Jenny Dahlstro(€)m a, *, Chuanyou Xia a, Xiangling Xing a, Xiaotian Yuan b, **,Magnus Bjo(€)rkholm a, Dawei Xu a
ABSTRACT
The activating-mutation of JAK2V617F drives the development of myeloproliferative neoplasms (MPNs). Several JAK2 inhibitors such as ruxolitinib and gandotinib (LY2784544) currently in clinical trials and, provide improvements in MPNs including myeloibrosis. However, JAK2 inhibitors are non-curative and murine experiments show that JAK2 inhibitors don’t eradicate MPN stem cells and it is currently unclear how they escape. We thus determined the effect of the speciic JAK2V617F inhibitor LY2784544 on leukemic stem (CD34+) cells (LSCs) using the JAK2V617F-bearing erythroleukemia cell line HEL. The LY2784544 treatment caused a transient proliferation inhibition and apoptosis of HEL cells, but a re- covery occurred within a week. Thereafter, the continuous LY2784544 exposure induced the accumu- lation of CD34+ LSCs, and the CD34+ cells increased from 2% to >90% by week 9, which was accompanied by increased clonogenic potentials. LY2784544 was capable of stimulating CD34 expression even in CD34- HEL cells, which indicated cellular de-differentiation. A signiicantly enhanced expression of the stem cell factor KLF4 was observed in LY2784544-treated HEL cells. Inhibiting KLF4 expression attenu- ated LY2784544-mediated accumulation of CD34+ LSCs. Moreover, the telomerase inhibitor GRN163L abolished the LY2784544-effect. JAK2 inhibitors thus cause enrichment of LSCs and are unlikely to cure MPN as a monotherapy. Simultaneously targeting JAK2V617F and KLF4 or telomerase may be a novel strategy for MPN therapy, which should be of signiicance both biologically and clinically.
Keywords:JAK2 mutation;KLF4;MPN;Leukemia stem cell;Telomerase;TERT
1.Introduction
Myeloproliferative neoplasms (MPNs) bearing no Philadelphia chromosomes are a group of clonal neoplastic diseases derived from myeloid stem cells in the bone marrow, consisting of polycythemia vera(PV), essential thrombocythemia (ET) and primary myeloibrosis (PMF)[1,2].The hotspot activating- mutation in exon 14 of the JAK2 gene (JAK2V617F ) is a driver for the pathogenesis of MPNs, which occurs in the vast majority of PVs, and approximately half of ETs and PMFs [2,3]. Therefore,JAK2 might be a rational therapeutic target for MPNs. Indeed, JAK2 inhibitors have been developed to treat JAK2V617F-positive MPNs. Clinical trials of JAK2 inhibitors in PMFs showed a reduction in spleen size and improvements in other symptoms [2,4e7]. However, this effect does not correlate with reduced JAK2V617F allele burden [8]. Xenotransplantation experiments revealed that JAK2 inhibitors target a population of progenitor cells in the human spleen while sparing another progenitor population and disease stem cells [9].Most JAK2 inhibitors for treatment of MPNs are generally not speciic to JAK2, and they may inhibit other targets including JAK1, JAK3 and TYK as well [10,11]. Certain side effects reported in clinical studies, such as severe anemia, are believed to be partially caused by the inhibition of other targets than JAK2 [4e7]. One inhibitor, LY2784544, is unique as it shows a dose-dependent selectivity for JAK2V617F [11]. The TEL-JAK3 cell based assay shows that LY2784544 has a negligible effect on the JAK3 signaling, and it also exhibits a greater degree of selection for JAK2 over JAK1 than other JAK2 inhibitors including ruxolitinib. Mechanistically, LY2784544 displays potent ATP-competitive inhibition of JAK2 tyrosine kinase [11]. Despite its highly selective inhibition of JAK2V617F cells, LY2784544 had minimal effects on JAK2V617F allele burden in clin- ical studies [5,8,12]. In general, the results of early phase 1 and 2 clinical trials with JAK2 inhibitors are not satisfying [4,13]. It is currently incompletely understood why MPN patients respond poorly to JAK2 inhibitors.
