased expression of cyclic adenosine monophosphate-regulated aldose reductase (AKR1B1) is associated with malignancy in human sporadic adrenocortical tumors. J Clin Endocrinol Metab 89:3010-3019

33. de Fraipont F, El Atifi M, Cherradi N, Le Moigne G, Defaye G, Houlgatte R, Bertherat J, Bertagna X, Plouin PF, Baudin E, Berger F, Gicquel C, Chabre O, Feige JJ 2005 Gene expression profiling of human adrenocortical tumors using complementary deoxyribonucleic acid microarrays identifies several candidate genes as markers of malignancy. J Clin Endocrinol Metab 90:1819- 1829

34. Spencer SJ, Rabinovici J, Mesiano S, Goldsmith PC, Jaffe RB 1992 Activin and inhibin in the human adrenal gland. Regulation and differential effects in fetal and adult cells. J Clin Invest 90:142-149

35. Ishimoto H, Muench MO, Higuchi T, Minegishi K, Tanaka M, Yoshimura Y, Jaffe RB 2006 Midkine, a heparin-binding growth factor, selectively stim- ulates proliferation of definitive zone cells of the human fetal adrenal gland. J Clin Endocrinol Metab 91:4050-4056

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.