Understanding PLZF

Abstract

PLZF can function as a transcriptional activator or as a transcriptional repressor. Recent studies have identified two direct transcriptional targets of PLZF, REDD1 and smooth muscle α-actin. REDD1 is activated by PLZF. It mediates the PLZF-dependent downregulation of TORC1 and is responsible for the maintenance of pluripotency in cultures of spermatogonial progenitor cells. This activity may extend to other stem-like cell types. The effect of REDD1 on TORC1 also raises the possibility that REDD1 controls cell growth, tumorigenicity and senescence. The regulatory loop extending from PLZF via REDD1 to TORC1 identifies REDD1 as a critical determinant of TOR activity. The transcription of smooth muscle α-actin is repressed by PLZF. In fibroblasts, this downregulation is accompanied by a change of cell shape and a dramatic reorganization of the cytoskeleton. It is also correlated with the acquisition of cellular resistance to oncogenic transformation. The resistance is selective, it works against some oncoproteins but not against others. The molecular mechanisms underlying the changes in the cytoskeleton and in the susceptibility to oncogenic transformation are unknown. However, these changes are dependent on the activity of RAS and thus probably involve the RAC/RHO family of proteins.

PLZF, a Versatile Transcriptional Regulator

PLZF (promyelocytic leukemia zinc finger) was originally identified in promyelocytic leukemia as one of several partner proteins fused by a reciprocal chromosomal translocation to the retinoic acid receptor RARα.^1,2 Both fusion products, PLZF-RARα and RARα-PLZF, play essential roles in the pathogenesis of the disease, acting as dominant negative mutants of RARα and of PLZF respectively.³ PLZF is a transcription factor belonging to the POZ-Krüppel (POK) family that binds to specific DNA sequences with its carboxy-terminal zinc fingers and suppresses transcription by recruiting co-repressors with its aminoterminal POZ domain. However, PLZF can also activate transcription.^4–6 The determinants of activator versus repressor function have not been defined.

PLZF affects diverse signaling, growth-regulatory and differentiation pathways. It is a key regulator of myeloid development.⁷ Recent reports have also demonstrated its role in the immune response.^5,8,9 In populations of stem cells it is essential for preserving pluripotency and the ability to self-renew.¹⁰ Additionally, it has contextdependent anti-oncogenic, tumor suppressive properties.^11,12 These diverse functions of PLZF most likely involve its transcriptional regulatory activities. Identifying direct transcriptional targets will therefore shed light on these functions. Although there are numerous genes that are differentially regulated by PLZF, the number of documented direct targets that involve interaction of PLZF with promoter or enhancer sequences of the target gene is small (Table 1). Of particular interest in the context of cell growth is MYC, which is repressed by PLZF but activated by the fusion protein PLZF-RARα.^13,14 Recent publications have revealed two such direct targets of potential significance for the activities of PLZF that affect growth control, self-renewal, senescence and tumor suppression. One of the targets, REDD1, is activated by PLZF at the transcriptional level; the other, α-actin, is repressed.

Table 1.

Direct Target Genes of PLZF¹

Gene	Product	Reference
	Repressed targets
ACTA2	Smooth muscle α-actin	20
BID	BH3 interacting domain death agonist	48
CCNA2	Cyclin A2	49
CDC6	Cell division 6 homolog	50
CEBPA	CCAAT/enhancer binding protein	4
CRABP1	Cellular retinoic acid binding protein 1	51
GFI1	Growth factor independent 1 transcription repressor	4
HOXB2	Homeobox B2	52
HOXD11	Homeobox D11	53
KIT	Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog	54
LEF1	Lymphoid enhancer-binding factor 1	4
MIR146A	MicroRNA 146a	55
MIR221	MicroRNA 221	56
MIR222	MicroRNA 222	56
MYC	Myelocytomatosis viral oncogene homolog	14
RER	Prorenin/renin receptor	57
VLA-4	Very late antigen 4	58
	Activated targets
DDIT4	REDD1—DNA damage-inducible transcript 4	15
DUSP6	Dual specificity phosphatase 6	4
ID2	Inhibitor of DNA binding 2	4
IFIT2	Interferon-induced protein with tetratricopeptide repeats 2	5
MPL	Myeloproliferative leukemia protein; thrombopoietin receptor	6
RSAD2	Radical S-adenosyl methionine domain containing 2	5

¹Direct interaction documented by ChIP or by EMSA or by both methods.

