Deregulated NKL Homeobox Genes in B-Cell Lymphoma

Recently, we have described physiological expression patterns of NKL homeobox genes in early hematopoiesis and in subsequent lymphopoiesis. We identified nine genes which constitute the so-called NKL-code. Aberrant overexpression of code-members or ectopically activated non-code NKL homeobox genes are described in T-cell leukemia and in T- and B-cell lymphoma, highlighting their oncogenic role in lymphoid malignancies. Here, we introduce the NKL-code in normal hematopoiesis and focus on deregulated NKL homeobox genes in B-cell lymphoma, including HLX, MSX1 and NKX2-2 in Hodgkin lymphoma; HLX, NKX2-1 and NKX6-3 in diffuse large B-cell lymphoma; and NKX2-3 in splenic marginal zone lymphoma. Thus, the roles of various members of the NKL homeobox gene subclass are considered in normal and pathological hematopoiesis in detail.


Hematopoiesis and B-Cell Development
The process of hematopoiesis is responsible for the production of all types of blood cells. Hematopoietic stem and progenitor cells (HSCs) generate common myeloid and lymphoid progenitors (CMP and CLP) which, respectively, represent the starting points for the myeloid and lymphoid cell lineages. The latter produces all types of lymphocytes comprising B-cells, T-cells, natural killer (NK)-cells and innate lymphoid cells. Early B-cell development, which includes the rearrangements of the B-cell receptor genes (immunoglobulin heavy chain, IGH), takes place in the bone marrow and begins with the CLP-derived B-cell progenitor (BCP). BCPs differentiate via the pro-B-cell and pre-B-cell stages into naïve B-cells. In contrast, early T-cell progenitors (ETP) migrate into the thymus to complete their differentiation. For the final differentiation steps to memory B-cells (memo B-cell) and plasma cells via the stage of germinal center (GC) B-cells, naïve B-cells migrate from the bone marrow into lymph nodes, the spleen and other lymphoid tissues [1][2][3] (Figure 1). In these compartments, additional molecular alterations occur, like somatic hypermutation and class switching of the IGH genes. These alterations are performed at the DNA level and the process of class switching is additionally connected with gene rearrangements.
The main regulatory steps of lymphopoiesis including B-cell development are controlled at the transcriptional level [3,4]. Accordingly, several transcription factors (TFs), like BCL6, EBF1, MYB, PAX5, PRDM1 (alias name: BLIMP1) and TCF3 (E2A), are members of a B-cell specific regulatory network which orchestrates basic differentiation processes [5][6][7]. TCF3 plays a prominent role in the development of all types of lymphocytes, while EBF1 and PAX5 are master factors of the B-cell lineage. BCL6 and PRDM1 inhibit each other and are involved in differentiation processes taking place in the GC. Provoked by aberrant chromosomal rearrangements or gene mutations, deregulations of these developmental TFs are thought to contribute to the generation of B-cell malignancies [8,9]. Abnormal rearrangements of the IGH genes represent a frequent mechanism of oncogene activation, while deregulated hypermutation is known to be responsible for many gene mutations.

Classification of Homeobox Genes
Homeobox genes encode TFs, which regulate fundamental processes in development and differentiation in both embryogenesis and the adult. They share the conserved 180 bp long homeobox, which encodes the homeodomain at the protein level. This domain consists of 60 amino acid residues and mediates specific interactions with DNA, chromatin, non-coding (nc)RNA and cooperating TFs, thus representing a common platform of their gene regulatory activities [10].
