«3. Medizinische Klinik und Poliklinik – Hämatologische Forschung Cks1 is a critical regulator of hematopoietic stem cell cycling, quiescence and ...»
Taken together, most of the experiments with p27-/- and DKO mice indicate that the observed effects in Cks1-deficient mice are not due to sole p27 accumulation and only the lineage distribution upon hematopoietic stress seem to be controlled through Cks1 in a p27 dependent manner.
Figure 30: Secondary Transplantations of whole bone marrow. Analysis of HSC/HPC Secondary transplantations were performed as described in figure legend 28.
Absolute numbers of donor delivered BM, MPP, LSK and CD150+ LSK cells at week 22 after secondary transplantation (n=8 secondary recipients of WT cells, n=8 secondary recipients of Cks1-/- cells, n=7 secondary recipients of p27-/- cells, n=8 secondary recipients of DKO cells; 2 independent experiments)
4.4. The role of Cks1 in a HSC disease In the course of this project an important role of Cks1 in HSC homeostasis was established. Considering that uncontrolled HSC/HPC regenerative activity is a hallmark of the prototypic HSC disorder chronic myeloid leukemia [185, 186], it was tempting to analyze the role of Cks1 in CML regulation. For this purpose, assays with a BCR-ABL expressing cell line, human samples and methylcellulose with mouse BCR-ABL expressing HSC/HPC cells were performed.
4.4.1. Cks1 transcript levels are regulated by BCR-ABL acitivity First, the expression of Cks1 (CKS1B in humans) in CML patients was assessed. CKS1B transcription from imatinib treated and untreated patients analyzed by real-time PCR revealed that CKS1B transcript levels were significantly elevated in untreated CML (Fig. 31A). Also, publicly available gene expression data with the accession numbers GSE24739  and GSE12211  was downloaded from the gene expression omnibus website and analyzed using the software program R-studio. In this case, CKS1B expression in chronic and accelerated phase (Fig. 31B) as well as in treated and untreated CML (Fig. 31C) was examined in association with the CKI p21 (CDKN1A), p27 (CDKN1B), p57 (CDKN1C) and p130 (RBL2). The CKI and p130 gene expression levels correlated negatively with CKS1B expression (Fig. 31B, C).
RESULTS Correlating with the results in Fig. 31A, CKS1B expression was significantly increased in chronic and accelerated phase CML versus normal control cells (Fig. 31B). Furthermore, the transcription levels of CKS1B were significantly suppressed upon imatinib treatment (Fig. 31C). This result was confirmed in vitro using BaF3-p210 cells, which ectopically overexpress BCR-ABL.
Immunoblot analysis with BaF3-p210 cell pellets demonstrate that Cks1 protein levels were reduced after treatment with imatinib (Fig. 31D). On the other hand, activation of BCR-ABL activity in the thermo-sensitive cell line BaF3-ts p210 (cultivating at 32°C) led to an increase in Cks1 transcript levels (Fig. 31E).
Thus, Cks1 expression is associated with BCR-ABL activity in the highly proliferative stem cell disorder CML.
4.4.2. BCR-ABL expression in Cks1-/- BM cells results in decreased colony forming capability Since expression of Cks1 is being regulated by BCR-ABL activity in CML, it was important to further evaluate the role of Cks1 in BCR-ABL mediated clonogenetic activity. For this purpose, 5-FU mobilized WT and Cks1-/- bone marrow was infected with BCR-ABL-GFP retrovirus. Methyl cellulose assays on medium without any supplemented factors were preformed either with sorted GFP+ cells, GFP+ MPP or GFP+ LSK (Fig. 32A). The medium is eligible only for transformed cells, since normal cells need growth factor in in vitro culture in order to survive and proliferate. In all cases, BCR-ABL expressing cells lacking Cks1 displayed significantly reduced ability to form hematopoietic colonies (Fig.
32B). This result points out that Cks1 is a mediator of BCR-ABL induced cytokine-independent clonogenetic activity, indicating a possible role of Cks1 in CML regulation.
Figure 31: Cks1 is up-regulated in CML.
(A) mRNA was isolated from peripheral blood mononuclear cells from healthy donors (ctrl), one imatinib treated patient in complete hematological remission (imatinib), or untreated CML patients (untreated), and analyzed for CKS1B transcript levels by real-time PCR. Shown is the relative CKS1B expression. (B) Expression of the indicated genes in control (ctrl), chronic and accelerated (acc) phase CML samples (source: www.oncomine.org). Left panel: heatmap; right panel: box plots. (C) Expression of the indicated genes in HSC/HPC cells of untreated or imatinib treated patients (source: www.oncomine.org). Left panel: heatmap; right panel: box plots. (D) Immunoblot analysis of BaF3-p210 cells treated with imatinib. (E) BaF3-ts p210 (active at 32°C) were cultured at 37°C (-12 to 0 hr) and transferred to 32°C. Shown is the relative expression of Cks1 assessed by real-time PCR.
