Effects of TCDD on Ah receptor, ARNT, EGF, and TGF-alpha expression in embryonic mouse urinary tract

Lipoprotein(a) [Lp(a)] biogenesis was examined in primary cultures of hepatocytes isolated from mice transgenic for both human apolipoprotein(a) [apo(a)] and human apoB. Steady-state and pulse-chase labeling experiments demonstrated that newly synthesized human apo(a) had a prolonged residence time (approximately 60 min) in the endoplasmic reticulum (ER) before maturation and secretion. Apo(a) was inefficiently secreted by the hepatocytes and a large portion of the protein was retained and degraded intracellularly. Apo(a) exhibited a prolonged and complex folding pathway in the ER, which included incorporation of apo(a) into high molecular weight, disulfide-linked aggregates. These folding characteristics could account for long ER residence time and inefficient secretion of apo(a). Mature apo(a) bound via its kringle domains to the hepatocyte cell surface before appearing in the culture medium. Apo(a) could be released from the cell surface by apoB-containing lipoproteins. These studies are consistent with a model in which the efficiency of post-translational processing of apo(a) strongly influences human plasma Lp(a) levels, and suggest that cell surface assembly may be one pathway of human Lp(a) production in vivo. Transgenic mouse hepatocytes thus provide a valuable model system with which to study factors regulating human Lp(a) biogenesis.     

Presecretory degradation of apolipoprotein [a] is mediated by the proteasome pathway

Plasma levels of atherogenic lipoprotein [a] (Lp[a]) vary over a 1000-fold range and are largely determined by the gene for its unique glycoprotein, apolipoprotein [a] (apo[a]). The apo[a] locus comprises more than 100 alleles, encoding proteins from <300 to >800 kDa. Using primary baboon hepatocyte cultures, we previously demonstrated that differences in the secretion efficiency of apo[a] allelic variants contribute to the variation in plasma Lp[a] levels. In the current study, we investigated the mechanism of apo[a] presecretory degradation. The proteasome inhibitors, acetyl-leucyl-leucyl-norleucinal and lactacystin, prevented apo[a] degradation and increased apo[a] secretion. Transfection with an HA-tagged ubiquitin construct demonstrated the accumulation of ubiquitinated apo[a] in the presence of lactacystin. These results suggest a role for the cytoplasmic proteasome in apo[a] proteolysis. Apo[a] that accumulated intracellularly in the presence of lactacystin remained sensitive to endo-B-N-glucosaminidase H, and apo[a] degradation was reversibly inhibited by brefeldin A, suggesting that transport to a post-endoplasmic reticulum (ER) pre-medial Golgi compartment is required for apo[a] degradation. Newly synthesized apo[a] bound to the ER chaperone calnexin and conditions that enhanced this interaction prevented apo[a] degradation, suggesting that calnexin can protect apo[a] from proteolysis. These studies provide further support for the role of the proteasome in endoplasmic reticulum quality control, and expand this role to one that influences plasma levels of the atherogenic lipoprotein Lp[a].-White, A. L., B. Guerra, J. Wang, and R. E. Lanford. Presecretory degradation of apolipoprotein[a] is mediated by the proteasome pathway.     

Inhibition of N-linked glycosylation results in retention of intracellular apo[a] in hepatoma cells, although nonglycosylated and immature forms of apolipoprotein[a] are competent to associate with apolipoprotein B-100 in vitro