The subgroup of NK-like homeobox genes, which were later called NKL homeobox genes, have been reported for the first time by Nirenberg and Kim (abbreviated as NK) in the fruit fly Drosophila. In this invertebrate organism, the subgroup members are arranged in a cluster consisting of six genes [11]. Additional orthologous genes have since been identified and the group extended. These studies indicated that this clustering represents the ancient gene order, which is now just barely visible in vertebrates [12]. Thus, in contrast to the still-clustered HOX-genes, NKL homeobox genes show only relicts of a clustered arrangement in humans. In addition to their conserved homeodomain, NKL homeodomain proteins share a short, conserved sequence in their N-terminal part, which has been termed the engrailed-homology motif (EH1) [13,14]. This motif performs physical interactions with corepressors of the groucho family, thus mediating transcriptional inhibition [15]. Figure 2 depicts a schematic structure of NKL homeodomain proteins. Most NKL homeobox genes are functionally associated with mesodermal development, possibly displaying their ancient function [16]. A systematic classification of all 235 human homeobox genes has generated a panel of eleven classes and several subclasses. Main classes are called antennapedia (ANTP) and paired box (PRD), comprising 150 genes. Other classes identified are CERS, CUT, HNF, LIM, POU, PROS, SINE, TALE, and ZF. Accordingly, NKL homeobox genes represent a subclass of the ANTP class and contain 48 members in humans [17].
Then, we analyzed the physiological activity of NKL homeobox genes in early hematopoiesis and T-cell development. This exercise revealed nine members, comprising HHEX, HLX, MSX1, NANOG, NKX2-3, NKX3-1, NKX6-3, TLX2, and VENTX ( Figure 3). They showed a specific expression pattern in stem cells, progenitor cells and immature thymocytes, but not in mature T-cells, which tested negative. We named this pattern and the respective genes the NKL-code [22]. This code demonstrated that most NKL homeobox oncogenes in T-ALL are ectopically expressed. For example, NKX2-5 is normally expressed in the developing heart and spleen but not in any hematopoietic cell [32]. Furthermore, MSX1 is normally expressed in hematopoietic progenitors, including CLP and BCP in addition to mature NK-cells, but not in the T-cell lineage ( Figure 3). Accordingly, MSX1 is an oncogene in T-ALL and a tumor suppressor in NK-cell leukemia [24,33,34]. HLX (HLX1, H2.0 or HB241) was the first described NKL homeobox gene that is physiologically expressed in hematopoietic cells, including myeloid and B-cells, but not in T-cells [35]. These data support the reported expression pattern of NKL-code members and highlight their functional role in leukemia.
Additional gene codes containing NKL homeobox genes have been published for the developing neural tube, pharyngeal region, and teeth [36][37][38]. The code members for the neural tube comprise DBX1, DBX2, NKX2-2, NKX6-1, and NKX6-2 [36]. In the pharyngeal region, the gene code consists exclusively of all six DLX family members, while in developing teeth, BARX1, DLX1, DLX2, MSX1, and MSX2 create a code [37,38]. Most of those NKL homeobox gene code members are regulated by signalling pathways and perform cross-reactivity. In the neural tube, the hedgehog-and BMP-pathways are regulated by ligand gradients which are created in opposite directions, thus regulating NKL homeobox gene activities [36]. Therefore, differentiation processes are frequently controlled by particular NKL homeobox genes, via formation of a code.
Two of these genes, HLX and HHEX (HEX or PRH), represent the first described non-HOX homeobox genes expressed in hematopoietic cells [35,40]. Expression analyses of these two genes revealed activity in B-cells and myeloid cells, while T-cells were described to be negative [35,[41][42][43]. Moreover, downregulation of HHEX was shown to be crucial for normal T-cell differentiation [44]. Of note, the data failed to detect HHEX in plasma cells, consistent with our screening data for the NKL-code [39,43]. Accordingly, analysis of HHEX-knockout mice showed reduced numbers of both mature and pre-B-cells, demonstrating the importance of HHEX for B-cell development [45,46]. Forced expression of HLX in hematopoietic progenitors enhanced myeloid differentiation but arrested the development of B-cells at the pro-B-cell stage and of T-cells at the CD4/CD8 double-positive stage [41,47]. In T H 1-cells, HLX is induced by the TF TBX21 (TBET) and thereby involved in the expression of interferon gamma IFNG [48,49]. However, in NK-cells, HLX performs negative regulation of IFNG, demonstrating context-dependent control of target genes [50]. Collectively, these data highlight HHEX and HLX as important regulators of hematopoiesis, including B-cell differentiation.
