Tissue‐specific regulatory regions of the PTH gene localized by novel chromosome 11 rearrangement breakpoints in a parathyroid adenoma

Abstract

Parathyroid adenomas can contain clonal rearrangements of chromosome 11 that activate the cyclin D1 oncogene through juxtaposition with the PTH gene. Here we describe such a chromosomal rearrangement whose novel features provide clues to locating elusive cis‐regulatory elements in the PTH gene and also expand the physical spectrum of pathogenetic breakpoints in the cyclin D1 gene region. Southern blot analyses of the parathyroid adenoma revealed rearrangement in the PTH gene locus. Analysis of rearranged DNA clones that contained the breakpoint, obtained by screening a tumor genomic library, pinpointed the breakpoint in the PTH locus 3.3 kb upstream of the first exon. Accordingly, highly conserved distal elements of the PTH 5' regulatory region were rearranged at the breakpoint approximately 450 kb upstream of the cyclin D1 oncogene, resulting in overexpression of cyclin D1 mRNA. Thus, PTH–cyclin D1 gene rearrangement breakpoints in parathyroid tumors can be located far from those previously recognized. In addition to expanding the molecular spectrum of pathogenetic chromosomal lesions in this disease, features of this specific rearrangement reinforce the existence of one or more novel cis‐enhancer/regulatory elements for PTH gene expression and narrow their location to a 1.7‐kb DNA segment in the distal PTH promoter. © 2010 American Society for Bone and Mineral Research.

Introduction

Parathyroid adenomas are benign neoplasms that are the most common cause of primary hyperparathyroidism. These neoplasms are monoclonal,1 indicating that they result from acquired mutations in key genes, thereby conferring a parathyroid cell with a selective growth advantage. Oncogenes are one such class of growth‐promoting genes, and they can be activated by several mechanisms, including intragenic mutations, chromosomal amplifications, and chromosomal rearrangements. Frequently, translocations or inversions of chromosomal segments cause the transcriptional activation of oncogenes by placing their coding regions in the vicinity of gene‐regulatory elements. Previous studies have demonstrated that a subset of parathyroid adenomas harbors a specific rearrangement of chromosome 11—inv(11)(p15;q13), a pericentromeric inversion—that positions the PTH gene's 5' regulatory region upstream of the cyclin D1 gene, thereby driving overexpression of cyclin D1.2, 3, 4 Analogous rearrangements of chromosomal locus 11q13 that activate cyclin D1 expression have been detected in other tumor types, including mantle cell lymphoma,5, 6 multiple myeloma,7, 8 and renal oncocytoma.9, 10, 11 The described locations of the breakpoints on chromosome 11q13 vary from a few kilobases to several hundred kilobases centromeric to the cyclin D1 gene. In the parathyroid adenomas described to date, these breakpoints occurred immediately adjacent to the 5' end of the cyclin D1 gene.2, 3, 4 Breakpoints on 11p15, at the PTH gene locus, have separated the entire 5' regulatory region, including noncoding exon 1, from the intact PTH coding region.

Cyclin D1 is overexpressed in 20% to 40% of parathyroid adenomas,12, 13, 14, 15 substantiating its role as a primary driver of parathyroid neoplasia. To further examine the effects of cyclin D1 overexpression in parathyroid cells, we developed a transgenic mouse model that mimics the effects of the rearrangement found in human parathyroid tumors.16 These mice harbor a transgene that contains 5.2 kb of the PTH gene's 5' regulatory region juxtaposed to the human cyclin D1 gene, thereby targeting overexpression of the cyclin D1 oncogene to parathyroid cells.16 This model manifests abnormal parathyroid cell proliferation and dysregulation of PTH secretion, leading to a phenotype of chronic biochemical primary hyperparathyroidism.16, 17 In addition to emphatically confirming the role of cyclin D1 as a driver oncogene in parathyroid neoplasia, these studies also provided important clues to the promoter‐enhancer elements necessary for PTH gene regulation. The ability of the 5.2‐kb region of the PTH promoter in this transgene to stimulate high‐level gene expression in parathyroid cells demonstrates that the necessary DNA sequences required for strong tissue‐specific expression of the PTH gene are contained within this segment.