Apolipoprotein[a] (apo[a]) is a highly polymorphic glycoprotein that forms a covalent complex with apolipoprotein B-100 (apoB-100), producing a lipoprotein species referred to as lipoprotein[a] (Lp[a]). We have studied the effects of alterations in glycosylation of apo[a] on its intracellular processing and secretion as well as its ability to associate with low density lipoprotein (LDL) apoB-100. HepG2 cells transfected with a 6 kringle IV (6 K-IV) apo[a] minigene were treated with tunicamycin, an inhibitor of N-linked glycosylation, which eliminated apo[a]-B-100 complexes from the media. Tunicamycin treatment also reduced secretion of the 6 K-IV apo[a] protein from transfected McA-RH7777 cells by approximately 50%, but completely eliminated secretion of apo[a] species containing 9 and 17 K-IV repeats. Mixing experiments, performed with radiolabeled media (+/-tunicamycin) from transfected McA-RH7777 cells, demonstrated no alteration in the extent of association of apo[a] with human LDL. Similar mixing experiments using culture media from glycosylation-defective mutant chinese hamster ovary (CHO) cells transfected with the same apo[a] minigene showed identical results. Apo[a] secretion was demonstrated in all mutant cell lines in the absence of either N- or O-linked (or both) glycosylation. The mechanisms underlying the reduced secretion of apo[a] from transfected hepatoma cells were examined by pulse-chase radiolabeling and apo[a] immunoprecipitation. Tunicamycin treatment altered the efficiency of precursor apo[a] processing from the ER by increasing its ER retention time. The increased accumulation of precursor apo[a] in the ER was associated with alterations in the kinetics of association with two resident endoplasmic reticulum (ER) chaperone proteins, calnexin and BiP. These findings suggest that the glycosylation state and size of apo[a] appear to play a role in regulating its efficient exit from the endoplasmic reticulum. However, neither N- nor O-linked glycosylation of apo[a] exerts a major regulatory role in its covalent association with apoB-100.     

Gemfibrozil significantly lowers cynomolgus monkey plasma lipoprotein[a]-protein and liver apolipoprotein[a] mRNA levels

Eight male cynomolgus monkeys (Macaca fascicularis) on a normal chow diet were orally administered gemfibrozil daily using a weekly rising dose protocol for 3 weeks (50, 125, and 200 mg/kg per day). At these drug doses, Lp[a] levels were reduced: 83.7% +/- 3.2 (SEM), (P < 0.024); 63.7% +/- 4.1 (P < 0.013); and 36.2% +/- 1.1 (P < 0.002), respectively, of pretreatment values. Lp[a] reduction was directly related to blood gemfibrozil concentration (range 36-428 microM, r = 0.969) and occurred without concomitant changes in apolipoprotein B. Three weeks posttreatment Lp[a] levels returned to pretreatment values. A specific ribonuclease protection assay demonstrated that liver apolipoprotein[a] (apo[a]) mRNA expression was decreased in all animals to an average of 19.1% +/- 3.0 (P < 0.0026), of pretreatment values after the 200 mg/kg treatment, whereas, albumin, apolipoprotein A-I, apolipoprotein E, and glyceraldehyde-3-phosphate dehydrogenase mRNAs were unchanged. Lp[a] levels were unaffected by gemfibrozil in HepG2 cells permanently transfected with an apo[a] 10-kringle cDNA construct containing partial 5'- and 3'-untranslated sequences and under control of a constitutive CMV promoter. However, both Lp[a] and apo[a] mRNA in primary cynomolgus monkey hepatocytes were coordinately lowered in a dose-dependent fashion by gemfibrozil. Thus, Lp[a] can be regulated by gemfibrozil at the level of apo[a] mRNA expression.     

Inhibitory effect of 9-(2-phosphonylmethoxyethyl)-adenine (PMEA) on human and duck hepatitis B virus infection

9-(2-Phosphonylmethoxyethyl)adenine (PMEA) was evaluated for its inhibitory effect on hepadnavirus replication in three different cell systems, i.e., human hepatoma cell lines HepG2 2.2.15 and HB611 (transfected with human hepatitis B virus (HBV)) and primary cultures of duck hepatocytes infected with duck hepatitis B virus (DHBV). PMEA inhibited HBV release from HepG2 2.2.15 cells and HB611 cells at a 50% inhibitory concentration (IC50) of 0.7 and 1.2 microM, respectively. Intracellular viral DNA synthesis was inhibited at concentrations equivalent to those required to inhibit virus release from the cells. DHBV secretion from duck hepatocytes was inhibited by PMEA at an IC50 of 0.2 microM. HBsAg secretion was inhibited by PMEA in a concentration-dependent manner in HB611 cells and DHBV-infected duck hepatocytes but not HepG2 2.2.15 cells. The 50% cytotoxic concentration, as measured by inhibition of [3H-methyl]deoxythymidine incorporation was 150 microM for the two human hepatoma cell lines and 40 microM for the duck hepatocyte cultures. In a pilot experiment PMEA was found to reduce the amounts of DHBV DNA in the serum of Pekin ducks.     