HHEX: In addition to their activity in developing B-cells, the four NKL-code members HHEX, HLX, MSX1, and NKX6-3 are involved in the differentiation of other tissues and organs as well. HHEX is expressed in parts of the early embryo, including the chorion and yolk sac, later in primordia of the liver and the thyroid, and then in fetal liver, thyroid and lung [51,52]. HHEX is also expressed in the 3rd pharyngeal pouch, which generates the thymus, and in primordia of the pancreas and the gallbladder [53]. Furthermore, HHEX plays a role in vascular and lymphatic development [54]. The expression of HHEX was also detected in normal breast tissue [55]. Accordingly, HHEX performs tumor suppressor activities in breast cancer [56]. In the embryonal anterior endoderm which generates the lung, thyroid, pancreas, and the liver, HHEX is regulated by a complex network which includes the BMP-and WNT-signalling pathways and the TFs LIM1, NODAL, OTX2, and VENTX [57]. Components of this network may play a regulatory role in the hematopoietic system, as well.
HLX: The embryonal expression pattern of HLX indicates some degree of overlap with HHEX, as shown in the foregut, liver, gallbladder, and lung. In addition, the splanchnic mesoderm and mesenteric tissues tested positive for HLX activity in the embryo [58]. HLX is also expressed in the placenta, and controls the switch from white to brown adipose tissue [59,60]. Furthermore, HLX is involved in the differentiation of embryonal stem cells into the intestinal lineage [61]. A screening in induced pluripotent stem cells (iPSCs) for genes regulating pluripotency revealed HHEX and HLX [62]. Thus, both genes are involved in pluripotency and reprogramming, highlighting their potential in cell differentiation processes [62].
MSX1: In early embryogenesis, MSX1 is expressed in the neural plate border region and is thus involved in the generation of neural crest cells and the preplacodal ectoderm [63][64][65][66]. These cells generate a multiplicity of tissues and structures which are fundamental for vertebrate development and evolution. Later on, MSX1 is expressed in neural crest cells and their derived tissues, including the teeth [67,68]. MSX1 also plays a role in the development of craniofacial structures and neural tissues, including the brain [69][70][71]. Accordingly, mutations in the MSX1 gene are frequently found to be connected to particular malformations of craniofacial tissues. Furthermore, MSX1 is expressed in the mammary gland [72]. Functionally, MSX1 is able to dedifferentiate cells. In muscle cells, it has been shown that MSX1 dedifferentiates myotubes by repressing the muscle master factor MYOD and the muscle-specific myogenin [73,74]. These data compellingly demonstrated the developmental potential of this TF.
NKX6-3: Finally, NKX6-3 is expressed in the developing stomach and hindbrain [75][76][77]. Depleting mutations of NKX6-3 in gastric epithelial cells result in the activation of APOBEC family members, which, in turn, enhance the generation of additional mutations and, subsequently, gastric cancer [76]. In addition to MSX1, NKX6-3 is expressed in the neural plate border region as well. Knockdown and overexpression experiments indicated a dominant role of NKX6-3 in the development of neural crest cells [78]. NKX6-3 was able to induce an ectopically neural crest when overexpressed, while NKX6-3 knockdown generated defects in the neural crest [78]. Taken together, experimental data of MSX1 and NKX6-3 highlight the importance and impact of NKL homeobox genes for these pluripotent cells.
Regulation of different developmental processes by the same pathways and TFs, including homeodomain proteins, is a frequent observation in embryogenesis [79]. For example, NKL homeobox gene NKX2-5 basically regulates the development of both heart and spleen [32,80,81]. Therefore, aberrant activities of developmental homeodomain TFs may recapitulate these tissue-specific operations ectopically or at the wrong stage of differentiation, which may lead to tumorigenesis. NKL homeobox gene NKX3-1 performs master gene activities in the development of the prostate and plays a role in the early stages of hematopoiesis, showing physiological activity in different tissues [22,28]. NKL homeobox gene NKX2-5 represents a master gene for the development of the heart in vertebrates and thereby regulates the expression of specific target genes, including MEF2C [80]. This function is reactivated in T-ALL and mediates the ectopic deregulation of MEF2C, which constitutes a major oncogene in this malignancy [18,25,82,83]. Thus, consideration of known physiological operations of aberrantly activated NKL-code members and non-members may reveal findings of clinical importance. The druggability of NKL homeobox genes in particular, and of TFs in general, is difficult to define, but their regulated genes may represent suitable and effective targets for therapy [84].

