The JAK2 V617F mutation occurs in hematopoietic ... - Semantic Scholar

2 downloads 0 Views 2MB Size Report
Apr 18, 2006 - stem cells in polycythemia vera and predisposes toward erythroid differentiation. Catriona H. M. Jamieson*, Jason Gotlib†, Jeffrey A. Durocher‡ ...
The JAK2 V617F mutation occurs in hematopoietic stem cells in polycythemia vera and predisposes toward erythroid differentiation Catriona H. M. Jamieson*, Jason Gotlib†, Jeffrey A. Durocher‡, Mark P. Chao†, M. Rajan Mariappan§, Marla Lay§, Carol Jones§, James L. Zehnder†§, Stan L. Lilleberg‡, and Irving L. Weissman§¶储 *Department of Medicine and Moores Cancer Center, University of California at San Diego, La Jolla, CA 92093; Departments of †Medicine and §Pathology, and ¶Institute for Stem Cell Biology and Regenerative Medicine and Comprehensive Cancer Center, Stanford University School of Medicine, Stanford, CA 94305; and ‡Transgenomic, Inc., Gaithersburg, MD 20878 Contributed by Irving L. Weissman, February 27, 2006

Although a large proportion of patients with polycythemia vera (PV) harbor a valine-to-phenylalanine mutation at amino acid 617 (V617F) in the JAK2 signaling molecule, the stage of hematopoiesis at which the mutation arises is unknown. Here we isolated and characterized hematopoietic stem cells (HSC) and myeloid progenitors from 16 PV patient samples and 14 normal individuals, testing whether the JAK2 mutation could be found at the level of stem or progenitor cells and whether the JAK2 V617F-positive cells had altered differentiation potential. In all PV samples analyzed, there were increased numbers of cells with a HSC phenotype (CD34ⴙCD38ⴚCD90ⴙLinⴚ) compared with normal samples. Hematopoietic progenitor assays demonstrated that the differentiation potential of PV was already skewed toward the erythroid lineage at the HSC level. The JAK2 V617F mutation was detectable within HSC and their progeny in PV. Moreover, the aberrant erythroid potential of PV HSC was potently inhibited with a JAK2 inhibitor, AG490. JAK 兩 signaling 兩 progenitors 兩 cell fate 兩 mutant allele

M

yeloproliferative disorders (MPDs), such as polycythemia vera (PV), are clonal hematopoietic disorders that share several characteristics (1–6), including the propensity to transform into myelofibrosis or acute myelogenous leukemia (7). Clonal derivation of PV from a primitive hematopoietic progenitor was first suggested by X chromosome linkage analysis that identified the same glucose-6-phosphate dehydrogenase allele in erythrocytes, granulocytes, and platelets of females with PV who were heterozygous for glucose-6-phosphate dehydrogenase alleles (8–10). PCR of the X-linked phosphoglycerate kinase gene was subsequently used to demonstrate the heterogeneity of clonal involvement of the myeloid and erythroid lineages in female PV patients (11). Subsequently, long-term PV marrow culture studies demonstrated unregulated proliferation of neoplastic progenitors, suggesting that primitive PV progenitors had an intrinsic defect that allowed them to bypass negative regulatory signals (12). However, the disease could have arisen in self-renewing HSC, in non-selfrenewing downstream multipotent progenitors, or even in myeloerythroid progenitors. Several recent reports provided critical insight into the genetic lesions involved in the development of PV by identifying a valineto-phenylalanine mutation at amino acid 617 (V617F) in the Janus kinase 2 (JAK2) tyrosine kinase gene in a substantial proportion of patients with PV as well as other MPDs, including essential thrombocythemia and idiopathic myelofibrosis (refs. 13–17 and reviewed in ref. 18). This recurrent somatic mutation results in constitutive activation of the JAK2 tyrosine kinase (15, 16). Expression of the mutant JAK2 gene in the Ba兾F3 factor-dependent cell line led to erythropoietin hypersensitivity and growth factorindependent survival (15). The in vivo relevance of the JAK2 V617F mutation to PV was tested in a mouse bone marrow transplant model (15). Marrow cells transduced with the JAK2 mutant allele 6224 – 6229 兩 PNAS 兩 April 18, 2006 兩 vol. 103 兩 no. 16