Telomerase, a RNA-dependent DNA polymerase responsible for telomere lengthening, is silent in human differentiated cells, while activated in the vast majority of malignancies [14e17]. Telomerase or its catalytic component telomerase reverse tran- scriptase (TERT) is required for the development and progression of human cancer by contributing to multiple cancer hallmarks. Telomerase inhibition has therefore been suggested as a therapy for human malignancies.For instance, GRN163L,a 13-mer deoxyribo-oligonucleotide, acts as a direct antagonist to the telomerase RNA template (TERC) to inhibit telomerase activity and has recently entered clinical trials for several human ma- lignancies including MPNs [18e20].Initial studies show that GRN163L inhibits the formation of megakaryocyte colonies in patients with ET and PMF and reduces the allele burden of JAK2V617F in megakaryocytes in some patients [21]. A promising eficacy of GRN163L in MPNs has been observed in several clin- ical investigations [22]. However, the drug has also been shown to cause severe myelosuppression in some patients [18e20]. A combined therapeutic strategy of GRN163L with other thera- peutic agents, might enhance the treatment eficacy and mitigate telomerase inhibitors’ side-effects [4].The aforementioned experimental and clinical studies raise a number of questions: First, why MPN patients have a limited response to speciic JAK2 inhibitors; second, how GRN163L ach- ieves its therapeutic eficacy in MPNs; and inally, whether the combined inhibition of both JAK2 and telomerase can have synergistic effects on MPNs. The present study was thus designed to address these issues. Speciically, we wanted to deine the effect of JAK2 and telomerase inhibitors on leukemic stem cells (LSCs) using HEL leukemic cells carrying the JAK2V617F mutation as a model.
2.Materials and methods
2.1.Cell culture
The erythroleukemia cell line HEL at cell density 0.5 millions/ml was treated with 0.5 mM of the JAK2 inhibitor LY2784544 (Selleck chemicals, Houston, TX) or/and 20 mM of the telomerase inhibitor GRN163L (Johnson-Johnson, New Brunswick, NJ). Further detailed treatment of cells is in supplementary materials.
2.2. Colony formation assay
Colony formation assay was performed with cells treated with LY2784544 and/or GRN163L following 4 and 8 week incubation. The detailed protocol is documented in supplementary materials.
2.3. RNA extraction and real time polymerase chain reaction (RT- PCR)
RNA was isolated using Trizol (Life technologies, Carlsbad, CA) according to the manufacturer’s protocol. RT-PCR and primer se- quences are detailed in supplementary materials.
2.4. Whole transcript expression analysis
Microarray analysis was performed using Affymetrix whole- transcript expression analysis and the WT assay gene ST 1.1 (Affy- metrix, Santa Clara, CA) in association with the Bioinformatics and Expression Analysis Core Facility (BEA), Karolinska Institutet. The expression proile was compared between cells with and without LY2784544 or GRN163L exposure and differentially expressed genes were identiied.
2.5.Western blot
Cells were lysed in RIPA buffer and protein concentration measured using DC protein assay (Bio-rad, Hercules, CA). The detailed protocol is documented in supplementary materials.
2.6. Flow cytometry
CD34 and apoptosis assays using flow cytometry are detailed in supplementary materials.
2.7. Lentiviral transfection
Lentiviral particles KLF4 ShRNA with green fluorescent protein (GFP) as a reporter and puromycin as a mammalian selection marker were purchased from Origene (TL316853V). The detailed protocol is documented in supplementary materials.
2.8.Telomerase activity assay
Telomerase activity was measured using a telomeric repeat ampliication protocol(TRAP)-ELISA KIT(Roche, Basel, Switzerland). The detailed protocol is documented in supplemen- tary materials.
