The endogenous proteinase inhibitors plasminogen activator inhibitor type 1 (PAI-1) and tissue inhibitor of metalloproteinase type 1 (TIMP-1) are two distinct proteins with separate molecular pathways. However, a close relationship between PAI-1 and TIMP-1 has been proposed indicating some degree of functional overlap due to their involvement in ECM turnover, tissue remodeling, and cellular migration and signaling. To study the housekeeping physiological implications of PAI-1 and TIMP-1, we generated a combined PAI-1 and TIMP-1 gene-deficient mouse model. We present the results on generating this specific mouse model with particular emphasis on phenotypical characteristics, blood leukocyte counts, histology, and gene expression studies of PAI-1 and TIMP-1 in various organs. We observed a significant deviation in segregation of offspring only in male mice (P < 0.01) predominantly caused by PAI-1 deficiency. In addition, the body weight in 3- and 20-wk-old male and 20-wk-old female mice was significantly different between genotypes (P ≤ 0.0008). Furthermore, blood leukocyte counts were significantly different between genotypes in 20-wk-old male mice (P ≤ 0.0002), whereas no significant differences were observed between genotypes in 20-wk-old female mice (P ≥ 0.13). Quantifying the relative expression of PAI-1 and TIMP-1 revealed upregulation of PAI-1 (P < 0.001) in male mice only. Our data highlight the complex roles of PAI-1 and TIMP-1 on physiological parameters such as segregation of offspring (embryonic development and survival), body weight (metabolism), blood leukocyte counts (immunity), and gene expression (regulatory redundancy). We conclude that PAI-1 and TIMP-1 seem to possess gender-dependent regulatory properties on various housekeeping physiological parameters and stress the potential implications in pathological conditions.
- body weight
- blood leukocyte count
- gene expression
- plasminogen activator inhibitor type 1
- tissue inhibitor of metalloproteinase type 1
the endogenous proteinase inhibitors, the tissue inhibitor of metalloproteinase type 1 (TIMP-1) and plasminogen activator inhibitor type 1 (PAI-1), are two distinct proteins with separate molecular pathways. However, a considerable amount of work points to a close physiological relationship between these two inhibitors reflecting some degree of functional overlap. TIMP-1 is described mainly by its properties of being an endogenous inhibitor of matrix metalloproteinases (MMPs). The MMPs degrade various components of the extracellular matrix (ECM) and are categorized into groups depending on their major substrates, e.g., collagenases, stromelysins, gelatinases, and matrilysins (44). Thus TIMP-1 is an important physiological regulator of ECM turnover. In contrast, PAI-1 is a serine protease inhibitor that exerts its physiological effect on the plasminogen activator (PA) system by inhibiting urokinase PA (uPA) and tissue-type PA (tPA). By reducing indirectly the generation of plasmin, PAI-1 is one of the primary regulators of the fibrinolytic system. In addition, PA and plasmin are implicated in numerous nonfibrinolytic processes leading to ECM degradation, either directly, by proteolytic cleavage of molecules such as fibronectin, laminin, proteoglycan, vitronectin, and collagen (14), or indirectly, through the activation of latent MMP precursors (9, 14). Thus TIMP-1 and PAI-1 contribute in concert to the regulation of ECM turnover, which constitutes the most obvious physiological link between the two systems. The presence of some degree of functional overlap between the MMPs and the PA system involving TIMP-1 and PAI-1 has already been established in studies on wound healing and tissue remodeling (12, 20, 25).
The ECM is a complex interacting and dynamic network of fibrillar proteins, which compose the scaffold for the various cellular compartments and determine the physiochemical environment for normal cellular homeostasis, growth, and migration. During many (patho)physiological processes (e.g., embryogenesis, angiogenesis, tissue remodeling, cancer, and inflammation), the level of both PAI-1 and TIMP-1 is altered (11, 40, 43). This alteration in proteinase inhibitor gene expression and protein levels most probably reflects the demand for tight regulation of PA/plasmin and MMP activity to avoid excessive and deleterious proteolytic turnover of ECM. The expression of PAI-1 and TIMP-1 is regulated at the transcriptional level under the influence of different biological mediators such as IL-1, IL-6, TNF-α, C5a, transforming growth factor (TGF)-β, ACTH, and glucocorticoid hormones (11, 14–16, 27, 35, 37). This substantiates the role of PAI-1 and TIMP-1 as regulatory proteins and potential biomarkers of proinflammatory and degenerative lesions (2, 18, 41), advanced cancerous conditions (13, 40), and metabolic disorders (17, 29).