Novel therapeutic strategy for atherosclerosis: ribozyme oligonucleotides against apolipoprotein(a) selectively inhibit apolipoprotein(a) but not plasminogen gene expression

BACKGROUND: Because mechanisms of atherosclerosis by lipoprotein(a) [Lp(a)] have been postulated in the decrease in active transforming growth factor-beta conversion by decreased plasmin, selective decrease in apolipoprotein(a) [apo(a)] independent of plasminogen may have therapeutic values. Although antisense can decrease apo(a), its application may be difficult because of very high homology of apo(a) gene to plasminogen. Thus we used ribozyme strategy that actively cleaves targeted genes to selectively inhibit apo(a) expression. METHODS AND RESULTS: We constructed ribozyme oligonucleotides containing phosphorothioate DNA- and RNA-targeted kringle 4 of the apo(a) gene that showed 80% homology to plasminogen. Transfection of human apo(a) gene produced Lp(a) in medium of HepG2 cells, whereas Lp(a) could not be detected in control cells. Cotransfection of ribozyme and apo(a) gene resulted in the decrease in mRNA of apo(a) but not plasminogen. Moreover, marked decrease in Lp(a) was also observed in the medium transfected with ribozyme and apo(a) gene compared with apo(a) gene alone (P<0.01), whereas there was no significant change in plasminogen level between ribozyme-transfected and control cells. Incubation of human vascular smooth muscle cells (VSMC) with conditioned medium from apo(a)-transfected HepG2 cells resulted in a significant increase in VSMC number, whereas addition of conditioned medium from cells cotransfected with ribozyme oligonucleotides and apo(a) gene resulted in no VSMC growth (P<0.01). DNA-based control oligonucleotides and mismatched ribozyme oligonucleotides did not have an inhibitory effect on Lp(a) production. CONCLUSIONS: Overall, our data revealed that transfection of ribozyme against the apo(a) gene resulted in the selective inhibition of the apo(a) but not the plasminogen gene, providing novel therapeutic strategy for treatment of high Lp(a), a risk factor for atherosclerosis.     

Obese Zucker (fa/fa) rats are resistant to insulin's inhibitory effect on hepatic apo B secretion

Hepatocytes derived from lean Zucker rats have reduced secretion of apo B and lowered cellular apo B in response to a physiologic range of insulin (0.1 nM-10 nM). Effects are attenuated in hepatocytes derived from Zucker obese rats and seen only at higher insulin concentrations (> 100 nM) with a significant shifting of the dose-response curve. Decreased sensitivity and responsiveness of hepatocytes derived from obese rats suggests insulin resistance and dose-response curves are consistent with coexistent binding and post-binding defects. Inability to inhibit hepatic apo B secretion in the presence of short-term high levels of insulin may have important implications to the balance of intestinal and hepatic triglyceride-rich lipoprotein secretion post-prandially.     

Differences in Lp[a] concentrations and apo[a] polymorphs between black and white Americans

Lp[a] concentrations in nmol/L and apo[a] size isoforms, expressed in terms of the relative number of apo[a] kringle 4 (K4) repeats, were determined in 3959 whites and blacks from four U.S. communities. Plasma Lp[a] analyses were performed by an ELISA method insensitive to apo[a] size heterogeneity and apo[a] size isoforms were determined by high resolution agarose gel electrophoresis. Allele frequencies were estimated by maximum likelihood methods in order to account for the presence of null alleles and coalescence of hands on gels. The apo[a] allele frequencies and phenotype distributions differed significantly between blacks and whites (P < 0.0001). Blacks had a higher relative frequency of the intermediate alleles K4(22) through K4(28) whereas whites had a higher relative frequency of the small alleles K4(17) through K4(24) and large alleles K4(29) through K4(33). The estimated frequency of the null allele was low in both blacks (1.0%) and whites (6.7%). The Lp[a] distribution was less skewed and Lp[a] concentrations were higher in blacks than whites (mean 94 nmol/L and 48 nmol/L, median 74 nmol/L and 20 nmol/L for blacks and whites, respectively). The relationship between apo[a] size and Lp[a] concentration also differed significantly between these two racial groups. For the large polymorphs (> 31 K4 repeats) both blacks and whites exhibited uniformly low Lp[a] values. For the intermediate isoforms K4(20) through K4(30), a considerable range of Lp[a] values was evident in blacks; the median Lp[a] for each isoform increased nearly linearly as the apo[a] size decreased. In contrast in whites there was little change in median Lp[a] concentrations for isoforms K4(20) through K4(30). For the small apo[a] size (< 20 K4) both blacks and whites exhibited high median Lp[a] levels and a wide variation of Lp[a] levels. The major difference in Lp[a] levels between the two racial groups occurred in the intermediate size isoform range of K4(20) through K4(25). In conclusion, whites and blacks differ significantly in Lp[a] concentrations, allele and phenotype frequencies, and in the relationship between apo[a] size isoform and Lp[a] concentration.     

