Derangement of transcription factor profiles during in vitro differentiation of HL60 and NB4 cells

Abstract

Sequential up- and down-regulation of a handful of critical transcription factors is required for proper neutrophil differentiation. Malfunction of transcription factors may lead to diseases such as acute myeloid leukemia (AML) and specific granule deficiency. In order to understand the molecular background for normal and malignant granulopoiesis, a good model system is required that faithfully mimics the in vivo transcription factor expression profiles. The two human leukemic cell lines HL60 and NB4 have been widely used as model cell lines for these purposes. Differentiation of HL60 and NB4 cells resulted in asynchronous differentiation to morphologically mature neutrophils over a period of 5–7 days. To obtain cell populations of more even maturity, cells at different stages of in vitro differentiation were purified by immunomagnetic isolation. This resulted in three cell populations that could be classified as promyelocytes, myelocytes/metamyelocytes, and mature neutrophils, respectively. Comparison of transcription factor mRNA profiles from these cell populations with those previously seen in normal human bone marrow, demonstrated that although all of the 14 transcription factors described in vivo, could be detected during in vitro differentiation, vast differences in their expression profiles was observed. These data illustrate the limitations of cell lines as models for normal granulopoiesis.

Introduction

Formation of the neutrophil granulocyte takes place in the bone marrow. A number of transcription factors play a critical role during this process and mutations of transcription factor genes may lead to aberrant maturation of neutrophils. Chromosomal translocations involving Runx1 (also called AML1) [1], and mutations or deregulation of C/EBPα [2], [3], [4] is frequently observed in acute myeloid leukemia (AML) patients. Furthermore, mutations of PU.1 have in some cases been associated with AML [5] and disruption of C/EBPɛ has been observed in patients with specific granule deficiency [6].

Gene disruption studies in mice have also identified several transcription factors essential for neutrophil development. Lack of definitive hematopoiesis is observed in Runx1^−/− or c-myb^−/− fetuses, demonstrating their importance for all hematopoietic lineages, including neutrophils [7], [8], [9]. Mice deficient in C/EBPα completely lack mature neutrophils and eosinophils, whereas all other hematopoietic lineages are present, and function normally [10]. Immature myeloid precursor cells are found in the blood of C/EBPα knock-out mice, demonstrating that C/EBPα is essential for neutrophil differentiation [10]. Conditional knock-outs of C/EBPα have demonstrated that C/EBPα is essential for the transition from the common myeloid progenitor (CMP) to the granulocyte/macrophage restricted progenitor (GMP), but not for GMPs to terminally differentiate to mature neutrophils [11]. C/EBPɛ-deficient mice also fail to develop fully mature neutrophils and eosinophils, only atypical hyposegmented neutrophils were present [12]. Transcripts for azurophil granule proteins were detected in the bone marrow, whereas an almost complete lack of transcripts for specific granule proteins was noted [12], [13], [14]. Mice deficient in PU.1 also generate neutrophils with a block in differentiation [15], [16], [17]. Conditional knock-outs have demonstrated that PU.1 is essential for the generation of CMPs, and for the terminal maturation of GMPs to mature neutrophils [18], [19].

Analysis of knock-out mice is valuable if one wish to demonstrate the requirement of a given transcription factor for myeloid differentiation. However, due to the inherent property of knock-out studies this method only allows one to determine the earliest point of differentiation where the transcription factor in question is essential. Furthermore, as it is becoming evident that not only the presence of a transcription factor, but also its overall concentration [20], [21] – as well as its concentration relative to that of other factor transcription factors [22] – is important for lineage decision and differentiation, model systems to investigate these issues are also required. Cellular systems to study transcription factor regulation during myeloid differentiation in vitro may therefore be of great value provided the models faithfully mimick the in vivo profile of transcription factor expression.

Two human leukemic cell lines, HL60 and NB4, are widely used as models for neutrophil differentiation. The HL60 cell line was established in 1977 from a patient suffering from AML, FAB M2 [23]. The cells grow continuously as myeloblasts and can differentiate to morphologically mature looking neutrophils after addition of all-trans retinoic acid (ATRA) and dimethyl sulfoxide (DMSO) [24]. The differentiation is not complete since specific granules and gelatinase granules are not formed and transcripts for matrix proteins contained in these granules are not detectable [25], [26]. Membrane proteins of these granules are synthesized but are routed to the plasma membrane [27]. Constitutive expression of the specific granule protein NGAL in HL60 cells results in an accumulation of NGAL in the azurophil granules when the cells are at the myeloblast stage. However, when differentiation of HL60 cells is induced, newly synthesized NGAL is routed extracellularly and does not accumulate in granules, reflecting the inability of more mature HL60 cells to synthesize granules and retain granule proteins [25], [27].

The NB4 cell line was established in 1991 from a patient suffering from acute promyelocyte leukemia (APL) having the t(15;17) translocation [28]. Morphologically, the cells are characterized as promyelocytes and can be stimulated to differentiate to neutrophils by ATRA. Like HL60 cells, specific and gelatinase granules are not formed during in vitro differentiation of NB4 cells and mRNA for specific granule matrix proteins cannot be detected [29]. Also like in HL60 cells, membrane proteins of specific and gelatinase granules are synthesized when neutrophil maturation is induced and routed to the plasma membrane [30].

We have previously used a method for separation of neutrophil precursors from human bone marrow into three populations of different maturity. These are myeloblasts (MB) + promyelocytes (PM), myelocytes (MC) + metamyelocytes (MM), and band cells (BC) + segmented neutrophil cells (SC), respectively. By analyzing these populations by Northern and Western blotting, a highly individualized expression of 14 transcription factors important for neutrophil differentiation was demonstrated during in vivo granulopoiesis [31].

The purpose of this study was to correlate the expression-pattern of transcription factor mRNAs in HL60 and NB4 cells during in vitro differentiation, to the in vivo pattern referred to above and to determine whether these two leukemic cell lines may be used as model systems for transcription factor regulation during granulopoiesis.