PAI-1 knockout (PKO) and TIMP-1 knockout (TKO) mice are viable, fertile, and without significant abnormalities in organogenesis and development compared with wild-type siblings (PWT and TWT) (3, 23). During normal conditions only scarce reports on significant phenotypic characteristics of PKO and TKO mice have been recorded (4, 33), suggesting some degree of physiological redundancy. Furthermore, no physiological implications of the TIMP-1 gene being X-linked have been described. However, a variety of pathological conditions reveal important phenotypic characteristics of PAI-1 or TIMP-1 deficiency in different mouse models (1, 19, 38, 39). Thus the availability of mice with targeted gene inactivation is a strong scientific tool that has shed considerable light on the biological functions of PAI-1 and TIMP-1.
To study in detail the relation between PAI-1 and TIMP-1 and to further uncover the intriguing phenomenon of functional overlap between these two proteins, we have generated a combined PAI-1 and TIMP-1 gene-deficient mouse model. We present the results of generating this specific mouse model with particular emphasis on phenotypical characteristics, blood leukocyte counts, histology, and gene expression studies of PAI-1 and TIMP-1 in various organs. We conclude that PAI-1 and TIMP-1 seem to possess sex-dependent regulatory properties on various housekeeping physiological parameters and stress the potential implications in pathological conditions.
MATERIALS AND METHODS
The PAI-1 and TIMP-1 single gene-deficient mouse strains were kindly donated by Peter Carmeliet and Poul Soloway, respectively, and generated as previously described (3, 42). In brief, PKO and TKO mice were generated by homologous recombination using a neomycin-containing gene-targeting vector in mouse embryonic stem cells completely abolishing functional gene transcripts. The PKO and TKO mice were backcrossed into the BALB/c mouse strain for at least eight generations. To generate the combined PAI-1 and TIMP-1 gene-deficient mouse strain (PKO/TKO), the two strains were crossed, and to obtain homozygous single- and double-gene KO and WT siblings, heterozygous mating was set up. The PKO/TKO mouse strain was subsequently backcrossed into the BALB/cJ background for six generations. All mice were contract bred and housed at Taconic Europe (Bomholt, Denmark) within environmentally and microbiologically controlled conditions on a 12:12-h day-night cycle with water and food ad libitum (NIH no. 31M rodent diet, 11.7 kJ/g; Taconic Europe). All animal procedures were conducted in accordance with institutional guidelines and on a license granted by The Danish Animal Experimental Board.
PAI-1 and TIMP-1 genotyping.
Genomic DNA was isolated from tail tip specimens sampled from 4-wk-old mice after experiments and from discarded breeding mice. Tail tips were degraded for 12 h at 55°C in lysis buffer (100 mM Tris, 200 mM NaCl, 5 mM EDTA, and 0.2% SDS) containing 0.1 μg/ml proteinase K (Sigma-Aldrich), and DNA was subsequently precipitated with isopropanol. Specific primers for murine PAI-1 and TIMP-1 were designed based on the sequence data in EMBL, and the neomycin reverse primer was designed based on the sequence data from the cloning vector pKONEO in GenBank (Table 1). PCR was performed in 25-μl reactions containing 1× HotStarTaq master mix (Qiagen Nordic, West Sussex, UK), 0.4 μM of each gene-specific primer, 2 mM MgCl2, and 1 μl (∼100 ng) of diluted genomic DNA template. The PCR amplification conditions were 15 min at 95°C, followed by 35 cycles of 94°C for 1 min, 64°C for 1 min, and 72°C for 1 min, and a final extension step of 72°C for 10 min. PCR products were electrophoresed on a 1.5% agarose gel (Fermentas, Helsingborg, Sweden) with 0.5 μg/ml ethidium bromide and 0.5× Tris-acetate-EDTA buffer, visualized with a UV transilluminator, and photographed on a Biospectrum imaging system (UVP, Upland, CA).
Phenotypic characteristics, blood leukocyte counts, and PAI-1 and TIMP-1 plasma measurements.