Impaired secretion of very low density lipoprotein-triglycerides by apolipoprotein E- deficient mouse hepatocytes

To explore mechanisms underlying triglyceride (TG) accumulation in livers of chow-fed apo E-deficient mice (Kuipers, F., J.M. van Ree, M.H. Hofker, H. Wolters, G. In't Veld, R.J. Vonk, H.M.G. Princen, and L.M. Havekes. 1996. Hepatology. 24:241-247), we investigated the effects of apo E deficiency on secretion of VLDL-associated TG (a) in vivo in mice, (b) in isolated perfused mouse livers, and (c) in cultured mouse hepatocytes. (a) Hepatic VLDL-TG production rate in vivo, determined after Triton WR1339 injection, was reduced by 46% in apo E-deficient mice compared with controls. To eliminate the possibility that impaired VLDL secretion is caused by aspecific changes in hepatic function due to hypercholesterolemia, VLDL-TG production rates were also measured in apo E-deficient mice after transplantation of wild-type mouse bone marrow. Bone marrow- transplanted apo E-deficient mice, which do not express apo E in hepatocytes, showed normalized plasma cholesterol levels, but VLDL-TG production was reduced by 59%. (b) VLDL-TG production by isolated perfused livers from apo E-deficient mice was 50% lower than production by livers from control mice. Lipid composition of nascent VLDL particles isolated from the perfusate was similar for both groups. (c) Mass VLDL-TG secretion by cultured apo E-deficient hepatocytes was reduced by 23% compared with control values in serum-free medium, and by 61% in the presence of oleate in medium (0. 75 mM) to stimulate lipogenesis. Electron microscopic evaluation revealed a smaller average size for VLDL particles produced by apo E-deficient cells compared with control cells in the presence of oleate (38 and 49 nm, respectively). In short-term labeling studies, apo E-deficient and control cells showed a similar time-dependent accumulation of [3H]TG formed from [3H]glycerol, yet secretion of newly synthesized VLDL-associated [3H]TG by apo E-deficient cells was reduced by 60 and 73% in the absence and presence of oleate, respectively. We conclude that apo E, in addition to its role in lipoprotein clearance, has a physiological function in the VLDL assembly-secretion cascade.     

Polymorphism of apolipoprotein E, lipoprotein(a), and other lipoproteins in children with type I diabetes

We assessed the effect of particular apolipoprotein (apo) E phenotypes, lipoprotein(a) [Lp(a)], and other lipoproteins on the development of dyslipoproteinemia in 450 patients with type I diabetes, ages 13-14 years. The control group consisted of 450 healthy school children of both sexes, ages 13-14 years. Both groups were found to be normolipidemic, but the concentration of Lp(a) was significantly (P < 0.05) higher in the diabetic children than in the control group. Apo E 3/2 and apo E 4/4 phenotypes were more frequent in the group of diabetics. Diabetics with the apo E 3/3 phenotype had higher concentrations of very-low-density lipoprotein (VLDL) and Lp(a), and lower concentrations of low-density lipoprotein (LDL) than the apo E 3/3 nondiabetics. For apo E 3/2 phenotypes, total cholesterol, LDL cholesterol, LDL, apo A-I, and Lp(a) concentrations were higher in the diabetic children than in the control group; for apo E 4/3 phenotypes, this was true for triglycerides and VLDL cholesterol. The distribution of Lp(a) lipoprotein concentrations between 0.01 and > 0.5 g/L indicated a more frequent occurrence of higher Lp(a) values in diabetic children than in the control group. Results of this study indicate that an increased concentration of Lp(a) lipoprotein and apo E 3/2 and apo E 4/3 phenotypes contribute to the expression of dyslipoproteinemia in type I diabetes in childhood.     