Deregulated NKL Homeobox Genes In B-Cell Malignancies
Developmental arrest is a main and widespread feature of cancer cells [85][86][87]. As shown by Allen and coworkers in 1995, forced expression of HLX in hematopoietic progenitors resulted in developmental arrest of pro-B-cells [47]. Therefore, this experiment represents the first hint of the oncogenic potential of NKL homeobox genes in B-cells. Ferrando and coworkers correlated aberrant expression of NKL homeobox gene TLX1 with the developmental arrest of malignant thymocytes at the double-positive stage in T-ALL [88]. Thus, the correlation of aberrantly expressed NKL homeobox genes with particular stages of differentiation from lymphocytes highlights the developmental impact of this gene subclass in lymphoid tumors. Recently, we screened deregulated NKL homeobox genes in a variety of B-cell malignancies, which included Hodgkin lymphoma (HL), follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL), hairy cell leukemia (HCL), mantle cell lymphoma (MCL), and splenic marginal zone lymphoma (SMZL). This study revealed 13 aberrantly overexpressed NKL homeobox genes (Table 1), supporting the relevance of this gene group for B-cell cancers [39]. In the following, we discuss in more detail seven selected NKL homeobox genes which have been studied by us and others in B-cell lymphomas, both in primary cells and in cell lines. This table is modified as described previously [39].
HLX: HLX is a member of the hematopoietic NKL-code and is expressed in early hematopoiesis and BCPs [22,39]. Aberrant expression of HLX has been found in HL, FL, MCL, and SMZL [39]. HLX is part of a regulatory network consisting of B-cell associated NKL-code members HHEX, MSX1 and NKX6-3 [39]. Aberrant overexpression of HLX in HL cell line L-540 mediated downregulation of MSX1 and NKX6-3, in addition to B-cell factors BCL11A and SPIB, and of pro-apoptotic factor BCL2L11 (BIM) [89]. These regulatory relationships may underlie the described differentiation arrest in immature B-cells after forced expression of HLX [47]. Furthermore, they enhance survival of the tumor cells which represent a main property of HL [89]. ChIP-seq data from ENCODE indicated HLX as target gene of STAT3, which is a transcriptional mediator of several signalling pathways [90]. Accordingly, aberrant activation of STAT3 by JAK2-mediated phosphorylation and/or HDAC-mediated deacetylation contributes to enhanced HLX expression in HL [89]. Acetylation of STAT3 was shown to regulate its nuclear localization, representing an additional level of gene regulation which depends on the activity of acetylating and deacetylating enzymes. Of note, HDACs are druggable enzymes and, thus, potential targets for rational therapies to inhibit STAT3-signalling. Taken together, HLX is embedded in major oncogenic disturbances of HL, including aberrant signalling and apoptosis.
MSX1: MSX1 is normally expressed in CLPs and BCPs and downregulated in the course of B-cell development [39]. Aberrantly overexpressed MSX1 in HL cell line L-1236 inhibited the expression of the B-cell factor ZHX2, probably using histone H1 as co-repressor [91,92]. Silencing of ZHX2 may contribute to the deregulated B-cell phenotype in HL. In ovarian cancer cells, MSX1 performs inhibition of cyclin D1 (CCND1) in addition to other cyclins and cell cycle regulators, while, in primary mesenchymal and epithelial progenitor cell types, MSX1 activates CCND1 [93,94]. These data reflect a balanced interplay between proliferation and differentiation. Accordingly, an inhibitory impact of MSX1 on CCND1 expression was lost in MCL cells containing chromosomal translocation t(11;14)(q13;q32). This genomic rearrangement separates the regulatory MSX1 binding site from the coding part of CCND1 [95]. Therefore, this oncogenic alteration disturbs the indicated balanced interplay and may support CCND1-mediated cell proliferation in MCL.
