Lipids in Health and Disease Human Apob Contributes to Increased Serum Total Apo(a) Level in Lpa Transgenic Mice

Background: The Lp(a) lipoprotein (Lp(a)) consists of the polymorphic glycoprotein apolipoprotein(a) (apo(a)), which is attached by a disulfide bond to apolipoprotein B (apoB). Apo(a), which has high homology with plasminogen, is present only in primates and hedgehogs. However, transgenic mice and rabbits with high serum apo(a) levels exist. Liver is the main site for apo(a) synthesis, but the site of removal is uncertain. To examine differences between transgenic mice expressing the LPA gene and mice capable of forming Lp(a) particles, LPA-YAC transgenic mice and hAPOB transgenic mice were crossed and their offspring examined.


Background
Association between Lp(a) lipoprotein (Lp(a) [1], and coronary heart disease was first demonstrated in 1974 [2]. Afterwards, numerous studies have confirmed that high plasma level of Lp(a) is an independent risk factor for atherosclerotic and thrombotic disease [3]. The intact Lp(a) molecule consists of the highly polymorphic glycoprotein apolipoprotein(a) (apo(a)), which is attached through a disulfide bond to apolipoprotein B-100 (apoB-100) of low density lipoprotein (LDL).
Gene encoding for apo(a) (LPA), which expression is mainly confined to the liver, is tightly linked to plasminogen (PLG) gene on chromosome 6 and is homologous to PLG [4,5]. This sequence similarity indicates that the human LPA gene evolved from a duplicated PLG gene during primate evolution [6,7]. Interestingly, in addition to humans, Old World monkeys, great apes and hedgehogs are the only species, which express LPA naturally [8][9][10].
Varying numbers of KIV structures account for the size variation of apo(a) whose size is inversely related to plasma Lp(a) level [11]. Furthermore, recent studies in transgenic animal models have shown that 50% of human Lp(a) is catabolized by the liver [12,13], but the mechanism for uptake is unclear. Study of WHHL transgenic rabbits expressing human apo(a) showed that defect LDL receptor (LDLR) lead to marked accumulation of plasma Lp(a) [14]. However, despite structural similarity of Lp(a) to LDL, the LDLR does not seem to be involved significantly in Lp(a) removal in humans [15][16][17].
Studies in female mice transgenic for LPA-yeast artificial chromosome (LPA-YAC), which contains a 40 kilobase (kb) LPA-PLG intergenic region [18], have shown that feeding of a high-fat diet reduces serum apo(a) and hepatic apo(a) mRNA level [19,20]. In contrast, serum apo(a) levels in mice transgenic for human apo(a) cDNA are stable despite high-fat feeding [21], which suggests that the effect of fat-rich diet on apo(a) levels in LPA-YAC transgenic mice is most likely exerted through regulatory elements located within the LPA-PLG intergenic region.
Recent studies on PPARα suggest that PPARα is of major importance for regulation of serum levels of apoB and production of apoB-containing lipoproteins [22,23], but the possible role of PPARα in determining Lp(a) serum level is unclear. The DNase I hypersensitive site (DH) II enhancer part of the LPA-PLG intergenic region contains the gene for peroxisome proliferators-activated receptor (PPAR) α binding site [24,25]. DH II increases the activity of the LPA promoter by over 7-fold in reporter-gene assays in HepG2 cells in vitro [26]. This together with recent cotransfection assays, which showed that PPARα increased the effect of the DHII enhancer on LPA transcriptional activity may indicate that PPARα plays a role also in determining plasma Lp(a) level [26].
In order to examine differences between transgenic mice only expressing the LPA gene and transgenic mice capable of forming Lp(a) particles, LPA-YAC transgenic and mice transgenic for the human APOB (hAPOB) gene were crossed and their offspring examined. The study revealed that serum total apo(a) levels in LPA-YAC trangenic mice also expressing the hAPOB gene were higher than in LPA-YAC transgenic animals lacking hAPOB. However, hepatic apo(a) mRNA level in LPA-YAC transgenic mice was higher than in LPA-YAC/hAPOB transgenic mice. Furthermore, we showed that feeding of a high-cholesterol/highfat (HCHF) diet to male LPA-YAC transgenic mice, which on a standard mouse diet exhibit much lower serum total apo(a) level than female mice, reduced serum total apo(a) and hepatic apo(a) mRNA level in a similar manner as previously shown in female mice [19,20]. The lack of significant variation in hepatic PPARα mRNA levels between wild type (WT) and LPA-YAC transgenic mice or between mice fed either a semi-synthetic mouse diet (AIN-76) or the HCHF diet, may indicate that neither LPA transgenicity nor high-fat feeding have a significant impact on expression of PPARα.

