differential requirement for steroidogenic factor-1 gene

Differential Requirement for Steroidogenic Factor-1Gene Dosage in Adrenal Development VersusEndocrine Function

MICHELLE L. BLAND, ROBERT C. FOWKES, AND HOLLY A. INGRAHAM

Department of Physiology, Biomedical Sciences Graduate Program, University of California, SanFrancisco, San Francisco, California 94143-0444

The importance of steroidogenic factor-1 (SF-1)gene dosage in endocrine function is evidenced byphenotypes associated with the heterozygousstate in mice and humans. Here we examinedmechanisms underlying SF-1 haploinsufficiencyand found a striking reduction (12-fold) in SF-1heterozygous (�/�) adrenocortical size at embry-onic day (E) 12. Loss of one SF-1 allele led to aselective decrease in adrenal precursors within theadrenogonadal primordium at E10.0, without af-fecting the number of gonadal precursors, asmarked by GATA-4. Beginning at E13.5, increasedcell proliferation in SF-1 �/� adrenals allows theseorgans to approach but not attain a normal size.Remarkably, neural crest-derived adrenomedullaryprecursors migrated normally in SF-1 �/� and null

embryos. However, later in development, medul-lary growth was compromised in both genotypes.Despite the small adrenal size in SF-1 heterozy-gotes, an unexpected elevation in steroidogeniccapacity per cell was observed in primary adultadrenocortical SF-1 �/� cells compared with wild-type cells. Elevated cellular steroid output is con-sistent with the up-regulation of some SF-1 targetgenes in SF-1 �/� adrenals and may partially bedue to an observed increase in nerve growthfactor-induced-B. Our findings underscore theneed for full SF-1 gene dosage early in adrenaldevelopment, but not in the adult adrenal, wherecompensatory mechanisms restore near normalfunction. (Molecular Endocrinology 18: 941–952,2004)

DURING EMBRYONIC DEVELOPMENT, elabora-tion of genetic programs controlling organ growth

is critical for optimal performance in the adult. Onceproper organ size is achieved, cellular function is reg-ulated to meet physiological demands to maintain ho-meostasis. In the adult endocrine system, circulatingpeptide hormones serve as trophic signals to influenceorgan size and simultaneously regulate tissue-specificgene expression. This is exemplified in the adrenal,where ACTH both maintains adrenal cortex weight andstimulates steroidogenesis. In the embryo, geneticpathways controlling the earliest stages of adrenalgrowth are presumed to function cell autonomouslyand without input from other endocrine organs. Onefactor known to be essential for adrenal developmentis the orphan nuclear receptor steroidogenic factor-1(SF-1, AD4BP, NR5A1). Indeed, deletion of SF-1 inmice results in adrenal and gonadal agenesis and

postnatal lethality due to severe adrenal insufficiency(1–4). In humans, three distinct partial loss-of-functionmutations in SF-1 are associated with XY sex reversaland severe adrenal insufficiency, demonstrating thatSF-1 acts in a dose-dependent manner (5–7). Similarlyin mice, loss of one SF-1 allele leads to adrenal insuf-ficiency due to hypoplastic and disorganized adrenalglands, underscoring the importance of full SF-1 genedosage during adrenal organogenesis (8).

Normal adrenal development is apparent at embry-onic day (E) 9.0 when a population of cells derivedfrom the coelomic epithelium of the intermediate me-soderm begin to express SF-1; these cells form theadrenogonadal primordium (9). Later at E11.0, cells inthe adrenogonadal primordium differentiate and giverise to both the adrenal cortex and gonad. Furtherdevelopment at E13.0 involves the migration and infil-tration of neural crest cells into the developing adrenalcortex; these cells give rise to the adrenal medulla andbecome completely enveloped by the cortex by E15.5(10). Although SF-1 expression is restricted to adre-nocortical cells, loss of both SF-1 alleles results inmassive apoptosis in the adrenal cortex, and conse-quently, agenesis of both the adrenal cortex and me-dulla (1). Whereas human and mouse genetics haveestablished a role for SF-1 in adrenal development, thedefective developmental processes contributing todecreased adrenal growth in SF-1 �/� mice have notbeen explored.

Abbreviations: BrdU, Bromo-deoxyuridine; 8Br-cAMP,8-bromo-cAMP; D�H-nLacz, dopamine �-hydroxylase-nu-clear LacZ; DTT, dithiothreitol; E, embryonic day; LRH, liverreceptor homolog; MC2R, melanocortin 2 receptor; NGFI-B,nerve growth factor-induced-B; SCC, side chain cleavage;Ser, serine; SF-1, steroidogenic factor-1; SR-B1, scavengerreceptor-B1; StAR, steroidogenic acute regulatory protein;TUNEL, terminal deoxynucleotidyl transferase-mediated de-oxyuridine triphosphate nick end-labeling.

Molecular Endocrinology is published monthly by TheEndocrine Society (http://www.endo-society.org), theforemost professional society serving the endocrinecommunity.

0888-8809/04/$15.00/0 Molecular Endocrinology 18(4):941–952Printed in U.S.A. Copyright © 2004 by The Endocrine Society

doi: 10.1210/me.2003-0333

941

on December 27, 2004 mend.endojournals.orgDownloaded from

http://mend.endojournals.org

In the adult adrenal, extensive in vitro studies havesuggested that SF-1 functions to coordinately regulatebasal expression of steroidogenic genes such as ste-roidogenic acute regulatory protein (StAR), scavengerreceptor-B1 (SR-B1), melanocortin 2 receptor (MC2R),and the steroid hydroxylases. In addition, SF-1 is pro-posed to mediate ACTH-stimulated up-regulation ofthese steroidogenic genes via the cAMP-protein ki-nase A pathway (11). The mechanism linking SF-1 withcAMP- and protein kinase A-mediated increases insteroidogenic gene expression has yet to be deter-mined. Several groups have proposed that posttrans-lational modifications of SF-1 in response to extracel-lular signaling modulate its activity (12–15). Despite anabundance of in vitro evidence supporting SF-1’s cen-tral role in regulating steroidogenic gene expression,recent findings showed a paradoxical increase in SF-1target gene expression in SF-1 �/� adrenals thatexpress low levels of SF-1 (8, 16). These results are atodds with SF-1’s dose-dependent activity during ad-renal development and raise questions as to whetherSF-1 is critical for activating steroidogenic gene ex-pression or whether functionally redundant pathwaysare induced in SF-1 heterozygous mice.

In this study, we investigated the mechanisms thatunderlie SF-1 haploinsufficiency beginning at the ear-liest stage of adrenal development. Specifically, weasked what mechanisms might account for reducedadrenal size in SF-1 heterozygotes, including: 1) an

increase in apoptosis, 2) a decrease in cell prolifera-tion, 3) a defect in homing of medullary precursors tothe developing adrenal cortex, and/or 4) a decrease inallocation of cells in the adrenogonadal primordium tothe adrenal. Moreover, we asked how a reduction inSF-1 gene dosage affected the steroidogenic capacityof adult SF-1 �/� adrenocortical cells. Our findingsdemonstrate that SF-1 gene dosage is most critical atthe onset of adrenal development within the adreno-gonadal primordium. We propose that compensatorypathways deployed later in adrenal development andin the adult allow SF-1 heterozygous adrenals to func-tion at high, albeit insufficient, levels in the adultmouse.

