Introduction
Splenic hemangiomas, although infrequent, represent the most common benign neoplasms of the spleen with an incidence of 0.03–14% found in large autopsy series (1,2). In the vast majority of cases, splenic hemangiomas are incidental surgical, radiologic, or autopsy findings, made during evaluation for other disorders (1–3). Arber et al (3) reported 7 localized splenic hemangiomas all of which were discovered incidentally in surgical patients. In a large series of 32 patients with splenic hemangioma, only 6 presented with abdominal symptoms and only 4 had a palpable spleen (1).
Because of the general absence of presenting symptoms and the typically incidental nature of splenic hemangiomas, the age at presentation varies considerably. Several non-autopsy, surgical series revealed an average age at detection/presentation of between 51 and 63 years (1,3). However, examination of autopsy data reveals a much younger average patient age (4) indicating that these benign lesions are likely to be present but remain undetected for long periods of time. No gender or race predilection has been reported (1).
Splenic hemangiomas are thought to be congenital in origin, arising from sinusoidal epithelium (5). Whether they are neoplastic or represent some other type of misgrowth, remains uncertain. They typically appear as circumscribed, non-encapsulated, honeycomb-like, red-purple masses that frequently blend imperceptibly into the surrounding splenic parenchyma (4). The spaces are composed of sponge-like tissue filled with blood and separated by fibrous septa. Occasional calcification may be seen, often in association with an organized infarct (6). Microscopically, the majority of hemangiomas are cavernous in nature, consisting of large interconnected, dilated, blood-filled spaces lined by a monolayer of cytologically bland endothelial cells separated by thin fibrous septa or splenic pulp tissue. Pure capillary architecture is less common. Instead, many lesions contain varying proportions of both cavernous and capillary components (4). Immunophenotypically, splenic hemangiomas show reactivity of endothelial lining cells for CD31, von Willebrand factor, Ulex europeaus, lectin I, and CD34. This pattern raises the possibility that splenic hemangioma may derive from a combination of splenic venous structures as well as from splenic sinusoidal cells (4).
Most splenic hemangiomas tend to be small in size (<4 cm) although lesions ≤36 cm have been reported (4). They need not be entirely without complications as larger lesions may rupture with resulting intra-abdominal hemorrhage (7–11). In some patients, they cause the Kasabach-Meritt syndrome (12).
The etiology and pathogenesis of splenic hemangiomas are unclear and no cytogenetic or molecular genetic information about the disease has been published. We here describe the cytogenetic analysis of a splenic hemangioma and the fusion gene corresponding to the chromosomal translocation thus found.
Materials and methods
Ethical approval
The study was approved by the Regional Committee for Medical and Health Research Ethics, SouthEast Norway (REK Sør) http://helseforskning.etikkom.no). Written informed consent was obtained from the patient. The consent included acceptance that the clinical details be published. The ethics committee's approval included a review of the consent procedure. All patient information has been anonymized and de-identified.
Patient
A twenty-nine-year-old woman was incidentally diagnosed with a splenic hemangioma during an ultrasound examination for cholecystitits. She had been without symptoms attributable to the splenic lesion, possibly except some pressure in the upper left abdomen. Because of continuous growth of the hemangioma, it was decided to do a splenectomy. Histological examination (Fig. 1) showed that the lesion was composed of large, blood-filled vessels lined by flat endothelium and separated by thin fibrous septa or splenic pulpa. Immunohistochemical analysis showed positivity for CD31 and ERG.
Control sample
The control sample was FirstChoice human spleen total RNA (Life Tehnologies, Carlsbad, CA, USA).
G-banding and karyotyping
Fresh tissue from a representative area of the tumor was received and analyzed cytogenetically as part of our diagnostic routine. The sample was disaggregated mechanically and enzymatically with collagenase II (Worthington, Freehold, NJ, USA). The resulting cells were cultured and harvested using standard techniques. Chromosome preparations were G-banded with Wright stain and examined. Peripheral blood lymphocytes stimulated with phytohemagglutinin (PHA) for 72 h were also karyotyped. The karyotypes were written according to the International System for Human Cytogenetic Nomenclature (ISCN) 2009 guidelines (13).