2.9.Telomere length analysis using flow-FISH
Telomere length was measured with flow-FISH as previously described [23]. The detailed protocol is documented in supple- mentary materials.
2.10.Statistical analysis
The data is expressed as means with error bars indicating the standard deviation (SD). For statistical analyses, Student’s t-test was used and a P-value <0.05 was considered signiicant.
3.Results
3.1.LY2784544 or GRN163L treatment of HEL cells induces apoptosis and inhibits proliferation
Treatment with LY2784544, a speciic JAK2 inhibitor, initially reduced both the number and viability of HEL cells (Fig. 1a and b). After an initial dip in viability seen after approximately one week, cells started to recover and at the end of the experiment (9 weeks) the viability was about the same as the control cells (Fig. 1b). A similar pattern was seen when assessing apoptosis after treatment with LY2784544 (Fig. 1c andd). Interestingly, this recovery was not seen in cells that were simultaneously treated with the telomerase inhibitor GRN163L (Figs. S3a and b). Treatment with GRN163L alone severely reduced the cell number and viability with no re- covery with time.
Fig. 1. The JAK2 inhibitor LY2784544 inhibits proliferation and promotes apoptosis of HEL cells. (a) Proliferation of HEL cells after treatment with 0.5 mM LY2784544 expressed as number of cell doublings/24 h. (b) Viability of HEL cells after treatment with 0.5 mM LY2784544 measured using nucleocounter. (c andd) Percentage of late and early apoptotic cells following treatment with 0.5 mM LY2784544, as analyzed using an Annexin V and 7-AAD staining kit. (e) Percentage of CD34+ cells after LY2784544 treatment compared to control, measured with flow cytometry. (f) Histograms of CD34-APC in HEL control and LY2784544 treated cells after 1, 3 and 7 weeks, respectively. Graphed data in igures aee are shown as mean and standard deviation.
3.2. CD34-positive cells accumulate following JAK2 inhibition in HEL cells
The fraction of CD34-positive (CD34+ ) cells drastically increased during LY2784544 treatment. The CD34+ fraction was assessed with flow cytometry every other week for 9 weeks and a gradually increasing fraction of CD34+cells was seen starting at approxi- mately 10% at 1 week and reaching 85% at 9 weeks of treatment (Fig. 1e and f). To assess the functional signiicance of these increased CD34+ cells, a colony formation assay was performed. Cells treated with LY2784544 indeed gave rise to more colonies, especially colonies containing >50 cells (Fig.1g). A slight (non- signiicant) decrease in the number of colonies was seen in cells treated with GRN163L compared to control cells. Simultaneous treatment with GRN163L was able to block the increase in CD34+ cells in the presence of LY2784544 (Fig. S3c) and colony formation potential as well (Fig.1g). To better understand ifLY2784544 causes an induction of CD34+ cells or if the CD34 enrichment is caused by a decreased sensitivity to the drug, HEL cells were then sorted into CD34 negative (CD34- ) and positive populations before treatment with LY2784544. An increase in CD34+ cell numbers was seen both in the negative and positive populations after treatment for 2 and 4 weeks, respectively (Fig. 2a). The CD34+ cells had a higher viability compared to the CD34-cells after treatment with LY2784544 (Fig. 2band c).
3.3.LY2784544-mediated up-regulation of KLF4 expression is responsible for the accumulation of CD34+ HEL cells
To deine the mechanism underlying LY2784544-mediated accumulation of CD34+ cells, we performed Affymetrix whole- transcript expression analysis in HEL cells treated with LY2784544 and/or GRN163L. The mRNA expression results are presented in Supplemental Table 1. A panel of genes was differentially expressed and we further selected them for veriication using QRT-PCR and/or Western blot. Altered expressions were seen for a number of genes intimately associated with hematopoiesis and stem cell phenotype; one of which was KLF4 (Fig. 2deg). KLF4 is essential for the gener- ation of induced pluripotent stem cells (iPSCs) and cellular de- differentiation, and we thus sought to determine whether KLF4 played a part in the LY2784544-mediated CD34 enrichment. To this end, we inhibited KLF4 expression in HEL cells using a shRNA spe- ciically targeting KLF4. Infection with the KLF4 ShRNA lentivirus suppressed KLF4 expression at both mRNA and protein levels in HEL cells(Fig.3aec),which subsequently attenuated LY2784544- mediated accumulation of CD34+ LSCs (Fig. 3a).