RESULTS

Early Adrenal Development Is SeverelyCompromised in SF-1 Heterozygous Mice

SF-1 functions in a dose-dependent manner in mice toaffect adrenal function, but the precise stage of adre-nal growth compromised in SF-1 heterozygotes re-mains unknown. Consistent with our previous findingsin adult mice, SF-1 �/� adrenals were clearly smallerthan �/� adrenals at late stages of embryonic devel-opment, E18.5 (Fig. 1A) (8). Earlier in development, atE13.5, we noted a more dramatic difference in SF-1

Fig. 1. Adrenal Size Is Decreased in SF-1 �/� Embryos throughout DevelopmentA, Genitourinary systems were dissected from E18.5 SF-1 �/�, �/�, and �/� embryos. Arrows point to adrenal glands. Bar,

1 mm; k, kidney; g, gonad; b, bladder. B, SF-1 immunoreactivity in transverse sections of E13.5 embryos shows decreasedadrenal cortex size (arrows) in SF-1 heterozygotes. Bar, 250 �m; sc, spinal cord; da, dorsal aorta; g, gonad. C, Adrenal size (crosssection area) is decreased in heterozygous embryos (�/�, black bars) compared with wild-type embryos (�/�, green bars) fromE12.0-E18.5, n � 3–5 embryos per group; **, P � 0.01 vs. �/�.

942 Mol Endocrinol, April 2004, 18(4):941–952 Bland et al. • Compensation for SF-1 Haploinsufficiency



�/� and �/� adrenal size (Fig. 1B). Quantification ofadrenal size throughout development revealed a 12-fold decrease in size at the earliest stages of develop-ment, whereas at late time points, heterozygous adre-nals were only 2-fold smaller than wild-type adrenals(Fig. 1C).

Increased Cell Proliferation at Late Stages ofSF-1 �/� Adrenal Development

We examined whether alterations in cell death and cellproliferation contributed to decreased adrenal size inSF-1 �/� embryos. Here we confirmed the results ofLuo and colleagues (1) that adrenal and gonadal agen-esis in SF-1 �/� mice is due to programmed celldeath at E12.0. However, rates of apoptosis did notdiffer between SF-1 �/� and �/� embryos at anytime point examined (Fig. 2, A and B, and data notshown). We next asked whether a proliferation defectcould account for decreased adrenal size in SF-1 �/�embryos and predicted that bromo-deoxyuridine(BrdU) labeling and histone-3 phosphorylation (indicesof S phase and mitosis, respectively) would be de-creased in SF-1 �/� embryos (17–20). Unexpectedly,although no apparent difference in BrdU labeling wasobserved between �/� and �/� embryos at E12.5, asignificant increase in BrdU-positive cells was de-tected in SF-1 �/� adrenals by E13.5 (Fig. 3, A and B).Two additional stages in development (E15.5 and

E17.5) also showed a significant increase in BrdU la-beling and histone-3 phosphorylation in �/� adrenalswhen compared with �/� adrenals (Fig. 3, C and D).

Given that SF-1 �/� and �/� mice exhibit abnor-mal adrenocortical development, we asked whethermedullary development was affected by loss of one orboth alleles of SF-1 (1, 8). To examine migration ofsympathoadrenal neural crest precursors [�-gal (�)] tothe adrenal cortex, SF-1 �/� mice were crossed withmice expressing LacZ under the control of the humandopamine-�-hydroxylase promoter (21). At E13.5, anequivalent number of �-gal (�) cells migrated to theadrenal cortex in SF-1 �/� and �/� embryos. Sur-prisingly, the same number of �-gal (�) cells alsomigrated to the same rostral-caudal location in SF-1null embryos, despite the lack of adrenocortical cells(Fig. 4A). However, by E15.0, no tissue correspondingto an adrenal medulla was found in SF-1 �/� em-bryos. In the presence of one functional SF-1 allele,growth of the adrenal medulla was diminished and noappreciable infiltration of �-gal (�) cells into the adre-nal was observed at E15.0 or E16.5 compared withwild type (Fig. 4B and data not shown).

Finally, we asked whether SF-1 haploinsufficiencyaffected the earliest stage of adrenal development,when adrenal and gonadal precursors are found in acommon primordium derived from the coelomic epi-thelium of the intermediate mesoderm at E9.0. This

Fig. 2. Increased Apoptosis Cannot Account for Decreased SF-1 �/� Adrenal SizeA, SF-1 immunoreactivity in genital ridges (white, left panels) and TUNEL staining (for detection of programmed cell death, right

panels) in adjacent transverse sections of SF-1 �/�, �/�, and �/� embryos at E12.0. Arrows indicate TUNEL-positive cells inthe SF-1 null embryo. Bar, 100 �m; da, aorta; m, mesonephros. B, TUNEL-positive cells per genital ridge area in wild-type (�/�,black bars), heterozygous (�/�, gray bars), and knockout embryos (�/�, white bars) from E11.0-E12.0. Rates of apoptosis didnot differ between wild-type and heterozygous embryos at any time point examined, although, as expected, rates of apoptosiswere increased in knockout embryos at E11.5 and E12.0 (**, P � 0.01 vs. wild type, n � 3–4 embryos per group).

Bland et al. • Compensation for SF-1 Haploinsufficiency Mol Endocrinol, April 2004, 18(4):941–952 943



precedes the stage when the adrenogonadal primor-dium splits, and the adrenal cortex and gonad begin todevelop separately (E11.0). Although SF-1 is requiredfor both adrenal and gonadal development, GATA-4 isrequired for gonadal development but is dispensablefor adrenal development (22–24). In comparing theexpression patterns of GATA-4 and SF-1 at E12.0, wefound that, unlike SF-1, GATA-4 marks only gonadalprogenitors (Fig. 5A). At E10.0, SF-1/GATA-4 double-positive cells were observed in the coelomic epithe-lium of the intermediate mesoderm, whereas SF-1 sin-gle-positive cells were observed dorsal and somewhatrostral to the SF-1/GATA-4 double-positive population(Fig. 5B). We found that the total area of the adreno-gonadal primordium did not differ between �/� and�/� embryos (data not shown). However, the percent-age of SF-1 single-positive cells in the adrenogonadalprimordium was decreased significantly in SF-1 �/�

embryos, suggesting that SF-1 gene dosage is mostcritical at the onset of adrenal development (Fig. 5C).

