ZNF764 Haploinsufficiency May Explain Partial Glucocorticoid, Androgen, and Thyroid Hormone Resistance Associated with 16p11.2 Microdeletion

Abstract

Context:

Nuclear hormone receptors exert their transcriptional effects through shared cofactor molecules; thus, defects in such intermediate proteins may be associated with multiple hormone resistance. Microdeletion of small chromosomal segments results in hereditary or sporadic diseases by affecting expression of residing genes.

Objectives:

We describe a 7-yr-old boy with partial resistance to glucocorticoids, thyroid hormones, and possibly androgens. He was diagnosed as being in the autism spectrum disorder and had developmental delay and several facial morphological manifestations. We explored genes responsible for multiple hormone resistance of this case.

Results:

We found in this patient an approximately 1.1-Mb heterozygous 16p11.2 microdeletion, which included an approximately 500-kb unique deletion along with the common, previously reported approximately 600-kb 16p11.2 microdeletion. The small interfering RNA-based screening revealed that knockdown of ZNF764, which is located in the deleted segment unique to our case, significantly reduced glucocorticoid-, androgen-, and thyroid hormone-induced transcriptional activity of their responsive genes in HeLa cells, whereas its overexpression enhanced their transcriptional activity. The activities of the estrogen and progesterone receptors, cAMP response element-binding protein, and p53 were not affected in these cells. ZNF764 (zinc finger protein 764) expression was reduced in the patient's peripheral blood mononuclear cells, whereas exogenously supplemented ZNF764 recovered responsiveness to glucocorticoids in the patient's Epstein-Barr virus-transformed lymphocytes. The effect of ZNF764 on the glucocorticoid receptor transcriptional activity was mediated through cooperation with a general nuclear hormone receptor coactivator, transcriptional intermediary factor 1.

Conclusions:

ZNF764 haploinsufficiency caused by microdeletion may be responsible for the partial multiple hormone resistance observed in our patient. ZNF764 appears to be involved in glucocorticoid, androgen, and thyroid hormone action.

Lipophilic hormones, such as the steroids and the thyroid hormones, exert physiological functions essential for life (1–3). These actions are mediated by the specific receptors of the nuclear hormone receptor (NR) superfamily. Once ligands bind, they modulate transcriptional activity of positively regulated genes by attracting protein complexes, called coactivators, that bridge the DNA-bound receptors and the transcription initiation complex as well as enzymatically modify the chromatin-bound histones and other related molecules (4). NRs share some coactivators for transactivation (4); thus, defects in one such protein may potentially influence the transcriptional activity of several NRs and could develop pathologies that span over multiple hormones. One family with resistance to glucocorticoids, mineralocorticoids, and androgens was reported previously, with speculation of a congenital coactivator defect as a primary cause (5, 6). Among such coactivators, the transcriptional intermediary factor 1(TIF1)β, also known as the tripartite motif-containing (TRIM) 28, acts as a general coactivator for many NRs, together with its homologous molecule TIF1α (TRIM24) (7–9). TIF1 interacts with NRs and participates in their transcriptional regulation (7, 10). Zinc finger proteins (ZNFs), which contain Kruppel-associated box (KRAB)-A and/or KRAB-B domains, and multiple zinc fingers in their N- and C-terminal portion, respectively, also bind TIF1s and contribute to their transcriptional activities (7, 11, 12).

Rearrangements of the genome, e.g. microdeletions and microduplications, usually occur at the chromosomal portions containing low-copy repeats or segmental duplications and cause hereditary or sporadic diseases (13). One such hot spot for rearrangement is located in chromosome 16 at the segment spanning from 16p11.2 to 16p12.2 (14) and causes deletions/duplications with sizes ranging from approximately 0.2–9 Mb (15–18). Patients harboring genetic rearrangements in this hot spot develop some or all of the following characteristic manifestations: cognitive impairment and developmental delays, schizophrenia, seizures, short stature, distinct facial features, orofacial clefts, heart defects, and hearing loss with frequent ear infections. Some cases harboring a common proximal approximately 600-kb deletion in 16p11.2 also demonstrate autism (19), whereas an approximately 200-kb deletion located in its distal region is associated with developmental delay and obesity (20). These manifestations appear to be caused by copy number changes in the dosage-dependent genes located in the affected chromosomal regions (13).