led to erythrocytosis after transplantation into lethally irradiated recipients (15). However, the role of the JAK2 V617F mutation in human PV pathogenesis and the stage of hematopoiesis at which it is expressed remained to be determined. In this study, we performed a targeted molecular analysis of cells with a hematopoietic stem cell (HSC) phenotype (CD34⫹CD38⫺CD90⫹Lin⫺) (19–22) and myeloid progenitors (23, 24) in PV patient samples to identify the hematopoietic progenitor compartment that harbors the JAK2 V617F mutation and their responses to a JAK2 inhibitor, AG490 (25, 26). Results PV Peripheral Blood Samples Have Increased Numbers of Cells with a HSC Phenotype. Phenotypic analysis of cells with a HSC phenotype

(CD34⫹CD38⫺CD90⫹Lin⫺) and progenitor populations was performed with the aid of FACS in PV and normal peripheral blood samples (19–24). Analysis of early-phase PV patient samples (Table 1) revealed increased numbers of cells with a HSC phenotype (Fig. 1A). When compared with PV patients with no leukocytosis (PV with normal WBC counts), patients who had PV with progressive myeloproliferation characterized by increasing leukocytosis (PV with high WBC counts) and splenomegaly had greater expansion of the HSC pool as well as an increase in the number of common myeloid progenitors (CMP) (Fig. 1 A) and the appearance of distinctive IL-3 receptor ␣-high cells (Fig. 1 A and B). PV Cells with a HSC Phenotype Exhibit Enhanced Erythroid Differentiation Potential. Hematopoietic progenitor assays revealed an

alteration in cell fate at the stem cell level in PV compared with normal controls. There were both quantitative (Fig. 2A) and qualitative (Fig. 2B) differences in the colony types derived from PV HSC compared with normal samples. PV HSC gave rise to a preponderance of large, abnormally shaped erythroid colonies compared with normal controls (Fig. 2B). Direct sequencing analysis of colonies derived from PV CD34⫹CD38⫺CD90⫹Lin⫺ (HSC) Conflict of interest statement: C.H.M.J. and I.L.W. have applied for U.S. patents entitled ‘‘Methods of Identifying and Isolating Stem Cells and Cancer Stem Cells’’ and ‘‘Methods of Diagnosing and Evaluating Blood Disorders’’ through the Stanford University Office of Technology and Licensing. I.L.W. receives consulting fees from and has equity ownership in Cellerant Therapeutics (San Carlos, CA). Freely available online through the PNAS open access option. Abbreviations: HSC, hematopoietic stem cell; CMP, common myeloid progenitor; GMP, granulocyte兾macrophage progenitor; MEP, megakaryocyte兾erythrocyte progenitor; PV, polycythemia vera; MPD, myeloproliferative disorder; JAK2 V617F, Janus kinase 2 valineto-phenylalanine mutation at amino acid 617; CFU, colony-forming unit; BFU, burstforming unit; BFU-E, BFU-erythroid; CFU-G, CFU-granulocyte; CFU-M, CFU-macrophage; CFU-Mix, mixed colonies composed of granulocytes, erythrocytes, megakaryocytes, and macrophages; CFU-Mega, CFU-megakaryocyte; CFU-GM, CFU-granulocyte兾macrophage. 储To whom correspondence should be addressed at: Department of Pathology, 279 Campus

Drive West, B257 Beckman Center, Stanford University School of Medicine, Stanford, CA 94305-5323. E-mail: [email protected]. © 2006 by The National Academy of Sciences of the USA

www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601462103

Table 1. Characteristics of patients with PV Patient no.

