Targeted Inactivation of the Mouse alpha(2)-Macroglobulin Gene

doi:10.1074/jbc.270.34.19778

HOME

HELP

FEEDBACK

SUBSCRIPTIONS

Volume 270, Number 34, Issue of August 25, pp. 19778-19785, 1995
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.

Targeted Inactivation of the Mouse -Macroglobulin Gene (*)

(Received for publication, April 5, 1995)

Lieve Umans , Lutgarde Serneels , Lut Overbergh , Kristin Lorent , Fred Van Leuven (§), , Herman Van den Berghe

From the Experimental Genetics Group, Department of Human Genetics, K. U. Leuven, Campus Gasthuisberg, B-3000 Leuven, Belgium

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

FOOTNOTES

ACKNOWLEDGEMENTS

REFERENCES

ABSTRACT

The mouse alpha (2) -macroglobulin gene was inactivated in embryonic stem cells by homologous recombination. Liver alpha (2) -macroglobulin mRNA and plasma protein was absent in homozygotes and reduced to 50% in heterozygotes. alpha (2) -Macroglobulin-deficient mice were viable and produced normally sized litters with normal sex ratio over 3 generations. Characterization of adult homozygotes included diets with different fat content, treatments with endotoxin, bleomycin, carbon tetrachloride, and ethionine to test for immune system, lung, liver, and pancreas toxicity, respectively. Knock-out mice were more resistant to endotoxin but more sensitive to a choline-free diet supplemented with ethionine. Regulation of murinoglobulin mRNA expression during pregnancy was analyzed as a possible back-up mechanism for the deficiency in alpha (2) -macroglobulin. In addition, expression of mRNA was studied, coding for alpha (2) -macroglobulin receptor/lipoprotein receptor-related protein, low density lipoprotein receptor, and very low density lipoprotein receptor and for some common ligands, i.e. apolipoprotein E, lipoprotein lipase, and the 44-kDa heparin binding protein. Their differential regulation in the knock-out mice relative to C57Bl mice was evident and is discussed. The impressive 15-fold increase in maternal liver murinoglobulin mRNA at partum in the knock-out mice indicated increased consumption, compared to only 4-fold in normal mice. Thus, murinoglobulin appears as the major proteinase inhibitor around partum, obviously solicited to a much greater extend in alpha (2) -macroglobulin-deficient mice.

INTRODUCTION

Mouse alpha (2) -macroglobulin (MAM) ()is a typical member of the proteinase inhibitors of the alpha (2) -macroglobulin (A2M) family, capable of inhibiting proteinases from all classes by a steric trapping mechanism(1, 2, 3) . Upon cleavage of a peptide bond in the bait region by the proteinase, A2M traps the proteinase by a major conformational change of the tetrameric A2M structure. Recognition sites for the alpha (2) -macroglobulin receptor thereby become exposed, allowing for the specific elimination of A2M-proteinase complexes(2) .

The precise role of A2M in vivo is far from clear. In contrast to some species, human A2M is not an acute phase protein, as changes in plasma levels are moderate and never diagnostic for any disease(4) . Decreased A2M concentrations, resulting probably from enhanced consumption and clearance of A2M-proteinase complexes, might occur in states associated with proteolytic problems, for example, pancreatitis(5) . On the other hand, the interaction of A2M with a wide range of cytokines and growth factors is not understood, although it might explain some of the immunochemical and growth promoting properties ascribed to A2M(6, 7) .

In humans, no individual with a complete A2M deficiency has ever been reported. This could mean that such a deficiency is either phenotypically silent or prohibitive for full-term development and lethal in utero. Evidence for either possibility can be obtained in an experimental, transgenic model. To define the function of the A2M system including the murinoglobulins and their receptor experimentally in vivo, we have embarked on characterizing this in the mouse in molecular detail. Previously, we have reported the cloning of the cDNA and the genes coding for mouse A2M(8, 9) , of three murinoglobulins, the single-chain proteinase inhibitors of the A2M type (10, 11) and the mouse A2MR/LRP cDNA (12) and its coding gene(13) .