REDD1

A recent study by Hobbs et al. has defined the role of PLZF in maintaining self-renewal of pluripotency in spermatogonial progenitor cells.¹⁵ In contrast to Plzf^+/+ cells, cultures of spermatogonial progenitors derived from Plzf^−/− mice progressively lose pluripotency. Plzf^−/− spermatogonial progenitors show hyperactivation of TORC1, resulting in an induction of a negative feedback that interferes with the function of growth factor receptors (Fig. 1). Notably, this feedback affects the expression of the receptor for glial cell-derived neurotrophic factor, a critical regulator and enabler of pluripotency. The high activity of TORC1 in Plzf^−/− cells is not caused by enhanced PI3K or ERK signaling but reflects a reduced level of REDD1. REDD1 intervenes in the PI3KTOR signaling pathway by acting on the TSC1/TSC2 complex and inhibiting RHEB-mediated activation of TORC1.¹⁶ The expression of REDD1 is modulated by developmental programs, stress, DNA damage or hypoxia.^17,18 But the analysis of Plzf^−/− spermatogonial progenitors has now also identified REDD1 as a direct transcriptional target of PLZF. ChIP assays document binding of PLZF to a specific site in the distal portion of the REDD1 promoter. This binding sequence is active when inserted into the promoter of a transcriptional reporter for REDD1, and that reporter activity also requires the transcriptional regulatory POZ domain of PLZF. PLZF functions as a positive regulator of REDD1, leading to an attenuation of TORC1 activity and thus eliminating the negative feedback to growth factor receptors and facilitating the maintenance of pluripotency.

Figure 1.

In spermatogonial progenitor cells from PLZF^+/+ mice, PLZF activates the transcription of REDD1, leading to an attenuation of TORC1 activity. The negative feedback activity from TORC1 is reduced allowing effective signaling from GCNF and other growth factors. Such signals are essential for allowing self-renewal of pluripotent cells. In spermatogonial progenitors from PLZF^−/− mice, hyperactive TORC1 generates negative feedback interfering with signals from growth factors that are essential for the maintenance of pluripotency.

Smooth Muscle α-Actin

Smooth muscle α-actin is an important component of the cytoskeleton; it is also involved in cellular motility and endocytosis.¹⁹ Overexpression of PLZF in chicken embryo fibroblasts has resulted in the identification of smooth muscle α-actin as a repressed transcriptional target.²⁰ Smooth muscle α-actin is abundantly expressed in these cells but is dramatically downregulated by PLZF at the mRNA and protein levels. PLZF binds to the smooth muscle α-actin promoter as documented by ChIP and EMSA and represses transcription in a dose-dependent fashion. A striking change in cell morphology from spindle-shaped to polygonal accompanies the repression of smooth muscle α-actin (Fig. 2). This altered cell shape reflects a PLZF-induced reorganization of the cytoskeleton. The parallel arrays of long stress fibers characteristic of fibroblasts are replaced by shorter fibers that are oriented in multiple directions. Dominant negative RAS inhibits these PLZF-induced morphological and cytoskeletal changes, suggesting involvement of RAC/RHO family proteins. The profound morphological changes observed in PLZF-expressing fibroblasts are probably not caused by lowered smooth muscle α-actin levels alone. It is more likely that several other, yet to be identified PLZF targets, act in concert to effect these changes in cell shape and cytoskeletal organization. PLZF-expressing chicken embryo fibroblasts also show strong resistance to oncogenic transformation induced by most, but not all oncoproteins. Notable exceptions are viral SRC and viral JUN; their oncogenic potency is not significantly affected by PLZF. Among the sensitive oncoproteins are MYC, PI3K, AKT, FOS and ABL. It is not known what makes an oncoprotein subject to PLZF-induced interference.

Figure 2.

Overexpression of PLZF induces a profound reorganization of the cytoskeleton in fibroblasts, accompanied by a change in cell shape from elongated to polygonal. The PLZF-expressing cells show resistance to the transforming activity of several oncoproteins.

