Androgen formation and metabolism in the pulmonary epithelial cell line A549: expression of 17beta-hydroxysteroid dehydrogenase type 5 and 3alpha-hydroxysteroid dehydrogenase type 3.

Surfactant synthesis within developing fetal lung type II cells is affected by testosterone and 5alpha-dihydrotestosterone (5alpha-DHT). The pulmonary epithelial cell line A549, isolated from a human lung carcinoma, like normal lung type II cell, produces disaturated phosphatidylcholines and has been widely used for studying the regulation of surfactant production. Androgen receptor has been detected in A549 cells; however, the capacity of these cells for androgen synthesis and metabolism has not been investigated at molecular level. This study was undertaken to identify the steroidogenic enzymes involved in the formation and metabolism of androgens from adrenal C19 steroid precursors in A549 cells. When cultured in the presence of normal FCS, A549 intact cells converted DHEA to androstenediol, androstenedione principally to testosterone, and 5alpha-DHT to 5alpha-androstane 3alpha,17beta-diol. High levels of 17beta-hydroxysteroid dehydrogenase (HSD) and 3alpha-HSD activities were detected in both cytosol and microsomes isolated from homogenates. Analysis of A549 RNA indicated the presence of 17beta-HSD type 4 and type 5, and of 3alpha-HSD type 3 messenger RNAs. Very low levels of 3beta-HSD type 1 and 5alpha-reductase type 1 messenger RNAs and activities were detected. With regard to active androgen formation, there was little or no capacity for the conversion of DHEA to 5alpha-DHT. In contrast, androstenedione was rapidly transformed to testosterone. The pattern of steroid metabolism was not affected by the use of charcoal-stripped FCS or by the synthetic glucocorticoid dexamethasone. Together, our findings show that A549 cells express a pattern of steroid metabolism in which 17beta-HSD type 5 and 3alpha-HSD type 3 are the predominant enzymes. The level of androgens is regulated at the level of catalysis in intact cells such that the intracellular level of testosterone is stabilized, whereas 5alpha-DHT is rapidly inactivated by reduction to 3alpha,17beta-diol. This pattern of androgen metabolism has implications for the relative importance of testosterone and 5alpha-DHT in normal lung development and surfactant production.

Length increase of the human alpha -globin 3'-untranslated region disrupts stability of the pre-mRNA but not that of the mature mRNA.

Polyadenylation increases the stability of mRNA molecules. By studying the effect of the length of 3'-untranslated region (UTR) on mRNA levels, we have found that alpha-globin pre-mRNA is stabilized by a mechanism that does not modulate the half-life of mature mRNA. The insertion of DNA fragments of various unrelated sequences into the 3'-UTR of the human alpha-globin gene strongly reduces mRNA abundance upon transfection into choriocarcinoma JEG-3 cells. We found an inverse relationship between mRNA levels and the length of the introduced fragments. In fact, mRNA levels as low as 1% were observed after inserting a 477-nucleotide (nt) fragment, whereas inserting a fragment of 86 nt at the same position had no effect on mRNA accumulation. DNA insertion induced no change in transcription rate or in half-life of mature mRNA. Semi-quantitative reverse transcription-polymerase chain reaction revealed that inserting a 477-nt fragment in the 3'-UTR resulted in decreased levels of nuclear pre-mRNA in proportion to that observed for mature mRNA. In contrast, the insertion of the 477-nt exogenous DNA in the last intron had no effect on mRNA levels despite the presence of intronic sequences in the pre-mRNA. This shows that the reduction of pre-mRNA level was not due to the insertion of putative ribonuclease cleavage sites or the insertion of a segment DNA that reduces the elongation efficiency. Taken together, our results strongly support the existence of a pre-mRNA stabilizing mechanism that can be disrupted by increasing the length of the 3'-UTR. The fact that the half-life of mature mRNA is not affected by DNA insertion is compatible with a pre-mRNA-specific stabilizing mechanism that acts specifically before polyadenylation.

Localization of type 1 17beta-hydroxysteroid dehydrogenase mRNA and protein in syncytiotrophoblasts and invasive cytotrophoblasts in the human term villi.

