Partial Hormone Resistance in Mice with Disruption of the Steroid Receptor Coactivator-1 (SRC-1) Gene

See allHide authors and affiliations

Science  20 Mar 1998:
Vol. 279, Issue 5358, pp. 1922-1925
DOI: 10.1126/science.279.5358.1922


The in vivo biological function of a steroid receptor coactivator was assessed in mice in which the SRC-1 gene was inactivated by gene targeting. Although in both sexes the homozygous mutants were viable and fertile, target organs such as uterus, prostate, testis, and mammary gland exhibited decreased growth and development in response to steroid hormones. Expression of RNA encoding TIF2, a member of the SRC-1 family, was increased in the SRC-1 null mutant, perhaps compensating partially for the loss of SRC-1 function in target tissues. The results indicate that SRC-1 mediates steroid hormone responses in vivo and that loss of its coactivator function results in partial resistance to hormone.

Sex steroid hormones have central roles in the control of puberty, sexual behavior, and reproductive functions. Their receptors belong to the nuclear receptor superfamily of ligand-dependent transcription factors (1,2). Upon hormone binding, steroid receptors undergo conformational change, bind to their cognate DNA response elements on nuclear target genes, and recruit coactivators and general transcription factors (GTFs) to form an active transcriptional complex, resulting in site-directed chromatin remodeling and enhancement of target gene expression (1, 3-8). SRC-1 is a coactivator for the steroid receptor superfamily; it functions in transcriptional activation through its histone acetyltransferase activity (HAT) and multiple interactions with agonist-bound receptors, other coactivators such as CBP or P300, other HAT such as p/CAF, and some GTFs such as TBP and TIFIIB (7, 9-11). SRC-1 is a member of a gene family that includes SRC-1, TIF2 (also termed GRIP-1 and SRC-2), and p/CIP (also termed RAC3, ACTR, AIB1, and SRC-3) (6, 9-16). Cell-free and in vitro transcription experiments have indicated that the SRC-1 family members enhance receptor-dependent transactivation of nuclear genes (6, 9-16). AIB1 is amplified and overexpressed in many breast cancers and thus could have a role in tumorigenesis (16).

To dissect the physiological role of SRC-1 in vivo, we used gene targeting to disrupt the endogenous SRC-1 gene in embryonic stem (ES) cells. The targeting vector contained 3.5-kb (5′) and 2.5-kb (3′) mouse SRC-1 genomic sequences flanking a GLVP cassette (17) and a neomycin-resistance gene (PGK-neo) (Fig.1A). In addition, the herpes simplex virus thymidine kinase (HSV-TK) gene was located outside of the SRC-1 sequence and served as a negative drug-selection marker. After correct recombination, the targeting event inserted an in-frame stop codon at the Met381 position and deleted ∼9 kb of genomic sequence extending downstream of Met381. Because the oligonucleotide sequence corresponding to Asp816 to Thr826 was detected in an Eco RV–Xba I fragment within the 9-kb region, targeting deleted the SRC-1 exon sequence encoding at least 446 amino acids (Met381 to Thr826). Except for the NH2-terminal basic helix-loop-helix and Per-Arnt-Sim (bHLH-PAS) domains, all SRC-1 functional domains for transcriptional activation, HAT activity, and interactions with nuclear receptors CBP, P300, and p/CAF were disrupted by the targeting event (6, 7, 9,10, 14, 18).