The gene coding for MAM, cloned from the 129/J mouse strain was used in a construct to target the MAM gene successfully in ES cells by homologous recombination(8) . We now report that the disrupted MAM gene was introduced into the germline and resulted in heterozygous and homozygous MAM-deficient offspring. The mice are viable, produced normally sized litters and showed no phenotypic abnormalities at the age of well over 1 year. Experiments with the MAM-/- mice demonstrated several potentially important differences in susceptibility to toxic agents known to cause malfunction of specified organs or systems. Tests of endotoxin(14) , carbon tetrachloride(15) , and bleomycin (16) are presented as well as the effect of diets with different fat contents (17) and a choline-deficient diet containing ethionine instead of methionine known to induce acute pancreatitis (18) .

MATERIALS AND METHODS

Targeting Vector and ES Cell Culture

The cloning of the MAM gene, the construct and the ES cell targeting was described(8) . In brief, MAM genomic sequences were isolated from a 129/J library (19) and a 7.5-kb EcoRI/SphI fragment encoding exons 16-19 was subcloned(8) . The hygromycin B phosphotransferase gene with the phosphoglycerate kinase promoter (20) was ligated into XhoI/ClaI sites of intron 17 replacing 0.7 kb of intronic sequences of the MAM gene. The construct was linearized at a unique BamHI site in exon 18 and electroporated into ES cell line E14(8, 21) . Twelve days after electroporation and selection in hygromycin containing medium, 240 colonies were picked, expanded, frozen, and analyzed by Southern blotting(8) . In the first screening round, StuI-digested DNA from 198 ES cell clones was blotted and hybridized with a genomic DNA fragment located at the 3` flank of the targeting vector (probe R), detecting 11.5- and 3.8-kb StuI fragments in wild type and targeted alleles, respectively. From 7 selected ES cell lines, confirmatory screening with 3 different probes resulted in 5 ES cell lines with a single copy of the MAM gene targeted as wanted.

Blastocyst Injection and Generation of Germ-line Chimeras

The 5 different ES cell lines were injected into C57Bl blastocysts (4.5 days) and these were subsequently implanted into pseudopregnant F1 (C57Bl ♀

CBA/J ♂) foster mothers. The resulting chimeric mice were mated to C57Bl mice and from one line, offspring with germline transmission, scored by coat color (agouti pups), were analyzed. The transmission of the targeted MAM gene was demonstrated by Southern blotting of tail DNA and heterozygotes were mated.

Northern Blot Analysis

Total liver RNA and poly(A) mRNA was isolated from mouse liver, uterus, and placenta, separated by electrophoresis, and blotted as described(22) . Filters were prehybridized for 6 h at 42 °C, in 5

SSPE, 5

Denhardt's solution (100 µg/ml polyvinylpyrrolidone, 100 µg/ml bovine serum albumin, 100 µg/ml Ficoll 400), 0.5% SDS, 50% deionized formamide, 100 µg/ml denatured sperm DNA, 50 µg/ml heparin and hybridized in the same solution supplemented with 10% dextran sulfate at 42 °C overnight with addition of 1 to 2 million cpm/ml of the indicated radiolabeled probe. The blots were washed in 0.3

SSPE, 0.5% SDS at 60 °C for 1 h. Stripping of the filters in 0.5% SDS, 40 mM Tris, pH 7.8, for 15 min at 80 °C, was performed between hybridization for MAM and MUG mRNA, since both characteristic transcripts are of the same size(9, 10) . Autoradiography was done by exposure to Hyperfilm MP (Amersham, United Kingdom) using intensifying screens, at -70 °C from 2 h up to 7 days as indicated.

cDNA Probes

cDNA probes for MUG, MAM, HBP-44, and VLDLR were generated by polymerase chain reaction amplification: MAM, position 1188-1758 (mouse cDNA)(9) ; MUG, position 1777-2262 (MUG1 cDNA)(10) ; HBP-44, position 8-1078 (mouse cDNA)(23) ; VLDLR, position 1086-2181 (mouse cDNA)(24) . The polymerase chain reaction fragments were isolated in low melting agarose or purified on silica.

The cDNA clone mLDLRc90 for mouse LDLR was kindly provided by M. Hofker (Sylvius Laboratory, Leiden, The Netherlands). The 700-base pair EcoRI insert was used as a probe(19) . The mouse ApoE cDNA clone, pmEUC18 was obtained from S. Tajima (Osaka, Japan) (25) . To detect A2MR/LRP mRNA, a 1.4-kb EcoRI cDNA fragment was used as a probe corresponding to positions 776-2197 of the mouse cDNA(12) . The cDNA probe used to detect LPL mRNA was a 1-kb PstI restriction fragment isolated from the coding region of rat LPL (provided by J. Auwerx, Lille, France). A 2.0-kb human beta -actin cDNA probe (Clontech) was used as a control in Northern blotting to allow for normalization of mRNA loading. Labeling of the probes was performed by hexanucleotide mediated incorporation of [P]dCTP, as described before(10, 22) .