Fig. 2. LY2784544 increases the CD34+ fraction in CD34¡ HEL cells. (a) Histograms of CD34-APC in control and LY2784544-treated cells after 2 and 4 weeks of incubation. HEL cells were sorted into CD34þ and – populations before starting treatment. (b) Viability in sorted CD34þ and – HEL cells with and without 0.5 μM of LY2784544. (c) Percentage apoptotic cells in CD34-sorted HEL after 72hincubation with 0.5 μM LY2784544. (d) mRNA expression of KLF4 in HEL cells after 48 hincubation with 0.5 μM LY2784544. (e) Protein expression of KLF4 in HEL cells after 48 hincubation with 0.5 μM of LY2784544. (f) mRNA expression of KLF4 after 48 hincubation with 0.5 μM LY2784544 in sorted CD34þ and – HEL populations. (g) mRNA expression of KLF4 in unsorted HEL cells after 48 hincubation with 0.5 μM LY2784544 and/or 20 μM GRN163L. The levelsofTERT or KLF4 mRNA were arbitrarily expressed as the ratio of target mRNA/β2-M. All graphed data are shown as mean and standard deviation. Asterisks indicates p-values *<0.05, **三0.01, ***<0.001.
3.4. JAK2 inhibition in HEL cells causes down-regulation of TERT
expression and telomerase activity but paradoxically telomere lengthening
TERT mRNA expression was signiicantly reduced following treatment with LY2784544 for 48 h (P = 0.013) (Fig. 4a). Telomerase activity was accordingly reduced following 1 week (P = 0.002), 4 Photocatalytic water disinfection weeks (P = 0.011) and 9 weeks (P = 0.003) of JAK2 inhibition (Fig. 4b). Telomere length following JAK2 inhibition was studied with flow-FISH after 1, 4, 7, and 9 weeks of treatment. Despite down-regulated TERT and telomerase expression, a signiicant in- crease in telomere length was observed at 4 (P = 0.03), 7 (P = 0.002) and 9 weeks (P = 0.003), which indicates a dissociation between telomerase activity and telomere length in LY2784544-treated HEL cells (Fig. 4c). As JAK2 inhibition leads to CD34þ cell accumulation, the question is whether CD34þ cells carry longer telomeres than CD34-cells, or if the observed telomere lengthening in LY2784554- treated cells is simply due to the increase of the CD34þ fraction. To address this issue, we further separated two cell populations and determined their telomere lengths. Telomeres in the CD34þ HEL cells were longer than those in CD34- HEL cells (P = 0.044) (Fig. 4d). We found that telomeres were signiicantly longer in CD34þ cells treated with LY2784544 for 8 weeks compared to CD34 positive cells without treatment (P = 0.048) (Fig. 4d). Telomerase activity was signiicantly reduced in both CD34þ and – populations after 8 weeks of incubation (P = 0.012 and 0.016, respectively). Telomerase activity was reduced in both populations already after 4 weeks of treatment with LY2784544, but only signiicantly lower in the CD34þ cells (P = 0.026) (Fig. 4e). In concordance with this, TERT mRNA expression was also signiicantly lower in both CD34þ and – cells treated with LY2784544 (at 4 weeks: P = 0.03 and 0.003, respectively; at 8 weeks: P = or <0.001 in both populations) (Fig. 4f). Taken together, both CD34 cell selection and other un- known mechanisms contribute to telomere lengthening in HEL cells treated with LY2784554.