Although the PTH gene has been cloned and characterized for several species,18, 19, 20, 21, 22, 23 little is known about the transcription factors and promoter/enhancer elements that regulate its expression in vivo. A major limitation to identifying such regulatory elements or confirming their functional relevance has been the lack of suitable parathyroid cell lines that faithfully represent the parathyroid chief cell phenotype, namely, the ability to synthesize and secrete PTH both at a basal level and in response to known physiologic regulators such as 1,25‐dihydroxyvitamin D₃ [1,25(OH)₂D₃] and ionic calcium. Previous examinations of the PTH promoter region have identified DNA response elements for the vitamin D receptor,24 calcium,25, 26, 27 cyclic AMP,28 and a yet unknown transcription factor.29 Recent studies have identified conserved DNA elements in the immediate upstream PTH promoter region that bind to transcription factors Sp1, Sp3, and nuclear factor Y.30, 31, 32 However, these studies have focused on the immediate upstream regions of the PTH gene and used non–parathyroid cell environments; thus the nature of the enhancer elements that regulate high‐level parathyroid tissue‐specific expression and may reside in the more distal parts of the upstream region remain unknown.

Here we report the discovery of novel chromosome breakpoint locations at both the cyclin D1 and the PTH gene regions in a parathyroid adenoma. The location of the break at the PTH gene locus provides crucial information localizing distal promoter/enhancer sequences upstream of the PTH gene.

Materials and Methods

Patient and samples

A 70‐year‐old man underwent surgical treatment for primary hyperparathyroidism, and a solitary parathyroid adenoma was resected. A portion of the tumor was flash frozen in liquid nitrogen and then stored at −80°C. Blood was obtained from the same patient as a source of his normal germ‐line DNA. Tumor and blood samples were obtained in accordance with institutional review board–approved protocols.

Analysis of chromosomal rearrangements and DNA sequencing

Genomic DNA was extracted from the tumor by standard methods, as described previously.2 For Southern blot analyses, the DNA was digested with BamHI, EcoRI, HindIII, BglII, and KpnI, separated by electrophoresis on a 0.8% agarose gel, and transferred onto nitrocellulose membranes. To detect rearrangement at the PTH gene locus, the membrane was hybridized with a 32P‐labeled DNA probe—a 745‐bp BglII fragment encompassing part of the PTH gene's exon 1 and its immediate upstream region (Fig. 1). Hybridization and wash conditions were as described previously.2

Figure 1.

Southern blot analysis of PTH gene rearrangement in parathyroid adenoma. Tumor genomic DNA (T) or control DNA from the same patient (C) was digested with the indicted enzymes. Rearrangement at the PTH gene locus was detected using a probe to the PTH 5' regulatory region. Tumor‐specific rearranged banding pattern is seen in the KpnI digest (arrow).

To clone the rearranged PTH allele, a tumor genomic DNA library was constructed using DASH II and EMBL‐3 phage vectors. Genomic DNA extracted from the tumor was digested with HindIII or BamHI and ligated into the HindIII and BamHI restriction sites of the DASH II and EMBL‐3 vectors, respectively. Ligated DNA was packaged into virions using the packaging extract and protocol provided by the manufacturer (Stratagene Corporation, La Jolla, CA, USA). Recombinant virions were plated onto a lawn of Escherichia coli on NZY agar plates to generate plaques. Plaques were transferred to nitrocellulose membranes. To identify recombinant virions that contained fragments of the rearranged PTH allele, membranes were hybridized with the 745‐bp BglII‐BglII fragment (probe 2, Fig. 2) and a 314‐bp HindIII‐PstI fragment of the PTH 5' regulatory region (probe 1, Fig. 2).

Figure 2.

