|
Introduction
VCAM-1 is a member of the immunoglobulin superfamily, and was first described on activated endothelium where its interaction with a4b1 on leukocytes is thought to facilitate the recruitment of these inflammatory cells during the immune response (Osborn, et al., 1989). In addition to their role in the immune system, we have found that an interaction between a4b1 and VCAM-1 is also important in skeletal muscle formation (Rosen, et al., 1992).
As reported above, VCAM-1 is strongly evident throughout the forming myocardium from as early as E7.5. Between E10 and E16, VCAM-1 expression becomes increasingly restricted to the muscular ventricular septum. This is particularly interesting because mouse embryos lacking VCAM-1 die by E12 with an underdeveloped, or in some cases no, ventricular septum (Kwee, et al., 1995). While VCAM-1 is clearly important to heart morphogenesis, it was not clear how the expression of VCAM-1 is restricted to the cardiomyocytes within the ventricular septum rather than being expressed uniformly throughout the myocardium.
The VCAM-1 promoter has three major regions: a core promoter within the first 32 basepairs; a group of cytokine responsive enhancers between -288 bp and -32bp; and several octamer sites over a kilobase upstream that can function as silencers (Iademarco et al., 1993). The 32 basepair core promoter has been shown to be active in almost all cultured cells, and contains a TATA box, and initiator, and an Interferon Regulatory Factor (IRF) binding site. While the IRF binding site is known to be essential for the activity of the core promoter in skeletal muscle (Jesse et al., 1998), little is known about its function in other lines. The region between -288 bp and -32 bp contains several NF-kB binding sites that are responsive to Tumor Necrosis Factor-a (TNFa) when transfected into endothelial cells (Iademarco et al., 1993). More enigmatic, the octamer sites between one and two kilobases upstream of the core promoter have been shown to act as repressors in unstimulated endothelium (Iademarco et al., 1993) as well as during differentiation of neuroepithelium into neurons (Sheppard et al., 1995). Based on these findings, our expectations in cardiomyocytes were to find similar core promoter activity to that seen in skeletal muscle, with differential control of expression lying either within the IRF site of the core promoter (again as seen in skeletal muscle) or within the octamer sites further upstream.
To address the question of differential expression of VCAM-1 on the myocardium, we began by culturing rat cardiomyocytes and comparing the activities of various fragments of the VCAM-1 promoter (Figure 3.3: top) upstream of the Chloramphenicol acetyltransferase (CAT) reporter gene. We also compare these results to those obtained in the H9c2(2-1) rat cardiomyocytic cell line, and find a strong correspondence between the cultures. This turned out to be important, since the H9c2(2-1) cells could be subcloned into VCAM-1 expressing and non-expressing populations, while the primary cardiomyocytes could not.
Using the subcloned populations of the H9c2(2-1) cell line, we identify the promoter elements responsible for the differential expression of VCAM-1 on cardiomyocytes, and further refine this activity to an octamer site 1180 bp upstream of the VCAM-1 transcription start site. By using gel supershift assays, we show the existence of Oct-1 in protein complexes bound to these octamer sites, and through further studies find Bob-1 to be present in these same complexes. Using antibodies to immunolabel embryonic mouse hearts, we demonstrate the colocalization of Bob-1 with VCAM-1 expression throughout cardiogenesis. Most importantly, we show that overexpression of Bob-1 in cardiomyocytes results in increased VCAM-1 expression, and we show that this overexpression not only derepresses of the octamer sites on reporter constructs, but increases activation when these site are present.
