Endocrine Effects of Relaxin Overexpression in Mice

0013-7227/06/$15.00/0 Printed in U.S.A.

Endocrinology 147(1):407– 414 Copyright © 2006 by The Endocrine Society doi: 10.1210/en.2005-0626

Endocrine Effects of Relaxin Overexpression in Mice Shu Feng,* Natalia V. Bogatcheva,* Aparna A. Kamat, Anne Truong, and Alexander I. Agoulnik Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, Texas 77030 Relaxin is a small peptide hormone with a variety of biological functions. To investigate the systemic endocrine effects of relaxin, we produced mice with transgenic overexpression of the Rln1 gene, Tg(Rln1), driven by rat insulin 2 promoter. The expression of relaxin was detected in the pancreas of the transgenic animals. An analysis of the sera from the transgenic animals revealed at least 20-fold elevation of the level of bioactive relaxin. Transgenic animals had normal viability and fertility in both sexes. Transgenic overexpression of Rln1 did not rescue the undescended testis phenotype in Insl3deficient males, suggesting that in vivo relaxin does not interact with the insulin-like 3 factor receptor, leucine-rich repeats-containing G protein-coupled receptor 8, Lgr8. Phenotypically, the excess of relaxin resulted in hypertrophic nipple development in virgin female mice. Deletion of the relaxin re-

ceptor, leucine-rich repeats-containing G protein-coupled receptor 7, Lgr7, in Tg(Rln1) animals abrogated the development of enlarged nipples in females, indicating that relaxin exerts its effect through Lgr7 alone. The levels of previously defined targets of relaxin signaling, such as matrix metalloproteinases 2 and 9, vascular endothelial growth factor, or nitric oxide, were similar in the sera of the transgenic and wild-type mice. However, the total plasma protein concentration in male Tg(Rln1) mice was lower than that in control animals. The livers of male Tg(Rln1) mice exhibited significantly higher hydroxyproline content, indicative of increased collagen deposition. Our results indicate that relaxin overexpression causes gender-specific changes in liver collagen metabolism. (Endocrinology 147: 407– 414, 2006)

R

ELAXIN IS A SMALL peptide hormone from the insulin superfamily, expressed in a number of reproductive and nonreproductive organs. The major source of circulating relaxin in females is the ovary; the expression of relaxin in males was detected in testis and prostate gland (1, 2). In the human genome there are two highly homologous RLN genes, RLN1 and RLN2, and in mouse there is a single Rln1 gene (2). Relaxin, initially recognized as a pregnancy hormone, is now known to have pleiotropic effects in different organs and tissues. Three approaches have been generally used to assess relaxin effects in vivo: the injections of exogenous relaxin (3–5), the neutralization of endogenous relaxin action either with specific antirelaxin antibodies (6 –9) or with peptide based on the ectodomain portion of relaxin receptor (10, 11), and, finally, phenotypic analysis of the mouse mutants with targeted deletion of Rln1 or its receptor (12–18). In vitro, porcine and human relaxins were shown to activate leucine-rich repeats-containing G protein-coupled receptor 7 (LGR7) and, with lower affinity, homologous receptor LGR8 (10). The latter receptor readily interacts with another insulin superfamily hormone, insulin-like 3 factor (INSL3) (19). Such an apparent indiscrimination inspired a series of mouse transgenic studies to determine the selectivity of Insl3/Lgr8 and relaxin/Lgr7 interaction (17, 20, 21). We have shown that the phenotypic effects of transgenic overexpression of Insl3 were abrogated in mice with deletion

of Lgr8, suggesting that Lgr8 is a single receptor for this peptide in vivo. It was reported that Rln1-deficient males had underdeveloped prostate (14); however, two independent Lgr7 mutants (17, 18) did not manifest such a phenotype, leaving the possibility that relaxin exerted its action via a receptor distinct from Lgr7. Although recent data indicated that the Lgr8 ablation did not account for the differences between Lgr7⫺/⫺ and Rln1⫺/⫺ phenotypes (17), the question regarding a possibility of Rln1-Lgr8 interaction in vivo required further elucidation. To address these questions, we have produced mice with transgenic overexpression of Rln1 and analyzed the resultant phenotype in mice deficient for Lgr7 or Insl3. The reported effects of relaxin in normal and pathological processes in reproductive and nonreproductive organs, such as internal organ fibrosis, regulation of vascular function (2, 22), or suggested role in tumorigenesis (23), raise the question regarding potential long-term effects of relaxin stimulation. It has been suggested that relaxin action might be mediated, among others, by such bioactive molecules as matrix metalloproteinases (MMPs), endotheline/nitric oxide (NO), and vascular endothelial growth factor (VEGF) (2, 22). Transgenic animals with overexpression of relaxin allowed us to mimic long-term pharmacological application of relaxin and to assess the effects of relaxin on the plasma content of bioactive molecules. Materials and Methods

First Published Online October 13, 2005 * S.F. and N.V.B. contributed equally to this work. Abbreviations: H&E, Hematoxylin and eosin; IBMX, 3-isobutyl-1-methylxanthine; INSL3, insulin-like 3 factor; LGR7, leucine-rich repeats-containing G protein-coupled receptor 7; MMP, matrix metalloproteinase; Rln1, mouse relaxin 1. Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the endocrine community.

Baylor College of Medicine Institutional Committee on Animal Care has approved all animal experiments described in this paper; all experiments were conducted in accordance with the accepted standards of humane animal care.