To assess the welfare and general health status of the F1 progeny, we measured the body weight of 3-wk-old male mice and 20-wk-old male and female mice and visually judged the general appearance. Blood samples from 20-wk-old male and female mice were collected from the retro-orbital venous plexus in heparin-coated capillary tubes. Complete white blood cell (WBC) counts and leukocyte differentials were performed using the Advia 120 hematology system automated cell counter (Bayer, Lyngby, Denmark) and by flow cytometry (FACSCalibur; Becton Dickinson, Fullerton, CA). Blood for blood leukocyte counting on Advia 120 and flow cytometry was obtained from the same blood sample. For flow cytometry, 10 μl of blood were hemolyzed in lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, and 1 mM disodium EDTA, pH 7.3) for 6 min and finally diluted 1:50 in isotonic PBS before analysis. WBCs were gated and categorized according to size and granularity by forward and sideward scatter plots. For absolute counting procedures, the instrument particle registration efficacy was calibrated and measured using TruCount controls (Becton Dickinson). Plasma for PAI-1 and TIMP-1 measurements was obtained from 16- to 20-wk-old mice by centrifugation of heparinized blood samples at 3,000 g for 10 min. To minimize cellular contamination, plasma was gently pipetted, leaving a guard zone of ∼2 mm. For the mouse PAI-1 total antigen assay kit (catalog no. MPAIKT-TOT; Innovative Research, Southfield, MI), plasma was diluted 1:4, and for the mouse TIMP-1 ELISA kit (catalog no. ELM-TIMP1-001; RayBiotech, Norcross, GA), plasma was diluted 1:50 according to the manufacturer's instructions. Whenever possible, all measurements were performed as duplicates.
For histological analysis, tissue specimens from 16- to 20-wk-old mice were fixed in 4% neutral buffered paraformaldehyde or formaldehyde for 24 h at 4°C, processed through graded concentrations of ethanol and xylene, and embedded in paraffin wax. Tissue sections of 3–5 μm were stained with hematoxylin and eosin (H&E) according to standard procedures. From 4 male mice from each homozygous genotype (PWT/TWT, PWT/TKO, PKO/TWT, and PKO/TKO), 11 different tissues were selected: lung, jejunum, colon, liver, skin, testis, striated skeletal muscle, cerebrum, femur including bone marrow, inguinal white adipose tissue (iWAT), and interscapular brown adipose tissue (iBAT). Because PAI-1 and TIMP-1 have been described to be involved in regulating adipocyte morphology (22, 26), we also selected iWAT and iBAT from 8 female mice (2 mice from each homozygous genotype). Assessment of histology was performed by an observer who was blinded to the genetic background.
RNA extraction and reverse transcription.
For RNA extraction, 16-wk-old mice were killed by cervical dislocation and organs were immediately removed. RNA was extracted from tissue samples using a spin column kit (SV total RNA isolation system; Promega, Madison, WI) according to the manufacturer's instructions. This procedure includes an on-column DNase treatment, minimizing the risk of DNA contamination. Before extraction, up to 60 mg of tissue sample were homogenized in 400 μl of lysis buffer. RNA was extracted from 175 μl of lysate, and the concentration of total RNA was measured spectrophotometrically. RNA (1.4 μg) was transcribed into cDNA using the First-Strand cDNA synthesis kit (Fermentas) in a total volume of 25 μl. The reaction consisted of 1× reaction buffer, 0.8 mM dNTPs, 20 units of RiboLock RNase inhibitor, 0.5 μg of oligo(dT) primer, 0.2 μg of random hexamer primer, and 40 units of Moloney murine leukemia virus reverse transcriptase. Samples were incubated at 25°C for 10 min, followed by 42°C for 1 h. The reaction was terminated by incubating at 95°C for 5 min, followed by cooling on ice.
PCR and quantitative real-time PCR.
To estimate the expression level of PAI-1 and TIMP-1, a quantitative real-time PCR (qPCR) assay determining relative expression levels was developed. β-Actin was used as reference gene. To avoid false positive results from contaminating genomic DNA, all primer sets used were intron spanning (Table 2). For PAI-1, forward exon 3 and reverse exon 4 primers were used. For TIMP-1, forward exon 5 and reverse exon 6 primers were used. qPCR was carried out using SYBR green I detection and the LightCycler 480 system (Roche Diagnostics, Hvidovre, Denmark) as previously described (34). PCR conditions for each primer set were optimized by determining the annealing temperature at which only the specific product was seen. Reactions were carried out in 20-μl volumes consisting of 1× LightCycler 480 SYBER green 1 master mix and 0.5 μM of the gene-specific primer. The amplification program was as follows: preincubation at 95°C for 5 min, followed by 45 amplification cycles (95°C for 10 s, 60°C for 10 s, and 72°C for 12 s). SYBR green fluorescence was acquired at 72°C in each amplification cycle. After the end of the last cycle, the melting curve was generated by starting the fluorescence acquisition at 65°C and taking measurements every 0.1 s until 95°C was reached. A standard curve for each primer set was generated using 10-fold dilutions of a pool of tissue cDNA. The standard curve was used to correct for differences in PCR efficiencies between primer sets. Dilutions of tissue cDNA samples were chosen for the generation of standard curves to ensure the same PCR efficiency in standards and samples. Relative quantification was done using Relative Quantification software (LightCycler 480; Roche). The relative expression level of PAI-1 and TIMP-1 in each tissue specimen was calculated as the mean value from two male and two female mice of each relevant homozygous genotype. The PWT/TWT tissue specimen with the lowest expression ratio was arbitrarily set to the value of 10, and all other tissue-specific expression ratios were subsequently related to this value. Finally, all data were log transformed and arranged in order of magnitude.