Age兾sex

Disease duration, years

WBCs per mm3

43兾M 59兾F 62兾M 68兾F 47兾F 63兾M 68兾F 30兾M 65兾F 48兾M 61兾F 46兾M 68兾M 80兾M 70兾F 76兾F

7 6 11 0.5 3 2 1 2 1.5 0.75 0.5 15 1 15 5.5 6.5

19,200 14,400 56,300 9,200 7,300 7,400 12,600 7,300 15,000 19,300 10,800 4,100 3,000 15,000 3,700 4,500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Hb (g兾dL)兾Hct (%)

Platelets per mm3

Disease-specific treatment at time of sample evaluation

Sample source

JAK2 V617F mutation status (PB)

14.7兾44.8 13.3兾41.8 12.4兾41.7 16.4兾51.6 13.4兾42.8 20.3兾60.6 16.2兾48.9 13.8兾40.8 15.5兾49.5 16.8兾51.5 16.1兾49.4 14.6兾43.9 18.7兾57.6 12.3兾40.4 12.2兾36.5 11.2兾32.0

687,000 248,000 903,000 511,000 567,000 194,000 734,000 244,000 666,000 230,000 1,302,000 308,000 461,000 853,000 558,000 24,000

PHB, HU PHB PHB, HU HU PHB PHB PHB PHB PHB PHB PHB, HU PHB, HU PHB PHB, HU PHB, HU BUS*

PB PB PB兾BM PB PB PB PB PB PB PB PB PB PB PB PB PB

⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

cells revealed that they harbored the JAK2 V617F mutation, suggesting that the JAK2 V617F mutation skewed differentiation of HSC toward an erythroid fate (Fig. 2B). The JAK2 V617F Mutation Occurs in PV Cells with a HSC Phenotype and Their Progeny. Of the 16 PV peripheral blood and bone marrow

mononuclear samples tested, 14 had the G 3 T mutation at nucleotide 1849 in exon 12 of the JAK2 gene, resulting in the V617F mutation (Table 1). In six of six patients identified as having the JAK2 V617F mutation in the mononuclear cell fraction, a more targeted molecular analysis demonstrated that the JAK2 V617F mutation was also detectable in purified HSC (Fig. 2C) that were sorted in sufficient numbers for accurate mutation detection. In patients with HSC involvement by JAK2 V617F, the mutation could also be detected in the CMP in four of five patients analyzed, in the

granulocyte兾macrophage progenitor (GMP) fractions in four of four patients analyzed, and in the megakaryocyte兾erythrocyte progenitor (MEP) fraction in four of four patients analyzed (Fig. 2C). In addition, the JAK2 V617F mutation was found in the unique IL-3 receptor ␣-high fraction of a PV sample but was absent in all HSC and progenitor populations derived from normal samples (n ⫽ 5). The mutant allele frequency was similar between HSC and more differentiated progeny, including the CMP, GMP, and MEP fractions (Fig. 2C), suggesting that the JAK2 V617F mutation is clonally transmitted by HSC to their progeny as had been suggested by PCR and glucose-6-phosphate dehydrogenase analysis but not previously established by using FACS-purified HSC and myeloid progenitors (8–11). Mutation analysis of colonies derived from PV HSC and committed progenitors revealed that not all colonies derived from these progenitors contained the JAK2 V617F muta-

Fig. 1. FACS-based progenitor profiling analysis demonstrated an increase in HSC in PV as well as a distinctive cell population with high IL-3 receptor ␣ expression. (A) Quantitative hematopoietic progenitor analysis. FACS analysis of primitive progenitors such as HSC and more committed progenitors including CMP, GMP, MEP, IL-3R␣⫹⫹CD45RA⫺, and IL-3R␣⫹⫹CD45RA⫹ cells revealed a statistically significant increase in the number of HSC per 105 mononuclear cells (P ⫽ 0.011, two-tailed Student’s t test, early PV versus Normal) in PV patients with normal WBC counts (PV normal WBC counts; n ⫽ 7) compared with normal peripheral blood samples (Normal; n ⫽ 4). PV associated with leukocytosis and兾or splenomegaly (PV high WBC counts; n ⫽ 3) was characterized by further expansion of the stem cell compartment (P ⫽ 0.006, statistically significant by unpaired two-tailed Student’s t test, advanced PV versus Normal) as well as an increase in CMP (P ⫽ 0.006, statistically significant by unpaired two-tailed Student’s t test, advanced PV versus Normal), IL-3 receptor ␣-high CD45RA⫺ (P ⫽ 0.04, statistically significant by unpaired two-tailed Student’s t test, advanced PV versus Normal), and IL-3 receptor ␣-high CD45RA⫹ (P ⫽ 0.02, statistically significant by unpaired two-tailed Student’s t test, advanced PV versus Normal) populations compared with normal peripheral blood samples (Normal). (B) PV versus normal peripheral blood progenitor profiles. Representative FACS progenitor profiles, obtained with the aid of a modified FACSVantage and FLOJO software, demonstrating that the percentage of the CD34⫹CD38⫹ lineage⫺ fraction composed of myeloid progenitors, including CMP, GMP, MEP, and IL-3 receptor ␣-high (IL-3-high) cells, in PV peripheral blood (Left) versus normal peripheral blood (Right).