Rln1 transgene and production of transgenic mice The genomic fragment containing two exons of the Rln1 gene was obtained by PCR of mouse genomic DNA with primers mRelaxinF,

407

408

Endocrinology, January 2006, 147(1):407– 414

5⬘-GACCGGAATGTCCAGCAGA-3⬘, and mRelaxinR, 5⬘-GCTCATTAGGCAATGGAGTCA-3⬘, using SuperTag DNA polymerase (Ambion, Austin, TX). The position of the 5⬘- and 3⬘-ends of the fragment corresponds to nucleotides ⫺7 to ⫹1130 of the Rln1 cDNA. The resultant fragment was cloned into pGEM-T vector (Promega, Madison, WI), Rln1-pGEM. Both exons of Rln1 gene were sequenced to verify the absence of mutations within the coding sequence. The 705-bp rat insulin 2 promoter was isolated from plasmid pRIP-1Tag (kind gift of Dr. Doug Hanahan, University of California, San Francisco, CA) by restriction digest with BamHI and XbaI and cloned into BamHI/XbaI sites of Bluescript KSII vector, producing the Ins2-KSII plasmid (Stratagene, Palo Alto, CA). To create Rln1 transgenic construct, the mouse genomic fragment was excised with NotI and SacII from Rln1-pGEM plasmid and subcloned into NotI/SacII sites of Ins2-KSII. Transgenic construct containing Rln1genomic DNA under rat Ins2 promoter was isolated from Tg(Rln1) plasmid by restriction digest with BamHI/SacII and purified by agarose gel electrophoresis (QIAGEN, Valencia, CA). The DNA was microinjected into fertilized FVB/N eggs to generate transgenic mice. Mice with Tg(Rln1) were selected by PCR with primers specific for rat Ins2 promoter (RatIns2F, 5⬘-CTCTGAGCTCTGAAGCAAGCA-3⬘, and RatIns2R, 5⬘-CCCAGGAGGCTGCAGTTGTT-3⬘). The transgenic line was established from the female by backcrossing to inbred FVB/N mice. All mice described in the experiments were hemizygous for the transgene; the control wild-type animals were siblings derived from the same crosses.

Mouse breeding

Feng et al. • Relaxin Transgenic Mice

females at the proestrus stage (25), derived from the same crosses, were used for these experiments. The animals were fasted overnight to normalize expression of the insulin promoter. 293T cells grown in D100 dishes were transfected with LGR7 expression cDNA construct (10) (kindly provided by Dr. Aaron Hsueh, Stanford University, Stanford, CA) and with pCDNA3.1 vector DNA (Invitrogen, Carlsbad, CA) as a control. Twenty-four hours after transfection, cells were plated on a 24-well plate and after 24 h were treated with different amounts of serum in the presence of 250 ␮m 3-isobutyl-1-methylxanthine (IBMX) for 20 min. The different amounts of porcine relaxin (kindly provided by Dr. David Sherwood, University of Illinois, Urbana, IL) were used to build the reference. Intracellular cAMP concentration was determined using Biotrak enzyme immunoassay system (Amersham) (21). All samples were analyzed in duplicate.

Fertility study Estimations of male and female fertility have been performed on 2to 5-month-old animals of different genotypes derived from the same litters. Males were bred to the four females within a 4-wk period. The number of pregnant females and the litter size were evaluated. Females were bred with the males of proven reproductive performance for 4 wk. Female pregnancy rate was evaluated as the number of successful pregnancies after the detection of the vaginal copulatory plugs.

Morphological measurements and histology

Total RNA from different mouse tissues was extracted with TRIzol reagent (Life Technologies, Rockville, MD). cDNA was synthesized using an oligo(dT) primer and RETROscript kit (Ambion, Austin TX.). The Rln1 expression was analyzed with primers derived from two Rln1 exons, mRelaxinF and mRelaxinR (see above). RT-PCR with primers from ubiquitously expressed ␤-actin gene, ␤-actinF, 5⬘-CCAAGGCCAACCGCGAGAAGATGAC-3⬘, and ␤-actinR, 5⬘-AGGGTACATGGTGGTGCCGCCAGAC-3⬘, was used to assess the quality of cDNA pools.

The third pair of nipples (from the top) was dissected from virgin 6-wk-old Tg(Rln1) and wild-type female mice and from lactating 3-month-old Tg(Rln1) and wild-type females 2 d postpartum. Similarly, the morphometric analysis of nipples of the Tg(Rln1) mice with and without Lgr7 allele was performed on 6-wk-old virgin females. The number of animals analyzed in each group is indicated in Results. The nipples were photographed under dissecting microscope; the area of each nipple on the photograph was assesses with KODAK 1D image analysis software. For regular histological examination, nipples were fixed and embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E). For the whole-mount examination, the fourth mammary glands were dissected from four 2-month-old Tg(Rln1) and wild-type females and fixed in freshly prepared Methacorn solution (60% methanol, 30% chloroform, and 10% acetic acid) for 48 h. Glands were dehydrated, skimmed with acetone for 48 h, rehydrated, and stained with iron hematoxylin for 2 h before washing and Histo-Clear conservation (National Diagnostics, Somerville, NJ). Parts of the tissues were fixed and embedded in paraffin, sectioned, and stained with H&E for regular histological examination. The development of alveoli, mammary duct, and adipose tissues was compared between three virgin transgenic and wild-type littermates. Reproductive tissues (testis, prostate, seminal vesicles, uterus, ovary, and cervix) were taken from 4-month-old Tg(Rln1) and wild-type males and virgin females at the time of blood collection (see above). The estrus cycle of the females was determined based on the external appearance of vulva and vagina (25); females at the proestrus stage were used in the analysis. The organs were examined for signs of any gross abnormalities, fixed in 10% buffered formalin, embedded into paraffin, sectioned, and stained with H&E. Three H&E sections were evaluated for each analyzed organ from at least three animals of transgenic and wild-type group. Ten random fields of each section at ⫻400 magnification were studied for pathological abnormalities. Testicular sections from Tg(Rln1) and wildtype littermate mice (at least three males in each group) were examined for tubule size, presence of normal spermatogenesis, and compactness. The H&E sections containing different parts of the prostate were evaluated for glandular epithelial cell morphology and presence of secretion. Testis position in Insl3⫺/⫺ males with and without relaxin transgene was evaluated in three animals in each group at 20 d and 2 months after birth.

