LRP1b shows restricted expression in human tissues and binds to several extracellular ligands, including fibrinogen and apoE-carrying lipoproteins.

OBJECTIVE: To investigate low-density lipoprotein receptor-related protein 1b (LRP1b) expression in human tissues and to identify circulating ligands of LRP1b. METHODS AND
RESULTS: Using two independent RT-PCR assays, LRP1b mRNA was detected in human brain, thyroid gland, skeletal muscle, and to a lesser amount in testis but absent in other tissues, including heart, kidney, liver, lung, and placenta. Circulating ligands were purified from human plasma by affinity chromatography using FLAG-tagged recombinant LRP1b ectodomains and identified by mass spectrometry. Using this technique, several potential ligands (fibrinogen, clusterin, vitronectin, histidine rich glycoprotein, serum amyloid P-component, and immunoglobulins) were identified. Direct binding of LRP1b ectodomains to fibrinogen was verified by co-immunoprecipitation. ApoE-carrying lipoproteins were shown to bind to LRP1b ectodomains in a lipoprotein binding assay. Furthermore, binding as well as internalization of very low density lipoproteins by cells expressing an LRP1b minireceptor was demonstrated. DISCUSSION: LRP1b expression in humans appears to be confined to few tissues, which could point out to specialized functions of LRP1b in certain organs. Most of the newly identified LRP1b ligands are well-known factors in blood coagulation and lipoprotein metabolism, suggesting a possible role of LRP1b in atherosclerosis.

The LDL receptor family comprises seven known receptors in mammals. All members share a common structure with a typical arrangement of ligand binding repeats and epidermal growth factor (EGF) receptor homology domains in their extracellular part. They fulfill a variety of different functions, ranging from the classical role in receptor-mediated endocytosis to integral roles in cellular signaling pathways [1]. Low-density lipoprotein receptor-related protein 1b (LRP1b) is one of three very large receptors of the family with a size of approximately 600 kDa and shares the greatest degree of homology (60% identical amino acid residuals) with LRP1. The unusually large LRP1b gene was discovered during studies of lung cancer cell lines, where alterations of the LRP1b gene, as e.g., the deletion of individual exons, were frequently observed. Therefore, LRP1b was originally termed LRP – deleted in tumors (LRP–DIT) and was postulated as a putative tumor suppressor [2]. The LRP1b mRNA encoded by 91 exons codes for a protein of 4599 amino acids, which comprises four ligand binding domain regions in the extracellular part. The expression of LRP1b in the mouse has been described previously. Murine LRP1b expression is highest in the brain, where the full-length receptor and an alternatively spliced form lacking exon 90 are present. The alternatively spliced form is also present in the adrenal gland and in the testis [3]. The expression of LRP1b in human tissues is controversial. In the first description of the receptor, a broad expression of LRP1b was reported (kidney, brain, lung, heart, liver) [2]. In a subsequent paper, LRP1b transcripts were reported to be present in human brain, thyroid gland and salivary gland only [4]. Later, LRP1b expression was reported in several human tissues (brain, adrenal gland, salivary gland, testis, skeletal muscle, lung, kidney, small intestine, prostate, thymus, heart, stomach) [5]. Independently, LRP1b expression was described in normal human urothel, smooth muscle cells of the arterial wall and recently in normal human gastric tissue [6-8]. The homologous LRP1 molecule is a broadly expressed multiligand receptor with more than 30 known ligands comprising apo E carrying lipoproteins, proteases/antiprotease complexes, and other molecules [9]. Some of these ligands, namely the receptor-associated protein (RAP), urokinase plasminogen activator (uPA), uPA receptor, plasminogen activator inhibitor type-1 (PAI-1), gp96, and pseudomonas exotoxin have also been shown to bind to LRP1b [4,10]. In addition, well known chaperones (RAP, gp96, sacsin, nedd7) and other proteins (synaptotagmin, GPR69a, laminin receptor precursor, beta-amyloid precursor protein) have been identified as LRP1b ligands [3,11]. Presently, the physiological role of LRP1b and possible functions of the receptor in diseases like cancer and atherosclerosis are largely unknown. In contrast to other LDL receptor family members [12], mice carrying a truncated form of LRP1b lacking the transmembrane region and therefore exclusively expressing a secreted extracellular domain appear phenotypically normal with normal plasma lipids [3]. Different from this finding, mice with more proximal truncations of the receptor are embryonically lethal, suggesting important functions of the extracellular part of LRP1b [13]. As stated above, LRP1b expression has been described in smooth muscle cells of the arterial wall. In addition, LRP1b was shown to modulate the expression of the uPA receptor and of the platelet derived growth factor receptor β in endothelial cells, thereby affecting the migration of smooth muscle cells [7,14]. These findings raise the possibility that LRP1b may play a role in the pathogenesis of atherosclerotic lesions. To further investigate the physiological functions of LRP1b, we set out to investigate LRP1b expression in human tissues and to search for circulating extracellular receptor ligands in human plasma.

