Identification of a Modified HOXB9 mRNA in Breast Cancer

First identified as a developmental gene, HOXB9 is also known to be involved in tumor biological processes, and its aberrant expression correlates with poor prognosis of various cancers. In this study, we isolated a homeodomain-less, novel HOXB9 variant (HOXB9v) from human breast cancer cell line-derived mRNA. We confirmed that the novel variant was produced from variationless HOXB9 genomic DNA. RT-PCR of mRNA isolated from clinical samples and reanalysis of publicly available RNA-seq data proved that the new transcript is frequently expressed in human breast cancer. Exogenous HOXB9v expression significantly enhanced the proliferation of breast cancer cells, and gene ontology analysis indicated that apoptotic signaling was suppressed in these cells. Considering that HOXB9v lacks key domains of homeobox proteins, its behavior could be completely different from that of the previously described variationless HOXB9. Because none of the previous studies on HOXB9 have considered the presence of HOXB9v, further research analyzing the two transcripts individually is warranted to re-evaluate the true role of HOXB9 in cancer.


Introduction
Homeobox (HOX) genes were initially characterized as developmental genes, which code for transcription factors that play critical roles in embryogenesis. Evolutionarily, they are highly conserved and share a high degree of homology, especially within the same paralog groups. All 39 mammalian HOX genes consist of two exons and a single intron. e homeobox domain, encoded in the second exon [1], includes the DNA binding site. e diverse and specific transcriptional activities of the HOX proteins often depend on key cofactors including PBX, MEIS, and PREP, which interact with hexapeptide motifs of HOX proteins [2].
HOX genes play key roles in both solid and hematological malignancies, including cancers of the colon, breast, prostate, lung, brain, thyroid, ovary, bladder, kidney, skin, and blood [3,4]. HOXB9, the ninth paralog in the HOX-B cluster, is associated with the growth and progression of multiple cancers. In lung adenocarcinoma, HOXB9 promotes metastasis by activating the WNT signaling pathway [5,6]. In breast cancer, the gene induces the expression of proangiogenic factors, increasing the cell motility and supporting epithelial-mesenchymal transition (EMT) [7,8]. HOXB9 also promotes the growth of colon cancer by activating IL6 signaling, inducing the secretion of angiogenic factors and increasing proliferation of tumor cells [9]. Similar observations are found in ovarian cancer and hepatocellular carcinoma [10,11]. us, HOXB9 activates the WNT signaling pathway and enhances the acquisition of capabilities critical to the transformation of normal cells to cancer, including EMTand the growth of new vasculature within the tumor microenvironment.
HOXB9 is also directly associated with cancer-induced patient mortality. e duration of disease-free and overall survival of patients with HOXB9-positive breast cancer is significantly shorter compared with patients with HOXB9negative breast cancer [12]. Increased HOXB9 expression significantly correlates with decreased overall survival for patients with colorectal cancer [9]. Patients of laryngeal squamous cell carcinoma, hepatocellular carcinoma, glioma, and endometrial cancer also present poor outcomes or tumor progression, resulting from aberrant HOXB9 expression [13][14][15][16]. e recent elucidation of the critical and diverse roles HOXB9 plays in various cancers have led us to explore the mechanism of this gene's role in cancer progression. In the present study, we identified and characterized a HOXB9 variant (HOXB9v) of mRNA from human breast cancer cell lines.
e sequence of HOXB9v largely differs from the previously known HOXB9 normal transcript (HOXB9n), and we found it lacks some important domains of HOX genes. Based on these findings, we inferred its role and function were different from HOXB9n. is study aimed to confirm the presence of HOXB9v in clinical breast cancer specimens and investigate the role of HOXB9v in breast cancer progression.

Patients and Samples.
Clinical specimens of human breast cancer (n � 14) were collected from patients with primary operable breast cancer who underwent total or partial mastectomy between July and November 2018 in Keio University Hospital (Tokyo, Japan). Patient-matched healthy breast epithelium samples (n � 6) were collected from healthy breast tissue from patients who underwent total mastectomy. Ethics approval for the present study was provided by the Ethics Committee at the Keio University School of Medicine (approval number: 20180090), and the study was performed in accordance with the provisions of the Declaration of Helsinki (as revised in Fortaleza, Brazil, October 2013). All included patients gave informed consent.

mRNA and Genomic DNA Extraction from Cell Lines and
Clinical Specimens. Total RNA and genomic DNA were extracted using the RNeasy Mini Kit and QIAmp DNA Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. Clinical specimens were homogenized using a Minilys homogenizer (Bertin Instruments, Bretonneux, France) using 2.4 mm metal beads prior to mRNA extraction. Total RNA was converted to cDNA using the High-Capacity RNA-to-cDNA Kit ( ermo Fischer Scientific).
After transformation into Competent Quick DH5α cells (Toyobo, Osaka, Japan) and culturing in kanamycin-containing LB plates, at least 4 colonies were selected for each sample.