Analysis of LGR7 receptor activation

NO determination

Sera of wild-type and hemizygous Tg(Rln1) 4-month-old mice were obtained from the blood, taken after animal decapitation, and used immediately to assess the biological activity of relaxin. In total, eight Tg(Rln1) and nine wild-type males and six Tg(Rln1) and seven wild-type

Plasma analysis was performed on wild-type and Tg(Rln1) mice at 4 months of age. Experimental animals were starved overnight before blood collection. EDTA plasma was obtained from the mouse blood by centrifugation at 400 ⫻ g, frozen immediately, and stored at ⫺20 C until

To prove that transgenic relaxin exerts its effects via Lgr7, we produced Tg(Rln1) transgenic mice with Lgr7 deficiency. For that, we crossed transgenic females with Lgr7⫺/⫺ males (17). The Tg(Rln1) Lgr7⫺/⫹ females were backcrossed to the homozygous Lgr7⫺/⫺ males, which produced Tg(Rln1) Lgr7⫺/⫹ and Tg(Rln1) Lgr7⫺/⫺ animals used in analysis. To ascertain the ability of transgenic relaxin to compensate the Insl3 deficiency, we produced Tg(Rln1) transgenic mice homozygous for Insl3 mutation. For that, we crossed homozygous and heterozygous Insl3⫺ females with Tg(Rln1) males. Insl3 mutants were kindly provided by Dr. Lois Parada, University of Texas, Southwestern Medical Center, Dallas, TX (24). Tg(Rln1) mice heterozygous for Insl3 mutant allele were intercrossed to generate Insl3-deficient mice with Tg(Rln1). All genotypes were defined by PCR of the tail DNA. Details of Lgr7 and Insl3 genotyping were described previously (17, 24).

Southern blot hybridization Genomic DNA from transgenic and wild-type mice was isolated by conventional phenol-chloroform method. The Southern blots were prepared with Nylon⫹ membrane (Amersham, Piscataway, NJ) according to manufacturer recommendations and hybridized with randomprimed radioactive labeled DNA fragment corresponding to the second exon of Rln1 in PerfectHyb Plus solution (Sigma Chemical Co., St. Louis, MO). Blots were washed three times under highly stringent conditions at 68 C with 0.1⫻ standard saline citrate/0.1% SDS and exposed to Kodak X-Omat film with one intensifying screen at ⫺80 C.

RT-PCR



409

assayed. Thawed plasma was diluted with reaction buffer (Assay Designs Inc., Ann Arbor, MI) 2-fold and filtered through the 10,000 molecular weight cutoff. Nitric oxide concentration was assessed with total nitric oxide and nitric oxide (nitrite/nitrate) assay kits (Assay Designs), using nitrate reductase for the nitrate-nitrite conversion and Griess reagents for the chromophore development. Each plasma sample was analyzed in duplicate. Sample absorbance was read at 490 nm with Bio-Rad plate reader (model 550) and compared with the absorbance of standards containing nitrate or nitrite, included in the kit.

VEGF ELISA Concentration of VEGF in mouse plasma was assessed using Quantikine mouse VEGF ELISA kit (R&D Systems Inc., Minneapolis, MN) according to the manufacturer’s recommendations. EDTA plasma was diluted five times with the calibrator diluent before the determination and assayed with anti-VEGF antibodies conjugated with horseradish peroxidase using tetramethylbenzidene as a substrate. Each plasma sample was analyzed in duplicate. Sample absorbance was read at 655 nm with Bio-Rad plate reader (model 550) and compared with the absorbance of standards containing recombinant mouse VEGF, included in the kit.

Detection and measurement of MMP2 and MMP9 by gelatinase zymography EDTA plasma, diluted 1:10 with PBS, was mixed with nonreducing SDS sample buffer and run through the 10% polyacrylamide gel, containing 0.1% gelatin. Gel was renatured in 2.5% Triton X-100 for 30 min and developed overnight in 50 mm Tris-HCl (pH 7.5), 150 mm NaCl, 5 mm CaCl2, 0.02% Brij 35. Gel was stained with 0.5% Coomassie Blue R-250 and destained with methanol/acetic acid-containing solution. Areas of gelatinase activity, appearing as clear bands against a blue background, were scanned using KODAK Gel Logic 200 imaging system and analyzed with KODAK 1D image analysis software. MMP2 appeared as latent and active bands of 72 and 66 kDa, respectively. MMP9 appeared as a predominantly latent band of 92 kDa (26). Band intensities were corrected according to plasma protein concentration, assessed with BCA protein assay kit (Pierce, Rockford, IL), and normalized to the values determined in wild-type animals.