The distribution of data was tested using the Kolmogorov-Smirnov test and residual plots, and if appropriate, data were analyzed using parametric tests; otherwise, nonparametric tests were used. Data are means ± SD or medians with interquartile ranges. Segregation of offspring was tested in 2 × 3 contingency tables by using χ2 tests with two degrees of freedom. Differences in body weight and relative levels of lymphocyte and granulocyte counts were tested using one-way analysis of variance (ANOVA) with Bartlett's test for equal variance and Bonferroni's multiple comparison test. The difference in breeding performance between TWT and TKO male mice and the differences in PAI-1 and TIMP-1 plasma levels between genotypes were tested using a two-tailed unpaired t-test. The blood leukocyte counting agreement between Advia 120 and flow cytometry was tested using a two-tailed, one-sample t-test. Differences in blood leukocyte counts were tested using the Kruskal-Wallis test with Dunn's multiple comparison test. Differences in PAI-1 and TIMP-1 expression levels were tested using the two-tailed Wilcoxon signed rank test. The level of statistical significance was P < 0.05.
Breeding of PAI-1 and TIMP-1 double-gene-deficient mice.
After the crossbreeding of the PAI-1 and TIMP-1 single-gene KO mouse strains (P−2) and subsequent backcrossing of female P−1 mice on P−2 male mice, we generated parental (P) heterozygous PAI-1 and TIMP-1 (PAI-1+/−/TIMP-1+/−) female breeding animals and PAI-1+/− and hemizygous TKO (TIMP-1−/0) male breeding animals (Fig. 1). Because the TIMP-1 gene is linked to the X chromosome, this breeding produced male F1 siblings of all possible PAI-1 and TIMP-1 gene variations; however, only homozygous male mice were used for further procedures (Fig. 1). A similar segregation of homozygous female F1 offspring was achieved only by generating half-siblings using TKO and TWT male breeding mice (pedigree not shown).
To obtain male F1 offspring with all four homozygous genotypes in each litter, we chose a breeding setup with parental breeding animals that would produce the expected genotypes (Fig. 2). Theoretically, both the PAI-1 and the TIMP-1 allele pair segregate independently during gamete formation. Statistically, this makes each outcome in the F1 progeny equally likely. Consequently, we tested whether the observed segregation in the F1 progeny followed the theoretical Mendelian inheritance (Fig. 2). The segregation of animals in the F1 male progeny did not statistically follow the Mendelian inheritance (P < 0.01), whereas we did not reject the hypothesis in the F1 female progeny (P > 0.05). However, only in the male progeny were all possible genotypes present. The relatively small numbers of male PKO/TWT (n = 16) and PKO/TKO mice (n = 16) contributed considerably to the χ2 test (χ2 = 6.2). In the female progeny, the relatively high number of PAI-1+/−/TKO mice (n = 77) also contributed notably to the χ2 test (χ2 = 3.5), however, not making the overall distortion statistically significant. In the male progeny, the effect of being either TWT (n = 95) or TKO (n = 103) did not contribute much to the distortion in the segregation of genotypes, whereas the effect of being PKO (n = 32) compared with PWT (n = 53) or PAI-1+/− (n = 113) had a much greater impact. Furthermore, we tested whether the two variables (PAI-1 and TIMP-1) were segregated independently in the F1 progeny. The two variables were statistically independent in both the male (P = 0.54) and female F1 progeny (P = 0.74).
Phenotypic characteristics, blood leukocyte counts, and PAI-1 and TIMP-1 plasma measurements.