Intracellular metabolism of human apolipoprotein(a) in stably transfected Hep G2 cells

Lipoprotein(a) [Lp(a)] consists of LDL and the glycoprotein apolipoprotein(a) [apo(a)], which are covalently linked via a single disulfide bridge. The formation of Lp(a) occurs extracellularly, but an intracellular assembly in human liver cells has also been claimed. The human apo(a) gene locus is highly polymorphic due to a variable number of tandemly arranged kringle IV repeats. The size of apo(a) isoforms correlates inversely with Lp(a) plasma concentrations, which is believed to reflect different synthesis rates. To examine this association at the cellular level, we analyzed the subcellular localization and fate of apo(a) in stably transfected HepG2 cells. Our results demonstrate that apo(a) is synthesized as a precursor with a lower molecular mass which is processed into the mature, secreted form. The retention times of the precursor in the ER positively correlated with the sizes of apo(a) isoforms. The mature form was observed intracellularly at low levels and only in the Golgi apparatus. No apo(a) was found to be associated with the plasma membrane. Under temperature-blocking conditions, we did not detect any apo(a)/apoB-100 complexes within cells. This finding was confirmed in HepG2 cells transiently expressing KDEL-tagged apo(a). The precursor and the mature forms of apo(a) were found in the ER and Golgi fractions, respectively, also in human liver tissue. From our data, we conclude that in HepG2 cells the apo(a) precursor, dependent on the apo(a) isoform, is retained in the ER for a prolonged period of time, possibly due to an extensive maturation process of this large protein. The assembly of Lp(a) takes place exclusively extracellularly following the separate secretion of apo(a) and apoB.     

Inducibility of enzyme activities associated with the cytochrome P-450 1A family, ethoxyresorufin O-deethylase, and methoxyresorufin O-demethylase in human hepatocyte lines derived from normal liver tissue

Three human hepatocyte cultures have been developed from specimens of normal human liver, in each case from an infant or child, by coculture with liver epithelial cells from 6-day-old rat pups in a complex growth medium. In the established cultures hepatocytes predominate and maintain typical hepatocellular morphology by light microscopy and albumin secretion into supernatant medium. The activity of ethoxyresorufin O-deethylase (EROD) and methoxyresorufin O-demethylase (MROD) basally and after treatment with polycyclic aromatic hydrocarbons was measured spectrofluorometrically in cell homogenates from each culture. Very low levels of EROD and MROD activity were found in each culture without induction [EROD: 1.78 +/- 0.71 (mean +/- SE) pmol/min/mg protein; MROD: 1.33 +/- 0.10 pmol/min/mg protein]. After treatment with 10 microM dibenz (a,h)anthracene x 48 hr, EROD and MROD activities rose approximately 20- to 50-fold. When the basal and induced enzyme activities were remeasured 2 months later, results were essentially the same. Incubation with 10 microM benz(a)anthracene x 48 hr also led to induction of EROD and MROD activities. We believe that these cultures can be regarded as human hepatocyte lines, which conserve human hepatic polycyclic aromatic hydrocarbon-inducible P-450s, most likely including P-450 1A2.     

Amyloid beta-peptide spin trapping. II: Evidence for decomposition of the PBN spin adduct

We investigated the effect of lipoprotein(a) (Lp(a)) on proliferation of human arterial smooth muscle cells (SMCs) and its mechanisms of action. Low density lipoprotein (LDL), Lp(a) and apolipoprotein(a) (apo(a)) significantly stimulated the proliferation of SMCs. Lp(a) and apo(a) reduced the amount of active transforming growth factor-beta (TGF-beta) with the mink lung epithelial cell bioassay, however LDL had no effect. Lp(a), but not apo(a), significantly stimulated the proliferation of SMCs even in the presence of a neutralizing antibody for TGF-beta. Our results suggest that Lp(a) stimulates the proliferation of SMCs via apo(a)-induced inhibition of TGF-beta activation and stimulation of SMCs by the LDL-particle of Lp(a).     