The 17beta-hydroxysteroid dehydrogenases (17beta-HSDs) play a key role in the synthesis of sex steroids. The hallmark of this family of enzymes is the interconversion, through their oxydoreductive reactivity at position C17, of 17-keto- and 17beta-hydroxy-steroids. Because this reaction essentially transforms steroids having low binding activity for the steroid receptor to their more potent 17beta-hydroxysteroids isoforms, it is crucial to the control of the physiological activities of both estrogens and androgens. The human placenta produces large amounts of progesterone and estrogens throughout pregnancy. The placental type 1 17beta-HSD enzyme (E17beta-HSD) catalyzes the reduction of the low activity estrogen, estrone, into the potent estrogen, estradiol. We studied the cell-specific expression of type 1 17beta-HSD in human term placental villous tissue by combining in situ hybridization to localize type 1 17beta-HSD mRNA with immunohistochemistry using an antibody against human placental lactogen, a trophoblast marker. Immunolocalization of E17beta-HSD was also performed. To ascertain whether other steroidogenic enzymes are present in the same cell type, cytochrome P450 cholesterol side-chain cleavage (P450scc), P450 aromatase, and type 1 3beta-hydroxysteroid dehydrogenase (3beta-HSD) were also localized by immunostaining. Our results showed that the syncytium is the major steroidogenic unit of the fetal term villi. In fact, type 1 17beta-HSD mRNA and protein, as well as P450scc, P450 aromatase, and 3beta-HSD immunoreactivities were found in these cells. In addition, our results revealed undoubtedly that extravillous cytotrophoblasts (CTBs), e.g. those from which cell columns of anchoring villous originate, also express the type 1 17beta-HSD gene. However, CTBs lying beneath the syncytial layer, e.g. those from which syncytiotrophoblasts develop, contained barely detectable amounts of type 1 17beta-HSD mRNA as determined by in situ hybridization. These findings, along with those from other laboratories confirm the primordial role of the syncytium in the synthesis of steroids during pregnancy. In addition, our results indicate for the first time that CTBs differentiating along the invasive pathway contain type 1 17beta-HSD mRNA.

Tissue- and site-specific gene expression of type 2 17beta-hydroxysteroid dehydrogenase: in situ hybridization and specificenzymatic activity studies in human placental endothelial cells of the arterial system.

Progesterone and estradiol are the most potent human sex steroid hormones of placental origin and are essential to the maintenance of pregnancy, the timing of parturition, the maturation of many fetal organs, and the preparation of the maternal reproductive system. Naturally, regulatory mechanisms must be in place to coordinate the synthesis and inactivation of these two hormones. We have previously shown that the highest levels of type 1 and type 2 17beta-hydroxysteroid dehydrogenase (17betaHSD) messenger ribonucleic acids (mRNAs) occur in the placenta, particularly in the villi. However, in contrast to type 1 17betaHSD mRNA, type 2 17betaHSD mRNA was not detectable in cell cultures of human cytotrophoblasts or syncytiotrophoblasts. Using in situ hybridization, we unequivocally identified endothelial cells as the only cell type expressing the type 2 17betaHSD gene in fetal villi. Moreover, type 2 17betaHSD mRNA was specifically detected in the endothelial cells of the arterial system, and at higher levels in the villi compared with endothelial cells of the cord arteries when the two tissue sections were cohybridized. In fact, both mRNA levels and enzymatic activity are at their highest levels in arterial endothelial cells. In conclusion, the endothelial cells of the villous arterioles are the primary site of type 2 17betaHSD gene expression. This suggests a regulatory role for these cells in the control of progestin, androgen, and estrogen levels during pregnancy, thus opening a whole new way of viewing regionalization and localization of steroidogenesis in the human villi.

Separation by thin-layer chromatography of the most common androgen-derived C19 steroids formed by mammalian cells.

Several methods have been developed in the past for the separation and identification of closely related steroid hormones. Although these methods are effective, most of them use HPLC-derived systems and are expensive, laborious, or time-consuming. In the course of our studies of the metabolism of dehydroepiandrosterone and androstenedione in tissues, we have modified a previously published technique in such a way that in one TLC step we can separate most of the androgen C19 steroid derivatives produced by mammalian cells. We have used this modified technique for the past 2 years with considerable success and reproducible results, and we find it to be rapid and relatively inexpensive.

Transfer of dehydroepiandrosterone- and pregnenolone-fatty acid esters between human lipoproteins.