Diagrammatic representation of the normal and rearranged PTH alleles along with rearranged DNA from the cyclin D1 region on 11q13 in the described tumor. Locations of the DNA probes used to screen tumor genomic libraries for rearrangement at the PTH locus are shown. The approximate extents of the inserts included in the clones isolated from the tumor genomic DNA phage libraries are shown beneath each normal or rearranged DNA diagram. Clone I represents the normal unrearranged PTH locus, and clones II and III contain DNA from each section of the reciprocal PTH‐11q13 rearrangement. Bold vertical arrows in the middle and bottom panels indicate the location of the breakpoint at which the PTH locus rearranged with 11q13 DNA. The critical 1.7‐kb region of the distal PTH regulatory region is indicated in the bottom panel. The locations of the KpnI sites used for southern blot analyses are shown on the normal allele and are situated outside the ends of clone I (diagram not drawn to scale).

Inserts from clones II and III, rearranged clones identified by screening recombinant virions, were subcloned into plasmid vectors to facilitate sequencing. An approximately 600‐bp BamHI‐StuI fragment of clone II was subcloned into plasmid vector pGEM3 and sequenced using Sp6 and T7 primers. Additionally, the entire 7.2‐kb insert from clone III was isolated and cloned into pGEM3, and the region adjacent to the PTH promoter region was sequenced.

Fluorescence in situ hybridization

Frozen tumor tissue was thawed on ice, minced, and incubated at room temperature with 0.9% sodium chloride. The tissue was fixed overnight in 10% buffered formalin phosphate, washed with 95% ethanol, and rehydrated in distilled water for 2 hours. The nuclei were extracted using Carlsberg solution [0.1% Protease XXIV (Sigma Aldrich, St. Louis, MO, USA), 0.07 M sodium chloride, and 0.1 M Tris] at 37°C for 2 hours. Nuclei then were dropped onto glass slides and fixed with a 3:1 methanol–acetic acid mixture. Screening a BAC library (Genome Systems, St. Louis, MO, USA) using PCR primers to the PTH gene's 5' region identified BAC clone 15d14 for use as a PTH locus probe. Cosmid clone 6.228 was kindly provided by Dr Ed Schuuring for use as the cyclin D1 region probe. The probes were labeled with rhodamine and fluorescein isothiocyanate (FITC), respectively, and hybridized onto tumor nuclei using standard protocols. Two hundred nuclei were evaluated for discrete colocalization of signals.

Analysis of cyclin D1 expression

For quantitative PCR analyses, total RNA was extracted from the tumor and three additional parathyroid tumors with no known PTH–cyclin D1 rearrangement. As positive control, we used RNA from the human breast carcinoma–derived cell line ZR‐75‐1, which has two‐ to fivefold amplification and marked overexpression of the cyclin D1 gene.33 To facilitate comparison between tumors by serving as a calibrator, we also analyzed cyclin D1 expression in human placental tissue. RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcribed using random hexamers. Cyclin D1 was amplified from each cDNA template sample using the following primers: forward primer 5'‐CAGAGGCGGAGGAGAACAAA‐3' and reverse primer 5'‐ATGGAGGGCGGATTGGAA‐3'. The housekeeping gene TATA‐binding protein (TBP) was used as endogenous control.34 Real‐time PCR reactions were done in duplicate on an ABI 7900HT real‐time PCR unit using the SYBR green detection and absolution quantitation method, as detailed in the Applied Biosystems' user bulletin (http://www3.appliedbiosystems.com/cms/groups/mcb_support/documents/generaldocuments/cms_041436.pdf). Expression of cyclin D1 was normalized to TBP expression. To facilitate comparison between tumors, the data are presented relative to expression levels in human placenta. Northern blot analyses for cyclin D1 were performed as described previously.3

In silico analyses

For identification of the chromosomal locus of the rearrangement partner, DNA sequences adjacent to the breakpoints were analyzed for homology with the publicly available human genome sequence35 using the BLAT program36 in the UCSC Genome Browser,37 available at http://genome.ucsc.edu/.

Cross‐species comparison of the PTH gene and its upstream promoter region were performed using GenomeVISTA, available at http://genome.lbl.gov/vista/index.shtml.38 The entire PTH gene and 5.2 kb of the upstream region were input into the program for comparison with orthologous sequences from whole‐genome assemblies of other organisms. For analysis of conserved sequences, we used the default parameters with the minimum percent conservation identity set at 70%.