Growth Control, Senescence, Self-Renewal and Tumor Suppression

The new transcriptional targets of PLZF, REDD1 and smooth muscle α-actin raise interesting questions concerning growth control, senescence, self-renewal and tumor suppression (Fig. 3). The attenuation of TORC1 activity by REDD1 is essential for the maintenance of pluripotency in cultures of spermatogonial progenitors. At first sight, this seems counterintuitive; it can only be understood in terms of negative feedback loops originating from TORC1 that have to be tempered in order to permit the generation of relevant signals. The effect of REDD1 on the maintenance of pluripotency is probably not unique to spermatogonial progenitors and extends to other cell types as well. But could the REDD1-mediated downregulation of TORC1 also affect more generally cell growth, senescence or oncogenicity in different contexts? TORC1 plays a somewhat ambivalent role in these processes, it promotes growth through its stimulation of protein synthesis, but it also generates powerful negative feedback loops via S6K1 and via MAPK pathways that are disabled by TORC1 inhibition.^21–25 Moreover, inhibition of the TORC1 complex by rapamycin can result in increased phosphorylation of AKT and enhanced cell survival.^26–30 Such inhibition is therefore an inherently two-edged sword that can lead to opposite results depending on context. In the clinic, responses to rapamycin as single agent in cancer therapy have been disappointing,^31–34 and there is unpublished evidence that at least in some situations rapamycin may even accelerate the growth of the cancer (N. Rosen, Sloan Kettering Cancer Center 2010, private communication).

Figure 3.

Summary of PLZF functions that are mediated through the transcriptional regulation of REDD1 and of α-actin.

Rapamycin also extends the life span of various organisms, including mammals.^35–38 The molecular mechanism of this effect is not fully understood but probably involves the TOR pathway.³⁹ In these scenarios, it is tempting to make a mental substitution of REDD1 for rapamycin. Could overexpression of REDD1 or any other TORC1-specific inhibitor interfere with PI3K-mediated oncogenic transformation, with negative feedbacks originating from TOR, or could it extend life span? Is longevity in humans correlated with constitutive activity of REDD1? Recently it was shown that REDD1 can also function as an oncoprotein by impeding apoptosis in prostate cancer cells,⁴⁰ and by mediating RAS-induced oncogenic transformation in ovarian epithelial cells.⁴¹ However, the mechanisms and the generality of these functions still need to be established. The expression and activity of indigenous REDD1 respond to various stimuli, amounting to very stringent regulation. This fine-tuning is then applied to TORC1 through its highly regulated effectors TSC1/2 and Rheb.^42–45 Such finetuning and the balance between TORC1 and TORC2 activation is necessary to prevent the deleterious consequences of a hyperactive and of an excessively inhibited TOR pathway.

As a structural protein, α-actin affects cellular activities mainly by its levels of expression and its multi-protein organization. The PLZF-induced downregulation of α-actin and the associated changes in the cytoskeleton are probably important aspects of tumor suppression by PLZF. Since oncogenic transformation leads to a restructuring of the cytoskeleton and to altered cell motility, it will also affect α-actin and in turn be affected by the state of α-actin. However, the relationship between the cytoskeletal effects of PLZF and cellular sensitivity to oncoproteins is not understood. It will be interesting to determine why some oncoproteins are susceptible to PLZF-induced resistance, whereas others are not. In PLZFexpressing cells, sensitive and resistant oncoproteins are produced at comparable levels, as they are generated from virtually identical retroviral vectors. Assuming such equal expression, it would be informative to identify the steps in the oncogenic process that are inhibited by PLZF. For instance, is the resistance to oncogenic PI3K due to an excessive downregulation of TORC1, mediated by REDD1? Is posttranscriptional modification of MYC by PLZF involved in resistance to transformation?⁴⁶ A puzzling observation is the sensitivity of FOS to PLZF-induced interference as compared to the resistance of JUN. Both are components of the AP1 transcription factor complex, and oncogenic transformation results from enhanced AP1 activity. However, JUN and FOS transform cells by dimerizing with different partner proteins.⁴⁷ Their transcriptional profiles are therefore not identical, and the differences could reveal targets that explain resistance versus sensitivity to PLZF.

PLZF is increasingly recognized as an important master regulator in the cell with effects on growth, oncogenesis, self-renewal and differentiation. Ultimately, the understanding of PLZF has to come in terms of transcriptional regulation. The identification of two direct targets, REDD1 and α-actin, both with links to the control of cell growth, oncogenesis and differentiation, constitutes a beginning. Many more targets remain to be identified.