Previously, when the single-gene PKO and TKO mouse strains were bred separately, we did not observe any significant distortion in the segregation of either of the progenies or any alterations in phenotypic characteristics such as litter size, distribution by sex, or body weight. In the present study, we measured the body weight of 3-wk-old homozygous male mice (n = 31) and 20-wk-old homozygous male (n = 58) and female mice (n = 37) from the F1 progeny (Fig. 3). The 3- and 20-wk-old male mice were different individuals. The difference in mean body weight between the four genotypes was highly significant in the 3-wk-old male (P = 0.0004) and in the 20-wk-old male (P = 0.0008) and female progeny (P < 0.0001). In 3-wk-old male mice, the PWT/TKO mice had the highest mean body weight, which was significantly higher compared with that of PWT/TWT (P < 0.001) and PKO/TKO mice (P < 0.01). In 20-wk-old male mice, the PWT/TKO mice also displayed the highest mean body weight, which was significantly higher compared with that of PKO/TWT mice (P < 0.001). In 20-wk-old female mice, the PKO/TKO mice had the lowest mean body weight, which was significantly lower compared with that of PWT/TKO (P < 0.01) and PKO/TWT mice (P < 0.001). Furthermore, we assessed the breeding performance of TWT and TKO male mice by generating a breeding setup using heterozygous female mice, as previously described, and male mice being PAI-1+/−/TWT or PAI-1+/−/TKO. The number of offspring in each litter was not statistically different between male TWT and TKO mice (P = 0.43). The mean litter size generated by TWT male mice was 7.1 ± 1.6 (n = 21), and that generated by TKO male mice was 6.7 ± 1.8 (n = 23). In addition, we assessed the breeding capacity of combined PAI-1- and TIMP-1-deficient mice by crossing female and male PKO/TKO mice. They produced live and fertile offspring of apparently normal litter sizes and an even distribution by sex (data not shown).
Blood leukocyte differentials in 20-wk-old male mice revealed highly significant differences in median WBC (P < 0.0001), lymphocyte (P < 0.0001), and granulocyte counts (P = 0.0002) between genotypes (Fig. 4). The WBC counts in PKO/TKO male mice were significantly lower compared with those in PKO/TWT (P < 0.05) and PWT/TKO mice (P < 0.001). The lymphocyte counts in PWT/TKO mice were significantly higher compared with those in PWT/TWT (P < 0.05), PKO/TWT (P < 0.05), and PKO/TKO mice (P < 0.001). The observed tendencies in WBC and lymphocyte counts did not apply for the granulocyte counts. The granulocyte counts were significantly higher in PKO/TWT mice compared with those in PKO/TKO (P < 0.001) and PWT/TWT mice (P < 0.01). Calculating the relative levels of lymphocytes and granulocytes (relative to WBC) in 20-wk-old male mice also revealed highly significant differences among genotypes (P ≤ 0.0001) (Fig. 4). In PKO/TWT mice, the relative lymphocyte counts were significantly lower compared with those in PWT/TWT (P < 0.01) and PWT/TKO mice (P < 0.001), whereas the relative granulocyte counts in PKO/TWT mice were higher compared with those in the other genotypes (P < 0.001). In 20-wk-old female mice, no significant difference in WBC, lymphocyte, or granulocyte counts among genotypes could be demonstrated (P ≥ 0.13, n = 28). Across genotypes, the mean blood leukocyte counts in 20-wk-old female mice were as follows (×109/liter): WBCs, 4.9 ± 1.0; lymphocytes, 3.3 ± 0.8; and granulocytes, 1.3 ± 0.3. Comparison of WBC, lymphocyte, and granulocyte counts on Advia 120 and flow cytometry was based on blood count values covering a wide physiological range, and the two methods had a high statistical correlation. WBCs: r2 = 0.97, n = 50 (median = 2.6 × 109/l, range = 0.4–10.2). Lymphocytes: r2 = 0.96, n = 50 (median = 1.0 × 109/l, range = 0.1–6.8). Granulocytes: r2 = 0.88, n = 50 (median = 0.8 × 109/l, range = 0.1–2.9). Differences in blood leukocyte counts between the two methods were relatively small and statistically insignificant (means ± SD): WBCs = 0.12 ± 0.42 (P = 0.76); lymphocytes = 0.03 ± 0.32 (P = 0.35) and granulocytes = 0.05 ± 0.22 (P = 0.55), n = 50.