Methods

For detailed methods, please refer to the online-only supplemental methods.

Cell lines and cell culture

HEK 293 cells (ATCC, Middlesex, UK), HEK 293S cells (ATCC) CHO cells (ldl-A7 subclone) [15], CHO ldl-A7 stably transfected with a plasmid coding for a LRP1b region IV minireceptor [3] (TR 3517) and hepatoma HepG2 cells (ATCC) were cultured in DMEM. HEK 293S cells were stably transfected with plasmids coding for LRP1b ectodomains II, III, and IV [3], grown in the same medium and selected with G418.

RNA isolation and cDNA preparation

Total RNA was isolated from cells using RNA STAT 60 from TEL-TEST Inc. (Friendswood, TX, USA). Two commercial panels of normal human total RNA from 20 different tissues each were purchased from Ambion (Austin, TX, USA; panel A) and Clontech (Mountain View, CA, USA; panel B), respectively, and transcribed into cDNA.

Conventional RT-PCR

Conventional RT-PCR was performed with specific primers to amplify the cytoplasmic part of LRP1b (exons 89–90). Control reactions were carried out using glyceraldehyde-3-phosphate dehydrogenase (G3PDH) specific primers. Primer sequences are given in the online-only supplemental methods.

Real time RT-PCR

For real time PCR, primers corresponding to the extracellular part of LRP1b close to the transmembrane region (exon–exon junction 87/88) were used. Results are shown as relative copy number in relation to a housekeeping gene (large ribosomal protein P0, RPLP0 [16]). Primer sequences are given in the online-only supplemental methods.

Affinity chromatography with immobilized LRP1b ectodomains

Affinity chromatography of potential ligands from human plasma was performed using immobilized recombinant LRP1b ectodomains as described previously [3]. Eluates were separated on 4–15% SDS gels and individual bands were cut out and analyzed by mass spectrometry.

Co-immunoprecipitation

Purified candidate ligands were incubated with recombinant LRP1b ectodomains in the presence of the respective antibodies at 4 °C over night. After centrifugation, the supernatants were removed and the pellets were washed and separated on 4–15% SDS gels. Western blotting was carried out using the M2 anti FLAG antibody (Sigma–Aldrich) to detect bound LRP1b ectodomains.

Lipoprotein binding assay

Very low density lipoproteins (VLDL) and high density lipoproteins (HDL) were isolated by ultracentrifugation from human plasma and labeled with biotin. Anti-FLAG agarose immobilized LRP1b ectodomains were incubated over night with different amounts of biotinylated lipoproteins. Binding was then quantified after incubation with streptavidin peroxidase.

Immunocytochemistry

CHO ldl-A7 and TR 3517 cells were incubated with biotinylated VLDL (250 μg/ml) at 37 °C or 4 °C for 4 h. After washing, fixed cells were stained with peroxidase-labeled streptavidin.

Immunofluorescence

CHO ldl-A7 and TR 3517 cells were fixed with 4% paraformaldehyde and stained with a rabbit polyclonal antibody directed against the carboxyl terminus of LRP1b described previously [3].