Hydroxyproline assay Quantification of hydroxyproline content (27) was performed on wild-type and Tg(Rln1) mice at 4 months of age. The whole organs were isolated, weighed, and homogenized in saline. Briefly, the method included the hydrolysis of homogenate in 2 n NaOH, oxidation of hydroxyproline with chloramine-T (Fisher Scientific, Pittsburgh, PA) and chromophore development with Ehrlich’s aldehyde (Sigma). Each organ homogenate was analyzed in duplicate. Absorbance of samples was read at 595 nm using a Bio-Rad plate reader (model 550) and compared with the absorbance of standards containing hydroxy-l-proline (Sigma). The values obtained were then corrected to the organ weight.

Statistical analysis The ANOVA was used to assess significance of differences among the different groups. P ⬍ 0.05 was considered significant.

Results Generation of Tg(Rln1) transgenic mice

The microinjection of Rln1 transgene construct (Fig. 1A) into the pronuclei of fertilized FVB/N mouse eggs rendered one male and one female carrying the Rln1 transgene. Transgenic progeny was not detected in the offspring of the male. Breeding the Tg(Rln1) female with the wild-type FVB/N male resulted in 50% of progeny with the transgene. These animals were used to establish the Tg(Rln1) transgenic line. Subsequent analysis of the transgene inheritance in reciprocal crosses with wild-type FVB/N animals showed simple Mendelian inheritance of the transgene (data not shown),

FIG. 1. Production of the Rln1 transgenic mice. A, Transgenic construct contains rat insulin 2 promoter linked to mouse genomic fragment containing two Rln1 exons. B, Southern blot analysis of genomic DNA isolated from wild-type (Wt) and Tg(Rln1) (Tg) mice indicates the presence of several extra copies of the Rln1 gene in transgenic mice. Probe is the Rln1 cDNA. Four different restriction enzymes indicated above the lanes were used in the analysis. C, Tg(Rln1) is expressed in the pancreas of the transgenic mice, shown by RT-PCR analysis of the Rln1 expression in the ovary and pancreas of the wild-type and transgenic females. No Rln1 expression is detected in the wild-type pancreas.

indicating that there was only one transgene integration site in the founder female. To estimate the number of transgene copies contained in the genome of Tg(Rln1) mice, we performed Southern blot analysis of their genomic DNA (Fig. 1B). Genomic DNA was digested with different restriction enzymes and hybridized with the probe corresponding to the second Rln1 exon. In the nontransgenic DNA, digested with different enzymes, this probe produced single bands corresponding to the endogenous Rln1 gene. The first restrictase (BamHI) did not cut

410


within the transgene DNA, and in BamHI-digested Tg(Rln1) genomic DNA an additional band of more than 14 kb was detected, corresponding to a single transgene integration site. Other enzymes produced from two (PstI) to six (EcoRI) additional fragments (Fig. 1B). Most frequently, the transgenic DNA is integrated into the genome as a block of several head-to-head or head-to-tail oriented copies. If there are no restriction sites within such a block, the level of hybridization signal corresponds to the number of copies. We have compared the intensity of the hybridization of the lower 8-kb band in the BamHI-digested DNA, corresponding to the endogenous Rln1 gene (two copies) and the 14-kb band, corresponding to the transgene. The intensity of hybridization measured by radioautography was almost two times stronger in the upper band, and thus, at least four copies of the transgene were integrated into the genome. The presence of several additional hybridization bands detected with the other enzymes indicated that there were several internal sites for these enzymes within the transgenic block and a possible rearrangement or partial deletion of some copies of the transgene. Transgene expression

To determine whether the Rln1 transgene was transcriptionally active, we first performed an RT-PCR analysis of Rln1 expression in Tg(Rln1) females. As shown in Fig. 1C, the expression of Rln1 was detected in total RNA isolated from the transgenic pancreas and ovary, whereas no expression could be detected in the wild-type pancreas. Thus, rat insulin promoter directed expression of the transgene to the latter organ. The transcript corresponded in size to the endogenous Rln1 RNA. The direct sequencing of the RT-PCR fragment derived from the transgenic pancreas indicated proper splicing of the transgene. To assess the presence of biologically active relaxin hormone in circulating blood, we have tested sera from transgenic and wild-type animals for the presence of relaxin activity in a functional assay. The sera were used to stimulate 293T cells, transfected with relaxin receptor LGR7 or vector DNA (Fig. 2). It has been shown that the activation of LGR7 by relaxin causes an increase in cAMP concentration (10). We did not detect the cAMP increase in 293T cells transfected with vector DNA challenged with the sera from wild-type or transgenic mice. No relaxin activity was detected in wildtype male serum. On the other hand, the analysis of intracellular cAMP level in LGR7-transfected cells, challenged with the mouse sera, demonstrated that the sera from the transgenic animals invoked at least a 20 times more potent response than the one from the wild-type females. The Rln1 activity was the same in male and female transgenic sera. Effects of transgenic relaxin on reproductive organs

No gross abnormalities were noted in anatomical or behavioral characteristics of the transgenic animals. Both males and females had normal fertility with a 100% pregnancy rate. The average litter size remained the same as in control wildtype females (9.75 ⫾ 0.97 pups at birth, n ⫽ 12, for Tg(Rln1) and 9.33 ⫾ 0.89, n ⫽ 12, for wild-type females). No differences in the adult body weight were found; the mean weight


FIG. 2. Stimulation of cAMP production in 293T cells, transfected with LGR7, by the sera of wild-type (Wt) or Tg(Rln1) (Tg) male and female mice. 293T cells transiently expressing wild-type human LGR7 were challenged with the sera of the wild-type or Tg(Rln1) mice for 20 min in the presence of IBMX. Indicated volume (␮l) of sera was added to DMEM (final volume of 250 ␮l), containing 250 ␮M IBMX, and used to stimulate two parallel wells of a 24-well plate. For the calibration curve, different concentrations of porcine relaxin (RLN) were used.

