- Research article
- Open Access
- Open Peer Review
HOX transcription factors are potential targets and markers in malignant mesothelioma
BMC Cancer volume 16, Article number: 85 (2016)
The HOX genes are a family of homeodomain-containing transcription factors that determine cellular identity during development and which are dys-regulated in some cancers. In this study we examined the expression and oncogenic function of HOX genes in mesothelioma, a cancer arising from the pleura or peritoneum which is associated with exposure to asbestos.
We tested the sensitivity of the mesothelioma-derived lines MSTO-211H, NCI-H28, NCI-H2052, and NCI-H226 to HXR9, a peptide antagonist of HOX protein binding to its PBX co-factor. Apoptosis was measured using a FACS-based assay with Annexin, and HOX gene expression profiles were established using RT-QPCR on RNA extracted from cell lines and primary mesotheliomas. The in vivo efficacy of HXR9 was tested in a mouse MSTO-211H flank tumor xenograft model.
We show that HOX genes are significantly dysregulated in malignant mesothelioma. Targeting HOX genes with HXR9 caused apoptotic cell death in all of the mesothelioma-derived cell lines, and prevented the growth of mesothelioma tumors in a mouse xenograft model. Furthermore, the sensitivity of these lines to HXR9 correlated with the relative expression of HOX genes that have either an oncogenic or tumor suppressive function in cancer. The analysis of HOX expression in primary mesothelioma tumors indicated that these cells could also be sensitive to the disruption of HOX activity by HXR9, and that the expression of HOXB4 is strongly associated with overall survival.
HOX genes are a potential therapeutic target in mesothelioma, and HOXB4 expression correlates with overall survival.
The HOX genes are a family of transcription factors characterized by highly conserved DNA- and co-factor binding domains. This conservation has been driven by their roles in some of the most fundamental patterning events that underlie early development . Most notable of these is the patterning of the anterior to posterior axis, for which a precise spatial and temporal order in the expression of HOX genes is required. This is achieved in part through a chromosomal arrangement whereby HOX genes are present in closely linked clusters allowing the sharing of common enhancer regions. In mammals there are four such clusters (A–D), containing a total of 39 HOX genes . The relative position of each HOX gene 3′ to 5′ within the cluster is reflected in a number of key attributes, including the spatial and temporal order of expression, whereby the 3′ most genes are expressed earlier than their 5′ neighbors. The nomenclature of the HOX genes reflects this precise chromosomal ordering, with members of each cluster being numbered with respect to the 3′ end, thus for example, the 3′ most member of cluster B is HOXB1 .
The 3′ to 5′ order of HOX genes is reflected not only in their expression patterns but also in their DNA binding specificities and co-factor interactions. For example, the products of the 3′ HOX genes (1 to 9) bind to another transcription factor, PBX, which modifies their binding specificity to DNA , influences their nucleocytoplasmic distribution , and also determines whether a HOX protein will activate of repress transcription of downstream target genes . This interaction with PBX is mediated through a highly conserved hexapeptide region on HOX proteins 1–9 that binds to a cleft in PBX [3, 5]. Once PBX has bound it can recruit other specific co-factors, including MEIS, which can then further modify HOX activity .
Although HOX genes were initially characterized as key developmental genes, they also function in adult stem cells to promote proliferation , and subsequently in their progeny to confer lineage-specific identities . Furthermore, HOX genes are strongly dys-regulated in cancer, and generally exhibit greatly increased expression. This differential change in expression in cancer may reflect the apparent ability of some HOX genes to function as tumor suppressors and some as oncogenes. Thus for example, HOXA5 acts as a tumor suppressor in breast cancer by stabilizing P53 , whilst forced expression of HOXB6 can immortalize fibroblast cells . Further examples of this phenomenon are listed in Table 1.