Dehydroepiandrosterone (DHEA)- and pregnenolone (PREG)-fatty acid esters (FA) are formed in plasma high density lipoproteins (HDL), whereas they accumulate in very low density lipoproteins (VLDL), low density lipoproteins (LDL), and HDL. We have hypothesized that these lipoidal steroids could be transferred from HDL to VLDL and LDL by the cholesteryl ester (CE) transfer protein (CETP), which mediates CE transfer activity in human plasma. In this study, we further investigated this hypothesis. Lipoproteins and lipoprotein-deficient plasma (LPDP) were purified and analyzed by Western blots. LPDP was rich in CETP, in contrast to lipoprotein preparations, which contained very low amounts. Using these preparations in in vitro transfer assays, CE transfer from radiolabeled steroid ester-HDL to VLDL or LDL was only observed in the presence of LPDP. In contrast, time- and temperature-dependent transfer of DHEA-FA and PREG-FA were observed in the absence of LPDP. The addition of LPDP had no effect on the DHEA-FA transfer rate, whereas the PREG-FA transfer rate was increased. Moreover, in the absence of LPDP, no decrease in the transfer levels of DHEA-FA and PREG-FA was observed after the removal of CETP from lipoprotein preparations by immunoaffinity column chromatography. The PREG-FA transfer activity of LPDP was studied using the anti-CETP monoclonal antibody TP-2, which is known to block the CE transfer activity of CETP. In the presence of LPDP, this antibody led to a dose-dependent decrease in CE transfer activity, whereas PREG-FA transfer activity was unaffected. In conclusion, we have shown that DHEA-FA and PREG-FA are transferred from HDL to VLDL and LDL by a CETP-independent mechanism. There is a major difference in the transport of lipoidal steroids by human lipoproteins compared to that of CE.

Effect of insulin on serum levels of dehydroepiandrosterone metabolites in men.

OBJECTIVE: Insulin was found to decrease the concentration of serum dehydroepiandrosterone (DHEA) and DHEA-sulphate (DHEAS) and recent data suggest that an increase in the metabolic clearance rate of DHEA (MCRDHEA) may be involved. In this study, we have investigated the effects of insulin on DHEA metabolism in men. PATIENTS: A total of 10 men were enrolled into the study, and all subjects completed the study. Subjects were healthy, non-obese, and 20-30 years old. DESIGN: DHEA was administered intravenously to subjects, alone or in combination with insulin. A hyperinsulinaemic-euglycaemic clamp was initiated for subjects with the insulin infusion and euglycaemia was maintained by checking blood glucose and adjusting the rate of a 25% dextrose infusion as needed. Serum was collected before DHEA infusion, during DHEA infusion after attaining steady state (3.5-4 h), and during DHEA plus insulin infusion (6-6.5 h) (steady state) and then assayed for DHEA, DHEA metabolites, and DHEA acylation by LCAT. RESULTS: Results showed rapid transformation of DHEA into androst-5-ene-3 beta, 17 beta-diol, DHEA fatty-acid esters (DHEA-FA), androstenedione and 5 alpha-androstan-3 alpha-ol-17-one glucuronide (androsterone glucuronide) whereas DHEAS, testosterone, androstane-3 alpha, 17 beta-diol glucuronide and oestradiol serum levels were not affected. When insulin was simultaneously infused once steady-state DHEA levels had been attained, we observed a decline only in DHEA, DHEA-FA and DHEAS levels, with no effect on other steroids examined. Although serum DHEA esterification was not affected by DHEA, a stimulation by insulin was observed. CONCLUSIONS: These results suggest that insulin increases the DHEA metabolic clearance rate by stimulating its conversion to DHEA-FA and by enhancing uptake of DHEA-FA by peripheral tissues.

Dehydroepiandrosterone-fatty acid esters in human plasma: formation, transport and delivery to steroid target tissues.