RT-PCR Analysis.
e expressions of HOXB9n and HOXB9v were detected by RT-PCR using the following primers: sense, 5′ TGTCCATTTCTGGGACGCTT 3′; antisense, 5′ CTACGGTCCCTGGTGAGGTA 3′. e genomic DNA of HOXB9 was detected by PCR using the following primers: sense, 5′ CGAGAGAGCTGCAAGTCGAT 3′; antisense, 5′ CTGCCGTCCGTCTACCAC 3′. e primers for genomic DNA were designed against exon 1 and the intron region of the HOXB9 gene to ensure that the pair will specifically amplify only genomic DNA and not cDNA derived from mRNA. e conditions applied for amplification were as follows: 94°C for 1 minute, followed by 35 cycles at 95°C for 5 seconds, 55°C for 5 seconds, and 72°C for 5 seconds, and run on the Life ECO thermal cycler (Hangzhou Bioer Technology, Hangzhou, China) using the SapphireAmp Fast PCR Master Mix (Takara Bio, Shiga, Japan) or KOD-Plus-Neo (Toyobo).

Public Data Reanalysis.
e RNA sequence data set (GSE119937) in FASTQ format was downloaded via SRA (SRP161704) using the SRA Toolkit (version 2.3.4-2). RNAseq reads were aligned by STAR (version 2.6.1b) against the hg38 reference genome. All reads mapped on the HOXB9 gene were visually confirmed by taking snapshots in IGV (version 2.4.15).

Establishment of Stable MCF7 Cell Lines Overexpressing
HOXB9n or HOXB9v.
e HOXB9n and HOXB9v sequences were amplified by PCR and subsequently cloned into the pBiT3.1-N [CMV/HiBiT/Blast] expression vector (Promega, Tokyo, Japan) at the Xhol and BamHI sites using Ligation high Ver.2 (Toyobo). e primers used for amplification were as follows: sense, 3′ ATACCTCGAGG TCCATTTCTGG 5′; antisense, 3′ CACGTCATACGGAT CCTCTTTG 5′. MCF7 cells were transfected with the HiBiT-tagged HOXB9n or HOXB9v vector using the Viafect Transfection reagent (Promega) and selected with 10 µg/mL of blasticidin for at least 4 weeks. e expression of HiBiTtagged HOXB9n and HOXB9v proteins was detected using the Nano-Glo HiBiT Blotting System (Promega) as per the manufacturer's instructions.

Quantitative Real-Time PCR.
Quantitative real-time PCR was run on ViiA7 ( ermo Fisher Scientific) using Fast SYBR Green Master Mix ( ermo Fisher Scientific). Preincubation was performed for 20 seconds at 95°C and amplification for 41 cycles (1 second of denaturation at 95°C and 20 seconds of annealing and extension at 60°C), followed by melt-curve analysis. GAPDH served as an internal control, and QuantStudio Real-Time PCR Software v1.2 ( ermo Fisher Scientific) was used for quantification. e relative standard curve method was used for linear regression analysis of unknown samples, and data are presented as fold change between samples. e primers used were as follows: HOXB9: sense, 5′ CGGTGGCTGTCGTGAAATT 3′; antisense, 5′ CGAGACAATCACCCCCAAAG 3′; GAP DH: sense, 5′ATCATCCCTGCCTCTACTGG 3′; antisense, 5′ TTTCTAGACGGCAGGTCAGGT 3′.
Cell proliferation in flat culture was measured using Cell Count Reagent SF (Nacalai Tesque). Briefly, 5000 cells/well in 96-well microtiter plates (Sumilon, Sumitomo Bakelite, Tokyo Japan) were incubated for 5 days or after 24 hour continuous exposure to either 30 nM HXR9 or 30 nM CXR9 for 4 days. e absorbance in the wells was measured on days 1, 3, and 5 using a Sunrise Rainbow-RC (TECAN, Männedorf, Switzerland) microplate spectrophotometer at 450 nm, using 600 nm as reference.

Microarray and Differential Expression Analyses.
Total RNA was isolated using the RNeasy Mini kit (Qiagen). Microarray was performed using the human Clariom S assay ( ermo Fisher Scientific) by the GeneChip Scanner 3000 7G system (Affymetrix, Santa Clara, CA, USA), and the results were analyzed by TAC 4.0 software ( ermo Fisher Scientific). Genes having a false discovery rate (FDR) under 0.05 and upregulated in HOXB9v samples were considered as upregulated differentially expressed genes (up-DEGs). DAVID Bioinformatics Resources 6.8 [17] was used for gene ontology and pathway analysis of DEGs. R ver. 3.5.0 software was used to draw the heatmap of 37 genes included in GO: 0043069 (negative regulation of programmed cell death). e sequence homology of the new transcript is shown in Figures 1(a) and S1.