Rocket Immunoelectrophoresis

Mouse plasma from wild type, heterozygous, and homozygous MAM-deficient mice was collected from anesthetized animals by cardiac puncture with heparinized needles. Rocket immunoelectrophoresis was carried out in 0.8% agarose, buffered in 89 mM Tris, 89 mM boric acid, 2 mM EDTA, pH 8.4, containing 2.5% (v/v) rabbit antisera directed against MAM or MUG(3, 9, 10) . After electrophoresis at 2.5 V/cm overnight, the gels were soaked, dried, stained with Coomassie Brilliant Blue and destained by diffusion.

Induction of Teratocarcinomas

Eight male and 8 female MAM-/- mice, aged 8 to 9 weeks, were injected with embryonic stem cells of the E14 cell line to test for the development of teratocarcinomas. About 4 million ES cells were resuspended in 100 µl of Glasgow ME medium and injected subcutaneously. After 3 weeks the teratocarcinomas were isolated and examined histologically.

Injection of Bleomycin

Male and female MAM-/- and C57Bl control mice were injected intraperitoneally twice a week for 5 weeks with a dose of 0.2 mg of bleomycin per mouse per injection. The lungs of each mouse were isolated, fixed in 4% paraformaldehyde, dehydrated through alcohol series, and embedded for sectioning. Five-µm sections were stained with hematoxylin and eosin.

Acute Toxicity of Carbon Tetrachloride

Acute liver injury was experimentally induced by administration of carbon tetrachloride mixed with paraffin oil in concentrations of 2 and 10%. MAM-/- and C57Bl mice were injected intraperitoneally and after 24 h the liver of each mouse was isolated, fixed, and sectioned for histology.

Injection of Endotoxin

Eight-week-old mice were anaesthetized with ether and injected in the hind footpads with 50 µg of endotoxin (Lipopolysaccharide W, Escherichia coli 0111:B4, LD

28.3 mg/kg; Difco Laboratories) in a volume of 25 µl of saline. After 4 days mice were sacrificed and the footpad tissue was isolated and fixed for 2 h in 4% paraformaldehyde, dehydrated, embedded, and sections were stained with hematoxylin and eosin. Other mice were injected intraperitoneally with 2 different doses of endotoxin corresponding to 1 and 5 times the LD

dose as determined by the manufacturer.

High and Low Fat Diets

MAM-/- mice of 8 weeks old were fed three different diets during a period of 9 weeks: the regular chow mouse diet (4352 Muracon G, Trouw, Ghent, Belgium), a diet high in fat and cholesterol (purified diet NK4021.01; Hope Farms, Woerden, The Netherlands), and a fat free diet (purified diet 4141.00; Hope Farms). In parallel, a control group of C57Bl mice received the high fat/cholesterol diet. The mice had free access to water and food. Mice were weighed twice a week over the 9-week period of the experiment. At this time, blood was collected by cardiac puncture in heparinized needles for determination of the level of MUG protein by rocket immunoelectrophoresis. The liver of each animal in the test was isolated either at the moment of spontaneous death or at the end of the experiment, its weight was determined before fixation in 4% paraformaldehyde. In each animal the gall bladder was isolated separately and examined to determine its size, the amount and color of gall fluid, and the presence of gallstones and their gross morphological appearance.

Induction of Pancreatitis

Female mice of 6 weeks to 4 months old were fed a diet deficient in choline and methionine (Harlan Teklad TD 90262) supplemented with 0.5% DL-ethionine (Sigma). The mice had free access to water and food. After 8 days on this diet the pancreas of all surviving mice were histologically examined.