In addition to dehydroepiandrosterone (DHEA) sulfate (S), the human plasma also contains a second form of DHEA ester: DHEA-fatty acid esters (DHEA-FA). In the human adult, the plasma concentrations of DHEA-FA, DHEA and DHEAS are in the range of 6, 12 and 2000 nM respectively. Although the adrenal is responsible for almost all production of DHEAS in the circulation, DHEA-FA is formed from DHEA by an enzyme present in the circulation. Our work has clearly demonstrated that lecithin-cholesterol acyltransferase, localized on high density lipoprotein, is responsible for DHEA-FA production. Once DHEA-FA is formed, it is subsequently transferred to very low density lipoprotein (VLDL) and low density lipoprotein (LDL), like cholesteryl esters. Plasma lipoproteins contain at least 90% of circulating DHEA-FA of which 40% are found in the LDL fraction. Analysis of the fatty acid composition of tritiated DHEA-FA-labelled LDL ([3H]DHEA-FA-LDL) indicated the prevalence of DHEA-linoleate/palmitoleate and DHEA-oleate. Treatment of [3H]steroid-FA-LDL with charcoal does not remove radioactivity, thus suggesting that the non-polar steroid is incorporated into the central non-polar core of the lipoproteins. Incubation of [3H]DHEA-FA-LDL with ZR-75-1 breast cancer cells produced a time-dependent increase in labeled non-conjugated steroids in the cell culture medium, whereas the levels of tritiated DHEA-FA decreased. Lipoidal radioactivity in cells increased with time, but non-conjugated radioactivity associated with the cells showed no such increase. HPLC analysis of the culture medium indicated the presence of tritiated DHEA and androst-5-one-3 beta, 17 beta-diol. Our study indicates that circulating DHEA-FA incorporated into lipoproteins may indeed act as a substrate for potent steroid formation following their entry into steroid target cells.

Formation of pregnenolone- and dehydroepiandrosterone-fatty acid esters by lecithin-cholesterol acyltransferase in human plasma high density lipoproteins.

Pregnenolone- (PREG-), and dehydroepiandrosterone- (DHEA-) fatty acid esters (FA) are present in human plasma, where they are associated with lipoproteins. Because plasma has the ability to form PREG-FA and DHEA-FA in vitro from their unconjugated steroid counterparts, we postulated that the LCAT enzyme might be responsible for their formation. Here we show that lecithin-cholesterol acyltransferase (LCAT) has PREG and DHEA esterifying activities. First, VLDL, IDL, LDL, and HDL were isolated by the sequential ultracentrifugation micromethod from the plasma of fasting men and women and tested for their ability to form PREG-FA, DHEA-FA, and cholesteryl esters in vitro from their respective unconjugated counterparts. The results showed that the three steroids were esterified only in HDL subfractions. The rate of tritiated PREG esterification was clearly higher than that of tritiated cholesterol and DHEA, both in total plasma and isolated HDL, and no gender difference was observed. Second, human and guinea pig LCAT were purified and used in phosphatidylcholine-reconstituted vesicles containing human apoAI to show their ability to esterify tritiated cholesterol, PREG, and DHEA in the absence of unlabeled steroid. The amount of cholesteryl ester, PREG-FA, and DHEA-FA increased after incubation as a function of time and amount of purified LCAT, showing that PREG is preferentially acylated by LCAT compared to cholesterol and DHEA. The PREG and DHEA esterifying activities of LCAT were cofactor-dependent, as shown by the absence of acylation without apoAI. Finally, we determined by HPLC the fatty acid moiety of PREG-FA and DHEA-FA formed in human plasma and guinea pig and rat sera in vitro after incubation with unconjugated tritiated PREG and DHEA. We showed that the fatty acid moieties of newly formed tritiated PREG-FA and DHEA-FA were similar to that reported for cholesteryl esters in the plasma of the three species. We conclude that LCAT has a lecithin-steroid acyltransferase activity and that PREG is probably the preferential substrate of this enzyme. In addition, the fact that the differences in the fatty acid moieties of cholesteryl esters of human, guinea pig, and rat plasmas are also observed for PREG-FA and DHEA-FA suggests that the LCAT is the sole circulating enzyme that has PREG and DHEA esterifying activities.

Guinea pig apolipoprotein D RNA diversity, and developmental and gestational modulation of mRNA levels.