The dys-regulation of HOX genes has been demonstrated in a range of cancers, and in some it has been shown to be a potential therapeutic target through the use of a peptide, HXR9. HXR9 prevents PBX binding to HOX and triggers apoptosis in malignant cells, whilst sparing normal adult cells [11–17]. Although these studies include non-small cell lung cancer (NSCLC) , they do not encompass mesothelioma, a malignancy of the mesothelium cells which is most frequently found in the lung and is associated with long term exposure to asbestos . Mesothelioma has limited treatment options and generally a very poor prognosis , and therefore finding novel therapeutic approaches in this disease is an important goal. In this study we show that HOX dys-regulation is present in cell lines derived from mesothelioma, and in primary tumors, usually with a significant increase in the expression of those HOX genes that behave as oncogenes. Furthermore, antagonism of the HOX / PBX interaction in these cell lines triggers apoptosis, with malignant cells generally being considerably more sensitive to HXR9 than cells derived from non-malignant mesothelium cells.
Cell lines and culture
The cell lines used in this study are listed in Table 2. They were obtained from the ATCC through LGC Standards Ltd (UK), and were cultured according to the instructions on the LGC Standards website.
Synthesis of HXR9 and CXR9 peptides
HXR9 is an 18 amino acid peptide consisting of the previously identified hexapeptide sequence that can bind to PBX and nine C-terminal arginine residues (R9) that facilitate cell entry. The N-terminal and C-terminal amino bonds are in the D-isomer conformation, which has previously been shown to extend the half-life of the peptide to 12 h in human serum . CXR9 is a control peptide that lacks a functional hexapeptide sequence but which includes the R9 sequence. The sequences of these peptides have been published previously . All peptides were synthesized using conventional column based chemistry and purified to at least 80 % (Biosynthesis Inc., USA).
Imaging of cell cultures
Cells were plated in 6-well plates using 2 ml of medium and allowed to recover for at least 24 h. When approximately 60 % confluent, cells were treated with the active peptide HXR9 (60 μM) or the control peptide CXR9 (60 μM) for 3 h.
Immunohistochemistry for HOXA4, HOXA9, and HOXB4
Expression of HOXA4, HOXA9, and HOXB4 in mesothelioma and normal mesothelium tissue was investigated using 3 μm-thick, formalin fixed, paraffin embedded tissue array sections (MS081, US Biomax, Rockville, MD, USA). Immunohistochemical analysis was performed using a monoclonal rabbit anti-HOXB4 antibody (ab676093, 1:100 dilution, Abcam, Cambridge, UK), a polyclonal rabbit anti-HOXA4 antibody (ab131049, 1:500 dilution, Abcam, Cambridge, UK), and a polyclonal rabbit anti-HOXA9 antibody (ab191178, 1:75 dilution, Abcam, Cambridge, UK). The ABC detection method with peroxidase block (DakoCytomation) was used for all of these primary antibodies. Antigen retrieval was performed using pH 9.0 Tris/EDTA buffer (DakoCytomation) and heating in a microwave for 23 min.
Analysis of cell death and apoptosis
Cells were treated with HXR9 or CXR9 as described above. Cell viability was assessed using the MTS assay (Promega) according to the manufacturer’s instructions. Cells were harvested by incubating in trypsin-EDTA (Sigma) at 37 °C until detached and dissociated. Apoptotic cells were identified using flow cytometry (Beckman Coulter Epics XL Flow) and the Annexin V-PE apoptosis detection kit (BD Pharmingen) as described by the manufacturer’s protocol. Caspase-3 activity was measured using the EnzCheck Caspase-3 Assay Kit (Molecular Probes), using the protocol defined by the manufacturer.
RNA purification and reverse transcription
Total RNA was isolated from cells using the RNeasy Plus Mini Kit (Qiagen) by following the manufacturer’s protocol. The RNA was denatured by heating to 65 °C for 5 min. cDNA was synthesized from RNA using the Cloned AMV First Strand Synthesis Kit (Invitrogen) according to the manufacturer’s instructions.