Results

LRP1b expression in human tissues

To detect LRP1b transcripts in human tissues, we applied both conventional as well as real-time RT-PCR using different primer pairs to amplify an intracellular (conventional RT-PCR) and an extracellular (real-time RT-PCR) region of the LRP1b receptor, respectively. As a positive control, we used CHO cells stably transfected with a region IV LRP1b minireceptor (TR 3517 cells). HEK 293 cells expressing endogenous LRP1b were used as a weak positive control. As a negative control, the HepG2 cell line which does not express LRP1b mRNA was used. Due to the design of the real time primers around exon junction 87/88, false-positive signals by contamination with genomic DNA can virtually be excluded. The integrity of the RNA was proven by control amplifications of G3PDH and RPLP0, respectively. Two commercially available RNA panels of normal human RNA from 20 different tissues were tested. As depicted in Figs. 1 and 2, LRP1b mRNA is abundantly expressed in human brain, thyroid gland, and skeletal muscle. Minor amounts of LRP1b mRNA were detected in testis in one, but not in the other RNA panel. In addition, LRP1b was found abundantly expressed in other central nervous system tissues (cerebellum, fetal brain, spinal chord, Table I of the online-only supplement). In all other tissues tested (adipose tissue, bladder, colon, esophagus, fetal liver, heart, kidney, liver, lung, placenta, prostate, salivary gland, small intestine, spleen, stomach, thymus, trachea, and uterus) virtually no LRP1b transcripts were detected. The results of the conventional RT-PCR were confirmed by sequencing. It is noteworthy that the alternatively spliced form lacking exon 90 described previously in mouse tissues [3] was not found in any of the tissues examined. This finding suggests that a predicted truncated alternative LRP1b transcript (LRP1b 003) comprising the carboxyl terminal part of LRP1b and lacking exon 90, that has been predicted in genomic databases [17], is apparently not efficiently expressed (see supplemental Figure I online).

Identification of LRP1b ligands in human plasma

To screen for potential circulating LRP1b ligands in human plasma, we used affinity chromatography on immobilized recombinant 3×FLAG-tagged LRP1b ectodomains containing ligand binding domain regions II, III or IV, respectively. After elution with a 3×FLAG peptide, potential LRP1b ligands were separated by SDS–PAGE and subsequently identified by mass spectrometry (Fig. 3). Fibrinogen β-chain and γ-chain were identified as ligands of ectodomain IV. Several potential ligands were found that bind to ectodomain II. These include fibrinogen α-chain, histidine rich glycoprotein (HRG), clusterin, vitronectin, serum amyloid P-component (SAP), and immunoglobulin chains (IGKV 1-5 and IGHA 1; see Table II of the online-only supplement). Using this technique, no ligands binding to ectodomain III were identified. To verify binding to ectodomains, co-immunoprecipitation was carried out using soluble recombinant ectodomains and purified potential ligands (fibrinogen and His-tagged clusterin). We found that fibrinogen binds strongly to ectodomain II and in a markedly lower intensity to ectodomains III and IV (Fig. 4). In contrast, we did not observe direct binding of His-tagged clusterin to LRP1b ectodomains using an analogous co-immunoprecipitation technique (data not shown). Since clusterin is known to be present on HDL lipoprotein particles, we set up a binding assay to assess lipoprotein binding to LRP1b ectodomains. To ensure specific binding, control experiments were performed using incubation with GST-RAP, known to inhibit binding of ligands to members of the LDL receptor family. As shown in Fig. 5, purified HDL bound to ectodomain II while VLDL bound to ectodomains II and IV. No significant binding to ectodomain IV could be detected for HDL. In addition, none of the lipoproteins showed binding to ectodomain III. To extend these findings we used a cell line stably expressing an LRP1b minireceptor comprising the ligand binding domain region IV (TR 3517). As shown in Fig. 6, TR 3517 cells efficiently bound VLDL particles at 4 °C and internalized VLDL particles at 37 °C compared to the parent cell line CHO ldl-A7.

Fig. 3

Affinity purification using FLAG-tagged LRP1b ectodomains: (A) 3×FLAG-tagged LRP1b ectodomains were immobilized on anti-FLAG agarose columns. Human plasma was passed through the columns and eluates were obtained with a 3× FLAG peptide. Potential ligands were separated on 4–15% SDS gels under reducing conditions and stained with a colloidal blue staining kit. Unbound fractions (flow through) and eluates from a column carrying 3×FLAG-labeled control protein (BAP) were loaded as controls. Single bands were then cut out as indicated and proteins were analyzed by mass spectrometry. (B) The structures of the N-terminal 3×FLAG-tagged ectodomains containing LRP1b ligand binding regions II, III and IV, respectively, are shown.

Fig. 4

Co-immunoprecipitation: purified human fibrinogen was incubated with 3×FLAG-tagged LRP1b ectodomains in the presence of anti-fibinogen antibodies and GammaBind sepharose at 4 °C over night. Bound fractions were separated on 4–15% SDS gels and Western blotting was carried out with the M2 anti-FLAG antibody to detect LRP1b ectodomains. Control lanes 1–3 show the different electrophoretic mobility of ectodomains II, III and IV, whereas the co-immunoprecipitation of LRP1b ectodomains is shown in lanes 4–6. Lane 7 is a control lane without ectodomains showing two unspecific bands which are also present in the other IP lanes. A representative blot from three independent experiments is shown.