RNA and DNA extraction
Tumor tissue adjacent to that used for cytogenetic analysis and histologic examination had been frozen and stored at −80°C. Total RNA was extracted using miRNeasy Mini kit according to the manufacturer's instructions (Qiagen Nordic, Oslo, Norway). Tumor tissue was disrupted and homogenized in Qiazol Lysis Reagent (Qiagen) using a 5-mm stainless steel bead and TissueLyser II (Qiagen). Subsequently, total RNA was purified using QIAcube (Qiagen). The RNA quality was evaluated using the Experion Automated Electrophoresis system (Bio-Rad Laboratories, Oslo, Norway). The RNA quality indicator (RQI) was 9.0. Genomic DNA was extracted using the Maxwell 16 Instrument System and the Maxwell 16 Tissue DNA Purification kit (Promega, Madison, WI, USA), and the concentration and purity of DNA were measured using NanoVue Plus Spectrophotometer (GE Healthcare Life Sciences, Oslo, Norway).
High-throughput paired-end RNA-sequencing
Three micrograms of total RNA were sent for high-throughput paired-end RNA-sequencing at the Norwegian Sequencing Centre, Ullevål Hospital (http://www.sequencing.uio.no/). The RNA was sequenced using an Illumina HiSeq 2000 instrument and the Illumina software pipeline was used to process image data into raw sequencing data. The regular TruSeq library preparation protocol (http://support.illumina.com/downloads/truseq_rna_sample_preparation_guide_15008136.ilmn) was used. The reads obtained had a length of 100 base pairs (bp). A total of 74 million reads were obtained. The quality of the raw sequence data was assessed using FastQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The software TopHat-Fusion was used for the discovery of fusion transcripts (14,15). To verify further the fusion gene which was found by TopHat-Fusion, the ‘grep’ command (http://en.wikipedia.org/wiki/Grep) was used to search the fastq files of sequence data (http://en.wikipedia.org/wiki/FASTQ_format). Our ‘specific expression’ was a sequence of 20 nucleotides at the fusion point, 10 bases upstream (5′-end gene), and 10 bases downstream from the junction (3′-end gene). The expression was ‘CGACCAATAGGTCCCCAAGT’. The sequences obtained by ‘grep’ were blasted against the human genomic plus transcript database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) as well as the sequences with accession numbers NM_024665.4 (TBL1XR1) and NM_002131.3 (HMGA1) and DB051170.1 (TESTI2 Homo sapiens cDNA clone TESTI2040990, 5′, mRNA sequence).
RT-PCR and genomic PCR analyses
The primers used for PCR amplification and Sanger sequencing are listed in Table I. For RT-PCR, 1 μg of total RNA was reverse-transcribed in a 20-μl reaction volume using iScript Advanced cDNA Synthesis kit for RT-qPCR according to the manufacturer's instructions (Bio-Rad Laboratories). The 25 μl PCR volume contained 12.5 μl Premix Ex Taq™ DNA Polymerase Hot Start Version (Takara Bio Europe/SAS, Saint-Germain-en-Laye, France), 1 μl of cDNA, and 0.4 μM of each of the forward and reverse primer. The primer sets TBL1XR1-229F1/DB051170-Intr2R1 and TBL1XR1-229F1/HMGA1-324R1 were used to detect possible TBL1XR1-HMGA1 fusion transcripts. The quality of the cDNA synthesis was examined by amplification of a cDNA fragment of the ABL1 gene using the primers ABL1-91F1 and ABL1-404R1 (16).
For genomic PCR, the 25 μl PCR volume contained 12.5 μl Premix Ex Taq™ DNA Polymerase Hot Start Version, 100 ng DNA, and 0.4 μM of each of the forward and reverse primers TBL1XR1-intron4-F1 and DB051170-intr2R1.