PAI-1 and TIMP-1 plasma levels in male and female mice were not statistically different among genotypes (P ≥ 0.17). The PAI-1 levels (ng/ml) in male mice were 3.8 ± 1.5 in PWT/TWT (n = 17) and 3.5 ± 1.4 in PWT/TKO (n = 18), and those in female mice were 3.6 ± 0.8 in PWT/TWT (n = 6) and 3.0 ± 0.7 in PWT/TKO (n = 8). TIMP-1 levels (ng/ml) in male mice were 7.9 ± 1.4 in PWT/TWT (n = 17) and 7.9 ± 1.6 in PKO/TWT (n = 10), and those in female mice were 7.0 ± 0.8 in PWT/TWT (n = 6) and 7.5 ± 0.6 in PKO/TWT (n = 7).
Histological examination of H&E-stained tissue sections from male mice revealed no apparent differences among genotypes. However, the morphological characteristics of adipocytes in iWAT from female mice had qualitative differences compared with male mice. iWAT from female mice representing all homozygous genotypes contained dispersed areas of cells having morphological characteristics of brown adipocytes (Fig. 5). These accumulations of iBAT-like cells were not identified in iWAT from either of the male genotypes.
Quantitative real-time PCR.
Based on the hypothesis that PAI-1 and TIMP-1 possess some degree of physiological redundancy, we generated cDNA from two PWT/TWT, two PKO/TWT, and two PWT/TKO male and female mice to quantify the difference in gene expression ratio of PAI-1 and TIMP-1 in selected tissue specimens between control (WT) and single-gene KO mice. The PAI-1 and TIMP-1 expression level in each tissue was quantitated relatively to the β-actin level. The assumption of equal β-actin expression levels among organs was assessed by analyzing the arbitrary β-actin expression level among organs in the different male and female genotypes. β-Actin expression levels in different organs between genotypes in male and female mice was not significantly different (P > 0.23). Thus we tested the relative PAI-1 (PWT/TKO vs. PWT/TWT) and TIMP-1 (PKO/TWT vs. PWT/TWT) expression levels in male and female mice (Fig. 6). In male PWT/TKO mice, the number of organs expressing relatively higher levels of PAI-1 was highly significant compared with that in PWT/TWT mice (P < 0.001; n = 15), whereas the relative level of PAI-1 expression in female PWT/TKO mice was not significantly different compared with that in PWT/TWT mice (P = 0.93; n = 16). The expression levels of TIMP-1 in male and female PKO/TWT mice compared with those in PWT/TWT mice were statistically insignificant (P = 0.12; n = 14, and P = 0.39; n = 16, respectively).
Generation of the combined PAI-1 and TIMP-1 KO mouse model had significant physiological implications affecting the expected distribution of offspring, body weight, and blood leukocyte differentials. The segregation of F1 offspring followed the expected distribution only in female mice, whereas the distribution of male mice displayed significant deviation (Fig. 2). The regulatory properties of PAI-1 and TIMP-1 on fibrinolysis, ECM turnover, cell proliferation, and migration point to an important role in trophoblast invasion, implantation, and placentation (5), which may affect embryonic survival and segregation of genotypes postpartum. In the male progeny PAI-1 significantly contributed to the deviation in segregation of offspring. The effect of PAI-1 was independent of TIMP-1, which contributed much less to the deviation. Thus the lack of PAI-1 in the PKO male mice progeny appeared unfavorable for the generation of live offspring compared with PWT and PAI-1+/− mice. This observation is in accordance with present knowledge of the relation between PAI-1 and early blastocyst implantation, which is very dependent on tight regulation of the fibrinolytic system for successful gestation (7). However, in the female F1 progeny, the adverse effect of being PKO compared with PWT and PAI-1+/− was much less pronounced, whereas the beneficial effect of being TKO compared with TIMP-1+/− seemed much more conspicuous than in F1 male siblings. Implications of TIMP-1 as a determinant factor for embryonic growth and survival in female offspring could be related to X chromosome imprinting early in development. Complete paternal X chromosome inactivation during early female mouse development until the 32-cell stage and selective X chromosome inactivation in the trophoectoderm (36) may affect embryonic growth, differentiation, and survival by selective maternal X chromosome TIMP-1 transcription in TIMP-1+/− mice. Acknowledging the impact of X chromosome TIMP-1 transcription in TIMP-1+/− female mice highlights the need for, and potential beneficial effects of, relatively high proteolytic activity during embryonic implantation, as may be the case in TKO mice (n = 139) compared with TIMP-1+/− (n = 110). However, the impact of X chromosome inactivation on TIMP-1 expression in female embryos did not affect the total number of offspring, since TWT and TKO male breeding mice produced equal litter sizes. This points to a negligible effect of TIMP-1 on the breeding capacity of male mice.