4.Discussion
The activating-mutation of JAK2V617F is a key genetic event to drive the development of MPNs, however, the therapeutic eficacy of JAK2 inhibitors in clinical trial has not lived up to expectations. The accumulated evidence indicates that JAK2 inhibitors fail to eradicate the disease clone [5], with a minimal effect of JAK2V617F allele burden on progenitors and stem cells [24]. To explore the underlying mechanism(s), we designed the present study. We show that JAK2V617F-carrying HEL cells undergo apoptosis and
Fig. 3.Inhibition of KLF4 expression attenuates the LY2784544 mediated increase in CD34+ cells. (aand b) Veriication of KLF4 suppression after transduction with lentivirus KLF4 ShRNA. (c) HEL cells were infected with scrambled control lenti-viral vector and ShKLF4 vector, respectively, and treated with 0.5 μM of LY2784544 for 2 weeks. Cells were then analyzed for KLF4 mRNA expression. The level of KLF4 mRNA was arbitrarily expressed as the ratio of KLF4 mRNA/β2-M. (d) GFP negative and positive populationsoh HEL cells represent shKLF4 non-transducedand transduced cells, respectively. Histogram of CD34-APC in cells treated with 0.5 μM of LY2784544 for 2 weeks. Upper right histogram show cells transduced with scrambled control ShRNA and lower right histogram show cells transduced with KLF4 ShRNA. All graphed data are shown as mean and standard deviation. Asterisks indicates p-values *<0.05, **<0.01, ***<0.001 diminished proliferation in the presence of the speciic JAK2 in- hibitor LY2784544 for the irst week, but recover gradually due to the CD34+ cell accumulation. Mechanistically, LY2784544 treat- ment led to the up-regulation of the stem cell factor KLF4 expres- sion. KLF4 inhibition signiicantly attenuated the enrichment of CD34+ cells mediated by LY2784544.Intriguingly, even in the CD34- HEL cell fraction, LY2784544 treatment still induced the generation and gradual accumulation of CD34+cells. It is thus evident from the present study that LY2784544 may promote stemness by inducing cellular de-differentiation.
KLF4 is one of the four transcription factors that when combined together can be manipulated to create iPSCs [25]. The role of KLF4 in hematopoiesis is not fully understood, but the ability of KLF4 together severe deep fascial space infections with other transcription factors to dedifferentiate mature
somatic cells into iPSCs suggests that it is involved in the mainte- nance of tissue-speciic stem cells. In addition, KLF4 is required for sustained cell proliferation by inhibiting cellular senescence [26]. KLF4 is not crucial for the development and maintenance of HSCs in the fetal liver, but this does not exclude its role in the regulation of the stem cell and progenitor niche in the adult bone marrow [27]. In this study, we do not address the mechanism by which inhibition of JAK2 in HEL cells increases the expression of KLF4; this has to be elucidated by further studies.Transcriptional proiling of CD34+cells from PV patients has shown that KLF4 is down- regulated and dependent on the action of JAK2V617F [28], which is consistent with our inding of KLF4 activation following JAK2 inhibition.JAK/STAT signaling has previously been suggested as a positive
Fig. 4. Alterations in telomere length and telomerase expression in HEL cells after treatment https://www.selleckchem.com/products/yo-01027.html with the JAK2 inhibitor LY2784544. (a) TERT mRNA expression in HEL cells treated with 0.5 μM ofLY2784544 for 48 h. (b) Telomerase activity in HEL cells analyzed with TRAP-ELISA after 1, 4, and 9 week incubation with 0.5 μM ofLY2784544, respectively. (c) Telomere length in HEL cells measured with flow-FISH after 1, 4, 7 and 9 weeks of incubation with 0.5 μM ofLY2784544. (d) Telomere length in sorted CD34+ and – populations after 4 and 9 weeks of incubation with 0.5 μM of LY2784544 as determined using flow-FISH. (e) Telomerase activity analyzed with TRAP-ELISA after 4 and 8 weeks of incubation with LY2784544 in sorted CD34+ and