Identification of a Novel
We will refer to the new truncated transcript as HOXB9v, to distinguish from the full-length HOXB9 transcript (HOXB9n). A 100-base deletion in exon 1 in the new transcript leads to a frameshift and the formation of a stop codon (TAG), which truncates the protein coding at AA167. Figure 1(b) is a schematic diagram of HOXB9 transcripts showing the exons and the deleted lesion.
e encoded protein will therefore possibly lack the homeobox domain, DNA binding domain, and the hexapeptide motif, a major player in cofactor interactions (Figure 1(c)). e HOXB9v sequence has been submitted to GenBank under Accession No. LC466645.
We next performed PCR analyses to verify the presence of the HOXB9v transcript and identify genomic DNA variations of the HOXB9 gene in human breast cancer cell lines. e primer target region included the deletion site of HOXB9v and the amplicon for HOXB9n was 643 bp and 543 bp for HOXB9v. We detected both HOXB9n and HOXB9v transcripts in breast cancer cell lines (Figure 2(a)). e PCR products were sequenced and were confirmed that each band corresponded to the exact sequence of the HOXB9n or HOXB9v (Figure 2(b)). By PCR of breast cancer cell lines' genomic DNA, no bands indicative of genomic DNA variations in HOXB9 were detected. e amplicon of variation-less HOXB9 genomic DNA was 569 bp (Figure 2(c)). Sequencing of the PCR products confirmed that they had no variations or deletions in genomic DNA ( Figure S2). ese findings show that the two transcripts (HOXB9n and HOXB9v) were produced from variation-less HOXB9 genomic DNA.
To determine the presence of HOXB9v transcripts in human breast cancer samples, we performed PCR analysis. HOXB9v was commonly detected from clinical breast cancer samples (Figure 3(a)), regardless of their hormone receptor and HER2 status. However, HOXB9v was not detected in normal mammary gland samples (Figure 3(b)).  us, to further confirm the presence of HOXB9v in human breast cancer samples, we reanalyzed a publicly available breast cancer RNA sequence data set (GSE119937) [18] and mapped the reads onto the HOXB9 gene sequence. We identified a region in exon 1 where the number of mapped reads was low in numerous samples; this region matched the deletion region of HOXB9v (Figure 4).
To determine the role of HOXB9v in breast cancer, we established stable MCF7 cell lines overexpressing HOXB9n or HOXB9v. Gene and protein expressions of HOXB9n and HOXB9v were verified in both cell lines (Figures 5(a) and  5(b)). Cell proliferation assays in both 3D culture and flat culture showed that HOXB9v overexpression increased MCF7 cell growth (Figures 5(c) and 5(d)). We also transiently overexpressed HOXB9n or HOXB9v in MDA-MB-468 cells, and the cell proliferation assay showed that HOXB9v overexpression increased MDA-MB468 cell growth (Figures 5(e) and 5(f )).

HXR9 and CXR9
Treatment. HXR9, an 18-amino acid peptide, competently inhibits the hexapeptide motif of HOX proteins and prevents HOX-PBX binding [19]. HOXB9n and HOXB9v expressing MCF7 cells were treated with HXR9 or a control peptide, CXR9 (Figures 6(a) and 6(b)). Although HOXB9v lacks the hexapeptide motif, which is known to interact with PBX proteins [20], HXR9 significantly inhibited proliferation of both cell lines. ere is a region where data are sparsely mapped (indicated with the red arrow), which matches the deletion region in HOXBv. HOXB9v overexpressing MCF7 and MDA-MB-468 cells proliferated faster compared with HOXB9n overexpressing cells.

Microarray and Gene Ontology Analysis.
To explore the reason behind the faster proliferation of HOXB9v overexpressing cells, we performed microarray and gene ontology analyses using HOXB9n and HOXB9v-expressing MCF7 cells. We chose 1056 genes as DEGs, and gene ontology analysis showed that up-DEGs between HOXB9n and HOXB9v expressing cells presented significant differences in pathways relevant to apoptosis suppression (GO:0060548∼negative regulation of cell death, GO:0043069∼negative regulation of programmed cell death, and GO:0043066∼negative regulation of apoptotic process) and steroid hormone response (GO:004 8545∼response to steroid hormone, GO:0032870∼cellular response to hormone stimulus, and GO:0071383∼cellular response to steroid hormone stimulus) ( Table 1). e genomic expression heatmap comparing 37 genes involved in the apoptotic process (GO:0043069∼negative regulation of programmed cell death) shows that apoptosis is highly suppressed in HOXB9v expressing cells (Figure 7).