Figure 1

Targeted disruption of SRC-1 gene. (A) Targeting strategy. The cloned mouse SRC-1 genomic DNA (27) was mapped by the indicated restriction enzymes. The marked ATG codon for Met381 was mutated to a stop codon by site-specific mutagenesis in the targeting vector. The vector was linearized by Bam HI digestion. The SRC-1 sequence on both sides of the GLVP cassette (17) and theneo gene are indicated. The position of the HSV-TK gene is also indicated. (B) Genotype analysis by Southern blotting. Genomic DNA was isolated from ES cells or tail samples from wild-type (+/+), homozygous mutant (−/−), or heterozygous (+/−) mice (19). DNA (15 μg) was digested by Xba I (A) and separated in a 0.7% agarose gel. The blot was analyzed by the 5′ probe A. The same blot was also analyzed by the 3′ probe B and the neo probe. (C) Genotype analysis by PCR. The same genomic DNA was used for the PCR template. The locations of specific primers (P1 to P4) are indicated in (A). The paired primers P1 (5′-caaccagcaaaggctgagtcca) and P2 (5′-agtacctcctgaggggttagag) detect a 309–base pair exon region that represents wild-type SRC-1. Primers P3 (5′-tgccgacgcgctagacgatttc) and P4 (5′-acacagcaaagaactggaggtg) detect a 687–base pair fragment located in the GLVP cassette, which represents the targeted SRC-1 locus. Results from PCR were consistent with those from Southern blot analysis whenever the same samples were assayed by both methods. (D) Absence of SRC-1 protein in homozygotes. Whole tissue lysates were prepared by homogenizing kidney (K) and liver (L) in 1× reducing sample buffer for SDS–polyacrylamide gel electrophoresis (7.5% gel). The blot was analyzed with a monoclonal antibody to the activation domain of SRC-1 (7). The endogenous mouse immunoglobulin G (IgG) from the tissues was detected by the secondary antibody, goat antibody to mouse immunoglobulin G conjugated to horseradish peroxidase (Bio-Rad).

After electroporation and drug selection with geneticin (G418) and FIAU (19), we identified 16 correctly targeted ES cell lines by Southern (DNA) analyses (Fig. 1, A and B). Three independent lines of chimeric founders were generated by microinjecting targeted ES cell lines into blastocysts donated from a C57 strain. We used two lines to inbreed with a 129Sv mouse strain and outbreed with a C57 strain to produce offspring. SRC-1 genotypic analyses by Southern blotting and polymerase chain reaction (PCR) confirmed that SRC-1–specific heterozygous and homozygous mutants were obtained (Fig. 1C). Heterozygotes appeared to be normal and were indistinguishable from wild-type mice. In homozygotes, both the 8.5- and the 6.5-kb SRC-1 mRNAs were absent when analyzed by Northern blot analysis (20). To confirm the absence of the COOH-terminus of SRC-1 protein in homozygotes, we used a monoclonal antibody (7) specific to the region of amino acids 840 to 947 in protein immunoblotting extracts prepared from kidney and liver. The SRC-1 protein was undetectable in assays with extracts from the homozygous mutant (Fig. 1D). Although the RNA encoding the bHLH-PAS domain was expressed in the mutants at a level similar to that in the wild-type mice (20), it would not have a dominant negative effect because this domain interacted with neither the full-length SRC-1 nor other SRC-1 family members such as TIF2 (21).

SRC-1 null mutants exhibited no obvious external phenotype. Both male and female homozygotes were fertile and showed growth rates similar to wild-type mice. We therefore examined steroid action in target organs including uterus, prostate, and mammary gland. Uterine response to mechanical traumatization (decidual stimulation) is a progesterone receptor (PR)–dependent process (22). We treated ovariectomized wild-type and mutant mice with a high dose of progesterone and a low dose of estrogen, followed by mechanical stimulation of the left uterine horn of each animal (22). The unstimulated right uterine horn served as a control. The decidual response, exhibited as an increase in uterine horn size, was consistently observed in the stimulated left uterine horn in wild-type mice, but the uterine horn of the SRC-1 null mutant revealed only a partial response (Fig. 2, A and B). We also examined estrogen-induced uterine growth in SRC-1 null mutants. Wild-type mice responded to estrogen treatment with a 4.3-fold increase in uterine wet weight. Uteri of homozygous SRC-1 mutants showed a smaller increase of about 2.8-fold (Fig. 2C). Thus, SRC-1 appears to be required for maximal uterine response to steroid hormone in vivo.