Experimental Procedures
Antibodies. M/K-2 (Miyake, et al., 1991; diluted 1:2) is a rat monoclonal antibody against mouse VCAM-1, and 5F10 (Chuluyan, et al., 1995; diluted 1:200) is a mouse monoclonal antibody against rat VCAM-1. PS/2 (Miyake, et al., 1991; diluted 1:5) is a rat monoclonal antibody against mouse a4. JLT-12 (Sigma-Aldrich; diluted 1:200) is a mouse monoclonal antibody against rabbit troponin T. MY-32 (Sigma- Aldrich; diluted 1:600) is a mouse monoclonal antibody against rabbit skeletal myosin heavy chain. Rabbit antisera against human FN (Collaborative Research; diluted 1:75) was also used. Bob-1, Brn-2, Oct-1, Oct-2, and Skn-1 a/i (Santa Cruz; diluted 1:1000) are rabbit antisera, and Oct-1(Ab-2) and Oct-1(Ab-3) (CalBiochem; diluted 1:500) are goat antisera. Fluorescein-labeled goat anti-rat and goat anti-mouse antibodies (both Jackson Immunochemicals; diluted 1:200) and rhodamine-labeled donkey anti-rabbit antibodies (Jackson Immunochemicals; diluted 1:200) were used as secondary antibodies for immunofluorescent detection, and horseradish Peroxidase-linked goat anti- mouse and donkey anti-rabbit antibodies (Jackson Immunochemicals; diluted 1:10,000) were used as secondary antibodies in conjunction with an ECL kit (Amersham) for chemiluminescent detection.
Histology. A colony of outbred mice was maintained by the laboratory. Breeding females were checked daily for plugs, and positive mice were separated and the day their plugs were found was recorded as embryonic day 0 (E0). Female mice came to term and dropped at E19. Timed-pregnant Sprague-Dawley rats were purchased from Taconic Farms. Embryos and tissue samples were covered in O.C.T. compound (Miles) and quick frozen in liquid nitrogen. A cryostat was used to cut 15 µm section which were mounted on Superfrost Plus slides (Fisher), and stored at -70°C. Sections were allowed to dry and fixed in methanol at -20°C, then rehydrated in phosphate-buffered saline (PBS). Fixed sections were then blocked with PBS containing 2% fish gelatin (Sigma-Aldrich) for 1 hour, followed by incubation with the indicated antibodies, diluted in 2% fish gelatin, overnight at 4°C. These slides were then washed several times in PBS before incubation with the appropriate secondary antibodies, also diluted in 2% fish gel, for 1 hour at room temperature. The slides were then covered in Vectashield (Vector Labs), coverslipped, and sealed with nail polish. Finished slides were examined microscopically under a Zeiss Axiophot epifluorescent microscope using both fluorescein and rhodamine filters, and pictures were taken on Kodak black and white TMAX film ASA 400 to document the data.
Tissue Culture. Timed-pregnant (E18) female Sprague-Dawley outbred rats (Taconic Farms) were euthanized and their embryos removed. Cells were isolated by a modification of Engelmann, et al., 1990. Hearts were collected from embryos, bisected with iris scissors, and rinsed in sterile Ad_s buffer (isotonic saline with 20 mM HEPES, 0.5 mM MgSO4, 0.1 % glucose, pH 7.4). Cardiac fragments were then repeatedly incubated in a shaker/incubator at 37°C in Ad_s buffer containing 150 units per mL Collagenase Type 2 (Worthington), to break apart the extracellular matrices. Every 20 minutes, supernatants were collected by decanting and collagenase solution was added back to the remaining heart tissue. The collected supernatants were centrifuged at 1000g for 5 minutes, and the pelleted cells were then resuspended in fetal calf serum (BioWhitaker) and pooled. Total cellular yield from the collagenase treatment was then placed over a pre-formed Percoll gradient (Pharmacia) and centrifuged at 2000g for 30 minutes to separate cardiomyocytes from other cell types. The cardiomyocyte layer was removed to a separate tube and washed several times in Ad_s buffer before being resuspended finally in Ham_s F-12 medium (Sigma) containing 10% fetal calf serum and 1% Penicillin/Streptomycin (Mediatech), and plated on gelatinized Petri dishes. After the cells attached to the substratum (approximately 3 days), the medium was replaced with fresh F-12 containing 10% fetal calf serum, 1% Penicillin/Streptomycin. Cardiomyocyte cultures were stable under these conditions for more than 2 weeks.