RESULTS

Targeting of the MAM Gene in ES Cells

The targeting construct contained a genomic 7.5-kb SstI/EcoRI fragment comprising exons 16-19 of the MAM gene (8) with a positive selection marker gene embedded in intron 17 (Fig. 1). Transfection into E14 embryonic stem cells and selection by hygromycin yielded about 850 colonies, 198 of which were analyzed by Southern blotting. The 7 positive clones were reanalyzed by hybridization with three different DNA probes (denominated R, L, and HYG in Fig. 1). The number of ES cell lines having one allele of the MAM gene targeted as wanted, resulting in the expected restriction patterns, was thereby reduced to 5 (Fig. 1). Hence, the overall frequency of recombination is about 2.5% (5/198).

Figure 1: Recombinant DNA construct used to target the MAM gene in ES cells. Top, the hatched box (PGK-HYG) represents the hygromycin marker gene which was embedded in intron 17. The construct was linearized at the unique BamHI (B) site in exon 18. StuI restriction sites are marked (S), pUC denotes the bacterial cloning vector. Middle, partial structure of the wild type MAM gene with the region used in the construct represented as a box. Restriction sites recognized by StuI (S) are 11.5 kb apart. R and L represent genomic DNA probes, located at the right and left relative to and externally of the genomic region used in the construct. Bottom, predicted structure of the targeted MAM gene. Insertion of the construct by homologous recombination will result in the loss of the wild type 11.5-kb StuI fragment and the generation of StuI fragments of 10.5 (L probe), 3.8 kb (R probe) and 8.5 kb (HYG probe). Inset, Southern blotting of one of the selected ES cell lines, digested with StuI and hybridized with the three different probes (R, L, and HYG).

Germline Transmission of the Targeted MAM Gene

The 5 recombinant ES cell lines were injected into C57Bl blastocysts and all resulted in coat color chimeric mice. However, only one cell line also resulted in germline transmission. Of 11 male chimeric mice with variable (30% to 80%) chimerism, 10 transmitted the targeted MAM gene while no germline transmission was obtained with 10 females. A total of 157 agouti pups analyzed by Southern blotting of tail tip DNA with probe R, yielded 47 female and 22 male heterozygous (MAM+/-) mice. Heterozygotes were mated and their offspring was genotyped 3 weeks after birth.

Viability and Fertility of MAM-/- Mice

In total, 207 pups of heterozygous MAM+/- couples were analyzed by Southern blotting of tail tip DNA, digested with StuI (Fig. 2). This identified 48 MAM+/+ (23%), 116 MAM+/- (56%), and 43 MAM-/- mice (21%). This is a rather normal Mendelian pattern. Subsequent mating of 21 pairs of MAM+/- mice demonstrated that 20 couples produced litters sized between 3 and 15 pups, with an average of 8.4 pups per litter. The sex ratio was 49/51% male/female.

Figure 2: Screening by Southern blotting of offspring of heterozygous matings. Genomic DNA, isolated from the tail tip, was digested with StuI and hybridized with the 0.5-kb MluI-EcoRI fragment (R probe). Among 18 pups, 4 were homozygous, indicated by the diagnostic 3.8-kb fragment, 9 were heterozygous (3.8- and 11.5-kb fragment), and 5 were wild type (11.5-kb fragment only).

At this moment 3 generations of MAM-/- mice have been obtained with the oldest homozygous MAM-deficient mice now 15 months old. They are kept in an open animal house on standard mouse chow, available ad libitum, and appear healthy without overt health problems.

Liver mRNA and Plasma MAM and MUG Protein

Homozygous MAM-deficient mice were analyzed at the mRNA level by Northern blotting and plasma proteins were measured by rocket immunoelectrophoresis. Poly(A) mRNA was isolated from liver of wild type, heterozygous, and homozygous MAM-deficient mice. Hybridization with the MAM cDNA probe revealed the specific 5-kb MAM mRNA in MAM+/+ mice. In MAM+/- mice, liver mRNA was about 50% lower, while it was completely absent in the liver of MAM-/- mice (Fig. 3). Subsequent hybridization with the MUG cDNA probe revealed the typical 5-kb mRNA in all mice. The mRNA level of MUG was lower in female mice, depending also on estrous cycle (Fig. 3), in accordance with previous results (26) . (

)

Figure 3: Northern blotting of poly(A) mRNA isolated from the liver of wild type C57Bl, heterozygous, and homozygous MAM-deficient mice. The blot was hybridized with a cDNA probe specific for MAM (9, 22) and subsequently with a cDNA probe specific for MUG(10, 22, 27) . Exposure was for 6 and 4 h at -70 °C, respectively, and the size of both mRNA species identified was 5 kb as indicated on the right. Overexposure for 5 days after hybridization with the MAM cDNA probe did not reveal the 5-kb band in the MAM-/- mice. Note the very low expression levels of MUG in certain females, depending on their hormonal status as reported recently (see text and (22) and (27) for details).