e HEL cell populations. (f) TERT mRNA expression in sorted CD34+ and – cells after 4 and 8 weeks incubation with 0.5 μM ofLY2784544. The level of TERT mRNA was arbitrarily expressed as the ratio of TERT/β2-M, and telomerase activity was expressed as absorbance assessed using a telomerase TRAP-ELISA kit. All graphed data are shown as mean and standard deviation. Asterisks indicates p-values *<0.05, **<0.01, ***<0.001 regulator of TERT expression by direct binding of STAT3 or STAT5 to the TERT promoter [29]. On the other hand, Zhang et al. showed that JAK2 inhibition was involved in senescence of hepatocellular carcinoma (HCC) cells, while the ectopic expression of TERT was capable of attenuating senescence by restoring the JAK2 activity [30]. These indings indicate a close interplay between JAK2 signaling and telomerase or TERT in oncogenesis. Consistent with the studies above, we observed that LY2784544 treatment of HEL cells signiicantly inhibited TERT expression and telomerase activ- ity. However, paradoxically, telomere length increased in HEL cells during JAK2 inhibition, uncouple with diminished levels of TERT and telomerase activity. Likely, telomeres were elongated in a telomerase-independent manner, or alternative lengthening of telomeres (ALT). ALT occurs frequently in sarcomas and high-grade astrocytomas [16,31]. It has been shown that malignant cells with telomerase activation are able to acquire ALT upon telomerase repression [32]. Because telomere over-erosion and dysfunction is widely present in MPN patient-derived myeloid cells [23,33], JAK2 inhibitor-mediated telomere lengthening may result in treatment failure [34]. Therefore, further studies are required to determine whether telomere elongation is seen in MPN cells from patients treated with JAK2 inhibitors, and if so, whether it will lead to treatment failure.The observations above provide a rationale to combine JAK2 inhibitors with telomerase inhibitors for MPN therapy.Indeed, GRN163L was able to abolish the CD34+ HEL cell accumulation resulting from LY2748544 treatment. We further demonstrated that telomere shortening did occur in GRN163L-treated HEL cells. Likely, shortened telomeres induced by GRN163L block the acqui- sition of stemness mediated by LY2784544. Defective telomere lengthening has been shown to promote HSC differentiation [35].An increase in CD34+ cells by LY2784544 treatment is also seen in the CD34- HEL population, which supports that the cells might have shifted to a more immature state. However, because the CD34+ fraction could not be eradicated completely using our sorting method, we were unable to fully exclude that such increase was due to an intrinsic resistance of the minimal CD34+ cells to LY2784544. Our results indeed suggest that those CD34+ cells are more tolerable to LY2784544 treatment than CD34-ones. Never- theless, LY2784544-mediated CD34+ cell accumulation provides a putative explanation for MPN patients’ poor response to JAK2 in- hibitors. Moreover, the identiication of KLF4 up-regulation by LY2784544 treatment not only gains mechanistic insights, but also suggests KLF4 as a new potential therapeutic target in MPNs. These indings should have important clinical signiicances. In summary, we report that treatment with the JAK2 inhibitor LY2784544 leads to an accumulation of CD34+cells via up- regulation of the stem cell factor KLF4 expression in a JAK2V617F- bearing leukemia derived cell line. Inhibiting LY2784544-mediated KLF4 induction signiicantly attenuated an increase in CD34+ cells. Therefore, targeting KLF4 together with JAK2 inhibitors may in- crease therapeutic eficacy for MPN patients. Furthermore, the accumulation of CD34+ cells resulting from LY2784544 was also blocked by simultaneous treatment with the telomerase inhibitor GRN163L. Both drugs used in this study have shown potential in the treatment of MPNs and our indings suggest that a combination of JAK2 and telomerase inhibitors might be effective in the treatment of MPNs.