Quantitative PCR was performed using the Stratagene MX3005P real-time PCR machine and the Brilliant SYBR Green QPCR Master Mix (Stratagene). The following primers were designed to facilitate the unique amplification of β-actin, c-Fos, and each HOX gene:
HsBeta-ActinF: 5′ ATGTACCCTGGCATTGCCGAC 3′
HsBeta-ActinR: 5′ GACTCGTCATACTCCTGCTTG 3′
HscFos1F: 5′ CCAACCTGCTGAAGGAGAAG 3′
HscFos1R: 5′ GCTGCTGATGCTCTTGACAG 3′
HsHOXA1F: 5′ CTGGCCCTGGCTACGTATAA 3′
HsHOXA1R: 5′ TCCAACTTTCCCTGTTTTGG 3′
HsHOXA4F: 5′ CCCTGGATGAAGAAGATCCA 3′
HsHOXA4R: 5′ AATTGGAGGATCGCATCTTG 3′
HsHOXA5F: 5′ CCGGAGAATGAAGTGGAAAA 3′
HsHOXA5R: 5′ ACGAGAACAGGGCTTCTTCA 3′
HsHOXA9F: 5′ AATAACCCAGCAGCCAACTG 3′
HsHOXA9R: 5′ ATTTTCATCCTGCGGTTCTG 3′
HsHOXB3F: 5′ TATGGCCTCAACCACCTTTC 3′
HsHOXB3R: 5′ AAGCCTGGGTACCACCTTCT 3′
HsHOXB4F: 5′ TCTTGGAGCTGGAGAAGGAA 3′
HsHOXB4R: 5′ GTTGGGCAACTTGTGGTCTT 3′
HsHOXB5F: 5′ AAGGCCTGGTCTGGGAGTAT 3′
HsHOXB5R: 5′ GCATCCACTCGCTCACTACA 3′
HsHOXB6F: 5′ ATTTCCTTCTGGCCCTCACT 3′
HsHOXB6R: 5′ GGAAGGTGGAGTTCACGAAA 3′
HsHOXB9F: 5′ TAATCAAAGACCCGGCTACG 3′
HsHOXB9R: 5′ CTACGGTCCCTGGTGAGGTA 3′
HsHOXC4F: 5′ CGCTCGAGGACAGCCTATAC 3′
HsHOXC4R: 5′ GCTCTGGGAGTGGTCTTCAG 3′
HsHOXC8F: 5′ CTCAGGCTACCAGCAGAACC 3′
HsHOXC8R: 5′ TTGGCGGAGGATTTACAGTC 3′
Mice and in vivo trial
All animal experiments were conducted in accordance with the United Kingdom Coordinating Committee on Cancer Research guidelines for the Welfare of Animals in Experimental Neoplasia and were approved by the University of Surrey Research Ethics Committee. The mice were kept in positive pressure isolators in 12 h light / dark cycles and food and water were available ad libitum.
Athymic nude mice were inoculated subcutaneously with a suspension of 2.5 × 106 MSTO-211H cells in culture media (100 μl). Once tumors reached volumes of approximately 100 mm3, mice were injected IP with PBS or 25 mg/Kg HXR9 in PBS (injection volume 100 μl), every 4 days. The mice were sacrificed after 36 days and the tumors were excised for RNA extraction, as previously described . Each treatment group contained ten mice. The mice were monitored carefully for signs of distress, including behavioral changes and weight loss.
Primary mesothelioma samples were obtained from 16 male and five female patients. The median patient age at diagnosis was 63.9 years (range, 38.2–79.53 years) and median survival was 9.04 months (range, 0.23–81.85 months). Recruitment was via a specialized multidisciplinary thoracic oncology clinic, involving thoracic surgeons, radiation oncologists, and medical oncologists. Histopathology and imaging review was undertaken for all patients. Patients underwent tumor resection at the Department of Thoracic Surgery, Guy’s & St Thomas’ NHS Foundation Trust. Tumor samples were confirmed as mesothelioma by pathological examination and categorized as a sarcomatoid, biphasic, or epithelial type using an antibody panel that included BerEP4, CEA, TTF1, Calretinin, WT1, CK5, MNF116, and EMA. Pseudoanonymised tissues and data were collected by the KHP Cancer Biobank, and subsequently released for this study in accordance with NHS REC approval number 07/H0804/91. Written informed consent was obtained from patients when they agreed to their tissue samples being included in the Biobank, it was not required for the specific use of these tissues in this project.
All values are given as the mean of three independent experiments and error bars show the standard error of the mean. Categorical variables were compared using Student’s t-test or a one-way ANOVA. Survival curves were generated using the Kaplan-Meier method and compared using the log-rank test. A p value < 0.05 was considered to be significant.