Fig. 5

Lipoprotein binding assay: M2 FLAG agarose-immobilized ectodomains II, III, and IV were incubated over night at 4 °C with biotinylated VLDL and HDL. After washing, bound lipoproteins were detected by streptavidin peroxidase followed by a color reaction. VLDL and HDL binding (–■–) are shown compared to control experiments in the presence of GST-RAP (- -□- -). Data represent at least three independent experiments.

Fig. 6

VLDL binding and internalization: CHO ldlA7 and TR 3517 cells were incubated with biotinylated VLDL (250 μg/ml) for 4 h at 4 °C (upper panels) or 37 °C (middle panels), respectively. Binding (at 4 °C) and internalization (at 37 °C) of VLDL particles was then shown by staining with steptavidine–peroxidase (200×). Efficient expression of the LRP1b minireceptor in TR 3517 cells is shown by staining with an LRP1b specific antibody (bottom panels; 400×). Scale bars represent 80 μm.

Discussion

In this report, we describe a restricted expression of LRP1b, one of the very large receptors of the LDL receptor family, in human tissues. Of the different tissues examined, consistent LRP1b expression was observed only in human central nervous system, thyroid gland, and skeletal muscle. Minor amounts of LRP1b transcripts were found in one of two mRNA samples from testis. In all other tissues including liver, lung, kidney, intestine, spleen, and urinary bladder, virtually no LRP1b transcripts could be detected. These results were obtained with two independent, highly sensitive RT-PCR methods. As stated above in the introduction, controversial reports exist regarding the expression of LRP1b in human tissues ranging from restricted [4] to broad expression [2,5]. Previously, we reported a similar restricted expression of LRP1b in the mouse (brain, adrenal gland, and testis). We cannot entirely explain the conflicting data in the literature, but the disparities may be due to different techniques, different sources of RNA and possibly alternative LRP1b transcripts that are listed in genetic databases [17]. Different from previous reports, we used two independent RT-PCR assays as well as two commercial mRNA panels from different suppliers to analyze LRP1b expression. We cannot entirely rule out very low-level expression or LRP1b in specialized epithelia or blood vessels, which may also have contributed to the equivocal results reported. However, our results of a restricted expression of LRP1b in contrast to the broadly expressed homologous LRP1 receptor, which is several fold higher expressed in virtually all tissues examined, is also supported by expressed sequence tags in the NCBI database (Table I of the online-only supplement). Since LRP1b was initially described as a tumor suppressor that is frequently inactivated in different tumors, the lack of expression in many tissues is surprising. To explain tumor-specific loss of LRP1b by mechanisms like homozygous deletions of exons, one would have to propose at least temporary expression of LRP1b during tumor development. Alternatively, LRP1b may be present at certain phases in normal development and later be silenced by epigenetic mechanisms, as has been described e.g., for esophageal tumors [18]. This hypothesis is supported by our finding of a low but reproducible expression of the receptor in human embryonic kidney cells (HEK-293) but apparent absence of LRP1b in adult human kidney tissue (Fig. 1). Interestingly, as already reported by Li et al. [5], we could not detect the alternatively spliced form of the cytoplasmatic LRP1b tail (Fig. 1) and therefore no evidence of the alternative shorter LRP1b-003 transcript predicted in the Ensembl database (Figure I of the online-only supplement). However, we cannot exclude the expression of this or other alternative LRP1b transcripts at low levels or in other tissues, as e.g., vascular smooth muscle cells. Comparing LRP1b to the broadly and several fold higher expressed, homologous LRP1 receptor, one could speculate that LRP1b with its restricted expression may fulfill specialized functions in certain organs. Using affinity chromatography and mass spectrometry, we were able to identify several potential LRP1b ligands binding to LRP1b ectodomains II and IV, but not ectodomain III. Similar binding characteristics have been described for LRP1 with ligand binding domain regions II and IV being responsible for the binding of nearly all known ligands [9]. Most of these ligands play important roles in blood coagulation, fibrinolysis, cell spreading, and adhesion. The influence of LRP1b on these mechanisms may in part explain the proposed role of the receptor in cell migration, as has been suggested by Tanaga et al. for vascular smooth muscle cells [7]. For example, all three fibrinogen chains were found to bind to FLAG-tagged LRP1b ectodomains II and