We have isolated and characterized two molecular types of guinea pig (GP) apolipoprotein D (apoD) cDNA. The sequences of cDNA clones GP APO D-20 and -38 are 100 % homologous in their putative exons 2-5, as determined by analogy within human apoD gene, but they differ totally in their putative exon 1. RNase protection assays showed the presence of both apoD RNA types 20 and 38 in cauda epididymis. Northern blot analysis revealed four polyadenylated apoD bands at 3.2, 2.7, 1.7, and 1.0 kb. Types 20 and 38 specific probes hybridized with the major 1-kb mRNA and two of the three other minor RNA transcripts, respectively. Southern blot analysis revealed that the guinea pig genome probably contains one apoD gene. Our data also demonstrated that the cauda epididymis and fallopian tubes had an apoD mRNA concentration 100-fold higher than the liver, suggesting that the apoD gene expression could be associated with the presence of steroids. The levels of the 1-kb mRNA increased in the fallopian tubes and ovaries during gestation and were lower in fetal reproductive tissues and liver than in mature animals. No positive correlation was found between apoD and 3 beta-hydroxysteroid dehydrogenase/delta5-delta4 isomerase (3 beta-HSD) mRNA levels in these tissues, thus suggesting that high amounts of apoD mRNA are not necessarily associated with in situ progesterone synthesis. Taken together, our results indicate that both the guinea pig epididymis and fallopian tubes are excellent models to study the local role of apoD in steroid target tissues.

Laminin-binding protein 120 from brain is closely related to the dystrophin-associated glycoprotein, dystroglycan, and binds with high affinity to the major heparin binding domain of laminin.

When brain proteins separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose are probe with 125I-labeled laminin, a single broad band of approximately 120 kDa binds laminin specifically. We show here by two-dimensional electrophoresis and protein microsequencing that this band consists of two distinct laminin-binding proteins. One of these is the amyloid precursor protein. The other, laminin-binding protein (LBP) 120, is closely related to the dystrophin-associated glycoprotein, dystroglycan (156 kDa); 5 peptides from purified bovine brain LBP120, ranging in size from 7 to 19 residues, are up to 100% identical to the predicted amino acid sequence of muscle dystroglycan (ibraghimov-Beskrovanaya, O., Ervasti, J. M., Leveille, C. J., Slaughter, C. A., Sernett, S. W., and Campbell, K. P. (1992) Nature 355, 696-702). These protein microsequence data support the data of Ibraghimov-Beskrovnaya et al., which suggest that the dystroglycan precursor is processed into 120/156- and 43-kDa proteins. Moreover, the data suggest a revision in the position of the proposed cleavage site of the precursor. The glycosylation and extracellular localization of LBP120/dystroglycan are consistent with it being a cell surface laminin receptor. LBP120/dystroglycan, either as a native protein, or following SDS-PAGE and transfer to nitrocellulose, binds with high affinity (Kd = 90 nM) to a proteolytic fragment of laminin (E3) containing the major heparin binding domain. This binding is Ca(2+)-dependent and inhibited by low concentrations of heparin. Thus, LBP120/dystroglycan is a major non-integrin laminin receptor whose high affinity interaction with laminin may reflect a structural role in brain and muscle.

Structure of the human apolipoprotein D gene promoter region.

A human genomic clone of 18.2 kbp encompassing the apolipoprotein D (apoD) exon-1 and 2 and 10 kbp of upstream sequence was isolated and characterized. DNA sequencing and primer extension analysis revealed a transcriptional initiation site located 27 bp downstream of a consensus TATA box sequence. The exon 1 was 66 bp long. Computer analysis of DNA sequence from positions -557 to +129 revealed some putative transcriptional regulatory elements including a stretch of (pyrimidine/purine)26 located from nucleotide -263 to nucleotide -212, which could potentially form Z-DNA. Steroid hormone regulatory elements were identified which may be related to the modulation of apoD gene expression by androgens and estrogens in vitro.

Localization of the major sites of rabbit apolipoprotein D gene transcription by in situ hybridization.

We have identified the sites of transcription of the apolipoprotein D (apoD) gene in the rabbit by in situ hybridization. We show here for the first time that 1) apoD mRNA production does not correlate with the sites of steroid hormone production in adrenal glands nor male genital tissues; and 2) the expression of the apoD gene is clearly higher in white than in gray matter throughout the central nervous system (CNS). Specifically, apoD mRNA was abundant near blood vessels and was expressed mostly in fibroblast-like cells, in particular in the testis, the efferent ducts, the ductus epididymis, the lung, and the subarachnoid space of the CNS. Other positive cell types were endothelial cells of adrenal sinusoidal capillaries and glial cells of the CNS. We detected apoD mRNA in both the adrenal cortex and medulla. White but not gray matter showed high levels of apoD mRNA throughout both the rabbit CNS and in human brain. The red pulp of spleen showed a strong hybridization. In prepubertal rabbits apoD mRNA levels were moderate in both testis and epididymis. Epididymal but not testicular expression increased with the onset of puberty and epididymal levels always exceeded those of the testes in animals showing spermatogenesis. Thus, the variation in levels of apoD mRNA among organs in vivo, that we and others have previously reported, can be explained by transcription being not only characteristic of cell type, with a few common cell types producing in each organ, but transcription also varied among cells of the same lineage.