We searched for putative transcription factor binding sites in the conserved region of the PTH 5' DNA sequences using the MAPPER database, available at http://bio.chip.org/mapper/mapper-main.39 To limit the number of hits, we set stringent parameters at a score of greater than 3 and an E value of less than 2.5. The higher the score, the better was the match between the target sequence and modeled binding site.39 The E value is a measure of false‐positive results. A smaller E value indicates a more significant hit.39

Results

Analysis of rearrangement at the PTH gene locus

To detect rearrangement at the PTH gene locus, genomic DNA from the parathyroid adenoma was examined by southern blot analysis using multiple restriction digests and a PTH gene probe. Tumor‐specific bands, not detected in control DNA from the same patient, documented a clonal rearrangement at the PTH gene locus in this tumor (Fig. 1). The approximately equal intensity of the normal and abnormal (rearranged) alleles also was consistent with the chromosomal abnormality being present in all tumor cells.

To clone DNA containing the rearranged PTH allele, phage libraries were constructed from the tumor genomic DNA and screened with DNA probes to the PTH upstream region. We identified three independent clones containing the PTH upstream region (Fig. 2). Based on restriction fragment analyses, we determined that clone I contained the normal, nonrearranged allele of the PTH gene and that clones II and III contained the rearranged PTH gene allele (Fig. 2). The genomic insert in clone II was approximately 14.5 kb and included the entire PTH gene, a portion of the 5' regulatory region, and in addition, approximately 200 bp of DNA from the rearrangement partner that was placed upstream of the PTH gene (Fig. 2). Clone III contained an approximately 7‐kb insert that included approximately 1 kb of the PTH gene's 5' regulatory region and approximately 6 kb of DNA from the rearrangement chromosomal partner (Fig. 2).

To identify the rearrangement partner across the breakpoint, the fragments of clones II and III adjacent to the breakpoint were subcloned, sequenced, and analyzed for homology to the human genome using the UCSC BLAT program. DNA sequences from the rearrangement partner mapped to chromosomal region 11q13, suggesting that the rearrangement was likely a pericentromeric inversion—inv(11)(p15;q13). Comparison of sequences from the rearranged allele with sequences from the PTH promoter region revealed that the breakpoint occurred 3.3 kb upstream of the PTH gene's exon 1. The rearrangement was associated with a microdeletion of 16 bp of the PTH 5' region and a 3‐bp deletion on 11q13. Fluorescent in situ hybridization (FISH) analysis of interphase nuclei from the tumor, using probes to both the PTH gene locus and to the 11q13 locus, confirmed the pericentromeric nature of this rearrangement (Fig. 3).

Figure 3.

Interphase FISH analysis of parathyroid adenoma. Representative tumor nucleus showing discrete signals from probes hybridized to cyclin D1 (green) and PTH (red) loci on the normal, nonrearranged allele and a fusion signal (yellow) from the rearranged allele.

We next determined the locations of known genes on 11q13 relative to the breakpoint. Pointedly, the established parathyroid oncogene cyclin D1 (PRAD1, CCND1), which is overexpressed in 20% to 40% of parathyroid tumors,12, 13, 14, 15 is located approximately 450 kb telomeric to the breakpoint, certainly a distance over which tissue‐specific enhancer elements can augment expression of a target structural gene. Indeed, quantitative real‐time PCR analyses revealed a dramatic increase in cyclin D1 mRNA from this tumor, approximately 75‐fold relative to that in the human placental tissue calibrator. In contrast, cyclin D1 expression in three comparison parathyroid tumors (without known PTH–cyclin D1 rearrangement) ranged from 0.3‐ to 3.8‐fold relative to the placental calibrator (Fig. 4). Thus cyclin D1 was expressed in the rearrangement‐bearing tumor at a level at least 20 times greater than the other parathyroid tumors; impressively, its expression was fourfold greater than even the ZR‐75‐1 breast cancer cell line, which is well known for its own amplification and overexpression of cyclin D1. Northern blot analysis confirmed the increased cyclin D1 expression in this tumor (data not shown). In contrast to cyclin D1, none of the other genes mapping near the breakpoint, namely, MYEOV and ORAOV1, have established roles in human neoplasia, although one cannot formally rule out their potential contribution in this clonal rearrangement.