Serum apo(a) level
On a regular RMI(E)SQC mouse diet serum total apo(a) level was significantly higher in LPA-YAC/hAPOB transgenic mice than in LPA-YAC transgenic mice (P < 0.05, figure 1) and serum total apo(a) levels were higher in female transgenic mice than in respective male transgenic mice (P < 0.005, figure 1).
Male LPA-YAC transgenic mice (n= 4) had a mean serum total apo(a) level of 7.6 ± 2.5 mg/dL at the start and of 6.4 ± 3.9 mg/dL at the end of feeding the AIN-76 diet for 7 weeks. Corresponding levels in LPA-YAC/hAPOB transgenic mice (n= 6) were 9.2 ± 1.0 mg/dL and 9.5 ± 2.2 mg/ dL, respectively. Serum total apo(a) level in male LPA-YAC/hAPOB transgenic mice was significantly reduced (P < 0.001, figure 2) after two weeks of HCHF feeding and the level remained low during 7 weeks of feeding. Feeding of the HCHF diet to male LPA-YAC transgenic mice resulted in reduced mean serum total apo(a) level from 2.9 ± 1.3 mg/dL to below background level, but because of the small number the difference did not reach statistical significance (P= 0.08).

Serum total cholesterol and triglycerides in male mice
LPA-YAC transgenic mice had similar serum total cholesterol (TC) and serum triglycerides (TG) levels as WT mice, independently of the diets (table 1). LPA-YAC/hAPOB transgenic mice fed the AIN-76 diet had higher serum TG level than either WT or LPA-YAC transgenic mice (P < 0.05). The HCHF fed LPA-YAC/hAPOB transgenic mice had significantly higher serum TC level than the AIN-76 fed mice (P < 0.005). In comparison, feeding of the HCHF diet to WT or LPA-YAC transgenic mice had no significant effect on serum TC levels compared to the AIN-76 fed mice. Serum TG levels were lower in the HCHF fed LPA-YAC and LPA-YAC/hAPOB transgenic mice compared to the AIN-76 fed transgenic mice (P < 0.05 and P < 0.005 respectively, for the two categories of mice). The comparable difference on serum TG level between the AIN-76 and the HCHF fed WT did not reach statistical significance.

Apo(a), PPARα and LDLR mRNA levels in male mice
Apo(a) mRNA was mainly detected in liver of LPA-YAC transgenic mice and to a very low degree in kidneys in 3 out of 4 mice (table 2). Apo(a) mRNA was not detected in spleen, heart, brain and intestine.
On the AIN-76 diet hepatic apo(a) mRNA levels were lower in LPA-YAC/hAPOB than in LPA-YAC transgenic mice (P< 0.05, figure 3). Furthermore, feeding of the HCHF diet to transgenic mice reduced hepatic apo(a) mRNA level significantly (P<0.05) in both strains.
Hepatic PPARα mRNA levels did not vary significantly between WT and transgenic mice, or between the AIN-76 Serum total apo(a) level in male and female LPA-YAC and LPA-YAC/hAPOB transgenic mice on a regular RMI(E)SQC mouse diet Figure 1 Serum total apo(a) level in male and female LPA-YAC and LPA-YAC/hAPOB transgenic mice on a regular RMI(E)SQC mouse diet. Values shown are mean ± SEM (mg/dL). Statistically significant difference between LPA-YAC and LPA-YAC/hAPOB transgenic mice is indicated by # (P < 0.05) or by ## (P < 0.01). Statistically significant difference between respective male and female mice is indicated by * (P < 0.005). Hepatic LDLR mRNA level did not vary significantly between the AIN-76 fed WT, LPA-YAC and LPA-YAC/ hAPOB transgenic mice, or between the HCHF fed and the AIN-76 fed LPA-YAC transgenic mice (figure 5). HCHF feeding reduced LDLR mRNA levels in WT and LPA-YAC/ hAPOB transgenic mice, compared to the levels in the AIN-76 fed mice (P < 0.05).

