Date Published: April 10, 2012
Publisher: Public Library of Science
Author(s): Daphna Levy, Ateret Davidovich, Shahar Zirkin, Yulia Frug, Amos M. Cohen, Sara Shalom, Jeremy Don, Irina Agoulnik. http://doi.org/10.1371/journal.pone.0034736
Potent survival effects have been ascribed to the serine/threonine kinase proto-oncogene PIM-2. Elevated levels of PIM-2 are associated with various malignancies. In human cells, a single Pim-2 transcript gives rise mainly to two protein isoforms (34, 41 kDa) that share an identical catalytic site but differ at their N-terminus, due to in-frame alternative translation initiation sites. In this study we observed that the 34 kDa PIM-2 isoform has differential nuclear and cytoplasmic forms in all tested cell lines, suggesting a possible role for the balance between these forms for PIM-2’s function. To further study the cellular role of the 34 kDa isoform of PIM-2, an N-terminally HA-tagged form of this isoform was transiently expressed in HeLa cells. Surprisingly, this resulted in increased level of G1 arrested cells, as well as of apoptotic cells. These effects could not be obtained by a Flag-tagged form of the 41 kDa isoform. The G1 arrest and apoptotic effects were associated with an increase in T14/Y15 phosphorylation of CDK2 and proteasom-dependent down-regulation of CDC25A, as well as with up-regulation of p57, E2F-1, and p73. No such effects were obtained upon over-expression of a kinase-dead form of the HA-tagged 34 kDa PIM-2. By either using a dominant negative form of p73, or by over-expressing the 34 kDa PIM-2 in p73-silenced cells, we demonstrated that these effects were p73-dependent. These results demonstrate that while PIM-2 can function as a potent survival factor, it can, under certain circumstances, exhibit pro-apoptotic effects as well.
Pim-2 is a member of a serine threonine kinase family of proto-oncogenes, which include also Pim-1 and Pim-3. The family was identified as a common proviral insertion site of MuLV (Moloney murine leukemia virus) in T and B cell lymphomas in mice –. Transgenic mice over-expressing either Pim-1 or Pim-2 are predisposed to T cell lymphomas, whereas both Pim-1 and Pim-2 act synergistically with c-Myc to accelerate development of B-cell tumors , –. Pim-1 or Pim-2 deficient mice show no in-vivo abnormalities –. However, Pim-1-Pim-2 double knockout, or even more so, a triple knockout of all three Pim genes, causes a mild phenotype of reduced body size, impaired proliferation of hematopoietic cells in response to a variety of growth factors, and an effect on the cell cycle entry of peripheral T cells in response to IL-2 and TCR activation .
Previously identified substrates of the PIM-2 kinase, share an oncogenic promoting function as either anti-apoptotic/survival factors, transcription factors that can increase expression of oncogenic proteins, or cell cycle regulators (review in ). Moreover, increased expression of PIM-2 was directly associated with various malignancies, such as non-Hodgkin’s lymphoma, CLL and prostate cancer –, further supporting the oncogenic function ascribed to PIM-2, and justifying targeting its kinase activity as a beneficial therapeutic approach , . In this study, however, we found that under certain circumstances PIM-2 might exert cell cycle arrest and pro-apoptotic effects, suggesting that PIM-2 may play a dual role and that targeting its activity might not always be therapeutically beneficial. Interestingly, PIM-1 expression has also been correlated with poor prognosis in haematopoietic malignancies and with good prognosis in other malignancies (review in ). It is possible that the differential function of PIM-2 is cell type dependent. However, given that similar results were obtained in more than one cell line, this phenomenon seems to be of a more general nature. Our finding that only the 34 kDa isoform, but not the 41 kDa isoform, could activate the cell cycle arrest and pro-apoptotic effects, suggest that the balance between these two isoformes might affects PIM-2’s overall effect. Moreover, the fact that the pro-apoptotic effect of the 34 kDa isoform was even intensified in a background of silencing the endogenous PIM-2 (Fig. S6) might suggest that the 41 kDa has a somewhat moderating effect on the activity of the 34 kDa isoform. This can be the outcome of both isoformes sharing the same substrates but with differential phosphorylation efficiencies, as has been reported in the mouse PIM-2 where the 41 kDa isoform is less active than the 34 kDa isoform . An additional potential regulatory level that can be considered is that the balance between nuclear and cytoplasmic PIM-2 (mainly the 34 kDa isoform) is important to determine PIM-2’s overall effect. This latter hypothesis is consistent with the results of Dai et al. , reporting a shift from a predominant nuclear expression of PIM-2 in normal prostate epithelium cells, to an increased cytoplasmic expression in prostate cancer cells (PCa). These authors further reported that increased nuclear expression of PIM-2 in perineural invasion (a common pathological phenomenon proposed to be the dominant pathway through which PCa spreads beyond the prostate) was associated with decreased proliferation. The experimental results reported herein, undoubtedly disrupted the cytoplasmic/nuclear balance, exposing nuclear substrates to increased PIM-2-dependent phosphorylation. This concept of a protein exhibiting rather opposing effects depending on its sub-cellular localization has been reported in other systems as well. One such example is the cyclin A-CDK2 complex in mouse mesangial cells, where in its nuclear form it functions to promote cell division, whereas following an apoptotic stimulus, it accumulates in the cytoplasm where it actively functions to promote apoptosis –. Whatever the reasons for the differential effects of PIM-2 are, this study suggests that while targeting PIM-2’s kinase activity might indeed be beneficial chemotherapeutically in certain malignancies, it might be devastating in other cancers in which activating PIM-2 might actually reduce the malignant phenotype. This of course necessitates in depth study of the molecular pathways through which each effect is exerted.