Figure 4.

Quantitative RT‐PCR analysis of parathyroid tumor RNA for cyclin D1 expression. Total RNA from the present case (case 1) and 3 additional parathyroid adenomas (cases 2, 3, and 4) was isolated, and cyclin D1 mRNA was quantified by real‐time PCR. Cyclin D1 expression was normalized for expression of the housekeeping gene TBP and is presented relative to expression in placental tissue. Numbers above the columns represent mean fold change relative to placenta. Breast carcinoma cell line ZR‐75‐1 is a positive control with known overexpression of cyclin D1.

Importantly, the known ability of the rearranged PTH upstream region to drive cyclin D1 gene expression in parathyroid adenomas with inv(11)(p15;q13) indicates that the essential enhancer elements that direct its parathyroid tissue–specific gene expression are contained in the 11p region that is juxtaposed on the portion of 11q13 that contains cyclin D1. We reasoned that these sequence elements likely would be conserved among species and thus performed a cross‐species comparison of the PTH gene's 5' regulatory region. Sequences of the entire human PTH gene and 5.2 kb of its upstream DNA were compared with the syntenic genomic regions from mouse, rat, dog, and horse using Genome Vista.38 Analyses were limited to this 5.2‐kb region upstream of human PTH because it has been shown previously to be sufficient to drive parathyroid cell–specific gene expression.16, 17 The PTH gene's coding exons were highly conserved across the species. Discrete blocks of human sequence upstream of the PTH gene were conserved across the species (Fig. 5). Two of these blocks were located in the 2 kb immediately upstream of exon 1. In addition, three defined blocks of sequence 98bp, 129bp, and 100bp in length were located approximately 3.5 and 4.5 kb upstream of exon 1. Strikingly, the latter three regions of conserved sequence were located immediately adjacent to the breakpoint that occurred in this tumor and thus were included in the PTH promoter region that was juxtaposed with the part of 11q13 on which cyclin D1 resides. The presence of conserved noncoding DNA sequence in this region further supports the importance of this DNA segment in regulating PTH gene expression. Further examination of the DNA sequence in this PTH 5' region showed the presence of putative transcription factor binding sites (Table 1).

Figure 5.

Analysis of cross‐species conservation of the PTH gene's 5' regulatory region. The human PTH gene is shown to the right. The scale to the left of the gene represents the distance in kilobases from the start of exon 1. The location of the breakpoint (arrow) and the region of the 5' regulatory region that is rearranged with cyclin D1 (black rectangle) are depicted below the scale. Noncoding conserved sequences are shown in pink. Conserved coding sequences are shown in purple.

Table 1.

Computational Identification of Putative Transcription Factor Binding Sites as Determined by the MAPPER Database

Factor	Score	E value
Cart‐1	6.6	1.7
Alx‐4	6.6	1.7
RORalfa‐2	5.5	1.9
RXR‐VDR	4.9	2.5
Tal1beta‐E47S	5.8	1.6
PAX6	8.3	1.2
LUN‐1	21.4	2.1e–4
Hunchback	5.4	1.1
RORalfa‐2	5.5	1.9
Octamer‐binding factor	3.1	2.5
p53	6.8	2.5
p53	5.4	1.8
COUP	5.2	1.9
HFH‐2	5.1	2.5
Broad‐complex_1	4.2	1.5
HNF‐1	5.3	2.1
HMG‐IY	4.0	2.4

