Haematopoiesis is the best studied stem cell differentiation process, where hematopoietic progenitors self-renew and differentiate into blood cells, or undergo apoptosis. Any failure to respond to stimuli regulating these processes may lead to leukaemia. Many transcription factors that regulate the haematopoietic processes have been described. However, the downstream pathways, including regulation of various factors by proteases - and control of the latter by specific inhibitors - are less well defined. In this study, we have used leukaemia cells as a model of myeloid progenitor cells. We have shown that the haematopoietic-specific bomapin (serpinb10) is a nuclear, redox-sensitive protein that enhances proliferation of myeloid leukaemia cells under normal growth conditions, and enhances apoptosis of the cells following growth factors withdrawal.
Bomapin has two cysteines: C68 which is located in the long CD-loop, and C395 located close to the C-terminus. Molecular modelling of bomapin suggests that these cysteines in the reduced form of bomapin are distant, solvent-exposed and predicted to be highly reactive. Due to a high potential flexibility/mobility of the CD-loop, the cysteines can be disulfide-linked to form intramolecular disulfide bond without major perturbations of the bomapin structure (Figure 1E). The high reactivity of both cysteines is supported by the fact that bomapin expressed in E. coli was present in the form of oxidized monomer (having the two cysteines connected by disulfide bond), and disulfide-linked dimers (Figure 1A). SDS-PAGE analysis of immunoprecipitated naturally expressed bomapin showed that majority of the protein existed in the oxidized monomeric form with intramolecular disulfide bond (Figure 1D). There is a theoretical possibility that bomapin was artificially oxidized during immunoprecipitation and SDS-PAGE, however, the fact that disulfide-linked conformation of bomapin is important for an enhanced cell proliferation (Figure 3C) strongly suggests that the oxidized monomers of bomapin existed already in intact cells. In contrast to the bacterial-expressed bomapin, we have not observed any oligomeric species for the naturally expressed bomapin in leukaemia cells. This could be due to low bomapin levels (Table 1) and crowding effect in nucleus, or known differences in redox environment between nucleus of human cells and bacterial cytoplasm.
Although it is generally known that cytosolic compartment of eukaryotic cells has a high reducing potential which prevents the formation of stable disulfide bonds in proteins, relatively less is known about redox potential of nuclear compartment. However, several reports show that nuclear proteins may exist in oxidized forms with disulfide bonds of functional importance. For example, lamins-A/C, which are structural proteins of nuclear envelope, have to be in the form of disulfide-stabilized dimer to be able to bind chromatin DNA ; formation of a single intramolecular disulfide bond in transcription factor Yap1 is responsible for redistribution of the protein into nucleus ; disulfide-linked form of clusterin is known to accumulate in nuclei of early apoptotic epithelial cells . Therefore, the existence of a disulfide bond in naturally expressed bomapin is consistent with its nuclear localization (Figure 1C). The disulfide bond also appears to be essential for the function of bomapin since only wt/oxidized bomapin, but not the C395S mutant, was capable of enhancing proliferation of K562 cells (Figure 3C). These data suggest that the oxidised conformation of bomapin is important for an interaction of the serpin with another nuclear protein, and that bomapin effect on cell proliferation might be dependent only on the redox status of bomapin. However, with the current knowledge, we can not exclude a scenario where bomapin inhibitory activity is also involved, but it becomes essential only after the oxidized bomapin binds to its nuclear partner. Bomapin is a second, after its closest homologue plasminogen activator inhibitor type 2 (PAI-2, serpinb2), redox-dependent intracellular serpin. For PAI-2, the formation of a single disulfide bond between cysteines located in the CD-loop and at the bottom of the serpin molecule, induces conformational changes which result in spontaneous non-covalent polymerization of the serpin [22, 23], but biological function of the conformational switch is unknown.
To study intracellular function of bomapin, we took advantage of the fact that the human K562 cells do not express bomapin naturally (real-time PCR and immunoprecipitation, data not shown; ), and stably transfected the cells with bomapin-EGFP fusion, or EGFP as a control. Consistent with previous studies on HeLa cells over-expressing GFP-bomapin , the bomapin-EGFP fusion in K562 cells had a dominant nuclear distribution (Figure 2A). Expression of bomapin-EGFP in K562 cells resulted in about 90% higher cell proliferation (Figure 2B and 2C), and a significant shortening of the cell cycle without changes in distribution of cells in different phases of cell cycle. Bomapin-EGFP expressing cells had also bigger nuclei than the control cells (Figure 2D). On the other hand, down regulation of bomapin expression in U937 cells by means of antisense oligonucleotides resulted in a decreased cell proliferation (Figure 2F), suggesting that the bomapin effect on cell proliferation was not specific for the K562 cells only. However, the effect of bomapin on cell proliferation was leukaemia/haematopoietic-specific because expression of bomapin-EGFP in the human fibrosarcoma HT1080 cells did not change proliferation of the cells (Figure 2G). This strongly suggests that bomapin needs a haematopoietic-specific partner protein to enhance cell proliferation. Two other serpins from clade B have been reported to influence cell proliferation. The first one is rat trespin which is believed to be a homolog of human bomapin, but it is expressed in various tissues whereas bomapin is bone marrow-specific [15, 24]; over-expression of trespin in human embryonic kidney epithelial cell line (Hek293) resulted in an increased proliferation of the cells . The second one is kidney-specific mouse megsin which is responsible for increased proliferation of messangial cells in megsin-transgenic mice . The mechanism(s) behind serpin-dependent enhancement of cell proliferation remains yet unknown.
Bone marrow haematopoietic progenitors, quiescent without stimulation, can be activated to proliferate and to differentiate by cytokines and growth factors. When growth factor levels decrease, the cells undergo mitotic arrest followed by apoptosis that leads to termination of cell expansion [3, 20, 26]. In contrast, leukemic cells cultured in the absence of growth factors can continue to proliferate and evade apoptosis for a long time. In the case of K562 cells, the aberrant Bcr/Abl fusion kinase activates both proliferation and anti-apoptotic signals that are responsible for relatively high proliferation rate of these cells, and their resistance to apoptosis . However, bomapin-EGFP expressing K562 cells cultured without serum showed an increased cell accumulation in S-phase and increased apoptosis, compared to the control cells expressing EGFP (Figure 4). Therefore, bomapin antagonise the anti-apoptotic properties of Bcr/Abl fusion and sensitizes K562 cells to apoptosis when growth factors are absent.