NKX2-3: NKX2-3 is normally expressed in HSCs, thus representing a hematopoietic stem cell factor [22]. In addition, NKX2-3 acts as a master gene for the embryonal development of the spleen [100]. Thus, NKX2-3 is closely associated with the differentiation of hematopoietic cells and tissues. However, in advanced hematopoietic stages or mature blood cells NKX2-3 is silenced. In SMZL, NKX2-3 is aberrantly activated by chromosomal translocation t(10;14)(q24;q32), juxtaposing the locus of this NKL homeobox gene to that of IGH [101]. Downstream analyses indicated aberrant activation of B-cell receptor signalling, enhanced expression of integrins, adhesion factor MADCAM1, and of chemokine receptor CXCR4 [101]. These features may underlie malignant transformation and homing of the tumor cells to the spleen and lymph nodes. In addition to SMZL, aberrant expression of NKX2-3 has been detected in DLBCL, FL, MCL, chronic lymphoid leukemia, and multiple myeloma (MM) [101]. Furthermore, deregulated expression of NKX2-3 has been associated with Crohn disease, ulcerative colitis and inflammatory bowel disease [102]. These diseases originate from immunological disturbances which may share certain pathological aspects with particular B-cell malignancies.
NKX6-3: NKX6-3 is a member of the NKL-code and normally expressed in BCPs, GC B-cells and plasma cells [39]. This pattern indicates important functions for the differentiation of B-cells and the status of plasma cells. Aberrant overexpression of NKX6-3 was detected in FL, DLBCL and MCL patients [39]. DLBCL cell line DOHH-2 showed overexpression of NKX6-3 that was mediated by aberrant BMP-signaling and enhanced activity of chromatin factor AUTS2 [39]. Interestingly, infection of this cell line with Epstein-Barr virus (EBV) resulted in enhanced expression of HLX via STAT3 which in turn repressed NKX6-3 [103]. The EBV-encoded factors LMP1 and LMP2A were shown to mediate STAT3 activation in this cell line. These data may explain the described malignant associations of EBV and B-cell lymphomas [104].
BARX2: BARX2 is normally expressed during embryogenesis in several tissues, including the nervous system, Rathke´s pouch and submandibular glands, and in the adult ovarian epithel [105][106][107]. However, BARX2 is not expressed in developing or adult hematopoietic cells or tissues. Aberrant activity of BARX2 has been found in HCL and MCL patients [39]. Furthermore, BARX2 and NKX2-3 were the only NKL homeobox genes aberrantly expressed in MM [101,108]. Interestingly, except for BARX2, MM and primary effusion lymphoma cell lines lacked activity of any NKL-code member, indicating aberrant downregulation of HHEX and NKX6-3. Thus, in mature B-cell malignancies, NKL homeobox genes may perform tumor suppressor activity.

Cell Lines as Models for Deregulated NKL Homeobox Genes
Cell lines represent experimental models for that tumor type from which they were derived. Of note, it is of fundamental importance to use authenticated, well characterized, and annotated cell lines to be able to extrapolate cell line data to particular cancers, including B-cell malignancies. We have evaluated and systematically listed hematopoietic cell lines which meet these criteria [109]. In addition, we reviewed cell lines derived from several entities of B-cell malignancies, including B-cell precursor-leukemia, MM, primary effusion lymphoma, primary mediastinal B-cell lymphoma, double-hit B-cell lymphoma, and HL, to highlight appropriate models for particular tumors [110][111][112][113][114][115]. Here, we presented data obtained from both patients and cell lines. Deregulated NKL homeobox genes were usually identified in patients and subsequently investigated in cell lines, analyzing mechanisms of deregulation and downstream activities. Table 2 shows malignant B-cell lines and corresponding aberrant NKL homeobox gene activities to help identify models which would be suitable for the type of cancer and/or the gene of interest.

Conclusions
NKL homeobox genes are physiologically expressed in hematopoiesis including B-cell development, in a specific pattern which we have termed the NKL-code. Aberrant activities of these basic developmental regulators are involved in the pathogenesis of B-cell malignancies. The knowledge of their pathophysiological activity and the understanding of their function may contribute to improved diagnostics and novel therapies in the future. Finally, in this field of research, validated cell lines represent informative models to explore the landscape of NKL homeobox genes.