HOX gene expression in mesothelioma-derived cell lines and primary tumors
In order to assess the expression of HOX genes in mesothelioma we used QPCR to measure RNA levels in four cell lines derived from this malignancy: NCI-H28, NCI-H2052, NCI-H226, and MSTO-211H, together with Met-5A which is derived from non-malignant mesothelium cells (Table 2). HOX gene expression was also studied in primary mesothelioma tumors. The expression of HOX genes within each cell line and between cell lines varied considerably, with MSTO-211H and Met-5A generally having far higher expression than the other cell lines. The only HOX genes expressed uniquely by a single cell line were HOXC12 and HOXD12, in Met-5A. Analysis of HOX genes that are known to have oncogenic or tumor suppressive functions (Table 1) likewise reveals considerable variation, although Met-5A showed higher expression of the potential tumor suppressor genes HOXA4 and HOXA5 compared to the malignant cell lines (Fig. 1a). We also assessed the expression of these HOX genes in 21 primary tumors using RT-QPCR, as well the protein expression of the three most strongly expressed, HOXA4, HOXA9, and HOXB4 at the protein level using immunohistochemistry (Fig. 1b).
High HOXB4 tumor expression is associated with poor overall survival
We looked for associations between the RNA expression levels of the different HOX genes and patient survival. The tumors of patients surviving less than 6 months had a significantly higher expression of HOXB4 (p = 0.0166; Fig. 1c), and likewise a Kaplan-Meier analysis of overall survival (OS) showed that high HOXB4 tumor expression was associated with a significantly shorter OS (p = 0.041; Fig. 1d).
HXR9 is cytotoxic to mesothelioma cells
Given the high level of HOX expression in the mesothelioma cell lines, we treated cells with the HOX / PBX inhibitor HXR9 that has previously been shown to block HOX / PBX interactions and trigger apoptosis in a number of other cancers [11–17]. Use of a fluorescently labeled version of HXR9 demonstrated that it can be taken up by the cell lines studied here (Fig. 2a), and the MTS assay for cell viability revealed that HXR9 is cytotoxic in all five cell lines (Fig. 2b,c; Table 2). The non-malignant line Met-5A is amongst the least sensitive with an IC50 of 98 μM, whilst the NCI-H28 cell line is the most sensitive with an IC50 of 18 μM (Fig. 2c, Table 2).
HXR9 triggers apoptosis
Previous studies have suggested that the mechanism of cell death when HOX function is blocked by HXR9 is primarily through apoptosis [11–17]. To establish whether this is also the case of the mesothelioma derived cell lines, a standard FACS based assay for apoptosis-associated cell membrane changes was used. This involves the use of Annexin V that binds to membrane components usually located on the cytoplasmic side but which relocate to the external surface during apoptosis , and a fluorescent dye (7AAD) which binds to DNA but can only enter cells when membrane integrity has been lost. This assay revealed that all the mesothelioma cell lines underwent apoptosis when treated with HXR9 at the relevant IC50 (Fig. 3), with the non-malignant cell line Met-5A showing the lowest level of apoptosis and NCI-H2052 the highest (Fig. 3c).
The induction of apoptosis by HXR9 is thought to depend, at least in part, upon a rapid increase in cFos expression , and QPCR analysis of the HXR9 treated cells correspondingly showed a significant increase in cFos in all of the cell lines, with the smallest increase in Met-5A and the largest increase in the most sensitive cell line, NCI-H28 (Fig. 4a). Correspondingly NCI-H28 also showed the greatest increase in Caspase 3 activity (a protease involved in the apoptotic pathway; Fig. 4b), whilst Met-5A failed to show any significant increase in caspase activity (Fig. 4c).
Sensitivity to HXR9 correlates with the expression of specific HOX genes
The expression of HOX genes with previously identified oncogenic or tumor suppressor properties (Table 1; Fig. 1), raises the possibility that the expression profile of these genes could determine the sensitivity of cells to HXR9. To assess this we divided HOX genes into two groups – those with potential oncogenic functions, and those with possible tumor suppressor functions. An expression ratio was obtained by dividing the total expression of genes in the former group with that in the latter (‘O/S ratio’). This revealed that the most sensitive cell line, NCI-H28, has the highest O/S ratio, whilst Met-5a and the least sensitive malignant line, NCI-H226, have the lowest O/S ratios (Fig. 5a). Plotting these ratios against the IC50 for each cell line suggest a positive correlation between the O/S ratio and sensitivity (Fig. 5b). Furthermore, the calculated O/S ratios for the primary mesothelium tumors indicate that these cells could also be sensitive HXR9 (Fig. 5b).