Discussion
Observation of higher serum levels of total apo(a) (figure 1 & 2) in LPA-YAC/hAPOB transgenic mice than in LPA-YAC transgenic mice are in agreement with the results of Callow et al [27]. Our finding of lower hepatic apo(a) mRNA levels (figure 3) in LPA-YAC/hAPOB transgenic mice than in mice having the LPA gene only, suggest that the elevated serum Lp(a) levels in LPA-YAC/hAPOB transgenic mice are not due to increased hepatic apo(a) mRNA transcription rate. Furthermore, analysis of LPA expression in different tissues (table 2) indicates that no other organs than the liver, can produce significant amounts of apo(a).
LDL can be removed from the circulation by the LDL receptor (LDLR), which binds to the apoB-100 part of LDL [28]. The efficient assembly of Lp(a) in mice transgenic for Effect of 7 weeks HCHF feeding on serum total apo(a) level in male LPA-YAC and LPA-YAC/hAPOB transgenic mice Figure 2 Effect of 7 weeks HCHF feeding on serum total apo(a) level in male LPA-YAC and LPA-YAC/hAPOB transgenic mice. Mice were fed the HCHF diet for 7 weeks. Mice were bled at the start, after 2 weeks of feeding and when the animals were sacrificed. Values shown are mean ± SEM (mg/dL) of serum total apo(a) level. According to manufacturer, values <1.2 mg/ dL reflect background noise. The * indicates statistically significant difference (P < 0.001) in comparison to the start level. Statistically significant difference (P < 0.05) between LPA-YAC and LPA-YAC/hAPOB transgenic mice is indicated by #. both LPA and hAPOB and higher efficiency of Lp(a) assembly with the human apoB-100 than with mouse apoB-100 [29,30] may suggest that measured total apo(a) level in LPA-YAC/hAPOB transgenic mice consist mainly of human Lp(a) particles.
Because mouse LDLR has low capacity to bind human apoB-100 containing particles [31], elevated serum total apo(a) level may theoretically be secondary effect of reduced clearance rate of human apoB containing particles via the LDLR dependent pathway in LPA-YAC/hAPOB transgenic mice. Support for this comes from study, which showed that overexpression of LDLR in mice leads to accelerated metabolism of injected human Lp(a) particles [32]. Furthermore, it is notable that serum Lp(a) levels in LDLR -/-/LPA transgenic mice are similar whether or not the human APOB gene is also present [33]. Since triglyceride-rich apoB particles have increased affinity to apo(a) in man [34,35], it seems likely that the human apoB part of the Lp(a) molecule in mouse serum may associate with TG containing particles, which may explain why hAPOB transgenicity also contributes to increased serum TG levels.
Suggestion for a role of LDLR in Lp(a) clearance comes also from studies in rabbits, where apoB, in contrast to mice, associates with human apo(a) forming Lp(a) particles [36]. A defect in LDLR in LPA transgenic WHHL rabbits is accompanied by accumulation of Lp(a) in plasma [14]. Since mice or rabbits do not normally have apo(a), the mechanism of Lp(a) removal in these animals may well differ from that in man, where the LDLR does not seem to affect Lp(a) metabolism [15][16][17].
The fall in serum apo(a) and hepatic apo(a) mRNA levels ( figure 2 and 3) in male transgenic mice on the HCHF diet agrees with observations by Acquati et al and Huby et al, who showed that hepatic apo(a) mRNA and serum apo(a) levels fell in female LPA transgenic mice after two weeks fat feeding [19,20]. Man, who naturally express LPA, do not have significant variation in plasma Lp(a) levels despite of changes in diet [37]. This indicates that man possesses LPA control mechanisms that are different from those in mice. This agrees with the fact that CYP7A1, which regulates the initial step in the pathway of cholesterol metabolism and bile acid synthesis, is increased during cholesterol feeding in rats and mice, but not in humans [38]. Rodents' capacity to convert cholesterol into bile acids by liver X receptor-α mediated stimulation of CYP7A1 transcription has been suggested to account for some of the differences in response to fat intake between humans and rodents [39]. Furthermore, PPARα, which binds to the DHII enhancer part of LPA-PLG intergenic region [24], and is of importance for production of apoB and apoB-containing lipoproteins [22], was recently shown to be regulated by bile acids in humans but not in rodents [40]. Since cholesterol feeding promotes bile acid synthesis in rodents [41], our results showing similar hepatic PPARα mRNA levels between the AIN-76 and HCHF fed mice may give further support to the notion that bile acids do not increase PPARα mRNA levels in mice. Further studies will be needed to elucidate whether or not the species-specific differences in cholesterol and bile acid metabolism is a key to understanding the mechanism(s), which cause the reduction on apo(a) transcription in HCHF fed transgenic mice.
The reduced hepatic LDLR mRNA level seen in HCHF fed WT mice and LPA-YAC/hAPOB transgenic mice (figure 5) is in agreement with previous results, which showed that hepatic LDLR level in mice is reduced in response to HCHF feeding [42]. This may suggest that the human apoB can interact sufficiently, although less than mouse Because of the small number of HCHF fed LPA-YAC transgenic mice in our study, the absence of an effect of feeding on LDLR mRNA level must be confirmed in further studies.
On the HCHF diet both WT mice and transgenic mice developed fatty livers (not shown). Since hepatic level either of housekeeping genes or PPARα mRNA was not reduced in response to the HCHF feeding, it seems clear that the decreased apo(a) mRNA level detected in the livers of HCHF fed transgenic mice was not due to a general decrease in liver function.