In this manuscript, we report a case harboring a heterozygous approximately 1.1-Mb microdeletion at chromosome 16p11.2, who demonstrated partial tissue resistance to glucocorticoids, thyroid hormones, and possibly androgens. Using small interfering RNA (siRNA)-based screening, we found that haploinsufficiency of the zinc finger protein ZNF764 gene caused by the patient's microdeletion may be responsible for his phenotype and partial multiple hormone resistance.

Subjects and Methods

Case report

Recruitment of the subject was approved by the institutional review board at the Eunice Kennedy Shriver National Institute of Child Health and Human Development, and a written informed consent was obtained from the patient's mother. The patient was born at 33 wk of gestation by emergency cesarean procedure (body weight at birth, 1270 g). At birth, he demonstrated ambiguous genitalia with small penis, bifid scrotum, undescended testes, and hypospadias; he underwent orcheopexy and further genitoplasty at 2 yr of age. He was diagnosed as suffering from autism spectrum disorder at 3 yr old. At 7 yr of age, his height and weight were 111 cm (25–50th percentile) and 17.2 kg (10–25th percentile), respectively. He had significant dolichocephaly, hypertelorism, micropenis, and small, underdeveloped scrotum (Fig. 1, A–C). He demonstrated elevated levels of morning serum cortisol (34.5 ng/dl, normal range = 6–23 ng/dl) and plasma ACTH (122 pg/ml, normal range = 9–52 pg/ml), which were not suppressed by dexamethasone (morning serum cortisol, 31.2 mg/dl, after administration of 2 mg dexamethasone in previous night). Serum TSH concentration was elevated at 6.88 mIU/liter (normal range = 0.4–4.0), whereas that of free T₄ was in a normal range at 1.9 ng/dl (normal range = 0.8–2.3). Serum concentrations of the LH, FSH, dehydroepiandrosterone, dehydroepiandrosterone-sulfate, free testosterone, and total testosterone were all within normal ranges for his age. The patient carried a heterozygous 16p11.2 microdeletion in chromosome 16, observed first in the array-based comparative genomic hybridization (aCGH) analysis using 1340 bacterial artificial chromosome (BAC) clone probes and then confirmed with fluorescent in situ hybridization (FISH) employing BAC CTD-3159M2 as a probe (Fig. 1, D and E). Quantitative PCR revealed that the patient's microdeletion spanned approximately 1.1 Mb, ranging from −307,000 kb (FBRS, GenBank accession number: NP_001098549) to −29,600 kb (SLC7A5P1, GenBank accession number: NR_002593) of chromosome 16 (Fig. 1F).

Fig. 1.

Clinical manifestations and an approximately 1.1-Mb heterozygous microdeletion at 16p11.2 of our patient. A–C, The patient has short stature and facial and genital abnormalities. Images of the patient (A) and his genital region showing a micropenis (B) and bifid scrotum (surgically repaired at 2 yr of age) (C) are shown. D, Ratio plot of patient's chromosome 16. The aCGH analysis using 1340 BAC clone probes identified a sharp ratio drop at the area of chromosome 16p11.2. E, Heterozygous loss of CTD-3159M2 in the FISH analysis. FISH was performed for the patient's chromosomes by using as a probe the CTD-3159M2 marker located in the deleted area. The affected chromosome 16 harboring the microdeletion and the intact chromosome 16 are indicated with arrowhead and arrow, respectively. F, The patient's 16p11.2 microdeletion spans approximately 1.1 Mb and contains 55 genes. Chromosome 16 (top), the area around the deleted segment found in the patient, and the genes harbored in it as well as results of quantitative PCR for representative genes are shown. The bottom red bar indicates the segment of deletion found in the patient. Bars indicate mean ± se values of relative dosage of indicated genes obtained by dividing levels of patient's copy number with that of normal controls. Dashed lines indicate dosages of 1.0 (presence of two alleles) and 0.5 (presence of only one allele), respectively. Chromosomal positions of some genetic markers used in the aCGH and FISH analysis are also shown.