HXR9 blocks the growth of mesothelioma tumors in vivo
In order to determine whether HXR9 could also block tumor growth in vivo, we established a xenograft mouse flank model using the MSTO-211H cell line. Mice were injected IP with either PBS or 25 mg/Kg HXR9 in PBS every 4 days after tumors had grown to a mean volume of 100 mm3. HXR9 significantly retarded tumor growth compared to PBS alone (Fig. 6a). In tumors from mice injected with PBS only, we found a significant, linear relationship between the expression of HOXB4 and final tumor size (r2 = 0.8278; p = 0.0321; Fig. 6b).
The dys-regulation of HOX genes in cancer is now well established, and in many cases a putative function for individual HOX genes has been established . Despite a high degree of sequence and regulatory conservation between HOX genes, there is apparently a wide range of cancer specific functions which include both oncogenic and tumor suppressing activities. Thus for example the fifth gene of the HOXA complex, HOXA5, acts primarily as a tumor suppressor in breast cancer through stabilizing p53 , whilst its closely related counterpart in the HOXB cluster, HOXB5, can be defined as an oncogene as it can immortalize fibroblast cells upon transfection .
None of these studies have as yet addressed whether HOX genes are dys-regulated in mesothelioma, but here we show that cell lines derived from mesothelioma as well as primary mesothelioma cells have distinctly different HOX expression patterns from the Met-5a cell line that is derived from normal mesothelium. One of the most striking differences is the expression of HOXC12 and HOXD12 by Met-5a but not by any of the mesothelioma cell lines. HOXC12 is repressed in follicular lymphoma through hypermethylation of its promoter, and has also been implicated in the differentiation of follicle cells , both of which suggest a possible function in tumor suppression. Likewise, the function of HOXD12 has not been defined, but it has been shown to be silenced in melanoma cells through the methylation of its promoter .
Another oncogenic HOX gene that we found to be up-regulated in primary mesothelioma tumors was HOXB4. High HOXB4 expression levels were associated with shorter OS, suggesting that HOXB4 expression is a potential prognostic factor in this malignancy. We also found that there was a positive, linear relationship between HOXB4 expression and tumor growth in a mouse model of human mesothelioma. Given the functional redundancy amongst HOX proteins, this finding that HOXB4 was the only HOX gene among the 39-strong family to have any prognostic significance seems unexpected. However, there are a number of other cancers for which a single HOX gene alone acts as a prognostic marker, and the identity of the HOX gene in each case varies from one malignancy to another. Examples include HOXC6 in gastric cancer, HOXB8 in ovarian cancer, and HOXD3 in breast cancer . This might reflect the embryonic origins of different cancer types, as HOX gene expression in adult cells tends to reflect their developmental origin . From a practical view point, there are currently no reliable markers of OS in mesothelioma , and the use of HOXB4 as a prognostic marker in this context therefore justifies further evaluation.
In this study we have found that the ratio of expression between HOX genes with a putative oncogenic function and those that have tumor suppressor activity (‘O/S ratio’) predicts which mesothelioma cell lines are most sensitive to HXR9, a peptide that prevents HOX proteins binding to PBX and has been shown to cause apoptosis in other malignancies [11–17]. The O/S ratio may indicate the degree to which malignant cells are dependent on the activity of oncogenic HOX genes for their proliferation and survival, a concept similar to the idea of ‘oncogene addiction’ , which would explain their sensitivity to HXR9. The extent to which this is true is yet to be determined, but at a more practical level the O/S ratio might act as a biomarker for the sensitivity of mesothelioma cells to HXR9, and could ultimately be used to select patients that might benefit from this therapeutic approach.
Our findings indicate that the HOX genes are widely dysregulated and often strongly upregulated in mesothelioma, and that elevated HOXB4 expression predicts shorter OS in mesothelioma patients. Targeting the interaction between HOX proteins and their PBX cofactor causes apoptosis in mesothelioma cells in vitro and retards tumor growth in vivo, indicating that HOX proteins are a potential therapeutic target in this malignancy.