The PCR amplifications were run on a C-1000 Thermal cycler (Bio-Rad Laboratories) with an initial denaturation at 94°C for 30 sec, followed by 35 cycles of 7 sec at 98°C, 30 sec at 55°C (58°C for genomic PCR), 1 min at 72°C, and a final extension for 5 min at 72°C. For amplification of the ABL1 cDNA fragment, the PCR cycling was an initial denaturation at 94°C for 30 sec followed by 35 cycles of 7 sec at 98°C and 2 min at 68°C, and a final extension for 5 min at 68°C. Three microliters of the PCR products were stained with GelRed (Biotium, Hayward, CA, USA), analysed by electrophoresis through 1.0% agarose gel, and photographed. The remaining 22 μl PCR products were purified using the MinElute PCR purification kit (Qiagen Nordic) and sequenced at GATC Biotech (Germany, http://www.gatc-biotech.com/en/home.html). The BLAST software (http://blast.ncbi.nlm.nih.gov/Blast.cgi) was used for computer analysis of sequence data.
Discussion
We describe the first cytogenetic and molecular genetic analysis of a splenic hemangioma. The tumor had an acquired chromosomal translocation, t(3;6)(q26;p21), which resulted in fusion of the TBL1XR1 (from 3q26) and HMGA1 (from 6p21) genes. The fact that an acquired genetic aberration was found in the splenic hemangioma cells argues strongly that the disease is neoplastic.
The protein encoded by the TBL1XR1 gene has sequence similarity with members of the WD40 repeat-containing protein family (18). The WD40 group is a large family of proteins which appear to have a regulatory function (19–21). WD40 repeats mediate protein-protein interactions and members of the family are involved in signal transduction, RNA processing, gene regulation, vesicular trafficking, cytoskeletal assembly, and they may also play a role in the control of cytotypic differentiation (19–21). TBL1XR1 is a core component of NCoR (nuclear receptor corepressor) and SMRT (silencing mediator of retinoic acid and thyroid hormone receptors) repressor complexes and is essential in targeting SMRT/NCoR corepressor complexes to the promoter of target genes. It is also required for transcriptional activation by nuclear receptors and other regulated transcription factors (18). TBL1XR1 is further essential in the activation of Wnt-β-catenin and NF-κB signaling pathways (18). De novo deletions and recurrent mutations have been identified in the TBL1XR1 gene and are linked to intellectual disability (18). TBL1XR1 plays an important role in tumorigenesis, invasion, metastasis, and the development of therapy resistance (18). TBL1XR1 is overexpressed in primary lung squamous cell carcinoma, breast cancer, cervical cancer, nasopharyngeal carcinoma, esophageal squamous cell carcinoma, and invasive prostate cancer (reviewed in ref. 18). Mutation of TBL1XR1 was found in primary central nervous system lymphomas (22,23). Deletions of TBL1XR1 were described in 15% of ETV6-RUNX1 positive acute lymphoblastic leukemias (24). In frame TBL1XR1 chimeric transcripts which code for chimeric proteins, were found in different neoplasias. A recurrent TBL1XR1-TP63 fusion gene was reported in diffuse large B-cell lymphoma, peripheral T-cell lymphoma, and follicular lymphoma which was the result of a chromosomal rearrangement between 3q26 (TBL1XR1) and 3q28 (TP63) (25,26). In most of the cases, the exons 1–7 of TBL1XR1 were fused in frame to exons 4–8 or 4–10 of TP63. Exon 14 or exon 4 of TBL1XR1 was involved in the remaining cases (25,26).
TBL1XR1 is fused to RARA in an acute promyelocytic leukemia carrying a variant t(3;17)(q26;q21) translocation (27). The TBL1XR1-RARA fusion protein was predominantly localized in the nucleus, formed homodimers or heterodimers with retinoid X receptor α, and acted as transcriptional activator in the presence of ligand. In the presence of pharmacologic doses of ATRA, TBLR1-RARA protein could be degraded, and its homodimerization was abrogated (27).