H9c2 (2-1) is a rat embryonic cardiomyocyte cell line (Kimes and Brandt, 1976) with a phenotype intermediate between cardiac and skeletal muscle (Hescheler, et al., 1991; Jin et al., 1995; Mejia-Alvarez et al., 1994) that was obtained from ATCC. H9c2 (2-1) cells were grown in DMEM containing non- essential amino acids, 10% FBS, and antibiotics including Penicillin and Streptomycin.
Plasmid Constructs. Fragments of various size from the VCAM-1 promoter were placed in front of the chloramphenicol acetyltransferase gene (CAT). The -2.1, -288, and -32 VCAM-1 promoter constructs and the octamer- pSV fusion promoter constructs have been described (Iademarco et al., 1992), and the -32 promoter mutants have also been described (Jesse et al., 1998). As a positive control for transfection, parallel assays were carried out using RSV-CAT, a construct containing the Rous Sarcoma Virus promoter elements in front of CAT that is ubiquitously active in all cell types. The Bob-1 mammalian expression vector, pCGN-OCA-B, and the corresponding empty vector, pCGN, were the generous gifts of Winship Herr (Babb et al., 1997). The Bob-1 mammalian expression vector, pCMV/neo/OCA-B, was the generous gift of Yan Luo (Luo et al., 1992).
Transfections. Once cells had attached to the substratum, about three days after plating, they were transfected for 12 hours with 10 µg of the reporter plasmid and 10 µg of carrier DNA, using the previously described calcium phosphate technique (Iademarco, et al. 1992). 48 hours after transfection, cells were scraped, and protein extracts were collected. Relative CAT activity was measured in 10 µg of each extract. CAT assays were allowed to proceed for 9 hours, and the extent of acetylation of [14C]chloramphenicol was determined by thin-layer chromatography and subsequent autoradiography.
For stable integration of constructs into cell lines, cells were transfected as above using the calcium phosphate technique, and then placed in growth medium containing 1 µg/ml G418 (Sigma). Survivors were passed once, and individual colonies were trypsinized and placed on Gelatinized 60 ml tissue-culture plates. The resulting clonal lines were grown under selection medium and assayed for protein expression by Western blot analysis.
Western Blot Analyses. For transcription factors, whole cell extracts were prepared from cultured cells or fresh tissues by quick freezing on dry ice and resuspending in Lysis Buffer (10 mM HEPES, 400 mM KCl, 1 mM EDTA, 0.5 mM DTT, 5% Glycerol, pH 7.9 with PMSF, Aprotinin, Leupeptin, and Pepstatin added to inhibit protease degradation). Alternatively, for membrane receptors, whole cell extracts were prepared from cultured cells or fresh tissues by resuspension in 10 mM Tris/0.1% SDS. Whole cell extracts were boiled in Laemmli sample buffer and then loaded onto SDS-polyacrylamide gels and electrophoresed. These samples were then transfered to PVDF membranes, which were probed with antibodies.
Gel Shift Assays. Whole cell extracts were prepared as above. 10 µg of whole cell extract was assayed for binding using 32P- 5_ end-labeled, double stranded oligonucleotides. The probe for the -1916 octamer site was 5'_TAGAAAATATAGGCATATTAATCA_3'; the probe for the -1554 octamer site was 5'_GTAGTGAATTTACATGATGATGA_3'; and the probe for the -1180 octamer site was 5'_TTCATGCCGTATTTACATATTATTG_3'; and the mutant octamer probe was 5'_GTAGTTAATTATTATGATGATGA_3' (underlined bases represent putative octamer sites). For gel shift assays, whole cell extracts were incubated with probe on ice for 30 min in Binding buffer (10 mM Tris, 50 mM NaCl, 1mM EDTA, 1 mM DTT, 5% glycerol, pH 7.9). For supershift analyses, whole cell extracts were incubated with 2 µl of antibody overnight at 0°C prior to the addition of probe. The sequences of the oligonucleotides used have been reported previously (Iademarco et al., 1992).