of the testes, uteri, cervices, or ovaries did not differ significantly in wild-type and mutant animals (Table 1). Histopathological analysis of the testes and prostate glands and examination of the whole-mount mammary gland of the transgenic animals failed to reveal any abnormalities associated with relaxin overexpression (data not shown). Tg(Rln1) virgin females displayed significantly larger nipples than their wild-type littermates (Fig. 3A). The increased size of the nipples in Tg(Rln1) females became clearly visible with sexual maturation. In 6-wk-old virgin females, the overall size of the nipples was at least three times bigger than in wild-type control. However, there was no difference between two genotypes in lactating females (Fig. 3A). No effect on the nipples in males was detected (data not shown). Histological evaluation of the virgin transgenic nipples showed a well-developed structure of the connective stromal tissues without any signs of epithelial hyperproliferation (Fig. 3B). To address the question whether phenotypic changes induced by relaxin overexpression are dependent on Lgr7, we have produced Tg(Rln1) mice deficient for Lgr7. As shown above, virgin female Tg(Rln1) mice demonstrated hyperdevelopment of the nipples. The comparison of the nipple size in virgin females with and without a wild-type allele of Lgr7 clearly demonstrated significant differences. Lgr7 deficiency abrogated development of the nipples in females (Fig. 4). Effect of transgenic relaxin on nonreproductive functions

No significant differences were found in the weight of adult nonreproductive internal organs (Table 1). To assess the potential effect of relaxin, exerted via circulating biomediators, we evaluated plasma levels of MMP2, MMP9, NO, and VEGF. As shown in Table 2, no alteration was noted in plasma VEGF or nitrite plus nitrate content of wild-type and Tg(Rln1) mice. The increase in MMP2 level, displayed by Tg(Rln1) animals, was not statistically significant. However, when total plasma protein concentration was evaluated, we found that Tg(Rln1) male mice exhibit significantly lower protein content than the wildtype males. Because hypoalbuminemia is often associated



411

TABLE 1. Weights (grams) of internal organs in control and Tg(Rln1) animals

Body weight Uterus Ovary Cervix Testis Heart Lungs Liver Kidney

Female WT

Female Tg

24.82 ⫾ 0.77 (n ⫽ 5) 0.066 ⫾ 0.009 (n ⫽ 3) 0.010 ⫾ 0.001 (n ⫽ 3) 0.036 ⫾ 0.005 (n ⫽ 3)

24.1 ⫾ 0.68 (n ⫽ 7) 0.070 ⫾ 0.005 (n ⫽ 5) 0.009 ⫾ 0.001 (n ⫽ 5) 0.037 ⫾ 0.005 (n ⫽ 5)

Male WT

0.125 ⫾ 0.007 (n ⫽ 5) 0.186 ⫾ 0.013 (n ⫽ 5) 1.014 ⫾ 0.102 (n ⫽ 4) 0.161 ⫾ 0.004 (n ⫽ 7)

0.116 ⫾ 0.003 (n ⫽ 7) 0.173 ⫾ 0.007 (n ⫽ 7) 1.06 ⫾ 0.036 (n ⫽ 8) 0.163 ⫾ 0.004 (n ⫽ 7)

Male Tg

29.39 ⫾ 0.28 (n ⫽ 10)

30.21 ⫾ 0.48 (n ⫽ 9)

0.10 ⫾ 0.005 (n ⫽ 10) 0.150 ⫾ 0.004 (n ⫽ 10) 0.207 ⫾ 0.015 (n ⫽ 10) 1.296 ⫾ 0.037 (n ⫽ 10) 0.235 ⫾ 0.006 (n ⫽ 10)

0.104 ⫾ 0.004 (n ⫽ 9) 0.146 ⫾ 0.005 (n ⫽ 9) 0.22 ⫾ 0.01 (n ⫽ 9) 1.287 ⫾ 0.034 (n ⫽ 9) 0.239 ⫾ 0.003 (n ⫽ 9)

Values represent means ⫾ SE. The number of mice in each group is indicated in parentheses. For the female reproductive organs measurement, only females in the proestrus were analyzed.

with liver or kidney dysfunction, we next assessed the level of fibrosis in these organs by determining collagen content. The liver of Tg(Rln1) males displayed signifi-

cantly higher hydroxyproline concentration compared with the liver of wild-type males, indicative of increased collagen deposition in this organ (Table 3).

FIG. 3. Increased nipple size in virgin Tg(Rln1) (Tg) compared with wild-type (Wt) females. A, Nipple size was increased in Rln1 transgenic mice; left, the nipples were dissected from 6-wk-old virgin Tg(Rln1) (top row) and wild-type (bottom row) females; right, the nipples were dissected from Tg(Rln1) (top row) and wild-type (bottom row) females on postpartum d 2. Values in the table represent nipple size (mm2) mean ⫾ SE. The number of mice in each group is indicated in parentheses. The virgin transgenic nipples were three times bigger than the wild-type controls (in bold, P ⬍ 0.001). B, Histology of nipples of 6-wk-old virgin Tg(Rln1) (right) and wild-type (left) females by H&E staining. Transverse section was made at the root of the nipple.

412



FIG. 4. Lgr7 deficiency causes reduction of the virgin nipple size in Tg(Rln1) females. The nipples were dissected from 6-wk-old virgin Tg(Rln1) Lgr7⫺/⫹ (top row) and Tg(Rln1) Lgr7⫺/⫺ (bottom row) females. Values in the table represent nipple size (mm2) mean ⫾ SE. The number of mice in each group is indicated in parentheses. The difference is statistically significant (in bold, P ⬍ 0.001).