SF-1 �/� Adrenals Have Increased Capacity forCorticosterone Production

We have previously shown that adult SF-1 �/� miceexhibit blunted corticosterone secretion in response tostress. Here we asked whether this impaired glucocor-ticoid secretion resulted from decreased adrenocorti-cal mass or a reduced steroidogenic capacity due tolower levels of SF-1. Surprisingly, SF-1 �/� adrenalscontain more corticosterone per milligram of adrenalweight compared with �/� adrenals (�/�: 70.0 ng/mg, �/�: 115.9 ng/mg); this finding predicts that cor-ticosterone secretion per cell would be increased inSF-1 �/� adrenals. To test this hypothesis, we mea-sured corticosterone secretion from equal numbers of

Fig. 3. Cell Proliferation Is Increased in SF-1 Heterozygous AdrenalsCell proliferation in embryonic adrenals was studied by measuring BrdU incorporation and histone 3 phosphorylation (indices

of S phase and mitosis, respectively). A, BrdU (red) and SF-1 (green) immunoreactivity in transverse sections of SF-1 wild-typeand heterozygous embryos at E12.5 and E13.5. White dotted lines outline the embryonic adrenal cortex. Bar, 100 �m; g, gonad.B, The ratio of BrdU and SF-1 double-positive cells (yellow) to the total number of SF-1 positive cells is increased in heterozygousembryos (�/�, gray bars) compared with wild type (�/�, black bars) at E13.5 (*, P � 0.05, n � 3 per group) but not at E12.5 (P �0.093, n � 4 per group). C, BrdU (red) and phospho-histone 3 (green) immunoreactivity in cross sections of SF-1 wild-type andheterozygous adrenals at E17.5. Bar, 100 �m. D, The number of BrdU positive cells per area of adrenal is increased inheterozygous embryos (�/�, gray bars) compared with wild type (�/�, black bars) at E15.5 and E17.5 (**, P � 0.01, n � 4 pergroup). Phospho-histone 3 labeling (positive cells per 104 �m2) is also increased in heterozygous adrenals compared with wildtype at both E15.5 (�/�: 1.20 � 0.03; �/�: 4.01 � 0.29; P � 0.01; n � 4 per group) and E17.5 (�/�, 0.99 � 0.03; �/�, 1.89 �0.15; P � 0.01; n � 4 per group).




SF-1 �/� and �/� adrenocortical cells stimulatedwith 8-bromo-cAMP (8Br-cAMP). SF-1 �/� adreno-cortical cells secreted significantly more corticoste-rone basally and in response to 0.1 mM, 0.5 mM, and 1

mM 8Br-cAMP compared with �/� adrenocorticalcells (Fig. 6A). Morphological inspection of cultured�/� and �/� adrenocortical cells revealed no majordifferences (Fig. 6B). Further analysis of SF-1 �/�

Fig. 4. Normal Medullary Development Requires a Full Dose of SF-1 in the CortexSF-1 heterozygous mice were crossed with D�H-nLacZ transgenic mice, and embryos were subjected to �-gal staining to

identify neural crest cells that give rise to the adrenal medulla and sympathetic ganglia. A, Equivalent numbers of medullaryprecursors (blue) arrived at the adrenocortical blastema or its approximate location at E13.5 in SF-1 �/�, �/�, and �/� embryos.Medullary precursors appeared to infiltrate the wild-type adrenal cortex at this stage. Black dotted lines outline the edges of theadrenal cortex (�/� and �/� embryos) or the ventral body wall (�/� embryo). Bar, 100 �m; da, dorsal aorta; sg, sympatheticganglia. B, By E16.5, wild-type adrenals contained a central medulla (arrow), whereas medullary cells had not appreciablyinfiltrated heterozygous adrenals. In SF-1 knockout embryos, no tissue corresponding to an adrenal medulla was found. Dottedlines outline the adrenal glands. Bar, 500 �m; sg, sympathetic ganglia; k, kidney.

Fig. 5. GATA-4 Immunoreactivity Defines a Subset of Cells in the Adrenogonadal PrimodiumA, GATA-4 (red) and SF-1 (green) immunoreactivity in transverse sections of E12.0 embryos. SF-1 and GATA-4 colocalize

(yellow) in the gonads but not in the adrenals (arrows). Note decreased adrenal size in the heterozygous embryo. Bar, 100 �m;g, gonad; da, dorsal aorta; ce, coelomic epithelium. B, GATA-4 and SF-1 immunoreactivity in longitudinal sections of E10.0embryos. Arrows point to SF-1-positive, GATA-4-negative cells (green) dorsal to the SF-1 and GATA-4 double-positive cells(yellow) that form the bulk of the adrenogonadal primordia. Bar, 100 �m; m, mesonephros. C, The percentage of SF-1-positive,GATA-4-negative cells in the adrenogonadal primordia is decreased in SF-1 heterozygous (�/�) embryos at E10.0. The area ofthe adrenogonadal primordia was calculated by measuring the area of all SF-1 immunoreactive cells in 15–25 sections per embryofor each genotype, n � 4 embryos per genotype.




adrenocortical primary cells revealed reduced expres-sion of SF-1, whereas expression of two rate-limitingproteins in steroidogenesis StAR and side chain cleav-age (SCC), was increased (Fig. 6C); these results areidentical with those obtained from whole �/� adrenals(8, 16). Collectively, these data suggest that factorsdistinct from SF-1 drive increased expression of ste-roidogenic genes, and thus, would account for theincreased cellular function observed in SF-1 �/� ad-renocortical cells.

Mechanisms Underlying Increased SF-1 TargetGene Expression in SF-1 �/� Adrenals

Given that ACTH is the primary regulator of steroido-genic gene expression and adrenal mass (25), ele-vated basal ACTH levels observed in SF-1 �/� micemost likely drive the increased function observed inSF-1 �/� adrenals (8). Indeed, inhibiting pituitaryACTH secretion with dexamethasone treatment re-sulted in significant decreases in both wild-type andheterozygous adrenal weight due to decreased corti-cal cell size (Fig. 7, A and B, and data in legend) (26).Dexamethasone treatment resulted in a marked re-duction in the levels of StAR, SCC, and SR-B1, withequivalent, low levels observed in wild-type and het-erozygous adrenals (Fig. 7, C and D). By contrast,SF-1 levels were unaltered after hormone treatment ineither genotype. Finally, dexamethasone treatmentalso suppressed corticosterone secretion and led tohigh, equivalent levels of CD4�CD8� thymocyte pro-grammed cell death in both genotypes (data notshown) (8). These results demonstrate that ACTH

maintains high expression levels of SF-1 target genesin SF-1 �/� adrenals.

Given that SF-1 is a primary regulator of basal andcAMP-stimulated steroidogenic gene expression inthe adrenal cortex, increased SF-1 target gene ex-pression and steroidogenic capacity observed inSF-1 �/� adrenals are paradoxical with reducedSF-1 levels. Therefore, we explored potential mech-anisms that would increase steroidogenic gene ex-pression downstream of ACTH signaling. One pos-sibility is that phosphorylation of SF-1 on serine(Ser) 203 may be increased in SF-1 �/� adrenalsbecause this posttranslational event is known toenhance SF-1’s ability to recruit coactivators (13).Using an antibody directed against phospho-Ser203 to supershift SF-1 bound to its response ele-ment in gel shift assays (15), we found that the basalphosphorylation state of SF-1 was not significantlydifferent between wild-type and heterozygous adre-nals (Fig. 8A and data in legend). Other potentialmechanisms may include altered expression of tran-scription factors that positively regulate steroido-genic genes or negatively regulate SF-1 in heterozy-gous adrenals. One such factor is the orphannuclear receptor liver receptor homolog-1 (LRH-1).LRH-1 shares high identity with SF-1 and thus couldpotentially regulate SF-1 target genes (27–29). How-ever, no appreciable levels of LRH-1 transcriptswere detected in either �/� or �/� adrenals usingNorthern and RT-PCR analyses (data not shown).