Rocket immunoelectrophoresis of plasma of the same mice and of many other MAM-/- mice confirmed the results obtained by Northern blotting. No MAM protein was present in the plasma of MAM-/- mice, while plasma levels were between 45 and 60% in MAM+/- mice, relative to wild type mice (Fig. 4A).

Figure 4: Rocket immunoelectrophoresis of mouse plasma. Plasma samples of wild type C57Bl (+/+), heterozygous (+/-), and homozygous (-/-) MAM-deficient female mice were run into antisera directed against MAM (panel A) and MUG (panel B). Plasma (2.5 µl) was applied undiluted in 3-mm wells punched in 0.8% agarose gels on glass plates. Some cross-reaction is evident with both antisera with unidentified plasma proteins. The most dense rocket in panel B corresponds to MUG as determined in separate experiments by addition of purified MUG1 protein, isolated from mouse plasma and authenticated by N-terminal amino acid sequencing (9, 10) .

The plasma levels of murinoglobulins in normal mice are more variable than those of MAM and, in female mice, are also dependent on the hormonal status of the animal. Nevertheless, comparing the MUG plasma levels of adult nonpregnant mice demonstrated no systematic difference with C57Bl control mice, in all MAM-/- and MAM+/- mice analyzed to date (Fig. 4B).

Extracted mRNA of intact MAM-/- embryos of 15 days post-coitus was analyzed by Northern blotting to confirm the absence of MAM mRNA. Further hybridization failed to demonstrate any expression of murinoglobulin at this embryological age, identical to our previous observations in normal mice(22) . Thus, murinoglobulins, which are normally expressed only in the second week postnatally, do not substitute for embryonic expression of MAM in the MAM-/- mice.

Expression of MUG in Pregnant Female Mice

In a recent study we have analyzed the negative response of MUG in maternal plasma during pregnancy in relation to MAM and to A2MR/LRP expression levels in fetal and maternal tissues.

In the present experiments, the plasma levels of MUG were measured in control C57Bl and in MAM-/- pregnant females at days 12 and 19 postcoitus and at day 1 postpartum. From each animal mRNA was isolated from uterus, liver, and placenta (days 12 and 19 only) and analyzed in Northern blotting.

The maternal hepatic MUG mRNA levels in control C57Bl mice were at parturition about 4-fold higher relative to the levels at day 12 postcoitus (Fig. 5). This increase is much more pronounced in the MAM-/- mice, since the maternal MUG mRNA levels are more than 14-fold higher at partum relative to day 12 postcoitus (Fig. 5). This impressive rise in maternal liver MUG mRNA levels is only partially reflected in the circulating MUG protein levels: at birth very similar concentrations of about 0.4-0.45 mg/ml were measured in the C57Bl and in MAM-/- mice (Fig. 5), pointing to extensive consumption of maternal MUG around birth in C57Bl mice, which is clearly much more evident in the MAM-/- mice.

Figure 5: Maternal plasma MUG protein levels and hepatic MUG mRNA levels in pregnant C57Bl and MAM-/- mice. Panel A, histogram representation of quantitative densitometric scanning of Northern blots of liver mRNA probed for MUG and normalized to the beta -actin signal. Liver mRNA was extracted from pregnant mice at 12 and 19 days postcoitus (pc) and at 1 day postpartum (pp). Each time point represents the calculated mean of determinations of three mice at each time point. The values are given in arbitrary units and are normalized to the signal obtained by hybridization with an actin probe of the same Northern blot (see ``Materials and Methods''). Panel B, histogram representation of quantitative rocket immunoelectrophoresis of MUG in maternal plasma of pregnant mice. Plasma was isolated from pregnant females at 12 and 19 days and from females 1 day postpartum. Each time point represents the calculated mean of measurement on three separate mice.

Expression of A2MR/LRP, LDLR, VLDLR, HBP-44, ApoE, and LPL during Pregnancy

The relative levels of mRNA coding for A2MR/LRP, LDLR, VLDLR, ApoE, and LPL were determined in liver, uterus, and placenta of pregnant MAM-/- mice. A2MR/LRP and the structurally and functionally related receptors LDLR and VLDLR, are partners in a complicated network of interactions in controlling proteolysis and lipid metabolism. They are subject to opposite regulation during pregnancy.