Studies on PAI-1- and TIMP-1-deficient male mice demonstrate an inhibitory effect on adipose tissue development, body weight, and obesity compared with WT control mice (22, 26). The impact of sex hormones on lipolytic activity in mice may, in combination with the PAI-1 and TIMP-1 constitution, help explain, at least in part, the sex-dependent differences in body weight observed among genotypes. Female sex hormones show stimulatory effects on lipolysis by altering the adrenergic receptor ratio, consequently increasing uncoupling protein 1 expression, whereas testosterone inhibits lipolysis in mouse brown adipose tissue (31). Thus sexual dimorphism in regional adipose tissue distribution and regulation of adipose tissue homeostasis probably affects thermoregulatory properties and body weight because of differences in response to adipokines (including PAI-1). Our observation that iBAT-like cells were identified in iWAT only in female mice supports the notion that sex may affect adipose tissue morphology and differentiation (Fig. 5). We propose that sex-related differences, including the capacity of brown adipose tissue recruitment, amplify the physiological implications of PAI-1 on body weight in male mice. This is substantiated in murine studies on diet-induced obesity in male mice; obesity was prevented only in PAI-1−/− mice (26), and administering a specific PAI-1 inhibitor induced a dose-dependent reduction in body weight (8). Thus the facts that PWT/TKO male mice (both 3 and 20 wk old) exhibited the highest mean body weight (Fig. 3) and that PWT/TKO male mice also displayed significant upregulation of PAI-1 transcription (Fig. 6A) may not be incidental and support the hypothesis that PAI-1 may not merely increase in response to obesity but may play a direct causal role. The difference in body weight between male genotypes was marginally more pronounced in 3-wk-old mice compared with 20-wk-old mice. The higher metabolic rate in young and growing animals could render the physiological implications of PAI-1 and TIMP-1 more distinct. Contradictory results on obesity in PAI-1-deficient mice (32) and in mice overexpressing PAI-1 (21) or TIMP-1 (10) emphasize the need for standardized animal models using proper controls and a robust experimental design. In the female progeny, the impact of PAI-1 on the mean body weight was not evident. This corresponds to the observation that PAI-1 transcription levels among female genotypes were very similar and highlights the sex-related impact of PAI-1. Female PKO/TKO mice displayed the lowest mean body weight compared with other genotypes, which was not the case for male PKO/TKO mice. This points to a stronger mutual dependency of PAI-1 and TIMP-1 in female mice, at least on body weight, indicating some degree of physiological redundancy. Furthermore, the observations that only iWAT from female mice contained areas of adipocytes with iBAT-like morphology and that PAI-1 expression in iBAT from female mice was less pronounced compared with male mice indicate the importance of sex on adipose tissue differentiation and PAI-1 expression. The reason why we were not able to identify quantitative differences in adipocyte size among genotypes, as described by others (22, 26), was probably because, in addition to mouse strain differences, mice were not fed high-fat diets and qualitative differences in adipocyte morphology in female mice may have obscured the interpretation of results.
The differences in blood leukocyte counts between genotypes in male mice demonstrate the physiological implications of PAI-1 and TIMP-1 on the number of circulating leukocytes during steady state. The relationship between obesity, PAI-1, and inflammation is primarily based on common regulatory mediators such as IL-1, IL-6, and TNF-α. Studies on humans have identified PAI-1 plasma levels and WBC counts as independent determinants of the metabolic syndrome (30). Thus our observation that PWT/TKO male mice have the highest mean body weight, the highest WBC median count, and the highest PAI-1 transcript upregulation parallel clinical findings. The observation that PKO/TKO male mice displayed the lowest blood leukocyte counts, PWT/TWT mice displayed intermediate levels, and PWT/TKO or PKO/TWT mice displayed the highest blood leukocyte levels indicates some degree of redundancy and a state of physiological dependency of PAI-1 and TIMP-1 on the level of circulating WBCs. Thus the positive effect of PAI-1 or TIMP-1 deficiency on the level of blood leukocyte counts becomes evident only with the other molecule being present. Inflammatory models using PAI-1- or TIMP-1-deficient mice have revealed a significant influence on early neutrophil migration and activation and a response to infection (19, 38). The high relative granulocyte count in PKO/TWT (36.9%) male mice compared with that in PWT/TWT (25.6%), PWT/TKO (23.9%), and PKO/TKO (28.5%) mice suggests that PAI-1 deficiency in combination with the presence of TIMP-1 regulates the level of circulating granulocytes, either directly by increasing the granulopoiesis in bone marrow or indirectly by shifting the equilibrium between marginated and circulating granulocytes. The latter is substantiated by results indicating that PAI-1 supports IL-8-mediated neutrophil transendothelial migration (28) and that TIMP-1 specifically inhibits granulocyte migration through basement membranes (6). Consequently, the high level of granulocytes in PKO/TWT mice had considerable impact on the relative level of lymphocytes (56.4%), which was significantly lower compared with other genotypes. In contrast, TIMP-1 deficiency on its own (PWT/TKO), as also indicated by Lijnen et al. (22), did not affect the granulocyte count, whereas the addition of PAI-1 deficiency in PKO/TKO mice led to the lowest median level of circulating granulocytes. Data on blood leukocyte differentials in female mice were insignificant. However, this accentuates our previous findings indicating sex-related implications of PAI-1 and TIMP-1 on different physiological parameters in mice.