Another candidate factor that binds to similar DNAbinding sites as SF-1 is the orphan nuclear receptornerve growth factor-induced-B (NGFI-B) (30). More-

Fig. 6. SF-1 �/� Adrenocortical Cells Have Increased Steroidogenic CapacityA, Wild-type and heterozygous cortical cells (20,000 cells per tube, two tubes per treatment) were incubated for 90 min with

vehicle or 8Br-cAMP at the concentration indicated. Corticosterone secreted into the media was measured by RIA. Heterozygouscells (�/�, gray bars) secreted more corticosterone than wild-type cells (�/�, black bars) in response to 0.1 mM 8Br-cAMP, *,P � 0.05. Differences in corticosterone secretion at 0.5 mM and 1 mM 8Br-cAMP did not reach statistical significance due tovariability (P � 0.12 and P � 0.11, respectively). B, SF-1 wild-type and heterozygous adrenocortical cells in primary culture showsimilar morphology (oil red O staining, gray; nuclei, black). Bar, 25 �m. C, Western blot analyses of SF-1, StAR, SCC, and actinlevels in wild-type and heterozygous primary adrenocortical cells.




over, NGFI-B expression is induced by ACTH, andNGFI-B, in turn, is thought to participate in ACTH-induced up-regulation of 21-hydroxylase (31–33).Northern blot analysis showed a 1.4-fold up-regula-tion of NGFI-B in adult heterozygous adrenals com-pared with wild type, whereas decreased SF-1 ex-pression was observed in �/� adrenals (Fig. 8, Band C). Up-regulation of NGFI-B expression wasconfirmed by Western blot analysis, with elevated

NGFI-B protein levels (2.3-fold on average) ob-served in SF-1 �/� adrenals compared with �/�adrenals. It should be noted that the broadly migrat-ing NGFI-B signal consists of multiple bands due tohyperphosphorylation, as previously reported (Fig.8D) (34). Finally, we excluded that diminished levelsof a putative repressor of SF-1, Dax1, might accountfor the up-regulation of SF-1 target genes. Instead,we find equivalent levels of Dax1 protein in wild-type

Fig. 7. Dexamethasone Treatment Normalizes Cellular Hypertrophy and Steroidogenic Gene Expression in SF-1 �/� AdrenalsAdrenal histology and gene expression were assessed in SF-1 wild-type and heterozygous mice treated with vehicle or

dexamethasone for 3 d. Dexamethasone treatment inhibited corticosterone secretion after 10 min of restraint stress in wild-typeand heterozygous mice (vehicle: �/�: 24.7 � 1.7 �g/dl, �/�: 18.3 � 2.9 �g/dl; dexamethasone: �/�: 0.8 � 0.2 �g/dl, �/�: 1.3 �0.4 �g/dl; P � 0.01 vs. vehicle treated). A, Toluidine blue staining of adrenal cross sections showed that in vehicle-treated mice,SF-1 heterozygous adrenocortical cells are significantly larger than wild-type cells (cells per 0.01 mm2: �/�, 98.4 � 0.4; �/�,73.2 � 0.3; P � 0.01 vs. wild type). Dexamethasone treatment reversed SF-1 �/� adrenocortical cellular hypertrophy andnormalized differences in cell size between wild-type and heterozygous adrenals (cells per 0.01 mm2: �/�, 100.7 � 0.6; �/�,108.3 � 0.6, P � 0.06 vs. wild type). Bar, 100 �m. B, Dexamethasone treatment (dex, gray bars) led to decreased adrenal weightin wild-type and heterozygous mice compared with vehicle treatment (veh, black bars), *, P � 0.05 vs. vehicle. SF-1 heterozygousadrenals weighed less than wild-type adrenals regardless of vehicle or dexamethasone treatment (**P � 0.01 vs. wild type; n �4 per group). C, Western blot analyses of SF-1, SCC, SR-B1, StAR, and actin levels in adrenals from vehicle- and dexamethasone-treated SF-1 wild-type and heterozygous mice (n � 2–3 per group). D, SF-1, SCC, SR-B1, and StAR levels were normalized toactin levels. Protein levels in vehicle-treated heterozygous mice (gray bars), dexamethasone-treated wild-type mice (black,stippled bars), and dexamethasone-treated heterozygous mice (gray, stippled bars) are expressed as fold increases or decreasesrelative to vehicle-treated wild-type mice (black bars) (n � 3–4 per group).




and heterozygous adrenals, with or without dexa-methasone treatment (Fig. 8E).

DISCUSSION

Our results demonstrate that loss of one SF-1 alleleresults in a dramatic reduction of adrenocorticalprecursors, which ultimately leads to adrenal insuf-ficiency in the adult. However, partial compensationfor decreased adrenal mass occurs both during lateadrenal development and in the adult adrenal. Theearly growth defect in the SF-1 �/� adrenocorticalprimordium is met with increased cell proliferationlater in development that allows �/� adrenals toapproach but not attain �/� adrenal size. Unex-pectedly, the loss of SF-1 protein observed in het-

erozygous adrenals leads to a marked increaserather than a decrease in steroidogenic capacity dueto elevated steroidogenic proteins. Indeed, on a percell basis, SF-1 �/� adrenocortical cells producemore corticosterone per dose of cAMP than �/�adrenocortical cells. Although these cellularchanges permit SF-1 heterozygotes to maintain rel-atively normal basal glucocorticoid secretion, theirability to secrete sufficient glucocorticoids duringsevere stress remains limited by the overall loss ofadrenal precursors during development (8).

SF-1 gene dosage is most critical during the earlieststages of adrenal development as evidenced by thesevere reduction of adrenocortical precursors in SF-1�/� embryos at E12.0. Whereas our immunocyto-chemical analysis of GATA-4 and SF-1-positive cellsat E10.0 revealed that the total size of the adrenogo-

Fig. 8. Up-Regulation of NGFI-B in SF-1 �/� Adrenal GlandsA, EMSA of SF-1 from wild-type and heterozygous adrenal nuclear extracts binding to the gonadotrope-specific element (GSE)

from the �-glycoprotein subunit promoter. The positions of the SF-1-DNA complex and free probe are indicated (arrows).Increasing amounts of phospho-Ser 203 antisera resulted in the formation of a larger protein complex (arrowhead). Quantitationof results of three separate experiments revealed that, at the highest antibody concentration used, the ratio of shifted to total SF-1did not differ between genotypes (�/�, 52 � 2%; �/�, 58 � 7%; P � 0.68). B, Northern blot analyses of NGFI-B, SF-1, and ActinmRNA levels in SF-1 wild-type and heterozygous adrenals. C, NGFI-B was up-regulated 1.4-fold, and SF-1 was expressed at1.8-fold lower levels in heterozygous adrenals compared with wild-type adrenals. D, NGFI-B was up-regulated in adult SF-1heterozygous adrenals compared with wild-type (n � 6 per genotype). NGFI-B migrates as a broad band due to hyperphos-phorylation. The fold induction of NGFI-B in each sample (relative to the lowest level observed) is indicated above each lane. E,Dax1 levels were equivalent in SF-1 wild-type and heterozygous adrenals from mice treated with vehicle or dexamethasone.




nadal primordium does not differ in �/� and �/�embryos, our findings indicated that fewer cells arededicated to the adrenal (SF-1 positive, GATA-4 neg-ative) in SF-1 heterozygotes at this stage of develop-ment. Our data suggest that SF-1 is important forexpansion of adrenal progenitors in the adrenogo-nadal primordium. Separation of adrenal and gonadalprecursors may also rely on SF-1. At this step, SF-1could act in concert with the signaling molecule Wnt4,which was recently shown to repress migration ofadrenal precursors into the developing gonad (35).Delineating SF-1’s role in the earliest steps of adrenaldevelopment will require identification of SF-1’s em-bryonic target genes.