aCGH and FISH analysis

aCGH analysis using 1340 BAC clone probes and FISH employing BAC CTD-3159M2 were performed as reported previously (21, 22).

siRNA-based screening for identifying the molecules potentially responsible for patient's resistance to multiple hormones

siRNA for the 38 genes located within the deleted chromosomal area of the patient and two control siRNA (those for human cyclophilin B and firefly luciferase), purchased from Thermo Fisher Scientific Inc. (Pittsburgh, PA) were transfected in duplicate into human cervical carcinoma HeLa cells plated on 24-well plates (Supplemental Tables 1 and 2, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). One day after the transfection, the cells were incubated with 10⁻⁶ m dexamethasone for 24 h, and total RNA was harvested. mRNA levels of the glucocorticoid-responsive glucocorticoid-induced leucine zipper (GILZ) and the control acidic ribosomal phosphoprotein P0 (RPLP0) were evaluated with SYBR Green-based real-time PCR. All experiments were repeated three times.

SYBR Green-based real-time PCR

Total RNA was reverse transcribed into cDNA, and the SYBR Green-based real-time PCR was performed using the primer pairs shown in Supplemental Table 3 (23). Obtained threshold cycle (Ct) values of each gene were normalized for those of RPLP0, and their relative mRNA expression was demonstrated as fold induction over the baseline.

Plasmids

ZNF-related plasmids were constructed by subcloning their coding sequences into pCDNA3.1His/A (Invitrogen, Carlsbad, CA) or pEGFP-C1 (Clontech, Palo Alto, CA). pCMV6-SPORT-TRIM28 (for human TIF1β), pCRE-Luc, and pGL4.73[hRluc/SV40] were purchased from OriGene (Rockville, MD), Agilent Technologies, Inc. (Santa Clara, CA), and Promega Corp. (Madison, WI), respectively. Plasmids expressing the indicated NRs, the cAMP response element-binding protein (CREB), p53, and the α-catalytic subunit of the protein kinase A, and corresponding reporter plasmids were described previously or gifted from Dr. M. L. Privalsky (University of California Davis, Davis, CA) (24, 25).

Cell culture and transfection

HCT116 cells were transfected with plasmids and/or siRNA, as previously described (23). HCT116 cells or peripheral blood mononuclear cells (PBMCs) obtained from the patient or three age- and sex-matched normal subjects were incubated with ligand for respective receptors for 24 h and were lysed for reporter assays or were subjected to purification of total RNA. Epstein-Barr virus (EBV)-transformed lymphocytes were transfected by using the Nucleofector system (Lonza Group Ltd., Basel, Switzerland) and the AMAXA Human T Cell Nucleofector Kit (Lonza Group Ltd.), as described previously (23). Firefly and renilla luciferase activities were determined as described previously (23).

Western blots for ZNF764 and NRs in PBMCs

PBMCs were lysed, run on 4–20% SDS-PAGE gels, and blotted, and ZNF764, glucocorticoid receptor (GR), androgen receptor (AR), thyroid hormone receptor (TR)-α, TRβ, and control β-actin were visualized with anti-ZNF764, -GR -AR, -TRα, -TRβ, and -β-actin antibodies (23). Anti-ZNF764 antibody was purchased from Sigma-Aldrich Corp. (St. Louis, MO), whereas antibodies for these NRs and β-actin were all purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

Statistical analysis

Statistical analysis was carried out by unpaired Student's t test with a two-tailed P value.