Figure 2

Uterine responses in SRC-1 mutant mice. (A and B) Uterine response to a decidual stimulus. The decidual response was measured as described (22). Eight-week-old females (10 +/+ and 11 −/− animals) were ovariectomized on day 0, treated with estradiol (0.1 μg per mouse per day) from day 10 to day 12, and treated with progesterone (1 mg per mouse per day) and estradiol (6.7 ng per mouse per day) from day 16 to day 23. Mechanical decidualization in the left uterine horn was done 6 hours after hormone injection on day 18. The whole uterus was dissected 6 hours after hormone injection on day 23 (A). The ratio of the weights of the stimulated to the unstimulated (control) horn was calculated. Statistical t test showed a significant difference (P < 0.01) in the ratios from +/+ and −/− mice. The weights of the control horns from either +/+ or −/− mice were similar. (C) Uterine growth stimulated by estrogen assessed as described (23). Eight-week-old females were ovariectomized on day 0 and treated with estradiol (E2) (0.8 μg/kg/day) or sesame oil (solvent control) from day 15 to day 17. Uterine wet weight was measured on day 18. The ratio of uterine weight to body weight was calculated. The t test showed a significant difference (P < 0.01) in E2-treated +/+ (n = 11) and −/− (n = 11) uteri. Data in (B) and (C) represent two independent experiments (mean ± SEM).

To assess androgen receptor function, we measured prostate growth in castrated male mice after they were treated with androgen. Eight days after castration, prostates in both wild-type and mutant mice regressed. Injection of testosterone for 7 days stimulated prostate growth in wild-type animals; a smaller response was observed in SRC-1 mutant mice. The ratio of the weight of prostate and urethra to body weight revealed a 34% reduction in steroid-stimulated growth in the absence of SRC-1 (Fig. 3A). Although histological analysis did not reveal a structural disorder, the testes were smaller in homozygotes. The average ratio of testis weight to body weight was 19% lower in the SRC-1 mutants (Fig. 3B). Smaller testes also were observed in 3-week-old null mutants. Thus tissue responses to testosterone are also reduced in mice lacking SRC-1.

Figure 3

Impaired response to testosterone in mice lacking SRC-1. (A) Stimulation of prostate growth. Twelve-week-old male mice (10 +/+ and 11 −/−) were castrated on day 0 and treated with testosterone (3 mg/kg/day) by subcutaneous injection during day 9 through day 15. The total weight of prostate and a section of urethra between bladder and penis was measured on day 16. The urethra section was included for technical reasons. Then the ratios of prostate and urethra weight to body weight were calculated. The t test showed a significant difference (P < 0.01) in hormone-treated +/+ and −/− mice. The data represent two independent experiments (mean ± SEM). Without hormone treatment, the ratios of regressed prostate and urethra weight to body weight were similar (∼5.5 × 10−4) in +/+ and −/− mice. (B) Smaller testes in SRC-1 null mutants. The body weight and testis weight were measured for 12-week-old male mice (23 +/+ and 19 −/−). The ratio of testis weight to body weight was calculated (P < 0.01 byt test). (C) Analysis of mRNA expression for TIF2 and p/CIP (RAC3). Total RNA (30 μg) from brains (B), mammary glands (M), testes (T), and uteri (U) of three +/+ or −/− mice was separated in each lane of the gel. Its blot was analyzed with 32P-labled human TIF2 or RAC3 cDNA probes. Cyclophylin (Cyc) served as a control for RNA quantity. By densitometry, the density ratios of TIF2 RNA bands to the Cyc band were 0.27 (+/+) and 0.45 (−/−) for brains and 0.71 (+/+) and 1.34 (−/−) for testes. The density ratios of RAC3 RNA bands to the Cyc band were 0.20 (B), 0.33 (M), 0.41 (T) and 0.23 (U) for +/+, and 0.29 (B), 0.23 (M), 0.42 (T), and 0.31 (U) for −/−.