Immunoprecipitation Assays. Whole cell extracts were prepared as above. These were mixed with IP buffer (50 mM HEPES, 250 mM NaCl, 1mM EDTA, 1mM DTT, 0.5% NP-40, X µg/ml dI/dC, pH 7.4) and 2 µl anti- Oct1 antibody. Samples were incubated at overnight 4°C. Bound complexes were then precipitated with Protein-A/Sepharose beads (Gibco- BRL) at 4°C for 45 minutes. The beads were washed several times with Lysis buffer before being boiled in Laemmli sample buffer to release the proteins. Immunoprecipitated proteins were detected by Western blot analysis.
Protein-DNA Complex Precipitations. Whole cell extracts were prepared as above, and mixed with Binding buffer (as above) and a biotinylated form of the -1180 double-stranded oligonucleotide, and then incubated on ice for 2 hrs. Protein/DNA complexes were precipitated with streptavidin/agarose beads (Invitrogen) on ice for 2 hrs. The beads were then washed several times in Binding buffer before proteins were released by boiling in Laemmli sample buffer and detected by Western blot analysis.
Results
Differential expression of VCAM-1 on myocardial cells is maintained in culture. As noted above, expression of VCAM-1 decreased on the compact myocardium and increased on the interventricular septum between E14 and birth. To determine whether this differential expression of VCAM-1 would be maintained in culture, cardiocytes were isolated from E18 embryonic rat hearts (VCAM-1 expression in the E18 rat is identical to that seen in the E14-E16 mouse, data not shown). The protocol used for the isolation of myocardial cells has been established in the rat, and rats were used because the numbers of cells recoverable from mice are limited. Antibodies against Troponin T, a late marker for muscle differentiation, labeled all cardiomyocytes but none of the few contaminating fibroblasts (Figure 3.1a). Most of the myocytes expressed relatively low levels of VCAM-1, with a smaller percentage (approximately 10%) expressing high levels of VCAM-1, presumably representing cells from the interventricular septum (Figure 3.1b). Expression of VCAM-1 was undetectable on cells from the non-myocytic fraction removed during percoll enrichment (data not shown). These results suggest that the myocardial cells maintained the differential expression of VCAM-1 in culture.
|
Differential expression of VCAM-1 on rat cardiomyocytes is recapitulated in H9c2(2-1) cells. Before testing for VCAM-1, expression of Troponin T on H9c2(2-1) cells was examined. The intense expression seen on some but not all of the cells demonstrated that this line represents a heterogeneous population (data not shown). As with the primary cardiomyocytes, most of the H9c2(2-1) cells expressed low levels of VCAM-1, while a small subpopulation (approximately 15%) expressed much higher levels. To distinguish between these subpopulations, we subcloned H9c2(2-1) cells which expressed high levels of both VCAM-1 (Figure 3.2a) and Troponin T (Figure 3.2b), and cells which expressed Troponin T (Figure 3.2d) but not VCAM-1 (Figure 3.2c).
|
Differential expression of VCAM-1 on cardiomyocytes is controlled by octamer sites. Primary rat cardiomyocytes and H9c2(2-1) cells were each transfected with a panel of reporter constructs including different fragments of the VCAM-1 promoter driving CAT. Assays of CAT activity from these transfections showed differential expression of CAT based on the inclusion of upstream elements from the VCAM-1 promoter region (Figure 3.3). Another construct using the core promoter region from the Rous sarcoma virus (RSV) to drive the CAT reporter gene gave the expected levels of expression, indicating reasonable transfection efficiency and acting as a positive control for the assay. The -32 bp promoter, when placed in front of CAT, gave activity equivalent to uncontrolled basal level expression in all cells tested. The -2.1 kb promoter, when placed before CAT, yielded almost no activity in cardiomyocytes (as did a -7.0 kb fragment, data not shown), although the -288 bp promoter was active in these cells, almost equivalent to the -32 bp promoter(Figure 3.3, black bars). Similar activities were observed in transfected H9c2(2-1) cells, using the heterogeneous population (Figure 3.3, grey bars). The VCAM-1 negative, subcloned line derived from the H9c2(2-1) cells showed the same activity as the mixed population. Conversely, the -2.1 kb and -288 bp promoters were almost equally active in the VCAM-1-selected subclones of H9c2(2-1) (Figure 3.3, white bars), similar to the activities of these constructs in the C2C12 skeletal myoblast cell line (Jesse et al., 1998).