Results

Our patient demonstrated partial target tissue resistance to glucocorticoids and thyroid hormone, as evidenced by elevated plasma ACTH and serum cortisol with a blunted response to dexamethasone, and elevated serum TSH with high normal levels of T₄. In addition, the patient showed some signs of possible androgen resistance, because he had ambiguous genitalia with small penis, undescended testes, bifid scrotum, and hypospadias, all of which indicate reduced action of androgens in utero. He harbored an approximately 1.1-Mb heterozygous microdeletion in chromosome 16p11.2, which contains at least 55 genes, including those encoding one hypothetical protein, one microRNA, five pseudogenes, and one non-protein-coding gene (Fig. 1F and Supplemental Table 1). The patient's PBMCs expressed less mRNA for three of five glucocorticoid-responsive genes (dual-specificity phosphatase 1 DUSP1, GILZ, and tristetraprolin), two of three androgen-responsive genes (inhibitor of κBα IκBα and serum/glucocorticoid-regulated kinase SGK1), and three of three thyroid hormone-responsive genes (Kruppel-like factor 9 KLF9, glucose-6 phosphate dehydrogenase G6PD, and malic enzyme 1 ME1), compared with the PBMCs obtained from six age- and sex-matched normal subjects (Fig. 2, A–C) (26–28). The patient's PBMCs also demonstrated blunted mRNA response of the glucocorticoid-, androgen-, and thyroid hormone-responsive genes (GILZ, SGK, and KLF9, respectively) to respective hormones (Fig. 2D). These results are consistent with the presence of partial target tissue resistance to these hormones suggested by the patient's manifestations and laboratory findings. Protein and mRNA of the GR, AR, and TRα and TRβ were similarly expressed in PBMCs obtained from the patient and six normal subjects (Fig. 3), indicating that reduced sensitivity to these hormones cannot be explained by decreased levels of their receptors.

Fig. 2.

mRNA levels of some glucocorticoid-, androgen-, and thyroid hormone-responsive genes and their responsiveness to respective hormones are reduced in patient's PBMCs. A–C, mRNA levels of some glucocorticoid-, androgen-, and thyroid hormone-responsive genes are reduced in patient's PBMCs. mRNA levels of known glucocorticoid-responsive genes (annexin A1, DUSP1, GILZ, PER1, and tristetraprolin) (A), androgen-responsive genes (FKBP5, IκBα, and SGK1) (B), thyroid hormone-responsive genes (KLF9, G6PD, and ME1) (C), and control RPLP0 were determined by SYBR Green-based real-time PCR. Bars represent mean ± se values of mRNA levels of the indicated genes corrected for those of RPLP0 by employing the levels of normal subjects as 1. *, P < 0.05; **, P < 0.01; n.s., not significant, compared with control. D, mRNA expression of the glucocorticoid-responsive (GILZ), androgen-responsive (SGK), and thyroid hormone-responsive (KLF9) genes demonstrated blunted response to their corresponding hormones in the patient's PBMCs. PBMCs obtained from the patient or three age- and sex-matched normal subjects were incubated in the presence or absence of 10⁻⁶ m dexamethasone (Dex), 10⁻⁸ m dehydrotestosterone (DHT), or 10⁻⁷ m T₃, and mRNA expression of GILZ, SGK, KLF9, and control RPLP0 were determined by SYBR Green-based real-time PCR. mRNA levels of respective genes were corrected for those of RPLP0, and fold expression was calculated by employing the values obtained in the absence of ligand as 1-fold, respectively, for the patient and control subjects. Bars represent mean ± se values of fold mRNA levels. **, P < 0.01, compared with control (the value of the control subjects obtained in the presence of indicated hormones). DUSP1, Dual-specificity phosphatase 1; FKBP5, FK506-binding protein 5; G6PD, glucose-6-phosphate dehydrogenase; IκBα, inhibitor of κBα; KLF9, Kruppel-like factor 9; ME1, malic enzyme 1; PER1, period 1; SGK1, serum/glucocorticoid regulated kinase 1.

Fig. 3.

Protein and mRNA levels of GR, AR, TRα, and TRβ are not altered in the patient. Protein (A) and mRNA (B) levels of GR, AR, TRα, and TRβ were determined with Western blots using specific antibodies or SYBR Green-based real-time PCR using total RNA extracted from PBMCs of the patient and six age- and sex-matched normal subjects. Bars represent mean ± se values of protein (A) and mRNA (B) levels of GR, AR, TRα, and TRβ corrected, respectively, for those of β-actin (A) and RPLP0 mRNA (B) by setting the levels of normal subjects as 1. *, P < 0.05; **, P < 0.01; n.s., not significant, compared with control. Representative gel images of the Western blot for GR, AR, TRα, and TRβ (top panels) and β-actin (bottom panels) are shown in the lower panels of A.