- O/S ratio:
ratio of oncogenic to tumor suppressor HOX gene expression
Hoegg S, Meyer A. Hox clusters as models for vertebrate genome evolution. Trends Genet. 2005;21:421–4.
Scott MP. A rational nomenclature for vertebrate homeobox (HOX) genes. Nucleic Acids Res. 1993;21:1687–8.
Phelan ML, Sadoul R, Featherstone MS. Functional differences between HOX proteins conferred by two residues in the homeodomain N-terminal arm. Mol Cell Biol. 1994;14:5066–75.
Pinsonneault J, Florence B, Vaessin H, McGinnis W. A model for extradenticle function as a switch that changes HOX proteins from repressors to activators. EMBO J. 1997;16:2032–42.
Morgan R, In der Rieden P, Hooiveld MH, Durston AJ. Identifying HOX paralog groups by the PBX-binding region. Trends Genet. 2000;16:66–7.
Knoepfler PS, Bergstrom DA, Uetsuki T, Dac-Korytko I, Sun YH, Wright WE, et al. A conserved motif N-terminal to the DNA-binding domains of myogenic bHLH transcription factors mediates cooperative DNA binding with pbx-Meis1/Prep1. Nucleic Acids Res. 1999;27:3752–61.
Antonchuk J, Sauvageau G, Humphries RK. HOXB4-induced expansion of adult hematopoietic stem cells ex vivo. Cell. 2002;109:39–45.
Shimamoto T, Ohyashiki K, Toyama K, Takeshita K. Homeobox genes in hematopoiesis and leukemogenesis. Int J Hematol. 1998;67:339–50.
Raman V, Martensen SA, Reisman D, Evron E, Odenwald WF, Jaffee E, et al. Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature. 2000;405:974–8.
Fischbach NA, Rozenfeld S, Shen W, Fong S, Chrobak D, Ginzinger D, et al. HOXB6 overexpression in murine bone marrow immortalizes a myelomonocytic precursor in vitro and causes hematopoietic stem cell expansion and acute myeloid leukemia in vivo. Blood. 2005;105:1456–66.
Daniels TR, Neacato II, Rodriguez JA, Pandha HS, Morgan R, Penichet ML. Disruption of HOX activity leads to cell death that can be enhanced by the interference of iron uptake in malignant B cells. Leukemia. 2010;24:1555–65.
Morgan R, Boxall A, Harrington KJ, Simpson GR, Gillett C, Michael A, et al. Targeting the HOX/PBX dimer in breast cancer. Breast Cancer Res Treat. 2012;136:389–98.
Morgan R, Boxall A, Harrington KJ, Simpson GR, Michael A, Pandha HS. Targeting HOX transcription factors in prostate cancer. BMC Urol. 2014;14:17.
Morgan R, Pirard PM, Shears L, Sohal J, Pettengell R, Pandha HS. Antagonism of HOX/PBX dimer formation blocks the in vivo proliferation of melanoma. Cancer Res. 2007;67:5806–13.
Morgan R, Plowright L, Harrington KJ, Michael A, Pandha HS. Targeting HOX and PBX transcription factors in ovarian cancer. BMC Cancer. 2010;10:89.
Plowright L, Harrington KJ, Pandha HS, Morgan R. HOX transcription factors are potential therapeutic targets in non-small-cell lung cancer (targeting HOX genes in lung cancer). Br J Cancer. 2009;100:470–5.
Shears L, Plowright L, Harrington K, Pandha HS, Morgan R. Disrupting the interaction between HOX and PBX causes necrotic and apoptotic cell death in the renal cancer lines CaKi-2 and 769-P. J Urol. 2008;180:2196–201.
van Meerbeeck JP, Scherpereel A, Surmont VF, Baas P. Malignant pleural mesothelioma: the standard of care and challenges for future management. Crit Rev Oncol Hematol. 2011;78:92–111.
Schostak M, Schwall GP, Poznanovic S, Groebe K, Muller M, Messinger D, et al. Annexin A3 in urine: a highly specific noninvasive marker for prostate cancer early detection. J Urol. 2009;181:343–53.
Shah N, Sukumar S. The Hox genes and their roles in oncogenesis. Nat Rev Cancer. 2010;10:361–71.