Figure 32: Bcr-Abl expression in Cks1-/- BM cells results in decreased colony forming capability.
(A) Scheme of the experiment: First, Phoenix-Eco 293T cell line was transfected with BCRABL-GFP expressing vector and used for virus production, which was used to infect BM cells from 5-FU treated WT and Cks1-/- mice. Subsequently, methyl cellulose assays in absence of cytokines with sorted BCR-ABL expressing GFP+ cells, GFP+ MPP or GFP+ LSK were performed. (B) Numbers of formed colonies from BCR-ABL expressing BM, MPP and LSK from WT or Cks1-/- mice in cytokine deficient methyl cellulose medium (n=4 for each genotype; 2 independent experiments).
5. Discussion In this study, genetic in vitro and in vivo evidence is provided, that the ratelimiting component of the SCFSkp2 ubiquitin ligase, Cks1, is a critical mediator of size and activity of the HSC/HPC pool at steady state and in stress hematopoiesis. CKI regulated by SCFSkp2 were shown to accumulate in the early hematopoietic subsets in mice lacking Cks1. One of the downstream SCFSkp2 targets, p27, was studied more closely in order to determine if the observed effects were due to disturbed p27 regulation. The experiments revealed that lineage differentiation was indeed controlled by p27. However, the role of Cks1 in HSC self-renewal and quiescence seemed to be p27 independent. Nevertheless, evidence was provided that Cks1 is being regulated by BCR-ABL, pointing Cks1 out as a possible regulator in the HSC disorder CML.
5.1. Cks1 controls the HSC/HPC pool at steady state in vivo and protects HPC from apoptosis in vitro The majority of HSC at steady state conditions is in G0 phase and only a small number is cycling and responsible for stem cell maintenance and mature cells production [31, 189]. It is known, that disruption of HSC quiescence leads to defects in self-renewal and may result in HSC exhaustion [12, 32]. On the other hand, if HSC are arrested in the G0 phase and enter the cell cycle more infrequently, accumulation in the early subsets and deficiency in the more mature HPC may appear. Such a scenario can be the explanation of the phenotype in Cks1-/- mice where lack of Cks1 resulted in reduction of HPC numbers, but an increase in the percentages of early HSC at steady state in vivo (Chapter 4.2.2.). The decline in the absolute BM and FL cell numbers (Chapter 4.2.1.) as well as in all HPC subpopulations (Chapter 4.2.2.) could be also related to the reduced body size of the Cks1-knockout animals [128, 177].
However, the reduction in the hematopoietic subsets was more enhanced compared to the body size decrease, pointing out the importance of Cks1 in the regulation of the hematopoietic subsets. Since the number of HSC/HPC is also controlled by their capacity to survive during homeostasis or under conditions of stress , in addition to the slower exit from the HSC stage, deregulated DISKUSSION apoptosis in the early Cks1-/- progenitors might also be playing a role in conserving LT-HSC. And indeed, in vitro investigation revealed higher susceptibility of Cks1-/- HPC to apoptosis compared to LT-HSC (Chapter 4.2.6.). An opposite effect was observed after cytokine withdrawal in p27-/- BM cell cultures , pointing p27 out as a possible negative regulator of survival in hematopoietic cells. This stands in contrast to the analysis of p27-/hematopoietic cells in this study. Even the contrary effect was observed, after cytokine withdrawal in Lin- cell cultures, p27-/- MPP were more susceptible to apoptosis than WT MPP (Chapter 4.3.2.). This contradiction may be due to different methodology used in both experiments. While Dijkers et. al. used Sca+ BM cells, excluding just the MPP fraction, in this project Lin- cells, including the Lin- c-Kit+ Sca1- MPP fraction were cultured prior apoptosis analysis.
Interestingly, ablation of both, Cks1 and p27 resulted in WT similar phenotype (Chapter 4.3.2.), suggesting that Cks1 regulates apoptosis in a p27 dependent manner. Although a discrepancy occurs in this observation since p27-/- cells did not show increased apoptotic resistance. Further analysis could provide a possible explanation of the observed effect. For better understanding of the role of Cks1 and p27 in apoptosis in hematopoietic cells, in vivo analysis, for instance using fluorochrom-labeled inhibitors of caspases  should be performed.