The above results suggest that the patient's multiple hormone resistance may be caused by reduced expression of a transcriptional coactivator shared by GR, AR, and TR. We therefore performed siRNA-based screening in which we transfected HeLa cells with siRNA for the genes located in the deleted segment and examined dexamethasone-induced mRNA expression of the glucocorticoid-responsive GILZ gene (Supplemental Table 1); we employed GR as a representative for affected receptors. We also used 38 available siRNA for the genes located in the deleted region and two control siRNA (those for human cyclophilin B and firefly luciferase mRNA). Among them, siRNA for the ZNF764 significantly reduced dexamethasone-stimulated GILZ mRNA expression (Supplemental Table 2 and Fig. 4A). ZNF764 siRNA also significantly reduced the transcriptional activity of AR and TRβ1 on their responsive reporters in HeLa cells (Fig. 4, B and C); thus, endogenous ZNF764 is required for the transcriptional activity of GR, AR, and TRβ.

Fig. 4.

ZNF764 knockdown reduces transcriptional activity of GR, AR, and TRβ. A, ZNF764 knockdown reduces dexamethasone-stimulated GILZ mRNA expression in HeLa cells. HeLa cells were transfected with ZNF764 or control siRNA, treated with 10⁻⁶ m dexamethasone (Dex) and mRNA levels of glucocorticoid-responsive GILZ, ZNF764, and control RPLP0 was measured with SYBR Green-based real-time PCR. Bars represent mean ± se values of GILZ (left panel) or ZNF764 (right panel) mRNA levels corrected for those of the control RPLP0 mRNA. **, P < 0.01, compared with the control values with the same treatment. B and C, ZNF764 knockdown reduces ligand-stimulated transcriptional activity of AR and TRβ in HeLa cells. HeLa cells were transfected with ZNF764 or control cyclophilin B siRNA together with AR-expressing plasmid (B) or TRβ1-expressing plasmid (C) and MMTV-Luc (B) or DR4-Luc (C) in the presence of pGL4.73[hRluc/SV40]. Bars represent mean ± se values of the firefly luciferase activity normalized for renilla luciferase activity in the presence or absence of 10⁻⁸ m dehydrotestosterone (DHT) (B) or 10⁻⁷ m T₃ (C) in the left panels, whereas those of ZNF764 mRNA levels corrected for those of the control RPLP0 mRNA measured in the aliquots of samples are shown in the right panels. **, P < 0.01, compared with the control values with the same treatment.

The patient's PBMCs demonstrated approximately 50% reduction of both ZNF764 protein and mRNA compared with those from six normal subjects (Fig. 5, A and B). We therefore normalized ZNF764 expression and tested GR-induced transcriptional activity in the patient's EBV-transformed lymphocytes. As expected, supplementation through cotransfection of small amounts of a plasmid encoding ZNF764 normalized reduced responsiveness of GILZ mRNA expression in response to treatment with dexamethasone (Fig. 5, C–E).

Fig. 5.

ZNF764 expression is reduced in the patient's PBMCs and exogenous supplementation of ZNF764 normalizes responsiveness of the patient's EBV-transformed lymphocytes to dexamethasone. A and B, Protein and mRNA levels of ZNF764 are approximately 50% lower in patient's PBMCs than in those of normal controls. ZNF764 protein (A) and mRNA (B) levels were determined, respectively, with Western blots using anti-ZNF764 antibody and with SYBR Green-based real-time PCR using total RNA extracted from PBMCs of the patient and six age- and sex-matched normal subjects. Bars represent mean ± se values of ZNF764 protein (A) and mRNA (B) levels corrected, respectively, for those of β-actin (A) and RPLP0 mRNA (B) by setting the levels of normal subjects as 1. *, P < 0.05; **, P < 0.01; n.s., not significant, compared with control. Representative gel images of the Western blot for ZNF764 (top panel) and β-actin (bottom panel) are shown in the lower panels of A. C–E, Exogenous supplementation of ZNF764 recovers GR-induced transcriptional activity in patient's EBV-transformed lymphocytes. EBV-transformed lymphocytes of the patient or three normal subjects were transfected with control or ZNF764-expressing plasmid and were treated with 10⁻⁶ m dexamethasone (Dex). Bars represent mean ± se values of GILZ mRNA levels corrected for those of the control RPLP0 mRNA. **, P < 0.01; n.s., not significant, compared with the control value obtained in the presence of dexamethasone.