Discussion
In the present study, we identified a novel modified transcript of HOXB9, in which a deletion in exon 1 causes a frameshift, formation of a stop codon, and truncation of protein coding. is leads to a defect in the homeodomain and hexapeptide regions, which are both crucial for HOX gene function. We confirmed that HOXB9v is widely present in human breast cancer cell lines and clinical breast cancer samples, however not in normal gland samples. Detection of HOXB9v in MCF10A may be attributed to the fact that this cell line is not karyotypically normal, while it maintains major characteristics of normal breast epithelium [21]. We further confirmed that HOXBv and the previously known HOXB9n were both produced from variationless HOXB9 genomic DNA. Two distinct mRNA products from a single variationless genomic DNA may indicate that HOXB9v is a novel splice variant of HOXB9. It may be assumed that HOXB9v has a typical splice site of GU as a 5′ donor site and AG as a 3′ acceptor site [22]; however, the single-nucleotide transitions at 248 and 258 and the singlenucleotide deletion at 254 must be taken into account. Further investigation and discussions are needed to decide whether HOXB9v is a splice variant or an mRNA modified by a different mechanism. Nonetheless, the modified transcript is certainly not a result of variations in genomic DNA. Additionally, HOXB9v results in a different stop codon site from that of HOXB9n and may become a target of nonsense-mediated mRNA decay (NMD), an mRNA surveillance mechanism that eliminates premature translation-termination codons. However, NMD is known to be initiated if an exon-junction complex is present more than 50-55 bases downstream of the stop codon [23]. HOXB9v forms its exon-junction complex upstream from its stop codon; therefore, we infer it escapes from NMD. Nevertheless, the production of splice variants of HOX genes lacking the homeodomain is likely a common phenomenon. Murine Hoxb9, which shares sequence similarity with human HOXB9, is reported to generate a splice variant without a homeodomain. Other HOX genes, including human HOXA1, HOXB6, HOXA9, HOXA10, murine Hoxb6 and Meis1, and Xenopus XlHbox2, have also been reported to generate splice variants lacking homeodomains [24][25][26][27][28]. Interestingly, HOXA9T, a homeodomain-less isoform of HOXA9, which is structurally similar to HOXB9v, has been demonstrated to act as an oncogene in leukemia without directly binding to DNA [24]. HOXB9 has been reported to promote tumorigenesis in various types of cancers; however, no previous research on HOXB9 has shown a mRNA variant of HOXB9. Considering that the structure of the HOXB9v protein differs from that of the previously studied HOXB9n protein and lacks important domains such as homeodomain and hexapeptide motif, its function is likely to differ from that of HOXB9n. Further study is required to re-evaluate the role of HOXB9 in  cancer. HOXB9n and HOXB9v should be assessed separately.
In the growth assay, MCF7 and MDA-MB-468 cells expressing HOXB9v at high levels presented more rapid proliferation than did those expressing HOXB9v. It is meaningful that similar results were observed in two cell lines: one with hardly any expression of HOXB9n and HOXB9v (MCF7) and the other with high expression of both (MDA-MB-468). Fostered proliferation by HOXB9v may be attributed to the suppressed apoptosis observed in our microarray studies. Gene ontology analysis of HOXB9v-expressing cells compared to HOXB9nexpressing cells indicated significant upregulation of pathways related to apoptosis suppression, further underscoring its role in this regard. Several HOX genes have been reported to regulate apoptosis in cancer [29,30], and HOX-regulated apoptosis is a general mechanism used during development to maintain metameric patterns [31,32]. However, the role of HOXB9 in apoptosis is yet to be investigated. e binding selectivity of HOX proteins is influenced by cofactors, including members of the PBX, MEIS, and PREP families [2]. Additionally, HOX-PBX interactions involve a short HOX protein motif, the hexapeptide, located upstream of the homeodomain [33]. HXR9 is an 18-amino acid peptide and suppresses tumor proliferation by inhibiting HOX-PBX binding [19]. In our proliferation assay, HOXB9v expressing cell lines were sensitive to the HOX-PBX inhibitor HXR9, even though HOXB9v lacks the hexapeptide.

Conclusions
We report a modified HOXB9 mRNA variant that results in a homeodomain defect. We confirmed its presence in human breast cancer cell lines and breast cancer clinical samples and also revealed that HOXB9v may promote breast cancer proliferation by suppression of apoptosis. Further research is warranted to analyze HOXB9v and HOXB9n individually and re-evaluate the true role of HOXB9 in cancer.

Data Availability
e HOXB9v sequence has been submitted to GenBank under Accession No. LC466645. e RNA sequence data set (GSE119937) used in public data reanalysis in the fastq format is available via SRA (SRP161704). Other datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.