The choice of mouse strain in animal experiments has substantial implications on physiological and immunological parameters, e.g., body weight and blood leukocyte differentials. The C57BL/6 and BALB/c mouse strains are regarded as prototypical Th1- and Th2-dominant mouse strains, respectively (45), and the choice of mouse strain may as such influence the outcome of many pathological conditions. Experimental studies on PAI-1 and TIMP-1 in C57BL/6 mice have revealed insignificant alterations in body weight and hematological parameters between genotypes during steady state (22, 26). However, higher levels of TNF-α, IL-12, and IFN-γ in BALB/c mice compared with C57BL/6 mice in response to an inflammatory stimulus (45) reveal obvious strain differences that may augment the physiological implications of PAI-1 and TIMP-1 in our model.
Finally, the relative expression of PAI-1 and TIMP-1 in different organs from PWT/TWT mice is in line with findings by others. The relatively high expression ratio of PAI-1 in heart and lungs compared with other tissues (Fig. 6, A and B), as also demonstrated by Oishi et al. (35), was independent of sex and probably reflects the physiological importance of tight regulation of the fibrinolytic system in organs with high endothelial stress. However, the sex-dependent difference in PAI-1 gene transcript upregulation in PWT/TKO male mice seems to have only local biological implications, since we were not able to detect differences in PAI-1 plasma levels among genotypes or between sexes. In addition, it seems likely that the adrenal gland, at least in male mice, is not merely a regulator of PAI-1 expression in heart and lungs (35) but also is self-implicated in pronounced PAI-1 expression probably regulated by auto- and paracrine effects of glucocorticoid hormones. Equally, TIMP-1 expression was relatively high in adrenal glands from male and female mice (Fig. 6, C and D). Collectively, these observations indicate a direct physiological link between ACTH and glucocorticoid hormones and the expression of PAI-1 and TIMP-1 in the adrenal glands during steady state. However, in response to different pathological conditions such as metabolic disorders and inflammation, ACTH and glucocorticoid hormones may also be involved in regulation of PAI-1 and TIMP-1 transcription in more distant organs (14, 35, 37). Furthermore, the high levels of PAI-1 expression in uterus and ovaries and of TIMP-1 in ovaries magnify the importance of PAI-1 and TIMP-1 in ovulation, endometrial remodeling, and implantation (24, 37).
In summary, we have presented the results of generating a combined PAI-1 and TIMP-1 gene-deficient mouse model with particular emphasis on phenotypical characteristics, blood leukocyte counts, histology, and gene expression studies of PAI-1 and TIMP-1 in various organs. The expression of PAI-1 seemed dependent on TIMP-1 in male mice only. This explains, at least in part, the sex-dependent distortion in segregation of offspring and differences in body weight and blood leukocyte differentials among genotypes in male mice. In conclusion, our data highlight the complexity of PAI-1 and TIMP-1, which seem to possess sex-dependent regulatory properties on various physiological parameters, ultimately stressing the potential implications in pathological conditions.
We are indebted to Peter Carmeliet and Poul Soloway for handing over the PAI-1 and TIMP-1 single-gene knockout mice, Bent Aasted at the Section of Immunology for help with flow cytometry, Natascha Errebo at the Laboratory of Clinical Pathology for help with blood leukocyte differentials, Annette Bartels, Hanne Hornemann Møller, Betina Andersen, and Lisbeth Kiørboe at the Laboratory of Pathology for technical assistance with preparation of tissue sections, and Vibeke Jensen, Lise Larsen, and Christine Skydsgaard Nielsen for technical assistance with genotyping.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society