Fusions of TBL1XR1 resulting in promoter swapping were also found (28,29). In the breast carcinoma MCB7 cell line, the untranslated (5′-UTR) exon 1 of TBL1XR1 is fused with the start codon-bearing exon 2 of RGS17, generating a fusion gene that encodes the full-length RGS17 protein under the control of TBL1XR1 gene promoter (28). The same 5′-UTR exon 1 of TBL1XR1 was also found to be fused with the PIK3CA gene in breast cancer and prostate adenocarcinomas. The result was again the expression of PIK3CA which came under the control of the TBL1XR1 promoter (29).
The HMGA1 gene encodes a non-histone protein involved in many cellular processes, including regulation of inducible gene transcription, integration of retroviruses into chromosomes, and the metastatic progression of cancer cells (30–32). The encoded protein preferentially binds to the minor groove of A+T-rich regions in double-stranded DNA. It has little secondary structure in solution but assumes distinct conformations when bound to substrates such as DNA or other proteins. The encoded protein is frequently acetylated and is found in the nucleus. Six transcript variants were identified encoding two different isoforms (http://www.ncbi.nlm.nih.gov/gene/3159). However, expressed sequence tags (EST) indicate that additional alternative splicing exists with various 5′-UTR sequences occurring in different tissues. For example, the EST with accession numbers DB051170 and DB03591 are found in testis, EST BM845429 in lymph node, CN350278 in embryonic stem cells, DC338100 in brain, and DA921912 is found in the small intestine. These splicing variants may be regulated by alternative promoters of HMGA1 (17).
HMGA1 is overexpressed in poorly differentiated cancers originating from all three germ layers, suggesting that it plays a fundamental role in tumorigenesis regardless of the cell type from which the tumor begins. High expression of the HMGA1 gene or high levels of HMGA1 protein were found in cancers of the thyroid, lung, breast, bladder, prostate, colon, pancreas, uterine corpus, uterine cervix, kidney, head and neck, nervous system, stomach, liver, and hematopoietic system. Expression studies reveal that HMGA1 overexpression correlates with adverse clinical outcomes in cancers (33–35).
The basis for HMGA1 overexpression is not well understood but it seems that HMGA1 expression is induced through different pathways. For example, HMGA1 expression was shown to be induced by cMYC and AP1 transcription factors (36,37). Recent studies suggest that microRNAs regulate HMGA1 in some types of cancer (38–40). Another recent study showed that HMGA1 was an immediate target of the β-catenin/TCF-4 signaling pathway in colon cancer (41).
Dysregulation of HMGA1 as a result of specific chromosomal rearrangements involving chromosomal region 6p21.3 has also been identified in a variety of common benign mesenchymal tumors such as lipomas, uterine leiomyomas, endometrial polyps, and pulmonary hamartomas (42–51). The breakpoints in most tumors are located 3′ of HMGA1, but 5′ or intragenic breakpoints have also been reported (42,43,46,47). The findings we present indicate that a similar pathogenetic mechanism may be operative also in splenic hemangioma. Whether any systematic tumorigenic differences exist between this and the other benign neoplasias is a question that can be answered only after many more splenic hemangiomas are genetically analyzed.
So far, only one HMGA1-fusion transcript has been described in the literature: an HMGA1-LAMA4 found in a pulmonary hamartoma as a result of inv(6)(p21q21) (46). Fusion of HMGA1 to the intergenic region between LPP and TPRG1 in a lipoma carrying a t(3;6)(q27;p21) has also been reported (52).
The present t(3;6)(q26;p21) chromosomal rearrangement would have molecular consequences for both TBL1XR1 and HMGA1. The TBL1XR1 gene would have only a single functional copy of the gene left in the cell, while the other, rearranged allele would produce a putative truncated form of TBL1XR1 protein containing the LiSH and F-box-like domains (NP_078941; LiSH domain: 4–32, F-box-like domain: 41–86). Thus, the truncated form of TBL1XR1, through the LisH and F-box-like domains, could participate in protein dimerization, affect protein half-life, and could influence specific cellular localizations (53,54).
With regard to HMGA1, the TBL1XR1-HMGA1 fusion transcript leads untranslated exons of HMGA1 to be replaced by the first 5 exons of the TBL1XR1 gene. The result is that the entire coding region of HMGA1 comes under the control of the TBL1XR1 promoter leading to dysregulation of HMGA1.