IV. Fibrinogen binding to LRP1b ectodomains was demonstrated independently by co-immunoprecipitation. Interestingly, LRP1 has been shown to bind fibrinogen and to inhibit fibrinogen accumulation on cell surfaces [19]. Besides its well-known function in coagulation, fibrinogen plays important roles in cell spreading, proliferation, and angiogenesis [20]. Another ligand binding to ectodomain II is vitronectin, also known as complement S protein. Vitronectin is a plasma glycoprotein with diverse biological functions, including blood coagulation, fibrinolysis, cell adhesion, and spreading. Together with clusterin, vitronectin is able to inhibit cytolysis after binding to the terminal complement complex [21]. Vitronectin plays also a role in thrombin degradation by promoting interaction of thrombin and PAI-1 followed by endocytosis of this complex by LRP1 and megalin [22]. HRG is an abundant plasma glycoprotein with many known ligands, including Zn2+, heparin, plasminogen, and fibrinogen. HRG is supposed to act as a extracellular adaptor protein regulating physiological processes like angiogenesis, cell adhesion and migration, fibrinolysis and coagulation [23]. Another identified potential LRP1b ligand is SAP, a protein belonging to the family of pentraxins, which comprises members as e.g., C-reactive protein. SAP plays an important role in the humoral arm of innate immunity, mediating e.g., pathogen recognition and apoptotic cell clearance [24]. We were also able to identify two immunoglobulin components as potential ligands for LRP1b: IGHA1 protein, an immunoglobulin heavy chain of IGA and IGKV1-5 protein, the variable region of immunoglobulin light chains. Whether LRP1b may function as an immunoglobulin receptor has to be investigated in further experiments. One ligand playing an important role in lipoprotein metabolism and other physiological pathways is clusterin, also known as apolipoprotein J. Clusterin is known to circulate in plasma bound to HDL particles [25]. Together with vitronectin, clusterin is a component of the terminal complement complex and mediates the folding of extracellular proteins as a chaperone [26]. We were not able to verify clusterin binding to LRP1b via immunoprecipitation, suggesting indirect binding of clusterin to LRP1b, probably via binding of HDL particles or after interaction with vitronectin. However, since binding of clusterin to two other LDL receptor family members, namely megalin [27] and apolipoprotein E receptor-2 [28], has been described, we cannot exclude direct binding of clusterin to LRP1b, which might have been inhibited by the His tag of the clusterin construct used in our experiments. Members of the LDL receptor family participate in the regulation of lipid metabolism, influencing atherosclerosis, neurodevelopment, regulation of nutrients and vitamins [29]. Most LDL receptor family members have been shown to bind apoE – carrying lipoproteins, which is probably true for all these structurally related molecules [30]. To formally prove lipoprotein binding by LRP1b, we set up a binding assay using biotin-labeled human lipoproteins which revealed binding of LRP1b ectodomains to apoE containing lipoproteins (HDL and VLDL). With our cell culture experiments using LRP1b minireceptor expressing TR 3517 cells we could prove that LRP1b can indeed bind and internalize lipoproteins like VLDL. However, given the restricted tissue distribution and relatively low expression compared to the homologous LRP1, it is not likely that LRP1b plays a major role in the clearing of circulating lipoproteins from plasma. Alternatively, lipoproteins and other ligands may induce signaling events mediated by the large LRP1b cytoplasmic tail. Further studies of the expression of LRP1b in normal arteries and atherosclerotic lesions as well as the interactions between LRP1b and these ligands will be necessary to gain more insight into the possible function of LRP1b in atherosclerosis. Given the multifunctionality of the homologous LRP1 receptor with more than 30 known ligands [9], it would not be surprising to see LRP1b having multiple functions as well. It has to be shown in which LRP1b expressing tissues and cells ligand binding is taking place and which mechanisms follow after receptor–ligand interaction.

Fig. 1

Conventional RT-PCR: total RNA derived from different human tissues (RNA panel A) as indicated was transcribed with reverse transcriptase and amplified by PCR using LRP1b specific primers to amplify a 400 bp fragment of the cytoplasmic tail. Control amplifications with G3PDH primers are shown in the lower panel. Representative agarose gels of a typical experiment repeated three times are shown.

Sources of funding

This work was supported by grant P 20825-B05 of the Austrian Science Fund (FWF) to P.M.

Disclosures

None.