We characterized the molecular action of ZNF764 on GR and other NRs. Green fluorescent protein-fused ZNF764 was exclusively localized in the nucleus in HeLa cells (Supplemental Fig. 1). Overexpression of ZNF764 enhanced dexamethasone-stimulated GR transcriptional activity on the glucocorticoid-responsive mouse mammary tumor virus (MMTV) promoter by approximately 3-fold in a dose-dependent fashion (Fig. 6A), whereas it shifted the dexamethasone titration curve of this promoter activity up- and leftward (Fig. 6B). These results indicated that ZNF764 acts as a coactivator for GR. ZNF764 overexpression also enhanced the transcriptional activity of mineralocorticoid receptor in addition to GR, AR, TRα, and TRβ, but not of estrogen receptor-α, progesterone receptor-A, CREB, and p53 (Fig. 6C). Coexpression of TIF1β potentiated ZNF764-induced enhancement of GR transcriptional activity, whereas knockdown of ZNF764 with its siRNA significantly reduced TIF1β-induced potentiation of GR transcriptional activity (Fig. 6, D and E). These results suggest that ZNF764 enhances GR-induced transcriptional activity by cooperating with TIF1β, possibly by influencing actions of the latter general NR coactivator. We also found similar results with TIF1α (data not shown).

Fig. 6.

ZNF764 acts as a coactivator for GR, AR, MR, and TR by cooperating with TIF1β. A and B, ZNF764 overexpression enhances GR-induced transcriptional activity. HCT116 cells were transfected with increasing amounts (A) or 0.05 μg/well (B) of ZNF764-expressing plasmid together with MMTV-Luc and pGL4.73[hRluc/SV40]. Bars and circles represent mean ± se values of the firefly luciferase activity normalized for renilla luciferase activity in the presence or absence of 10⁻⁶ m dexamethasone (Dex) (A) or indicated concentrations of this steroid (B). C, ZNF764 enhances GR-, AR-, TRα-, TRβ-, and mineralocorticoid receptor (MR)-induced transcriptional activity but not of progesterone receptor (PR)-A, estrogen receptor-α (ERα), CREB, and p53. HCT116 cells were transfected with indicated steroid hormone receptor- or CREB-expressing plasmid together with their responsive promoter-driven luciferase gene and pGL4.73[hRluc/SV40] and were treated with indicated compounds or transfected with the plasmid expressing the α-catalytic subunit of the protein kinase A (PKA) or p53. Bars represent mean ± se values of the firefly luciferase activity normalized for renilla luciferase activity in the presence or absence of indicated compounds, PKA, or p53. **, P < 0.01; n.s., not significant, compared with the control value obtained in the presence of indicated steroids, PKA, or p53. D, Coexpression of TIF1β increased ZNF764-induced enhancement of GR transcriptional activity in HCT116 cells. HCT116 cells were transfected with increasing amounts of ZNF764-expressing plasmid together with MMTV-Luc and pGL4.73[hRluc/SV40] in the presence or absence of 0.01 μg/well of TIF1β-expressing plasmid. Bars represent mean ± se values of the firefly luciferase activity normalized for renilla luciferase activity in the presence or absence of 10⁻⁶ m dexamethasone (Dex). E, ZNF764 knockdown attenuated TIF1β-induced enhancement of GR transcriptional activity in HCT116 cells. HCT116 cells were transfected with the control cyclophilin B or ZNF764 siRNA together with MMTV-Luc and pGL4.73[hRluc/SV40] in the presence or absence of 0.1 μg/well of TIF1β-expressing plasmid. Bars represent mean ± se values of the firefly luciferase activity normalized for renilla luciferase activity in the presence or absence of 10⁻⁶ m dexamethasone (Dex).