Low density lipoprotein receptor-negative mice expressing human apolipoprotein B-100 develop complex atherosclerotic lesions on a chow diet: no accentuation by apolipoprotein(a)

We have generated mice with markedly elevated plasma levels of human low density lipoprotein (LDL) and reduced plasma levels of high density lipoprotein. These mice have no functional LDL receptors [LDLR-/-] and express a human apolipoprotein B-100 (apoB) transgene [Tg(apoB+/+)] with or without an apo(a) transgene [Tg(apoa+/-)]. Twenty animals (10 males and 10 females) of each of the following four genotypes were maintained on a chow diet: (i) LDLR-/-, (ii) LDLR-/-;Tg(apoa+/-), (iii) LDLR-/-;Tg(apoB+/+), and (iv)LDLR-/-;Tg(apoB+/+);Tg(apo+/-). The mice were killed at 6 mo, and the percent area of the aortic intimal surface that stained positive for neutral lipid was quantified. Mean percent areas of lipid staining were not significantly different between the LDLR-/- and LDLR-/-;Tg(apoa+/-) mice (1.0 +/- 0.2% vs. 1.4 +/- 0.3%). However, the LDLR-/-;Tg(apoB+/+) mice had approximately 15-fold greater mean lesion area than the LDLR-/- mice. No significant difference was found in percent lesion area in the LDLR-/-;Tg(apoB+/+) mice whether or not they expressed apo(a) [18.5 +/- 2.5%, without lipoprotein(a), Lp(a), vs. 16.0 +/- 1.7%, with Lp(a)]. Histochemical analyses of the sections from the proximal aorta of LDLR-/-;Tg(apoB+/+) mice revealed large, complex, lipid-laden atherosclerotic lesions that stained intensely with human apoB-100 antibodies. In mice expressing Lp(a), large amounts of apo(a) protein colocalized with apoB-100 in the lesions. We conclude that LDLR-/-; Tg(apoB+/+) mice exhibit accelerated atherosclerosis on a chow diet and thus provide an excellent animal model in which to study atherosclerosis. We found no evidence that apo(a) increased atherosclerosis in this animal model.     

Effects of fetal calf serum and disruption of cadherin function on the formation of bile canaliculi between hepatocytes

The polarization of hepatocytes to form a connected network of bile canaliculi (BC) is necessary for the function of the liver. Hepatocyte polarization may be controlled by soluble factors and/or physical interactions between cells. Monolayer cultures of embryonic chicken hepatocytes in DMEM supplemented with ornithine, dexamethasone, and insulin express BC-specific antigens for at least 7 days. However, BC-specific antigen expression is lost within 3 days of culture initiation in DMEM containing 10% fetal calf serum. The dedifferentiating effects of fetal calf serum (FCS) can be reversed. Furthermore, cultures in medium containing ornithine, dexamethasone, insulin, and 10% FCS appear identical to cultures grown in 10% FCS alone. Thus FCS contains a soluble inhibitor of hepatocyte polarization. Aggregate cultures grown in suspension maintain hepatocyte polarization for 10-12 days. This may be due to the increased cell-cell contact between hepatocytes in aggregate culture or to more normal contact with the extracellular matrix. We have evaluated the role of cadherin-mediated interactions on hepatocyte polarization. Anti-E-cadherin Fab' fragments disrupted the formation of long networks of BC in monolayer cultures but did not stop polarized expression of BC-specific antigens. The BC antigens in anti-E-cadherin-treated cells were concentrated in small areas between cells and were present at lower levels uniformly on the cell surface. These results indicate that E-cadherin is required for the formation of extended BC networks, but that other factors are responsible for maintaining the synthesis and localization of BC-specific antigens.     

The science and art of radiation oncology after a century

Troglitazone is a new oral hypoglycemic agent that reduces insulin resistance in non-insulin-dependent diabetes mellitus (NIDDM). However, this agent increases serum lipoprotein(a) [Lp(a)], which is known as an atherogenic lipoprotein. The relationships between the response of Lp(a) to troglitazone and the apolipoprotein(a) [apo(a)] phenotype were investigated in this study. Nineteen NIDDM patients were treated with troglitazone for 4 weeks. Lp(a) increased significantly from 20.1+/-16.5 mg/dL to 44.1+/-31.9 mg/dL (P<.001) in all study patients. Lp(a) increased from 25.7+/-34.2 mg/dL to 50.1+/-38.7 mg/dL (P = .03) in patients with smaller apo(a) phenotypes (S1S4 to S2S4). Lp(a) also increased from 17.5+/-12.0 mg/dL to 41.3+/-29.6 mg/dL (P<.01) in patients with larger apo(a) phenotypes (S3 to S4). Therefore, the increase of Lp(a) by troglitazone may be independent of the apo(a) phenotype.