Harris DM, Hazan-Haley I, Coombes K, Bueso-Ramos C, Liu J, Liu Z, et al. Transformation of human mesenchymal cells and skin fibroblasts into hematopoietic cells. PLoS One. 2011;6, e21250.
Shang L, Pruett ND, Awgulewitsch A. Hoxc12 expression pattern in developing and cycling murine hair follicles. Mech Dev. 2002;113:207–10.
Furuta J, Nobeyama Y, Umebayashi Y, Otsuka F, Kikuchi K, Ushijima T. Silencing of Peroxiredoxin 2 and aberrant methylation of 33 CpG islands in putative promoter regions in human malignant melanomas. Cancer Res. 2006;66:6080–6.
Shaoqiang C, Yue Z, Yang L, Hong Z, Lina Z, Da P, et al. Expression of HOXD3 correlates with shorter survival in patients with invasive breast cancer. Clin Exp Metastasis. 2013;30:155–63.
Morgan R. Hox genes: a continuation of embryonic patterning? Trends Genet. 2006;22:67–9.
Mineo TC, Ambrogi V. Malignant pleural mesothelioma: factors influencing the prognosis. Oncology (Williston Park). 2012;26:1164–75.
McCormick F. Cancer therapy based on oncogene addiction. J Surg Oncol. 2011;103:464–7.
Mohankumar KM, Xu XQ, Zhu T, Kannan N, Miller LD, Liu ET, et al. HOXA1- stimulated oncogenicity is mediated by selective upregulation of components of the p44/42 MAP kinase pathway in human mammary carcinoma cells. Oncogene. 2007;26:3998–4008.
Abramovich C, Humphries RK. Hox regulation of normal and leukemic hematopoietic stem cells. Curr Opin Hematol. 2005;12:210–6.
Hayashida T, Takahashi F, Chiba N, Brachtel E, Takahashi M, Godin-Heymann N, et al. HOXB9, a gene overexpressed in breast cancer, promotes tumorigenicity and lung metastasis. Proc Natl Acad Sci U S A. 2010;107:1100–5.
Miller GJ, Miller HL, van Bokhoven A, Lambert JR, Werahera PN, Schirripa O, et al. Aberrant HOXC expression accompanies the malignant phenotype in human prostate. Cancer Res. 2003;63:5879–88.
Klausen C, Leung PC, Auersperg N. Cell motility and spreading are suppressed by HOXA4 in ovarian cancer cells: possible involvement of beta1 integrin. Mol Cancer Res. 2009;7:1425–37.
Adwan H, Zhivkova-Galunska M, Georges R, Eyol E, Kleeff J, Giese NA, et al. Expression of HOXC8 is inversely related to the progression and metastasis of pancreatic ductal adenocarcinoma. Br J Cancer. 2011;105:288–95.
Lechner JF, Tokiwa T, LaVeck M, Benedict WF, Banks-Schlegel S, Yeager Jr H, et al. Asbestos-associated chromosomal changes in human mesothelial cells. Proc Natl Acad Sci U S A. 1985;82:3884–8.
Bepler G, Koehler A, Kiefer P, Havemann K, Beisenherz K, Jaques G, et al. Characterization of the state of differentiation of six newly established human non-smallcell lung cancer cell lines. Differentiation. 1988;37:158–71.
Fan D, Yano S, Shinohara H, Solorzano C, Van Arsdall M, Bucana CD, et al. Targeted therapy against human lung cancer in nude mice by high-affinity recombinant antimesothelin single-chain Fv immunotoxin. Mol Cancer Ther. 2002;1:595–600.
The authors gratefully acknowledge the support of the British Lung Foundation, grant number ICAPPG10-1. KJH acknowledges support from the ICR/RM NIHR Biomedical Research Centre.
The authors declare that they have no competing interests
RM designed and oversaw the study and wrote the manuscript draft. GS conducted the in vivo study. SG conducted the cell culture experiments and assays. CG oversaw the collection of tumour samples and helped analyse the data. ZT advised on the design and interpretation of the cell culture studies. JS oversaw the collection of tumour samples and helped analyse the data. KJH helped design the study and write the manuscript. HSP helped design the study, write the manuscript, and analyse the data. All of the authors have and approved the final version of the manuscript.