|
Demarcation of the elements responsible for repression of promoter activity. To determined which of the octamer sites might play a role the repressive activity of the -2.1 kb to -288 bp fragment, artificial constructs were generated which place subfragments of this region directly upstream of a CAT reporter construct driven by the core promoter region of simian virus 40 (SV40). Constructs containing the -1180 octamer site were able to repress SV40 promoter activity when transfected into H9c2(2-1) cells, while constructs containing only the -1554 octamer site could not (Figure 3.4). This repression was seen in both the mixed population (see Figure 3.10c) and the VCAM-1 negative subclones (Figure 3.4, black bars) but not in the VCAM-1 positive subclones (Figure 3.4, white bars), suggesting specificity of the repression.
|
Protein binding activities of the octamer elements upstream of VCAM- 1. To further examine the relevance of the upstream octamer sites, Gel Shift Assays were performed using each of the putative octamer elements as probes. Using whole cell extract from the H9c2(2-1) cells, each of the three octamer elements, -1180, -1554, and -1916, was able to form a complex with the same mobility as seen using a known octamer element from the immunoglobulin heavy-chain (IgH) promoter (data not shown), while an oligonucleotide containing a mutated form of this site could not. Cross competition with unlabeled probes showed that the -1180 element has the highest affinity for the complex, easily outcompeting the -1554 and -1916 probes, while being only partially competed away by unlabeled oligonucleotides containing either of the other two elements (Figure 3.5).
|
Expansion of gel shift assays to examine murine heart development. To understand the relevance of the gel shift data, it was important to compare the H9c2(2-1) cell line to murine cardiac muscle. Extracts were collected from the hearts of adult rats, as well as from the dissected hearts of rat embryos at E11 and E16. Gel shift analyses using the -1180 probe showed the presence of the same specific band in each of these extracts as was seen in the H9c2(2-1) cell extracts, although the band seemed less intense in the extracts from the adult heart (Figure 3.6). Furthermore, a second, faster migrating band that was evident in the H9c2(2-1) extracts was only apparent in extracts from the E16 hearts. To examine the possible importance of this faster migrating band, its affinity as compared to the slower major band was examined by diluting the - 1180 probe relative to the H9c2(2-1) extract. Interestingly, this faster migrating band showed a higher affinity for the -1180 probe than the major band, actually increasing in intensity as the probe was diluted 1:1000 relative to normal gel shift conditions.
|
Identification of proteins involved in binding the -1180 element. To identify the proteins present in the major gel shift band, antibodies against the known POU proteins, Skn-1 a/i, Brn-2, Oct-1, and Oct-2, were used to supershift the complex. Only one antibody, directed against Oct-1, was successful at supershifting the -1180 major band in H9c2(2-1) extracts (Figure 3.7a). Further gel shift assays demonstrated the ability of this antibody to supershift extracts from each of the heart tissues (data not shown). Furthermore, Western blot analyses using this and other antibodies against Oct-1 confirmed its presence in embryonic heart tissue as well as in the H9c2(2-1) cell line (Figure 3.7b).
|
Identification of a binding partner for Oct-1 in cardiomyocytes. The two best characterized binding partners for Oct-1 are VP16 (a strong transcription activator from herpes simplex virus) and Bob-1 (B-cell Oct-1 Binding factor-1, also called OCA-B and OBF-1). Since VP16 is a viral protein, it made more sense to examine the possibility that Bob-1 was involved in the regulation of VCAM-1 expression. Although preincubation with an antibody against Bob-1 had no effect on the mobility of any band in gel shift assays using each of the extracts, Western blot analyses of these extracts showed a strong presence of Bob-1 in embryonic heart extracts as well as in the H9c2(2-1) cell line (Figure 3.7b).