Overexpression of relaxin does not affect testicular descent

Males with Insl3 deficiency are presented with high intraabdominal cryptorchidism as a result of a failure of gubernacular differentiation (24). If relaxin overexpression can compensate the Insl3 deficiency in vivo, Tg(Rln1) Insl3⫺/⫺ males will exhibit gonadal descent associated with onset of gubernacular development. Analysis of Tg(Rln1) Insl3⫺/⫺ male phenotype revealed intraabdominal cryptorchid testes located below the kidneys and a complete lack of gubernacular differentiation, as in Insl3⫺/⫺ littermates. Additionally, no defects in the position of the ovaries were detected in Tg(Rln1) females. Discussion

To study the significance of relaxin endocrine effects, we have created a novel mouse transgenic strain, expressing Rln1 in pancreas, where no endogenous Rln1 expression was detected. The analysis of the resultant Rln1 transgenic phenotype revealed overdeveloped nipples in the virgin females and lower plasma protein concentration in males. The malespecific phenotype was also associated with increased liver collagen deposition. The effects of Rln on liver followed the deviations in liver panel parameters previously reported in Lgr7-deficient mice (18). Effects on nipple development was completely abrogated in females with deletion of Lgr7, indicating that Lgr7 is the only receptor for Rln in vivo. Overexpression of relaxin did not compensate the effects of Insl3

deficiency, thus suggesting that Rln and Insl3 signaling pathways do not overlap in mice in vivo. The expression of the Tg(Rln1) transgene was directed by insulin 2 promoter in pancreatic ␤-cells. Pancreatic cells can produce biologically active hormone in vitro after transfection with Rln1 expression vectors (21). The ability of sera from transgenic mice to activate LGR7 receptor indicates that the transgenic peptide was successfully processed into an active hormone and secreted into the circulation. The level of biologically active peptide measured in sera of Tg(Rln1) mice was at least 20 times higher than that in the wild-type females. To ascertain whether circulating relaxin is able to exert its action via Insl3 receptor, Lgr8, we created Tg(Rln1) mice with Insl3 deficiency. Although we did not perform the detailed morphometric analysis of testis migration during development, the comparison of Insl3⫺/⫺ and Tg(Rln1)Insl3⫺/⫺ phenotype revealed that relaxin overexpression does not improve adult testicular position, impaired by Insl3 deficiency. Thus, an elevated level of circulating relaxin does not elicit Lgr8-mediated effects. Although human and porcine relaxin were shown to stimulate LGR8 with lower efficiency than LGR7 (10, 19), recent data suggest that rodent relaxin is unable to activate LGR8 in vitro (21, 28). Our results confirm the previous observation that differences in Rln1⫺/⫺ and Lgr7⫺/⫺ phenotype cannot be explained by Lgr8 signaling (17). These studies allow us to prove unequivocally that in rodents, relaxin/Lgr7 and Insl3/Lgr8 pathways are fully separated. The question, however, remains whether in nonrodent species relaxin is able to act via LGR8. Analysis of the reproductive phenotype of Tg(Rln1) mice revealed hypertrophic development of mammary gland nipples in virgin females. These data correlate with the earlier observation that immunoneutralization of relaxin in rats abrogates the nipple development (6). Accordingly, the genetic ablation of relaxin (12, 13) or relaxin receptor Lgr7 (17, 18) was associated with severely underdeveloped nipples and inability to feed pups during lactation. However, no differences in the size of lactating nipples, the weight of the reproductive organs, pregnancy rates, and overall female fertility were found in Tg(Rln1) mice. It should be noted that rodent females experience up to a 100-fold surge of relaxin levels during late pregnancy (2). It is possible, therefore, that relatively less transgenic overexpression of relaxin could elicit effects only when the endogenous relaxin level is low, as in virgin females. The fact that Tg(Rln1) males do not exhibit enlarged nipples suggests that relaxin likely requires other gender-specific factors to potentiate nipple growth. Enhanced virgin nipple development, caused by relaxin

TABLE 2. Biochemical parameters of plasma of control and Tg(Rln1) animals Parameter

Female WT

Female Tg

Male WT

Male Tg

VEGF (pg/ml) NO3/NO3 (␮M) MMP2 (r.u.) MMP9 (r.u.) Plasma protein (mg/ml)

77.59 ⫾ 5.78 (n ⫽ 6) 53.87 ⫾ 5.57 (n ⫽ 6) 1.0 ⫾ 0.157 (n ⫽ 6) 1.0 ⫾ 0.168 (n ⫽ 6) 44.8 ⫾ 1.65 (n ⫽ 6)

66.56 ⫾ 6.19 (n ⫽ 8) 59.86 ⫾ 4.73 (n ⫽ 7) 1.084 ⫾ 0.099 (n ⫽ 8) 0.81 ⫾ 0.103 (n ⫽ 8) 43.61 ⫾ 1.6 (n ⫽ 8)

68.66 ⫾ 10.71 (n ⫽ 8) 42.05 ⫾ 6.56 (n ⫽ 9) 1.0 ⫾ 0.131 (n ⫽ 6) 1.0 ⫾ 0.09 (n ⫽ 6) 44.59 ⴞ 1.19 (n ⫽ 10)

75.58 ⫾ 18.15 (n ⫽ 8) 34.72 ⫾ 3.69 (n ⫽ 10) 1.325 ⫾ 0.145 (n ⫽ 7) 1.036 ⫾ 0.211 (n ⫽ 7) 41.19 ⴞ 0.96 (n ⫽ 10)

Values represent means ⫾ SE. The number of mice in each group is indicated in parentheses. Values different with P ⬍ 0.05 are in bold. Plasma protein is total plasma protein concentration. r.u., Reference units.