Conclusions
In conclusion, we have shown that serum total apo(a) level is higher in LPA-YAC/hAPOB transgenic mice than in mice that are transgenic for the LPA gene alone. The higher level is not caused by increased apo(a) synthesis. Lower hepatic apo(a) mRNA level in LPA-YAC/hAPOB than in LPA-YAC transgenic mice may suggest that the increase in total apo(a) level is a result of apo(a) accumulation in serum. Our results on reduced serum total apo(a) and hepatic apo(a) mRNA level in HCHF fed male LPA-YAC transgenic mice confirm earlier findings in females, and show that there are no sex difference in mechanisms for lowering apo(a) level in response to HCHF feeding.

Mice
Mice possessing a 270 kb YAC with a human genomic DNA clone containing the intact LPA gene and a 70 kb LPA-like gene with extensive 60 kb flanking regions on both sides and mice possessing the intact hAPOB gene in FVB genetic background have been described previously [18,27]. To develop mice capable of forming Lp(a) particles, LPA-YAC transgenic mice and hAPOB transgenic mice were crossed. Sentinel animals were used to run a full felasastyle health-monitoring scheme and mice were found to be healthy.
The presence of the LPA gene and the hAPOB gene was demonstrated by PCR analysis [43]. For the analysis of hAPOB transgenicity sense primer TGGAAACGGA-GAAATTATGGA and antisense primer CACTTGGCAAAT-ACAATTCCTG were used. Mice were housed in a room with 12 h light/dark cycle and 55% relative humidity at 21°C. Mice were kept on the regular RMI(E)SQC mouse diet (Special Diet Services, Witham, Essex, England) until the start of the feeding experiment.
Four 1-year old male LPA-YAC transgenic mice and two 1year old male wild type mice were sacrificed for the study of LPA expression pattern. For the study of serum total apo(a) level prior to the feeding experiment, six female and seven male LPA-YAC transgenic as well as five female and fifteen male LPA-YAC/hAPOB transgenic mice were bled.