Jamieson et al.

PNAS 兩 April 18, 2006 兩 vol. 103 兩 no. 16 兩 6225

DEVELOPMENTAL BIOLOGY

M, male; F, female; WBC, white blood cells; Hb, hemoglobin; Hct, hematocrit; PB, peripheral blood; BM, bone marrow; PHB, phlebotomy; HU, hydroxyurea; BUS, Busulfan. *Administered up to 2 months before sample evaluation.

Fig. 2. The JAK2 V617F mutation occurs in PV HSC, is transmitted without alteration in mutant allele frequency to committed progenitors, and predisposes PV HSC toward erythroid differentiation. (A) Quantitative analysis of differentiation potential of normal versus PV HSC in vitro. HSC (CD34⫹CD38⫺CD90⫹Lin⫺) derived from normal peripheral blood, bone marrow, or cord blood (Normal; n ⫽ 11) or PV (PV; n ⫽ 11) bone marrow or peripheral blood were FACS sorted with the aid of a modified FACSVantage onto 35-mm plates containing methylcellulose supplemented with recombinant human cytokines (Methocult GF⫹ HH35; StemCell Technologies). Colonies including CFU-Mega, CFU-GM, CFU-G, CFU-M, BFU-E, and CFU-E as well as CFU-Mix colonies were scored with the aid of a Nikon Eclipse TS100 inverted microscope on day 14. Results are expressed as the number of colonies per 100 cells plated. (B) Qualitative analysis of altered in vitro differentiation potential of PV HSC. (Upper) Representative phase-contrast photomicrographs of colonies derived from FACS-sorted HSC revealed increased erythroid differentiation potential of PV HSC compared with normal HSC. (Magnification: ⫻100.) HSC derived from normal peripheral blood, bone marrow, or cord blood (Normal; n ⫽ 11) or PV (PV; n ⫽ 11) bone marrow or peripheral blood were FACS-sorted onto methylcellulose supplemented with cytokines. (Lower) Mutation analysis performed on individual HSC colonies derived from normal (n ⫽ 4) or PV samples (n ⫽ 4) demonstrated that normal HSC colonies lacked the G 3 T mutation at nucleotide 1849 (black), whereas PV HSC colonies frequently harbored this JAK2 mutation (red), resulting in the valine-to-phenlyalanine substitution at position 617 (V617F). (C) PV HSC and their progeny harbor the JAK2 V617F mutation. Sequencing analysis revealed that PV peripheral blood and bone marrow samples derived from six of six patients with the JAK2 V617F mutation in their peripheral blood mononuclear cells harbored the same G 3 T mutation (black arrow) at nucleotide 1849 of JAK2 (red) in HSC as well as their progeny including CMP in four of five patients, GMP in four of four patients, and MEP in four of four patients analyzed, indicating clonal transmission of the mutation. (D) Expression of JAK2 V617F in normal versus PV colonies. Sequencing analysis was performed to detect the presence of the JAK2 V617F mutation (JAK2⫹) versus the absence of JAK2 V617F mutation (JAK2⫺) in normal (Normal; n ⫽ 3) and PV (PV; n ⫽ 4) HSC, CMP, GMP, or MEP colonies. This analysis revealed that not all PV progenitor-derived colonies from patients with the JAK2 mutation in peripheral blood mononuclear cells harbored the JAK2 V617F mutation. Results are expressed as the number of JAK2⫺ and JAK2⫹ colonies out of the total number of colonies analyzed that were derived from HSC, CMP, GMP, and MEP.