A consequence of impaired adrenocortical develop-ment in SF-1 �/� and �/� embryos is disruptedadrenomedullary development. We found that the ini-tial migration of neural crest cells to the adrenal cortex(or its approximate location) occurs normally regard-less of SF-1 dosage, showing that neural crest cells donot rely on signals from adrenocortical cells for properhoming. However, at later stages of development,medullary cells are lost in SF-1 �/� embryos and themedulla fails to infiltrate the small SF-1 �/� adrenalcortex. Our data show that normal growth and survivalof the adrenal medulla depends on SF-1 function in theadrenal cortex. It will be of interest to determinewhether the requirement for SF-1 is indirect or direct.For instance, normal medullary growth may simplyrequire sufficient adrenocortical mass. Alternatively,SF-1 might directly regulate genes that serve as para-crine growth factors for medullary cells. This occurs inskin, where target-derived growth factors supportexpansion and correct localization of neural crest-derived melanocyte precursors (36).

How can the early growth defects in SF-1 �/�adrenals be reconciled with their impressive increasein cell proliferation later in development? Perhaps de-creased adrenal mass is somehow sensed at E13.5and leads to increased adrenal growth factor levelsthat drive cell proliferation in SF-1 heterozygotes. Thisstrategy would parallel the compensatory response toSF-1 haploinsufficiency in the adult adrenal in whichelevated ACTH levels lead to increased steroidogeniccapacity per cell. Another anterior pituitary hormone,pro-�-MSH, stimulates adrenal cell proliferation afterundergoing local proteolysis (carried out by adrenalsecretory protease) that produces a shorter form withmitogenic activity (26, 37, 38). This signaling pathwayis thought to account for the compensatory adrenalgrowth response after removal of one adrenal. Thelack of this growth response in SF-1 �/� mice mightsuggest that SF-1 participates in pro-�-MSH signaling(39). It will be of interest to determine the interplaybetween SF-1, ACTH, and/or pro-�-MSH signalingpathways during early adrenal development.

Ultimately, increased cell proliferation in SF-1 �/�adrenals cannot compensate for the early deficits inadrenal development. SF-1 �/� adrenals do not se-crete enough corticosterone to support normal phys-

iological responses to stress, but elevated ACTHlevels in �/� mice stimulate steroidogenic gene ex-pression and raise steroidogenic capacity per cell,ensuring relatively normal basal corticosterone secre-tion (1, 8). Are heterozygous levels of SF-1 sufficient tomaintain SF-1 target gene expression at supernormallevels in �/� adrenals or are other mechanisms suchas posttranslational modification, ligand availability, orup-regulation of other transcription factors employed?Although it is known that phosphorylation of SF-1 onSer 203 promotes SF-1 transcriptional activity, we didnot detect significant differences in the ratio of phos-phorylated to total SF-1 in �/� and �/� adrenals.Increased ligand availability is also an unlikely mech-anism because neither an exogenous nor an endoge-nous obligatory ligand has been identified for SF-1,consistent with the fact that SF-1 is constitutively ac-tive in a variety of steroidogenic and nonsteroidogeniccell lines (40). Furthermore, biophysical and structuralevidence suggests that members of nuclear receptorsubfamily V (SF-1 and LRH-1) adopt an active confor-mation in the apparent absence of ligand (15, 41).

Increased or decreased activity of other transcrip-tion factors may underlie the paradoxical increases inSF-1 target gene expression in SF-1 �/� adrenals.For example, the orphan nuclear receptor Dax1 re-presses SF-1 activity in vitro (42–44). However, loss ofDax1 does not lead to further increases in SF-1 targetgene expression in SF-1 �/� adrenals (16). Indeed,we found that Dax1 levels were similar in SF-1 �/�and �/� adrenals, regardless of circulating ACTH lev-els. Our data show that NGFI-B levels are significantlyelevated in SF-1 �/� adrenals, supporting a compen-satory role for NGFI-B in the regulation of steroido-genic gene expression. However, the normal adrenalresponses to stress in NGFI-B null mice suggest thatNGFI-B does not normally regulate steroidogenicgenes, including 21-hydroxylase (32). Taken together,these data indicate that NGFI-B does not serve as theprimary regulator of steroidogenic genes but may as-sume a more important role when SF-1 dosage isreduced.

Although our data confirm and extend the essen-tial role of SF-1 in early embryonic adrenal develop-ment, they also lead us to question SF-1’s preciserole in late adrenal development and in the adultadrenal gland. The increase in steroidogenic capac-ity in SF-1 �/� adrenocortical cells is particularlyunexpected given the central role that SF-1 hasbeen thought to play in coordination of adrenal ste-roidogenesis. Our study strongly suggests that al-ternative molecular mechanisms exist to increaseexpression of many SF-1 target genes and illus-trates the differential requirement for transcriptionfactor function in development and the adult. Futureexperiments aimed at generating a temporal-specific SF-1 deletion in the adult will be essentialfor dissecting SF-1’s well-established role in adrenaldevelopment from its less-defined role in regulatingadult adrenal steroidogenesis.




MATERIALS AND METHODS

Animal Experiments

SF-1 �/� and �/� mice obtained from The Jackson Labora-tory (Bar Harbor, ME), were maintained on a C57BL/6J � FVBbackground, and cared for in accordance with National Insti-tutes of Health (NIH) guidelines. Experimental procedures wereapproved by the University of California, San Francisco Labo-ratory Animal Research Committee. Mice were kept on a 12-hlight, 12-h dark cycle (lights on 0600–1800 h) and were givenfood and water ad libitum. Male mice 6–8 wk old were used forall experiments unless otherwise noted. Dopamine �-hydroxy-lase-nuclear LacZ (D�H-nLacz) transgenic mice (from Dr. R.Kapur, University of Washington) were crossed with SF-1 �/�mice. Animals were genotyped as described previously (45). Fordexamethasone experiments, �/� and �/� mice were injectedip with saline or dexamethasone sodium phosphate (0.5 mg/kg,Sigma, St. Louis, MO), twice per day (0900 and 1730 h) for 3 d(n � 5 per group). On the fourth day beginning at 0800 h, micewere exposed to 10 min of restraint stress and decapitated.Thymus cells from vehicle- and dexamethasone-treated micewere analyzed by flow cytometry as described previously (8).Plasma and adrenal corticosterone were measured using acommercially available (ICN Pharmaceuticals, Costa Mesa,CA) kit.