Apolipoprotein D transcription occurs specifically in nonproliferating quiescent and senescent fibroblast cultures.

We studied apolipoprotein D (apoD) mRNA in primary cultures of human diploid fibroblasts (HDF). In early-passage HDF no apoD mRNA was detected in replicating cells in sparse culture, but the gene was expressed in quiescent cells in confluent and in serum-starved cultures. In contrast, late-passage HDF expressed apoD mRNA in sparse culture, but the level increased after attainment of confluence. Thus fibroblasts, the common cell-type expressing apoD mRNA in vivo, express this characteristic following growth-arrest. The same pattern of activation was found in another fibroblast cell line deficient in apoB/E (LDL) receptors, excluding a role for cellular cholesterol delivery by the LDL-receptor pathway controlling apoD expression.

The epitopes of apolipoprotein A-I define distinct structural domains including a mobile middle region.

The definition of epitopes on apoA-I provides evidence for a very dense packing of the peptide chain and supports the proposed supersecondary structure, made of repetitive antiparallel helices. Screening with overlapping synthetic hexapeptides shows that only a few epitopes are continuous and that all continuous contact sequences identified within the epitopes coincide or are contiguous to a putative beta-turn, such as residues 1-8, 48-55, 98-104, 118-123, and 135-140. On the N-terminal half of apoA-I we identified 6 overlapping tertiary discontinuous epitopes, uniquely constituted by amino acids and discontinuous sequences on helical segments that are far apart and which define a particular region with a complex tertiary structure. Among these are the epitopes for antibody 5G6, which reacts with residues 45-51, 83-92, 119-126, and 136-143; for A16, which reacts with 14-19, 23-28, and 60-82; and for r-FC1, which reacts with 1-8, 29-35, 78-83, and 98-121. The very far apart discontinuous sequences included in these epitopes can be explained by the predicted turns and coiled domains, and thus provide evidence for such a tertiary structure. Alternatively, these results could also be explained by intermolecular epitopes involving the N-terminal region. In contrast, in the middle of apoA-I, all identified epitopes are shorter and discontinuous within the secondary structure and are constituted by residues forming a beta-turn with all or part of an adjacent alpha-helix. We hypothesize that these multiple epitopes, that are mostly limited to a single helix, reflect the existence of a very mobile domain, possibly with hinged pairs of adjacent alpha-helices. On the C-terminal half of apoA-I, several monoclonal antibodies react with overlapping epitopes located between residues 149 and 186, which probably reflects 2 antiparallel alpha-helices interrupted by a beta-turn.

Molecular characterization and differential mRNA tissue distribution of rabbit apolipoprotein D.

We report for the first time the quantification of relative apolipoprotein D (apoD) mRNA concentrations in a wide selection of organs and a detailed characterization of the rabbit protein. ApoD cDNA clones were isolated from a rabbit testis cDNA library by screening with a human apoD cDNA-derived RNA probe. The 912 nucleotide sequence of rabbit apoD cDNA contains a unique reading frame coding for a protein sharing 80% homology with human apoD. The two sequences have two potential asparagine-linked glycosylation sites at the same positions, almost superimposable hydrophobicity plot, and the antigenic proteins show similar charge polymorphism, Mr, and lipoprotein distribution. This high degree of similarity shows that the rabbit system can be used as a model for apoD studies. Moreover, the two consensus sequences of the hydrophobic ligand carrier (alpha 2-microglobulin) family present in human apoD are also found in the rabbit protein and these sequences coincide with the most conserved regions. The distribution of apoD mRNA among rabbit organs was determined by Northern blot and quantitative dot blot analysis. The highest levels of mRNA were found in spleen, adrenal glands, lungs, brain, testis, and kidneys. Moderate or low concentrations were detected in all the other organs tested including liver and small intestine. Thus, our results show that the apoD gene is expressed mainly in peripheral organs, with levels as high as 59-fold that of the liver, unlike other apolipoproteins. We suggest that apoD exerts its main function locally in peripheral organs.