Expression of Lipoprotein Receptors

In the liver, MAM-/- mice and C57Bl mice express the 15-kb A2MR/LRP mRNA(12, 22)

and the 5-kb LDLR mRNA (28) to comparable levels, while no mRNA transcript coding for VLDR was detected (Fig. 6) as expected(24, 29) . During pregnancy the differential regulation of hepatic expression of the lipoprotein receptors is evident: in C57Bl mice both A2MR/LRP and LDLR mRNA levels are down-regulated, respectively, more than 3- and 2-fold, while in MAM-/- mice the levels are maintained fairly constant during the course of pregnancy (Fig. 6).

Figure 6: Histograms of levels of mRNA coding for A2MR/LRP, LDLR, VLDLR, HBP-44, ApoE, and LPL in liver, placenta, and uterus in MAM-/- and C57Bl mice. Extraction of mRNA from liver (A and D), placenta (B and E), and uterus (C and F) of MAM-/- (A, B, and C) and C57Bl (D, E, and F) mice was consecutively hybridized with cDNA probes specific for A2MR/LRP, LDLR, VLDLR, HBP44, ApoE, LPL, and beta -actin. The histograms represent, in arbitrary units, the quantitation by densitometric scanning of the Northern blots, normalized to the beta -actin signal on each blot. Each time point, i.e. 12 and 19 days postcoitus (pc) during pregnancy and 1 day postpartum (pp), as indicated, represents the calculated mean of determinations of three individual mice at each time point that were analyzed on different filters. Abbreviations of each of the mRNA species probed for is indicated, eventually with each mRNA species detected and identified by its size.

In placenta, A2MR/LRP and LDLR are regulated oppositely in C57Bl mice (Fig. 6) confirming that the placenta shifts toward A2MR/LRP mediated uptake of ApoE-VLDL lipoproteins with progressing pregnancy. This is much less marked in the MAM-/- mice, which in conjunction with the different hepatic expression pattern of these receptors, indicated that lipoprotein metabolism is regulated differently in the MAM-/- mice. The possible involvement of or even compensation by another receptor of this family, i.e. the VLDL-receptor (VLDLR), was supported by the results: from day 12 to day 19 postcoitus placental VLDLR mRNA levels increased in C57Bl mice but not or hardly in MAM-/- mice (Fig. 6).

In the uterus of both C57Bl and MAM-/- mice, the constant expression levels of A2MR/LRP mRNA contrasted sharply with the 8-fold down-regulation of LDLR mRNA at parturition (Fig. 6). In C57Bl mice this is eventually compensated for by increased expression, at least of the 3.9 kb, the smaller of the two transcripts coding for VLDLR, an increase which is not observed in either mRNA transcript of the VLDLR in the uterus of MAM-/- mice (Fig. 6).

Lipoprotein Receptor Ligands

The two transcripts coding for HBP-44 or the 39-kDa receptor associated protein are expressed at a practically constant level in liver and uterus of MAM-/- mice throughout pregnancy (Fig. 6). In placenta, levels of both transcripts were increased about 8-10-fold at 19 days postcoitus relative to 12 days postcoitus and this was not different from control mice (Fig. 6).

The 1.2-kb ApoE transcript (25) is abundantly detected in both control and in MAM-/- mice, in liver, placenta, and uterus. With progressing pregnancy, the uterus and especially the placenta are characterized by a vast increase in ApoE mRNA levels, more than 30-fold in MAM-/- mice, while the liver ApoE mRNA levels decreased (Fig. 6). This is in liver again an opposite regulation as seen in the C57Bl strain, in which liver ApoE mRNA levels increased from mid pregnancy toward parturition (Fig. 6).

The 4-kb LPL mRNA (30, 31) was up-regulated about 10-fold in placenta and uterus as pregnancy progresses from day 12 to 19 postcoitus in MAM-/- mice (Fig. 6) very similar to observations in C57Bl mice. The regulation of expression of liver LPL mRNA was again the exception, since in MAM-/- mice it increased very importantly by about 13-fold, while it was somewhat decreased in C57Bl mice (Fig. 6).