Discussion

Parathyroid adenomas are benign neoplasms that are the most common cause of primary hyperparathyroidism. Several molecular genetic abnormalities have been described previously. To date, cyclin D1 is the only established driver oncogene in sporadic parathyroid tumorigenesis. Cyclin D1 is overexpressed in 20% to 40% of parathyroid adenomas. In a subset of these tumors, this overexpression is a consequence of genomic rearrangements, where the cyclin D1 gene is juxtaposed with the PTH gene's promoter region. In the rearrangements reported to date, these breakpoints have occurred immediately upstream of the cyclin D1 gene. Here we report the genomic anatomy of a parathyroid adenoma where the breakpoint occurred approximately 450 kb upstream of cyclin D1. This finding underscores the fact that enhancer elements can exert their effect on gene transcription from large distances and has important implications for the study of DNA sequences that regulate gene expression. Furthermore, previous analyses of cyclin D1 gene rearrangements in parathyroid adenomas have used traditional southern blotting techniques with DNA probes in the near upstream region of the cyclin D1 gene2, 4 and thus would have been unable to detect breakpoints that occurred at greater distances from cyclin D1. This study shows that such breakpoints do indeed exist in parathyroid adenomas and that past analyses likely underestimated the true prevalence of PTH–cyclin D1 rearrangements in parathyroid adenomas. Indeed, using interphase FISH, which can detect breakpoints over a large chromosome segment, cyclin D1 rearrangements can be detected in as many as 8% of parathyroid adenomas.40 Such rearrangements are associated with increased cyclin D1 expression, further emphasizing cyclin D1's role as the target oncogene in this region.

An equally important contribution from the examination of this tumor is provided by the analysis of the breakpoint at the PTH gene locus. In previously reported rearrangements in parathyroid adenomas, the breakpoints occurred in the immediate vicinity of the PTH gene's exon 1. Here we show that the breakpoint in the PTH promoter region occurred 3.3 kb upstream of the PTH gene's first exon. Thus this novel rearrangement juxtaposed the more distal region of the PTH gene's 5' regulatory elements with the cyclin D1 side of the 11q13 breakpoint region. The ability of the rearranged PTH 5' region to drive expression of the cyclin D1 gene located on its rearrangement partner indicates that promoter/enhancer elements necessary for parathyroid‐specific gene transcription are located in this segment of the PTH gene's upstream regulatory region. Combined with our previous demonstration that the 5.2‐kb region immediately upstream of the human PTH gene is sufficient to drive parathyroid‐specific gene expression in a transgenic mouse model,16, 17 our present data powerfully narrow this critical region to just a single 1.7‐kb DNA segment (Figs. 2 and 5). Complementing this finding, our cross‐species analysis of the PTH gene's 5' regulatory region showed that two blocks of DNA sequence within this 1.7 kb are highly conserved in the mouse, rat, canine, and bovine genomes, further supporting the conclusion that this DNA segment likely plays an important role in regulating PTH gene expression. Analysis of this region revealed the presence of several putative transcription factor binding sites (Table 1), including one for the vitamin D receptor. This information provides a framework for future in vitro analyses to investigate the potential functional significance of these candidate factors in tissue‐specific PTH gene regulation.

The ability to even more precisely identify these DNA elements has been severely hampered by the lack of suitable cell lines that faithfully replicate the parathyroid chief cell phenotype. However, recent findings may carry some promise that this aim could be advanced in the future. Specifically, a case of ectopic PTH production by a metastatic neuroendocrine tumor of the pancreas was described.41 Cells cultured from this tumor successfully supported transcription of a reporter gene driven by the 5.2‐kb fragment of the human PTH promoter, raising the interesting possibility that these cultured tumor cells may recapitulate critical aspects of PTH gene regulation.41 In addition, a very recent report has demonstrated the use of cultured primary human parathyroid cells from patients with chronic renal disease, transfected with lentiviral‐based reporter gene constructs, to study transcriptional activity of the glial cell missing‐2 transcription factor.42 The validation and application of such in vitro approaches could facilitate the precise identification of the minimal enhancer elements that regulate tissue‐specific PTH gene expression, which then would be subjected to in vivo confirmation in transgenic animals. An improved understanding of the molecular mechanisms of PTH gene regulation, to which our results have contributed significantly, has important implications for treating human disease. For example, it could aid in the development of novel therapeutic agents or stem cell approaches to treat hyperparathyroidism or hypoparathyroidism. Furthermore, given the role of PTH as an anabolic agent, targeted approaches to increasing PTH gene transcription might be of contributory value combined with agents that intermittently release stored PTH from parathyroid cells in treating patients with osteoporosis.

Disclosures

All the authors state that they have no conflicts of interest.