413

TABLE 3. Hydroxyproline content (␮g/g wet weight) in organs of control and Tg(Rln1) animals Organ

Female WT

Female Tg

Male WT

Male Tg

Liver Kidney

927.4 ⫾ 48.1 (n ⫽ 4) 656.3 ⫾ 51.2 (n ⫽ 7)

989.1 ⫾ 68.0 (n ⫽ 8) 629.8 ⫾ 28.5 (n ⫽ 7)

1113 ⴞ 29 (n ⫽ 10) 590.2 ⫾ 15.1 (n ⫽ 10)

1310 ⴞ 49 (n ⫽ 9) 561.2 ⫾ 20.2 (n ⫽ 9)

Values represent means ⫾

SEM.

The number of mice in each group is indicated in parentheses. Values different with P ⬍ 0.05 are in bold.

overexpression, is not manifested in Tg(Rln1)Lgr7⫺/⫺ females, proving that the relaxin effect is exerted via Lgr7. We did not detect any abnormalities in reproductive characteristics of male mutant mice. Despite earlier reports that demonstrated a critical role for relaxin in the development of the male reproductive tract (14, 18), no differences were noted in the morphology, histology, or weight of the mutant testes and prostate in our experiments. Although it is possible that the circulating level of transgenic relaxin might be negligible compared with the local level of relaxin synthesized in testes and prostate, the results of this study corroborate our previous data showing that Lgr7⫺/⫺ phenotype is not associated with the abnormalities of the male reproductive system (17). To further ascertain the effects of endocrine relaxin, we assessed the content of MMPs, VEGF, and NO conversion products in blood plasma of transgenic mice. Zymography did not reveal major changes in the gelatinase (MMP2 and MMP9) activity. The MMP2 level was slightly higher in both transgenic males and females; however, these differences did not reach statistical significance. Neither VEGF nor nitrite/ nitrate plasma content were noticeably altered by relaxin overexpression. The only parameter significantly deviating in Tg(Rln1) animals was the plasma protein concentration. It is interesting that this aberration was detected only in males. We assumed that lower protein concentration (hypoalbuminemia) in Tg(Rln1) males was associated with either impaired albumin synthesis by the hepatic cells or by increased albumin loss as a result of nephropathy. Because relaxin was shown earlier to control fibrotic processes in liver and kidney, we next focused our attention on the level of collagen deposition in these tissues. In accordance with lower plasma protein concentration in Tg(Rln1) males, we found that livers of these animals display significantly higher hydroxyproline content, indicative of liver fibrosis. Interestingly, Lgr7⫺/⫺ male mice have been reported to have significantly different plasma levels of the liver panel enzymes (alanine aminotransferase, aspartate aminotransferase, and alkaline phosphatase) and displayed slightly higher plasma protein and plasma albumin concentration than their wild-type littermates (18). Based on these data, we propose that relaxin signaling may be important for the control of liver function in male mice, although we cannot explain the male-specific manifestation of the observed effects. Similarly, only Rln1deficient males, but not the females, were affected by renal fibrosis (15) and increased myocardial collagen deposition, associated with ventricular diastolic dysfunction (16). It is possible, therefore, that relaxin signaling is affected by certain gender-specific factors. Relaxin was shown to induce extracellular matrix degradation in other experiments (2, 12), causing significant weight increase in several nonreproductive organs in aging Rln1-deficient mice (13–16). Based on these observations, we

expected to observe decreased collagen deposition in transgenic organs. However, we were not able to detect any differences in the liver, heart, lung, and kidney weights in Rln1-transgenic mice. It is possible that the analysis of the older mice or the transgenic mice on a different genetic background would reveal such an effect. However, increased collagen deposition in transgenic liver might indicate the existence of some compensatory mechanism, counteracting the collagen-degrading action of relaxin. The increased collagen deposition in the liver of Tg(Rln1) males, caused by the moderate and long-term relaxin stimulation, may point to the potential side effects of relaxin treatment, if relaxin is to be considered as a therapeutic agent. Mice with relaxin overexpression may be used to study the role of relaxin signaling in certain pathological processes. Data accumulated to date implicate relaxin in cancer cell growth, migration, and tumor neovascularization (23). Intercrossing of relaxin transgenic mice with mice prone to cancer development can clarify the effects of this hormone in tumor growth and invasion in vivo. These and other aspects of relaxin physiology are currently under investigation in our laboratory. Acknowledgments We thank Dr. L. Parada for Insl3-deficient mice, Dr. D. Hanahan for pRIP-1Tag plasmid, and Dr. B. Korchin for the assistance with mammary gland analysis. Received May 24, 2005. Accepted October 6, 2005. Address all correspondence and requests for reprints to: Dr. Alexander I. Agoulnik, Department of Obstetrics and Gynecology, 6550 Fannin Street, Baylor College of Medicine, Houston, Texas 77030. E-mail: [email protected]. A.A.K. was a National Institutes of Health BIRCWH Fellow (Grant 5K12 HD01426). This work was supported by National Institutes of Health Grants R01 HD37067 and P01 HD36289 (to A.I.A.). Current address for A.A.K.: Department of Gynecologic Oncology, UT-MD Anderson Cancer Center, Houston, Texas 77230-1439. The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.