We also tested whether endocrine back control systems were affected. We measured estradiol, progesterone, and testosterone concentrations in serum from age-matched wild-type and SRC-1 null mutants. Estradiol and testosterone concentrations in female null mutants were 1.2 and 1.5 times those in wild-type animals, respectively. The SRC-1 null mutation does not elicit the typical hormonal changes exhibited in animals with disrupted estrogen receptor or PR (22-24). These results may reflect redundant coactivator function among multiple SRC-1 family members (14,25). Indeed, a twofold overexpression of TIF2 mRNA was detected in certain tissues such as brain and testis in SRC-1 null mutants, but expression of the p/CIP (RAC3) gene was unchanged (Fig. 3C). Thus TIF2 might compensate partially for the loss of SRC-1 in the null mutants.

Mammary development is tightly regulated by steroid hormones. Although prenatal morphogenesis of female mammary gland is relatively independent of steroid hormones, extensive growth of mammary gland during puberty requires estrogen. Physiologically, both estrogen and progesterone are essential for alveolar development during pregnancy (26). In 8-week-old wild-type females, mammary ducts grew extensively and occupied almost the entire mammary fat pad. In contrast, the extent of mammary ductal branching as well as the number of branches was substantially reduced in the mammary glands of age-matched SRC-1 null mutants. The ductal tree occupied only half the area of the mammary fat pad (Fig. 4, A and B). By day 18 of pregnancy, alveolar structures in wild-type mammary glands were highly developed and appeared on all ductal sections, filling the interductal spaces. In the SRC-1 mutant mammary glands, alveoli were much less developed in terms of number and size of alveoli, and very few alveoli were observed at the ends of ducts at the same stage of pregnancy (Fig. 4, C through F). Although mammary glands of SRC-1 null mutants can still produce milk, our results suggest that SRC-1 is required for normal mammary ductal elongation and alveolar development in vivo.

Figure 4

Mammary gland development in SRC-1 null mutant mice. Whole mounts of mammary glands were prepared and stained as described (22). (A and B) The fourth pair of mammary glands from 8-week-old virgins with the indicated SRC-1 genotypes. (C and D) The fourth pair of mammary glands from mice pregnant for the first time with the indicated genotypes. (E and F) Higher magnification of the ducts and alveolar structures of the mammary glands in (C) and (D), respectively. (G and H) The fourth pair of mammary glands from 13-week-old females treated with progesterone and estradiol. Eight-week-old females were ovariectomized on day 0 and then treated with progesterone (1 μg per mouse per day) and estradiol (50 μg per mouse per day) from day 14 to day 34. Whole mounts of mammary glands were prepared on day 35. (I andJ) Higher magnification of the mammary ducts and alveolar structures from (G) and (H), respectively. Scale bar in (A) also applies to (B); scale bar in (C) also applies to (D), (G), and (H); and scale bar in (E) also applies to (F), (I), and (J).

We also analyzed mammary gland development in response to estrogen and progesterone treatment in overiectomized adult mice. Estrogen and progesterone stimulate a complex ductal arborization and extensive alveolar formation in mammary glands of wild-type mice (22). This differentiated phenotype mimics a stage of mammary gland development in early pregnancy. In the mammary glands of SRC-1 mutant mice, only partial ductal growth was observed after hormone treatment (Fig. 4, G through J). Thus SRC-1 is required for efficient proliferation and differentiation of the mammary gland in response to estrogen and progesterone.

Our results demonstrate that a steroid receptor coactivator (SRC-1) is required for efficient steroid hormone action in vivo. Loss of coactivation function in the SRC-1 null mutants may be partially compensated by increased expression of the closely related coactivator TIF2. Certain clinical syndromes of partial hormone resistance in which receptors are intact might be explained by impairment of nuclear receptor coactivators.

  • * To whom correspondence should be addressed. E-mail: berto{at}


Navigate This Article