Characterization of the expression of the Oct-1 and Bob-1 transcription factors during murine cardiogenesis. Data collected from immunostaining of mouse hearts from E8 through E18 (Figure 3.8) demonstrate that VCAM-1 is expressed early throughout the myocardium but is later restricted to the compact myocardium and interventricular septum before being completely lost postnatally. These data also show Bob-1 expression throughout the myocardium. While this expression correlates with early VCAM-1 expression, Bob-1 expression remains constant on the entire myocardium throughout later cardiogenesis, even as VCAM-1 expression becomes progressively restricted.
|
Identification of Bob-1 binding to the -1180 octamer element. To confirm the ability of Bob-1 and Oct-1 to interact with the -1180 octamer element, an oligonucleotide probe similar to those used for the previous gel shift assays was biotinylated and used to precipitate proteins extracted from embryonic murine heart tissue. In these experiments, the relevant protein extracts were incubated in the presence or absence of specific competitors and then with the biotinylated probe. The results from these blots demonstrate that both Oct-1 and Bob-1 form protein:DNA complexes with the -1180 element at endogenous levels found in extracts prepared from either embryonic heart tissue or cultured H9c2(2-1) cells (Figure 3.9).
|
Demonstration of Bob-1 regulation of VCAM-1 promoter activity. The Bob-1 mammalian expression vector, pCGN-OCA-B, was transiently cotransfected into the H9c2(2-1) rat cardiomyocyte cell line along with the above CAT reporter constructs, in order to assess the effect of Bob-1 overexpression on the activity of the VCAM-1 promoter. The overexpression of Bob-1 was able to overcome the repression of the VCAM-1 promoter by the upstream octamer sites, relative to the CAT activity from cells cotransfected with the reporters and the empty pCGN vector (Figure 3.10A).
To further examine this derepression, the Bob-1 expression construct, pCMV/neo/OCA-B, was stably transfected into the H9c2(2-1) cell line. Western blots made from extracts of these cell lines demonstrate an increase in endogenous VCAM-1 protein expression in clones expressing the highest levels of Bob-1 (Figure 3.10B). Two such lines, H9B-1 and H9B-10, were transfected with the reporter constructs containing portions of the VCAM-1 silencer domain upstream of the SV40 promoter described above. Like the VCAM-1 positive subclones, these clonal lines showed no repression compared to normal H9c2(2-1) cells (Figure 3.10C). On the contrary, the H9B- 10 line showed overactivation of constructs containing either both octamer sites or the sites without intervening sequences.
|
To continue examining this element, gel shifts were performed using extracts from the stable expression lines and compared to extracts from normal H9c2(2-1) cells as well as subcloned H9c2(2-1) populations that uniformly expressed VCAM-1. There appeared to be much less of the slowest migrating band present in extracts from either the Bob-1 expressing clones or the VCAM-1 positive subclones, as compared to normal H9c2(2-1) cells (Figure 3.11). In all cases, an antibody against Oct-1 was able to supershift the specific bands (data not shown), indicating that the complexes formed by Bob-1 overexpression and the complexes driving expression in the VCAM-1 positive subclones contain Oct-1. Taken together, these data suggest that the Bob-1 expression seen concurrently with VCAM-1 during cardiogenesis is sufficient to drive VCAM-1 expression in the myocardium.
|
Discussion
Previously, we defined the expression of VCAM-1 on the myocardium during murine embryogenesis, and our findings were supported by VCAM-1 knock-out mice (Gurtner et al., 1995; Kwee et al., 1995) which failed to develop complete ventricular septa. Here we have examined the elements responsible for the expression of VCAM-1 on the developing heart muscle. We determined that the differential expression of VCAM-1 could be maintained in culture on primary cardiocytes from E18 embryonic rat hearts, and found the H9c2(2-1) cell line to be sufficiently similar in its expression of VCAM-1 to serve as a model for embryonic myocardium. We were able to subclone the H9c2(2-1) cells, and isolate colonies that were either positive or negative for VCAM-1 expression, giving us an opportunity not available with the primary cardiomyocytes to examine this difference in expression.