References 1. Bathgate RA, Samuel CS, Burazin TC, Gundlach AL, Tregear GW 2003 Relaxin: new peptides, receptors and novel actions. Trends Endocrinol Metab 14:207–213 2. Sherwood OD 2004 Relaxin’s physiological roles and other diverse actions. Endocr Rev 25:205–234 3. Fields PA, Larkin LH 1980 Enhancement of uterine cervix extensibility in oestrogen-primed mice following administration of relaxin. J Endocrinol 87: 147–152 4. Zaleski HM, Winn RJ, Jennings RL, Dziuk PJ, Sherwood OD 1995 Effects of relaxin administration in early gestation or prior to mating on uterine length and fetal survival in gilts. Biol Reprod 52:1389 –1394 5. Hayes ES, Curnow EC, Trounson AO, Danielson LA, Unemori EN 2004 Implantation and pregnancy following in vitro fertilization and the effect of recombinant relaxin administration in Macaca fascicularis. Biol Reprod 71: 1591–1597

414


6. Hwang JJ, Lee AB, Fields PA, Haab LM, Mojonnier LE, Sherwood OD 1991 Monoclonal antibodies specific for rat relaxin. Endocrinology 129:3034 –3042 7. Summerlee AJ, Ramsey DG, Poterski RS 1998 Neutralization of relaxin within the brain affects the timing of birth in rats. Endocrinology 139:479 – 484 8. Steinetz BG, Sherwood OD, Lasano S, Horton L, Bosland MC 2004 Immunoneutralization of circulating relaxin does not alter the breast cancer-protective action of parity in MNU-treated rats. J Exp Ther Oncol 4:59 – 68 9. Zhao S, Sherwood OD 2004 Induction of labor with RU 486 (mifepristone) in relaxin-deficient rats: antepartum administration of relaxin facilitates delivery and increases pup survival. Am J Obstet Gynecol 190:229 –238 10. Hsu SY, Nakabayashi K, Nishi S, Kumagai J, Kudo M, Sherwood OD, Hsueh AJ 2002 Activation of orphan receptors by the hormone relaxin. Science 295: 671– 674 11. Hsu SY, Nakabayashi K, Nishi S, Kumagai J, Kudo M, Bathgate RA, Sherwood OD, Hsueh AJ 2003 Relaxin signaling in reproductive tissues. Mol Cell Endocrinol 202:165–170 12. Zhao L, Roche PJ, Gunnersen JM, Hammond VE, Tregear GW, Wintour EM, Beck F 1999 Mice without a functional relaxin gene are unable to deliver milk to their pups. Endocrinology 140:445– 453 13. Zhao L, Samuel CS, Tregear GW, Beck F, Wintour EM 2000 Collagen studies in late pregnant relaxin null mice. Biol Reprod 63:697–703 14. Samuel CS, Tian H, Zhao L, Amento EP 2003 Relaxin is a key mediator of prostate growth and male reproductive tract development. Lab Invest 83: 1055–1067 15. Samuel CS, Zhao C, Bond CP, Hewitson TD, Amento EP, Summers RJ 2004 Relaxin-1-deficient mice develop an age-related progression of renal fibrosis. Kidney Int 65:2054 –2064 16. Du XJ, Samuel CS, Gao XM, Zhao L, Parry LJ, Tregear GW 2003 Increased myocardial collagen and ventricular diastolic dysfunction in relaxin deficient mice: a gender-specific phenotype. Cardiovasc Res 57:395– 404


17. Kamat AA, Feng S, Bogatcheva NV, Truong A, Bishop CE, Agoulnik AI 2004 Genetic targeting of relaxin and insulin-like factor 3 receptors in mice. Endocrinology 145:4712– 4720 18. Krajnc-Franken MA, van Disseldorp AJ, Koenders JE, Mosselman S, van Duin M, Gossen JA 2004 Impaired nipple development and parturition in LGR7 knockout mice. Mol Cell Biol 24:687– 696 19. Kumagai J, Hsu SY, Matsumi H, Roh JS, Fu P, Wade JD, Bathgate RA, Hsueh AJ 2002 INSL3/Leydig insulin-like peptide activates the LGR8 receptor important in testis descent. J Biol Chem 277:31283–31286 20. Bogatcheva NV, Truong A, Feng S, Engel W, Adham IM, Agoulnik AI 2003 GREAT/LGR8 is the only receptor for insulin-like 3 peptide. Mol Endocrinol 17:2639 –2646 21. Feng S, Bogatcheva NV, Kamat AA, Agoulnik AI 2005 Genetic targeting of relaxin and Insl3 signaling in mice. Ann NY Acad Sci 1041:82–90 22. Conrad KP, Novak J 2004 Emerging role of relaxin in renal and cardiovascular function. Am J Physiol 287:R250 –R261 23. Silvertown JD, Summerlee AJ, Klonisch T 2003 Relaxin-like peptides in cancer. Int J Cancer 107:513–519 24. Nef S, Parada LF 1999 Cryptorchidism in mice mutant for Insl3. Nat Genet 22:295–299 25. Silver L 1995 Mouse genetics. New York: Oxford University Press; 66 26. Pepper MS 2001 Role of the matrix metalloproteinase and plasminogen activator-plasmin systems in angiogenesis. Arterioscler Thromb Vasc Biol 21: 1104 –1117 27. Reddy GK, Enwemeka CK 1996 A simplified method for the analysis of hydroxyproline in biological tissues. Clin Biochem 29:225–229 28. Halls ML, Bond CP, Sudo S, Kumagai J, Ferraro T, Layfield S, Bathgate RA, Summers RJ 2005 Multiple binding sites revealed by interaction of relaxin family peptides with native and chimeric relaxin family peptide receptors 1 and 2 (LGR7 and LGR8). J Pharmacol Exp Ther 313:677– 687

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