Measurements of Cellular Proliferation and Cell Death

For BrdU labeling, timed-pregnant mice received an ip injec-tion of BrdU (50 mg/kg, Sigma). After a 1-h pulse, wholeembryos or fetal adrenals were collected, fixed overnight in4% paraformaldehyde, cryoprotected in 30% sucrose, andfrozen in OCT compound (Tissue Tek Sakura, Torrance, CA).Cryostat sections (10 �m) were treated with 2 N HCl at 37 Cfor 20 min to denature DNA, blocked in 10% normal goatserum, incubated overnight at 4 C with rat anti-BrdU antisera(1:10, Harlan Sera-Lab) and either rabbit anti-SF-1 (1:1000) orrabbit antiphosphorylated histone 3 (phospho-His3, 1:1000,Upstate Biotechnologies, Waltham, MA), washed, and incu-bated for 2 h at room temperature with goat antirabbit Alexa488 and goat antirat Alexa 546 secondary antibodies (1:200each, Molecular Probes, Eugene, OR). Images were collectedwith a confocal microscope, and adrenal cross-section areaand the number of BrdU-positive (�) and phospho-His3 (�)cells per section were measured with the NIH Image pro-gram. The number of digitally counted BrdU (�) and phos-pho-His3 (�) cells was confirmed by visual assessment toensure appropriate parameter settings. Apoptosis was de-tected using an in-house terminal deoxynucleotidyl trans-ferase-mediated deoxyuridine triphosphate nick end-labeling(TUNEL) assay as described previously (45). For each sec-tion, the number of TUNEL (�) cells was divided by the areaof SF-1 immunoreactivity. For SF-1 �/� embryos, the aver-age number of TUNEL (�) cells per section was divided bythe average area of SF-1 immunoreactivity in SF-1 �/�genital ridges.

For measurement of adrenogonadal primordia size, em-bryos were staged by counting somites. Embryos were fixedas described above. Cryostat sections (10 �m) were blockedfor 30 min in 10% normal donkey serum, and incubatedovernight at 4 C with rabbit anti-SF-1 (1:200) and goat anti-GATA-4 (1:500, Santa Cruz Biotechnology, Inc., Santa Cruz,CA). The next day, sections were washed, incubated for 2 hat room temperature with donkey antirabbit Alexa 488 (1:200,Molecular Probes) and donkey antigoat Cy3 (1:200, JacksonImmunoResearch, West Grove, PA). Images were collectedwith a confocal microscope and the total areas of all SF-1 (�)cells and SF-1 (�), GATA-4 (�) cells were measured with NIHImage.

Histological Analysis

Detection of �-gal in D�H-nLacZ transgenic embryos wasperformed as described previously (21). Adrenal sections (10�m) were stained with toluidine blue O. Cellular hypertrophywas assessed by counting Hoechst-stained nuclei per 0.01mm2 in four sections per �/� and �/� adrenal (n � 4adrenals per group).

Western and Northern Analysis

Western analyses were carried out as described previously(8). For Western analysis, each lane represents a separateindividual, n � 3–6 per genotype. Additional antibodies usedin this study were: rabbit anti-SR-B1 (1:20,000, Novus Bio-logicals, Littleton, CO), mouse anti-NGFI-B [1:10,000, a kindgift of Dr. J. Milbrandt (Washington University, St. Louis,MO)], and goat antimouse horseradish peroxidase (1:10,000,Bio-Rad, Hercules, CA). For quantification of protein levels,scanning densitometry was performed on blots developedwith chemiluminescence (ECL, Amersham Biosciences, Pis-cataway, NJ). These levels were confirmed by quantifyingradioactive signals from Western blots performed with a ra-diolabeled goat antirabbit secondary antibody (NEN, Boston,MA). For Northern analyses, total RNA (20 �g) prepared fromSF-1 �/� and �/� adrenals was separated by formalde-hyde-gel electrophoresis, transferred to nylon membranes,and hybridized overnight at 42 C with random-primed, la-beled DNA probes for fragments of the mouse SF-1, mouseLRH, rat NGFI-B, and rat actin cDNAs. Membranes werewashed at medium stringency (0.2� sodium chloride sodiumcitrate, 0.1% sodium dodecyl sulfate at 42 C) and exposed toX-OMAT film (Kodak, Rochester, NY). For Northern blot andEMSA (see EMSA) experiments, radioactive signals werequantified with ImageQuant Mac software (Amersham Bio-sciences after exposure to a phosphorimager screen (Storm,Amersham Biosciences).

Primary Cell Culture

Adrenals from female mice were dissected free of fat,minced, and washed in culture medium (M-199 with 4 mg/mlBSA plus penicillin and streptomycin). Cells were incubatedin dispersal medium (M-199 containing 20 mg/ml BSA pluspenicillin and streptomycin, 2.5 mg/ml type I collagenase(Invitrogen, Carlsbad, CA), and 10 �g/ml deoxyribonucleaseI) for 30 min at 37 C with shaking and were dissociated byrepeated pipetting every 10 min, filtered over 70-�m nylonmesh, washed twice by centrifugation, and resuspended inculture medium. Equivalent numbers of SF-1 �/� and �/�cortical cells were incubated in 950 �l culture medium at 37C with 5% CO2. After 1 h, H2O or 8-bromo-cAMP (Sigma) atfinal concentrations of 0.1, 0.5, and 1 mM were added in avolume of 50 �l. Cells were incubated for 90 min, pelleted bycentrifugation, and the supernatant was removed for corti-costerone measurements. Cells were lysed with 2% sodiumdodecyl sulfate, 100 mM dithiothreitol (DTT), and 60 mM Tris-HCl (pH 6.8), and Western blot analyses was carried out asdescribed above. Small aliquots of �/� and �/� cells wereplated on tissue culture slides coated with collagen and in-cubated overnight at 37 C, 5% CO2 in culture medium plus10% fetal calf serum. The next day, cells were fixed with 4%paraformaldehyde and stained with oil red O.