Primary rat cardiomyocytes, H9c2(2-1) cells, and each of the subcloned lines were transfected with the VCAM-1 promoter CAT constructs, and the data from these experiments demonstrate that while the -32 bp promoter yielded basal level expression, a region between -2.1 kb and -288 bp upstream of the VCAM-1 promoter was acting as a silencer in the mixed populations, as well as in the VCAM-1 negative cells. Conversely, the -2.1 kb fragment and -288 bp fragment were almost equally active in the VCAM-1 positive subclones, similar to the activities of these constructs in the C2C12 skeletal myoblast cell line (Jesse et al., 1998).
These data suggest that differential expression of VCAM-1 in cardiomyocytes is under the control of elements between -288 bp and -2.1 kb which repress promoter activity during differentiation from myoblasts into myocytes. Several potential octamer sites have been identified previously in this region of the VCAM-1 promoter in the context of neural differentiation of P19 cells (Sheppard, et al., 1995). Octamer sites have been characterized as the binding sites for POU-domain homeotic transcription factors, which are known to act as repressors of transcription of several gene targets (Jacobson, et al., 1997). Conversely, there are other elements within this region which weakly match Retinoic Acid Response Elements (RARE's), although expression of the retinoic acid receptor, RARa, in the heart precedes the downregulation of VCAM-1 on the myocardium by several days. It is noteworthy, that experiments knocking out RARa result in the malformation of the compact myocardium and ventricular septum (Kastner, et al., 1994), and our own work in P19 embryonal carcinoma cells showed induction of VCAM-1 expression several days after retinoic acid treatment (Sheppard, et al., 1995); suggesting that retinoic acid and VCAM-1 expression are part of a common differentiation mechanism.
Attempts at identifying the elements responsible for the basal activity found in the first 32 bases of the VCAM-1 promoter, distinguished control of expression in heart muscle from that of skeletal muscle. Unlike skeletal muscle which depends mostly on an IRF binding site for the activity of the core promoter, cardiomyocytes appear to rely solely on the Inr site (data not shown). However, mutational analysis shows a slightly different set of bases to be important in H9c2(2-1) cells compared to in C2C12 cells. Our lab has shown that Inr usage is related to IRF binding in skeletal myoblasts (Jesse et al., 1998), and we suggest that the lack of IRF binding in cardiomyocytes is the source of the change in Inr usage. Furthermore, we have tested the NF-kB sites between -288 bp and -32 bp, and found them to have no effect on the activity of the promoter, even when transfected cells are treated with TNFa (data net shown). These results may warrant further investigation to understand the differences in core promoter activity between skeletal and cardiac myoblast cell lines.
By truncating the upstream fragments of the VCAM-1 promoter and placing them in front of an SV40 promoter, we have mapped the silencer activity to two or more octamer elements. Only one of these elements, at -1916 bp, completely matches the consensus for an octamer site (Jacobson, et al., 1997), while two other sites, at -1554 and -1180, have only one base pair mismatched. Our constructs demonstrated a reproducible repression of SV40 promoter activity by the -1180 octamer site but not by the -1554 octamer site (the -1916 site was not tested) when transfected into H9c2(2-1) cells. This repression was seen in both the mixed population and the VCAM-1 negative subclones but not in the VCAM- 1 positive subclones, suggesting specificity of the repression.
We used the elements identified this way as probes for gel shift assays using extracts from the H9c2(2-1) cells. These experiments showed that while each of the three octamer elements formed the same complexes, the -1180 element demonstrated the highest affinity for these complexes. Similarly, when extracts were made from hearts dissected from adult rats and rat embryos at E11 and E16, gel shift analyses using the -1180 probe showed the presence of the same specific bands as were seen in the H9c2(2-1) cell extracts. Also, we found that antibodies against Oct-1 could supershift the -1180 major band in each of the extracts tested. Unfortunately, our Bob-1 antibody was unable to shift or block any of the complexes visible on our assays - probably because of inaccessibility of the epitope in the particular complexes present in heart tissue.