Discussion

The patient that we reported here demonstrated partial tissue resistance to glucocorticoids, thyroid hormone, and possibly androgens. He carried an approximately 1.1-Mb heterozygous microdeletion located at 16p11.2, which is not found in the general population as a copy-number variant according to the databases. Using the siRNA knockdown approach and subsequent molecular characterization, we found that haploinsufficiency of the ZNF764 gene and subsequent reduction of its expression may be responsible for the patient's resistance to these hormones. ZNF764 is located at the position −30,570 kb of chromosome 16p, the area approximately 300 kb proximal to the end of the common approximately 600-kb segment, which is frequently deleted in the patients with 16p11.2 microdeletion and appears to be responsible for autism and other behavioral problems, mental retardation, developmental delay, and some facial manifestations in these reported cases and our patient (17, 19). Previously reported cases harboring this common 16p11.2 microdeletion did not show hormone resistance; also, we do not know of any reports of genital abnormalities or any other evidence of hormonal abnormalities within the cohorts of these patients (17, 19), further supporting that ZNF764, which is located outside of this common segment, may be responsible for our patient's multiple hormone resistance. Our siRNA-based screening covered 84% of the genes located in the deleted area that express either proteins or RNA; we did not rule out the possibility that sequences not included in the screening may also have regulatory activity on the affected receptors; these include miRNA4518, noncoding LOC100271831, and others.

Our patient demonstrated manifestations associated with reduction of androgen action during early fetal life, whereas resistance to glucocorticoids and thyroid hormone were relatively mild at 7 yr of age and were recognized only by laboratory tests. These observations may suggest that, despite apparent mild effects of ZNF764 when haploinsufficient, its full expression supported by the two alleles of the ZNF764 gene is essential for the biological actions of AR required for proper development of a male phenotype in utero. ZNF764 appears to act in this case as a critical dosage-dependent gene for fetal androgen action, especially in genital androgen-responsive tissues. ZNF764 is likely to be less important for the actions of these hormones after birth or childhood, which might be explained by compensation by other, functionally overlapping, coactivator molecules possibly expressed in later life. It remains to be seen whether our patient will have a normal pubertal development and what the androgen levels will be at that time.

ZNF764 appears to interact only with a subset of the NRs we studied, possibly incorporated in the transcriptional complex formed by TIF1 on their responsive promoters and modulating the activities of the latter general NR coactivators. This hypothesis, including protein-protein interaction between ZNF764 and TIF1, has never been tested previously, and thus, is a potential target for future research. Because many KRAB domain-containing proteins physically interact with TIF1 (12), multiple ZNF may act through TIF1 as transcriptional regulators for several NRs and, probably, other transcriptional factors with distinct specificities. The ZNF genes locate in clusters at the chromosome area close to centromeres (29). Indeed, the area around the segment deleted in the patient contains 10 ZNFs, ZNF48, -629, -646, -668, -688, -689, -747, -768, -771, and -785, in addition to ZNF764. Some of these molecules show a structure similar to ZNF764 (Supplemental Fig. 2A). We did examine their effects on GR-induced transcriptional activity, but only ZNF764 had a regulatory effect (Supplemental Fig. 2B), confirming that ZNFs have functional specificity despite sharing similar molecular structures. ZNF764 may act as a phase- and tissue-specific coactivator for some NRs (such as for AR on its actions in fetus), fine-tuning their transcriptional activity on responsive genes in particular tissues by modulating regulatory actions of the general coactivators TIF1s. Because the patient demonstrated the significant genital manifestations that indicate androgen resistance in utero, ZNF764 may be expressed more preferentially in fetal androgen-responsive tissues.

In conclusion, we report a patient with combined NR resistance due to haploinsufficiency of the ZNF764 gene. These results are important for elucidating functional coordination of multiple NRs; ZNF764 is apparently one of the molecules that adjust the actions of some NRs in a phase- and/or tissue-specific fashion. We hope our findings will facilitate future research on such NR function-tuning molecules, which could provide more detailed insight to the regulation of NR functions physiologically and pathologically.