EMSA

Nuclear extracts were prepared from adrenal glands collectedunder basal conditions at 1730 h. Adrenals were cleaned of fat,homogenized in cold PBS, and centrifuged at 4000 rpm for 5min. Cells from eight to 12 adrenals were resuspended in 400 �lbuffer A [10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1




mM EGTA, 1 mM DTT, and 0.5 mM phenylmethylsulfonyl fluoride]and incubated on ice for 15 min before the addition of 25 �l of0.1% Nonidet P-40 in buffer A. Nuclei were vortexed for 10 secand centrifuged at 11,000 rpm for 30 min. Nuclei were resus-pended in 50 �l buffer C [20 mM HEPES (pH 7.9), 0.4 M NaCl, 1mM EDTA, 1 mM EGTA, 1 mM DTT, and 1 mM phenylmethylsul-fonyl fluoride], rotated at 4 C for 15 min and then centrifuged for5 min at 11,000 rpm. The supernatant was subjected to EMSAsas follows: oligonucleotides encoding the SF-1 response ele-ment in the human glycoprotein hormone �-subunit promoter(forward: 5�-GCTGACCTTGTCGTCAC-3�, reverse: 5�-GTGAC-GACAAGGTCAGC-3�) were annealed and radiolabeled as de-scribed (14). In each binding reaction, 1–3 �g of adrenal nuclearprotein extracts were mixed with the labeled probes in 20 �lvolume of 20 mM Tris (pH 8.0), 60 mM KCl, 2 mM MgCl2, 1.2 mM

DTT, 12% glycerol, 2.5 �g poly (deoxyinosine-deoxycytosine),1% (wt/vol) BSA, incubated at room temperature for 5 minbefore the addition of 2 �l of probe (200,000 cpm) and incuba-tion for 15 min at 30 C. Typically, 8 �l of the reaction mixturewere resolved on a 5% native acrylamide gel, dried and visual-ized by autoradiography. For all antibody gel-shift experiments,0.1–3.0 �l of anti-phospho-SF-1 antiserum was added to thereaction minus probe and incubated on ice for 60 min.

Statistical Analysis

Data are presented as means � SEM. Unpaired two-tailed ttests and ANOVA were used to determine statisticalsignificance.

Acknowledgments

We wish to acknowledge Drs. Mary Dallman and MarionDesclozeaux [University of California, San Francisco (UCSF)]for discussions. We are especially grateful to Dr. C. Jamieson(University of California, Los Angeles) for measurement ofthymocyte apoptosis and Dr. R. Kapur (University of Wash-ington, Seattle, WA) for the generous gift of the D�H-nLacZtransgenic mice. We thank Drs. W. Miller (UCSF) for StAR andSCC antibodies, K. Morohashi (National Institute for BasicBiology, Okazaki, Japan) for the SF-1 antibody, and J. Mil-brandt (Washington University, St. Louis, MO) for the NGFI-Bantibody.

Received August 29, 2003. Accepted January 8, 2004.Address all correspondence and requests for reprints to:

Holly A. Ingraham, Department of Physiology, Box 0444,University of California, San Francisco, San Francisco, Cali-fornia 94143-0444. E-mail: [email protected].

This work was supported by the American Heart Associ-ation (Predoctoral Fellowship to M.L.B.) and by National In-stitutes of Health-National Institute of Diabetes and Digestiveand Kidney Diseases (RO1 to H.A.I.).

REFERENCES

1. Luo X, Ikeda Y, Parker KL 1994 A cell-specific nuclearreceptor is essential for adrenal and gonadal develop-ment and sexual differentiation. Cell 77:481–490

2. Luo X, Ikeda Y, Schlosser DA, Parker KL 1995 Steroido-genic factor 1 is the essential transcript of the mouseFtz-F1 gene. Mol Endocrinol 9:1233–1239

3. Sadovsky Y, Crawford PA, Woodson KG, Polish JA,Clements MA, Tourtellotte LM, Simburger K, Milbrandt J1995 Mice deficient in the orphan receptor steroidogenicfactor 1 lack adrenal glands and gonads but expressP450 side-chain-cleavage enzyme in the placenta and

have normal embryonic serum levels of corticosteroids.Proc Natl Acad Sci USA 92:10939–10943

4. Shinoda K, Lei H, Yoshii H, Nomura M, Nagano M, ShibaH, Sasaki H, Osawa Y, Ninomiya Y, Niwa O, MorohashiK-I, Li E. 1995 Developmental defects of the ventrome-dial hypothalamic nucleus and pituitary gonadotroph inthe Ftz-F1 disrupted mice. Dev Dyn 204:22–29

5. Achermann JC, Ito M, Hindmarsh PC, Jameson JL 1999A mutation in the gene encoding steroidogenic factor-1causes XY sex reversal and adrenal failure in humans.Nat Genet 22:125–126

6. Achermann JC, Ozisik G, Ito M, Orun UA, Harmanci K,Gurakan B, Jameson JL 2002 Gonadal determinationand adrenal development are regulated by the orphannuclear receptor steroidogenic factor-1, in a dose-de-pendent manner. J Clin Endocrinol Metab 87:1829–1833

7. Biason-Lauber A, Schoenle EJ 2000 Apparently normalovarian differentiation in a prepubertal girl with transcrip-tionally inactive steroidogenic factor 1 (NR5A1/SF-1) andadrenocortical insufficiency. Am J Hum Genet 67:1563–1568

8. Bland ML, Jamieson CA, Akana SF, Bornstein SR, Eisen-hofer G, Dallman MF, Ingraham HA 2000 Haploinsuffi-ciency of steroidogenic factor-1 in mice disrupts adrenaldevelopment leading to an impaired stress response.Proc Natl Acad Sci USA 97:14488–14493

9. Ikeda Y, Shen WH, Ingraham HA, Parker KL 1994 Devel-opmental expression of mouse steroidogenic factor-1,an essential regulator of the steroid hydroxylases. MolEndocrinol 8:654–662

10. Morohashi K 1997 The ontogenesis of the steroidogenictissues. Genes Cells 2:95–106

11. Parker KL, Schimmer BP 1997 Steroidogenic factor 1: akey determinant of endocrine development and function.Endocr Rev 18:361–377

12. Sewer MB, Waterman MR 2002 Adrenocorticotropin/cyclic adenosine 3�,5�-monophosphate-mediated tran-scription of the human CYP17 gene in the adrenal cortexis dependent on phosphatase activity. Endocrinology143:1769–1777

13. Hammer GD, Krylova I, Zhang Y, Darimont BD, SimpsonK, Weigel NL, Ingraham HA 1999 Phosphorylation of thenuclear receptor SF-1 modulates cofactor recruitment:integration of hormone signaling in reproduction andstress. Mol Cell 3:521–526

14. Fowkes RC, Desclozeaux M, Patel MV, Aylwin SJ, KingP, Ingraham HA, Burrin JM 2003 Steroidogenic factor-1(SF-1) and the gonadotrope-specific element (GSE) en-hance basal and pituitary adenylate cyclase-activatingpolypeptide (PACAP)-stimulated transcription of the hu-man glycoprotein hormone �-subunit gene (�GSU) ingonadotropes. Mol Endocrinol 17:2177–2188

15. Desclozeaux M, Krylova IN, Horn F, Fletterick RJ, Ingra-ham HA 2002 Phosphorylation and intramolecular stabi-lization of the ligand binding domain in the nuclearreceptor steroidogenic factor 1. Mol Cell Biol 22:7193–7203

16. Babu PS, Bavers DL, Beuschlein F, Shah S, Jeffs B,Jameson JL, Hammer GD 2002 Interaction betweenDax-1 and steroidogenic factor-1 in vivo: increased ad-renal responsiveness to ACTH in the absence of Dax-1.Endocrinology 143:665–673

17. Miller MW, Nowakowski RS 1988 Use of bromodeoxyuri-dine-immunohistochemistry to examine the proliferation,migration and time of origin of cells in the central nervoussystem. Brain Res 457:44–52

18. Wei Y, Yu L, Bowen J, Gorovsky MA, Allis CD 1999Phosphorylation of histone H3 is required for properchromosome condensation and segregation. Cell 97:99–109