Feeding experiment
Ten male WT, seven LPA-YAC transgenic and fourteen LPA-YAC/hAPOB transgenic mice were divided into groups and fed with a semi-synthetic mouse diet (AIN-76, ICN, Asse-Relegem, Belgium) or a high-cholesterol/highfat diet (Custom high-fat diet, ICN, Asse-Relegem, Belgium) for seven weeks (table 3). The HCHF diet contained 1.25% cholesterol, 18.4% regular butter, corresponding to approximately 9% saturated fat, and 1% corn oil, while the AIN-76 diet contained 5% corn oil. Corn oil contains approximately 1.7% saturated fat. Water and food were given ad libitum. In addition to the blood sampling before feeding, blood samples from the saphenous vein were drawn after two weeks feeding and when the animals were sacrificed after seven weeks feeding.

Analysis of serum total apo(a), total cholesterol and triglycerides level
Serum total apo(a) level was measured using a radioimmuno assay (RIA) kit (Mercodia AB, Uppsala, Sweden). This method is based on direct sandwich technique in which two monoclonal antibodies are directed against separate antigenic determinants of apo(a). Results are given in units per liter (U/L). According to the manufacturer, results must be multiplied by 0.07 to obtain values in mg/dL, and levels below 1.2 mg/dL (17 U/L) are considered as background noise. For details of test's reliability to measure serum apo(a) level see Berg et al [43].
Serum TC and TG level was measured using enzymatic, colorimetric methods (Modular Analytics, Roche Diagnostics, F-Hoffmann-La Roche Ltd, Basel, Switzerland) at Fürst AS, Oslo, Norway.

Tissue preparation for mRNA analysis
For RNA preparations, tissue samples were washed in icecold saline, submerged in RNAlater™ buffer (Ambion Diagnostics, Cambridgeshire, England) and stored at -70°C until analysis.

Quantitative PCR
Level of mRNA was determined by quantitative PCR (qPCR), which was performed using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Primers and probes (table 4) were designed using Primer Express software (Applied Biosystems, Foster City, CA, USA).
Total RNA from 100-300 mg of tissues was isolated after homogenisation in 1 ml Trizol reagent according to the manufacturer's recommendations (Invitrogen Co, Carlsbad, CA, USA). RNA was treated with Dnase I (Ambion Ltd, Cambridgeshire, England) and 5 µg of total RNA was reverse-transcribed using pd(N) 6 random primers (First strand cDNA synthesis kit, Amersham Biosciences, Little Chalfont Buckinghamshire, England).  LPA primers and probe (Eurogentech, Seraing, Belgium) were designed to bind specifically to the KV part of the LPA gene. Because of the high degree of sequence homology between the LPA and PLG genes, specificity of LPA primers and probes was tested by running PCR reactions with different concentrations of plasmid containing PLG gene or LPA plasmid construct without KV (a kind gift from Dr. S. Frank, University of Graz, Austria). There was no detectable PCR product when these plasmids were used as templates in the PCR reaction. cDNA made from RNA isolated from the liver of WT mouse was used as a negative control when analyzing the transgenes.
Because of poor RNA quality, two samples were excluded from analysis.

Statistical analyses
The non-parametric Mann-Whitney test was used to analyse differences between groups. All data from serum analysis are presented as mean ± SEM. In the case of mRNA data, results are presented as median value of normalized mRNA level.