PRH in haematopoiesis and leukaemia
PRH is required for haematopoietic cell differentiation [56] at multiple stages of the differentiation process, reviewed in [20, 57]. Studies using embryoid body differentiation and blastocyst complementation have demonstrated critical roles for PRH in the development of definitive haematopoietic stem cells (HSCs) and B-cells [58–61]. Conditional knockout mice (Mx-Cre and Vav-Cre) revealed broadly similar results namely that HHEX is not required for maintenance of adult HSCs and myeloid lineages either postnatally, or at later time points, but is essential for the commitment of multiple lymphoid lineages at the stage of the common lymphoid progenitor. Although there are some differences between the two models presumably because of differences in Cre induction and the timing of knockout, both models show decreased expression of cyclin D1 and effects on lymphopoiesis [62, 63]. This decrease in cyclin D1 expression with HHEX knockout is also observed in ES cell derived haematopoietc colonies [61]. Overexpression of cyclin D1 in progenitor cell populations after inducible knock out (Mx-Cre) of HHEX rescues the B-cell developmental potential of PRH-null lymphoid precursors. Thus, PRH appears to regulate early lymphoid development by increasing cyclin D1 expression [62, 63]. Interestingly under conditions of stress haematopoiesis, that is, after sub-lethal irradiation, deletion of HHEX results in an inability of bone marrow cells to contribute to all bone marrow lineages as well as alterations in the proportion of long term and short stem cells, increased proliferation in vivo of stem cells and progenitors, and defects in T-cell populations [63].
PRH is implicated in several subtypes of leukaemia. Confocal microscopy and Western blot analysis demonstrate a substantial decrease in PRH protein levels and decreased nuclear localization of PRH in 13 of 13 primary AML (French American British classification M4/M5) and seven of seven blast crisis CML (bcCML) specimens but not in 11 of 11 M1/M2 AML, seven of eight acute lymphoid leukemia (ALL) specimens, or two of two chronic-phase CML [64]. Over-expression of eIF-4E transforms rat embryo fibroblasts and increased levels of this protein have been found in AML M4/M5 subtypes [64]. PRH binds to eIF-4E disrupting eIF4E nuclear bodies and repressing mRNA transport of eIF-4E targets such as cyclin D1 mRNA; in U937 human leukemic cells this results in the inhibition of cell proliferation [52, 64]. Thus PRH is a post-transcriptional repressor of cyclin D1 protein expression in leukaemic cells. The positive regulation of cyclin D1 mRNA expression by PRH inferred from knock out mouse experiments [61–63] is in contrast to the negative regulation of cyclin D1 protein expression in leukaemic cell lines. This apparent contradiction may be related to the lineage/differentiation state of the transformed cells compared to untransformed progenitors in vivo or it may indicate a subtle mechanism for the fine-tuning of cyclin D1 expression by PRH.
In keeping with an inhibitory role of PRH on the proliferation and transformation of leukaemic cells of myeloid origin, [44, 64–67] PRH directly represses the transcription of multiple genes involved in VEGF-signalling in leukaemic K562 cells. PRH knockdown in K562 cells results in increased transcription of these genes and increased VEGF autocrine signalling leading to increased cell survival [44]. These genes are regulated by PRH in other cells types: in endothelial cells, PRH has been shown to repress VegfR2 transcription although, in this case, repression is via an interaction with the transcription factor GATA-2 [47]. The derepression of these genes following PRH down-regulation is likely to be important in tumourigenesis.
Perturbation of the subcellular localisation of endogenous PRH has been proposed to be involved in the development of acute promyelocytic leukaemia (APL). The PML tumour suppressor protein regulates cellular signalling pathways controlling cell proliferation, apoptosis and senescence; PML can also interact with and negatively regulate eIF4E [68]. Chromosomal rearrangements observed in APL produce fusion proteins between PML and retinoic acid receptor ? (RAR?) [69]. PRH interacts with PML independently of its interaction with eIF-4E, and PRH over-expression disrupts PML nuclear bodies [52, 66]. Since PRH interacts with PML as well as with PML-RAR? [66], PML-RARa is likely to interfere with both endogenous PRH and PML activity, promoting leukaemogenesis [57].
More recently it has been demonstrated that PRH can function as an oncogene in AML. PRH mRNA expression has been noted to be elevated in microarray studies from a variety of human AML samples. Moreover high PRH mRNA expression correlates with poor survival [70]. Importantly, PRH knockout in a mouse model of AML where the AML is initiated by expression of a MLL-ENL fusion protein showed that PRH is required for the initiation and maintenance of the leukaemia and functions alongside HOXA9-Meis1 as a transforming oncoprotein [70]. PRH recruits Polycomb co-repressor complexes to bring about the repression of a set of genes with protein products including the cell cycle inhibitors p16-INK4 and p19-ARF. This repression is essential to maintain the leukaemic blasts [70]. Hence targeting PRH is a potential therapeutic approach for MLL-ENL dependent AML and may also be relevant for other AML where PRH mRNA expression is elevated.
Thus it appears that in some situations alteration of endogenous PRH protein levels or nuclear localisation can contribute to AML whereas in others oncogenic transformation leading to elevated PRH mRNA expression can result in AML. The cell type that is transformed, that is whether transformation occurs in a differentiated cell or a progenitor/stem cell might account for these apparent contradictions. The complexity of PRH involvement in AML is underscored by the finding that PRH is directly involved in the initiation of at least one AML without involvement of additional transforming proteins [65]. In this case a cytogenetic abnormality generated the fusion of nucleoporin 98 and PRH (Nup98-HHEX), where the N-terminal domain of PRH was substituted by that of Nup98, culminating in the emergence of a leukaemogenic gene expression profile [65]. Transplantation of murine bone marrow cells expressing Nup98-HHEX into transgenic mice resulted in acute leukaemia though with a latency period of 9Â months, suggesting that the translocation is a pre-requisite for disease induction but not in itself sufficient for leukaemogenesis [65].
PRH can also act as an oncogene in a subtype of T-cell acute lymphoblastic leukaemia known as early T-cell precursor-like ALL (ETP-ALL). PRH can phenocopy the Lmo2 oncoprotein in inducing self-renewal when overexpressed in mice and elevated PRH causes a T-cell leukemia, which is strikingly similar to that caused by Lmo2. These results suggested that PRH is an important mediator of Lmo2-driven T-cell self-renewal and leukemia [41, 67, 71]. Moreover Lmo2-transgenic mice with conditional deletion of Prh/HHEX showed a significantly delayed onset of the T-cell leukemia [41]. However, recent work showed that deletion of Prh/HHEX does not always block Lmo2-induced leukemia indicating that these proteins can act via parallel pathways [72]. Interestingly, deletion of PRH in the thymus of Lmo2 transgenic mice did result in a reduction in the transplantation capacity and the radioresistance of Lmo2-transgenic thymocytes but did not inhibit the development of the leukaemia [72].