19. Goto H, Tomono Y, Ajiro K, Kosako H, Fujita M, SakuraiM, Okawa K, Iwamatsu A, Okigaki T, Takahashi T, Ina-gaki M 1999 Identification of a novel phosphorylation site




on histone H3 coupled with mitotic chromosome con-densation. J Biol Chem 274:25543–25549

20. Gurley LR, D’Anna JA, Barham SS, Deaven LL, Tobey RA1978 Histone phosphorylation and chromatin structureduring mitosis in Chinese hamster cells. Eur J Biochem84:1–15

21. Kapur RP, Hoyle GW, Mercer EH, Brinster RL, PalmiterRD 1991 Some neuronal cell populations express humandopamine �-hydroxylase-lacZ transgenes transientlyduring embryonic development. Neuron 7:717–727

22. Ikeda Y, Swain A, Weber TJ, Hentges KE, Zanaria E, LalliE, Tamai KT, Sassone-Corsi P, Lovell-Badge R, Cam-erino G, Parker KL 1996 Steroidogenic factor 1 andDax-1 colocalize in multiple cell lineages: potential linksin endocrine development. Mol Endocrinol 10:1261–1272

23. Ketola I, Rahman N, Toppari J, Bielinska M, Porter-TingeSB, Tapanainen JS, Huhtaniemi IT, Wilson DB, Heikin-heimo M 1999 Expression and regulation of transcriptionfactors GATA-4 and GATA-6 in developing mouse testis.Endocrinology 140:1470–1480

24. Tevosian SG, Albrecht KH, Crispino JD, Fujiwara Y,Eicher EM, Orkin SH 2002 Gonadal differentiation, sexdetermination and normal Sry expression in mice requiredirect interaction between transcription partners GATA4and FOG2. Development 129:4627–4634

25. Simpson ER, Waterman MR 1988 Regulation of the syn-thesis of steroidogenic enzymes in adrenal cortical cellsby ACTH. Annu Rev Physiol 50:427–440

26. Dallman MF 1984 Control of adrenocortical growth invivo. Endocr Res 10:213–242

27. Wang ZN, Bassett M, Rainey WE 2001 Liver receptorhomologue-1 is expressed in the adrenal and can regu-late transcription of 11�-hydroxylase. J Mol Endocrinol27:255–258

28. Sirianni R, Seely JB, Attia G, Stocco DM, Carr BR, PezziV, Rainey WE 2002 Liver receptor homologue-1 is ex-pressed in human steroidogenic tissues and activatestranscription of genes encoding steroidogenic enzymes.J Endocrinol 174:R13–R17

29. Schoonjans K, Annicotte JS, Huby T, Botrugno OA, Fa-yard E, Ueda Y, Chapman J, Auwerx J 2002 Liver recep-tor homolog 1 controls the expression of the scavengerreceptor class B type I. EMBO Rep 3:1181–1187

30. Wilson TE, Fahrner TJ, Milbrandt J 1993 The orphanreceptors NGFI-B and steroidogenic factor 1 establishmonomer binding as a third paradigm of nuclear recep-tor-DNA interaction. Mol Cell Biol 13:5794–5804

31. Wilson TE, Mouw AR, Weaver CA, Milbrandt J, Parker KL1993 The orphan nuclear receptor NGFI-B regulates ex-pression of the gene encoding steroid 21-hydroxylase.Mol Cell Biol 13:861–868

32. Crawford PA, Sadovsky Y, Woodson K, Lee SL, Mil-brandt J 1995 Adrenocortical function and regulation of

the steroid 21-hydroxylase gene in NGFI-B-deficientmice. Mol Cell Biol 15:4331–4316

33. Davis IJ, Lau LF 1994 Endocrine and neurogenic regu-lation of the orphan nuclear receptors Nur77 and Nurr-1in the adrenal glands. Mol Cell Biol 14:3469–3483

34. Fahrner TJ, Carroll SL, Milbrandt J 1990 The NGFI-Bprotein, an inducible member of the thyroid/steroid re-ceptor family, is rapidly modified posttranslationally. MolCell Biol 10:6454–6459

35. Jeays-Ward K, Hoyle C, Brennan J, Dandonneau M,Alldus G, Capel B, Swain A 2003 Endothelial and steroi-dogenic cell migration are regulated by WNT4 in thedeveloping mammalian gonad. Development 130:3663–3670

36. Wehrle-Haller B, Weston JA 1995 Soluble and cell-boundforms of steel factor activity play distinct roles in mela-nocyte precursor dispersal and survival on the lateralneural crest migration pathway. Development 121:731–742

37. Bicknell AB, Lomthaisong K, Woods RJ, Hutchinson EG,Bennett HP, Gladwell RT, Lowry PJ 2001 Characteriza-tion of a serine protease that cleaves pro-�-melanotropinat the adrenal to stimulate growth. Cell 105:903–912

38. Lowry PJ, Silas L, McLean C, Linton EA, Estivariz FE1983 Pro-�-melanocyte-stimulating hormone cleavagein adrenal gland undergoing compensatory growth. Na-ture 306:70–73

39. Beuschlein F, Mutch C, Bavers DL, Ulrich-Lai YM, Enge-land WC, Keegan C, Hammer GD 2002 Steroidogenicfactor-1 is essential for compensatory adrenal growthfollowing unilateral adrenalectomy. Endocrinology 143:3122–3135

40. Mellon SH, Bair SR 1998 25-Hydroxycholesterol is not aligand for the orphan nuclear receptor steroidogenic fac-tor-1 (SF-1). Endocrinology 139:3026–3029

41. Sablin EP, Krylova IN, Fletterick RJ, Ingraham HA 2003Structural basis for ligand-independent activation of theorphan nuclear receptor LRH-1. Mol Cell 11:1575–1585

42. Crawford PA, Dorn C, Sadovsky Y, Milbrandt J 1998Nuclear receptor DAX-1 recruits nuclear receptor core-pressor N-CoR to steroidogenic factor 1. Mol Cell Biol18:2949–2956

43. Ito M, Yu R, Jameson JL 1997 DAX-1 inhibits SF-1-mediated transactivation via a carboxy-terminal domainthat is deleted in adrenal hypoplasia congenita. Mol CellBiol 17:1476–1483

44. Nachtigal MW, Hirokawa Y, Enyeart-VanHouten DL,Flanagan JN, Hammer GD, Ingraham HA 1998 Wilms’tumor 1 and Dax-1 modulate the orphan nuclear receptorSF-1 in sex-specific gene expression. Cell 93:445–454

45. Tran PV, Lee MB, Marin O, Xu B, Jones KR, Reichardt LF,Rubenstein JR, Ingraham HA 2003 Requirement of theorphan nuclear receptor SF-1 in terminal differentiationof ventromedial hypothalamic neurons. Mol Cell Neuro-sci 22:441–453

Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremostprofessional society serving the endocrine community.




differential requirement for steroidogenic factor-1 gene

Technology

adrenal growth

adult adrenal

adrenocortical sf

adrenal precursors

serine sf

deletion of sf

adrenal medulla

adrenal organogenesis