However, western blot analyses of each extract using antibodies against Oct-1 and Bob-1 confirmed the presence of both in each tissue as well as in the H9c2(2-1) cell line. To confirm this, we immunostained mouse hearts from E8 through E18 with antibodies against VCAM-1, Oct-1, and Bob-1. As is evident from the staining, Bob-1 is expressed evenly throughout the myocardium, while VCAM-1 becomes increasingly restricted to the interventricular septum. Concurrently, Oct-1 appeared to be expressed at higher levels in tissues surrounding the myocardium than on the myocardium itself. Although Bob-1 expression correlated with the expression of VCAM-1 during early cardiogenesis, it is interesting to note that Bob-1 expression remained constant throughout cardiogenesis, even as VCAM-1 expression became progressively restricted, suggesting that while Bob-1 may well be responsible for the upregulation or enhancement of VCAM-1 expression in the forming myocardium, it is the influence of a repressive transcription factor separate from Bob-1 (but binding the same octamer element) that effects its progressive downregulation.
With Bob-1 expressed concurrently with VCAM-1 on embryonic heart and Bob-1 protein interacting with the VCAM-1 promoter, it became evident that this transcription factor could be directing VCAM-1 expression in some way. To test this, we overexpressed Bob-1 in H9c2(2-1) cardiomyocytes, both transiently and stably, and found that this overexpression could overcome the repression of the VCAM-1 promoter by the upstream octamer sites, as compared to control constructs. Extracts from Bob-1 overexpressing cells demonstrated an increase in endogenous VCAM-1 protein expression, as well as altering the pattern of protein complexes on our gel shift assays, similar to the pattern seen in extracts from our VCAM-1 positive H9c2(2-1) subclones. Also, cells transfected with the Bob-1 expression vector, either transiently or stably, showed derepression of the octamer sites when placed upstream of the SV40 promoter. These data demonstrate that Bob-1 expression is sufficient to drive the expression of VCAM-1 in cardiomyocytes.
The ability of Bob-1 to activate VCAM-1 expression in cardiomyocytes and its overlapping expression on the myocardium during development, point to Bob-1 as a major factor in the upregulation of VCAM-1 during the early stages of cardiogenesis. The fact that Bob-1 remains high throughout the myocardium while VCAM-1 becomes increasingly restricted to the interventricular septum suggests that while Bob-1 is responsible for the upregulation of VCAM-1 throughout the myocardium, there must be another, as yet undefined, transcriptional repressor that binds the same octamer sites as the Bob-1/Oct-1 complex. This repressor appears to both deactivate Bob-1 and repress the basal promoter. The presence of near-consensus nuclear hormone binding sites, in the form of RARE's, within the -2.1 kb to -288 bp fragment offers a tempting model for repression of the VCAM-1 promoter in non-septal myocardial cells: however, our own studies using fragments of this region suggest that the silencer activity lies within the octamer sites themselves, and is not associated with these putative RARE's.
It is also noteworthy that Bob-1 expression has not been previously described in the heart (Ryan et al., 1997). Although it is possible that Bob-1 expression in the myocardium exists solely to promote VCAM-1 expression, it is more than likely that other proteins involved in heart formation are influenced by Bob-1 expression. While Bob-1 knock-out mice have been generated, these mice displayed little anatomical abnormality outside of reduced germinal centers in the lymph nodes (Schubart et al., 1996; Kim et al., 1996). While VCAM-1 is expressed in the lymph nodes (Sheppard et al., 1994), it is more likely the lack of mature B-cells that is responsible for this phenotype. It has been suggested, regarding this lack of gross phenotypic changes in the Bob-1 knock-out mice, that there may be a redundant factor equivalent to Bob-1 which can compensate for its loss (Kim et al., 1996). It is also possible that removing Bob-1 may alter expression of Oct-1 such that this factor can compensate for its diminished ability to activate promoters. We suggest that a better way to test the importance of Bob-1 in development would be to overexpress it. Our data indicate that overexpression of Bob-1 in the developing myocardium could have profound effects on heart formation.
Chapter 1 | Chapter 2 | Chapter 3 | Chapter 4 | References