The CDK inhibitor p27 is a critical regulator of cell growth and proliferation and increased body size is a significant feature of the p27-/- phenotype that highlights the key role of p27 in maintenance of the G0 state, tissue homeostasis and prevention of malignant growth [28, 192-194]. Concomitant loss of Cks1 in p27-/- animals rescues the size defect back to wild type levels indicating a p27 dependent function of Cks1. But in the case of HPC pool at steady state, DKO mice exhibit a phenotype similar to p27-/- and possess increased total HPC numbers narrowing the role of Cks1 in regulating p27 protein turnover in those cells (Chapter 4.3.1). Also, when monitored in the presence of cytokines, MPP and LSK cells were better sustained in both, p27-/and DKO cultures (Chapter 4.3.2). These findings support other studies, which define p27 as a regulator of the HPC alone  and of the HSC, together with p57 . However, the increase in the percentage of early LSK observed in Cks1-/- mice was even more prominent in DKO mice (Chapter 4.3.1), clearly DISCUSSION indicating a p27 independent role of Cks1 at least in the regulation of the HSC pool.
5.2. Downstream targets of the SCFSkp2-Cks1 complex are strongly involved in self-renewal and quiescence balance in HSC The cell cycle is perhaps the most tightly regulated part of HSC regulation. The balance between self-renewal and differentiation is critical for the maintenance of the HSC throughout life [31, 52] and deviations from that balance can contribute to hematopoietic insufficiencies and the development of hematopoietic malignancies . Phosphorylation through the activity of CDK and targeting for protein degradation through ubiquitin ligases such as SCF complexes are the most prominent factors involved in the regulation of HSC cycling [8, 196]. The present study demonstrates that several substrates of SCFSkp2-Cks1 (the CKI p21, p27, p57 and the Rb family member p130) are strongly regulated by the ubiquitin ligase in the HSC/HPC subsets (Chapter 4.1.1). These CKI [94, 99-101] and p130  have been identified as important regulators of HSC/HPC homeostasis by means of gene knock-out studies [32, 62]. CKI loss results in increased HSC cycling, an effect opposite to that observed in the Cks1-/- (Chapters 4.2.8 and 4.2.13) or Skp2-/- genotype [197, 198]. For instance, Cheng et al. demonstrated, that HSC in p21-deficient mice are increased in numbers at steady state due to accelerated proliferation .
However, more recent studies show p21 to have rather limited function in HSC regulation at steady state [47, 98], diminishing the possibility, that p21 alone is responsible for the observed effects in Cks1-/- mice. Thus, it is most likely that other SCFSkp2 targets are involved in the regulation of HSC/HPC homeostasis.
Supporting this hypothesis, studies with double knockout of p57 and p27  and triple knockout of pRb, p107and p130  show, that these cell cycle inhibitors cooperate to maintain HSC quiescence. No hematopoietic phenotype was observed in p130 deficient mice , but conditional deletion of all tree Rb family members results in severe defects in self-renewal . Besides, Rb family members collectively seem to maintain balance between lymphoid and myeloid cell fates in the hematopoietic system . Analysis of p27-deficient mice suggest rather limited functions of p27 in HSC and more pronounced role in regulation of HPC . On the other hand, p57 is shown to be critical for the DISKUSSION maintaining of HSC quiescence . Interestingly, the defect in p57-/- HSC quiescence is enhanced by concomitant deletion of either p21 or p27 and only p27 overexpression compensates for the loss of p57 [100, 101]. The central finding in this project that p21, p27 and p57 accumulate most strongly in the Cks1-/- CD150+ LSK (Chapter 4.1.2.) and Cks1 deficiency causes an opposite effect (Chapters 4.2.10.-4.2.13.) than p27/p57 deletion  is in accord with the idea that Cks1 regulation is a determining factor in the early hematopoiesis upstream of CDK inhibitors. These findings clearly indicate that the control of CKI by SCFSkp2 is of paramount importance for maintenance of HSC selfrenewal and quiescence. Complex studies with additional knockout animals and respectively double or triple knockouts, can provide definitive evidence for the link between Cks1 and CDK inhibitors. Another option to test if the observed effects in Cks1-/- HSC are due to p27 and p57 accumulation is by following the downstream targets of these CKI in the quiescent machinery. Zou et. al demonstrated, that concomitant loss of p27 and p57 results in nuclear import of an Hsc70/cyclin D complex and Rb phosphorylation , hence Cks1 deficiency might show cytoplasmic retention of Hsc70/cyclinD and reduced Rb phosphorylation in HSC.
Additionally, the fact that p21, p27 and p57 can compensate for each other could explain the results in this study with p27-/- and double knockouts for Cks1 and p27 (Chapter 4.3). Although some effects (body size, lymphoid differentiation after 5-FU treatment) seem entirely p27 dependent, the majority of the tests confirmed that in absence of p27, Cks1 loss affects most likely other downstream targets.