tion (Fig. 2D), and therefore these mutant lineage cells cohabitate with normal stem and progenitor cells in PV bone marrow and peripheral blood. Alternatively, it is possible that erythropoietinindependent marrow cells without the JAK2 V617F mutation coexist with JAK2 V617F-positive cells in PV. Inhibition Assays with the JAK2 Inhibitor AG490. To ascertain

whether mutant JAK2 signaling played a role in the skewed differentiation potential of PV versus normal samples, hematopoietic progenitor assays were performed in the presence or absence of AG490, a well characterized JAK2 inhibitor, at a dose (50 ␮M) that was reported to preferentially induce apoptosis in leukemic compared with normal cells (25, 26). When purified normal HSC, CMP, and MEP were exposed to AG490 in hematopoietic progenitor assays, erythroid and mixed colony formation were partially inhibited, whereas myeloid colony formation was not affected (Fig. 3A). In PV samples, however, there was a significant decrease in the number of mixed colonies and a diminution in erythroid colonies derived from HSC (Fig. 3 B and C) after exposure to AG490, whereas the effects were not as dramatic for CMP and MEP (Fig. 3B), suggesting an increased sensitivity of PV to JAK2 inhibition at the stem cell level. Mutation analysis of HSC-derived colonies from 6226 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601462103

four patients with PV revealed that the JAK2 V617F mutation persisted in some colonies that survived treatment with AG490 (Fig. 3D). Further JAK2 V617F mutation analysis of individual HSC-derived colonies from two separate patients revealed differences in individual patient sensitivity to JAK2 inhibition at the stem cell level, indicating that in some patients, HSC clones harboring the JAK2 V617F mutation are less sensitive to inhibition with AG490 (Fig. 3E). These studies suggest that initially, PV patients have a proliferation of cells with a HSC phenotype, in part as a consequence of JAK2 V617F mutation expression, resulting in an increase in terminally differentiated progeny, whereas patients with evidence of progressive disease such as increasing WBC counts and splenomegaly also have an expansion of the CMP compartment and the production of a unique IL-3 receptor ␣-high population. Discussion We have previously proposed (19, 27–30), and in some cases shown (23, 31), that, in the progression of myelogenous leukemias (and by analogy many cancers), the self-renewing stem cell (here HSC) must be the first cell to sustain the genetic or epigenetic event that initiates the MPD, and in the myeloid lineage, only HSC self-renew (32, 33). However, it is possible that the first event, in some cases, Jamieson et al.

occurs in a progenitor that cannot self-renew. In that view, successive clonal progeny from the altered HSC clone could accumulate progression events, so that at the end of the process, a multiply altered clone could enable poorly regulated self-renewal by turning on the hematopoietic self-renewal pathway(s) (19). A corollary of that hypothesis is that the initiating events in these MPDs that can Jamieson et al.

progress to acute leukemias sustain the events in HSC. Previous studies have demonstrated that PV arises from an oligopotent or multipotent hematopoietic progenitor (8–12). Consistent with this observation, a previously unrecognized mutation at position 617 of the JAK2 signaling molecule (JAK2 V617F mutation) was detected in individual colony-forming unit (CFU)-granulocyte兾macrophage PNAS 兩 April 18, 2006 兩 vol. 103 兩 no. 16 兩 6227

DEVELOPMENTAL BIOLOGY

Fig. 3. Aberrant PV HSC erythroid differentiation potential is inhibited but JAK2 V617F-positive colonies are not completely eliminated by AG490, a JAK2 inhibitor. (A) Effect of JAK2 inhibition on normal HSC differentiation potential in vitro. HSC, CMP, or MEP derived from normal bone marrow, peripheral blood, or cord blood (n ⫽ 7) were FACS sorted onto methylcellulose supplemented with or without AG490 in addition to cytokines. Colonies of CFU-G, CFU-M, BFU-E, and CFU-E, as well as CFU-Mix, were scored with the aid of a Nikon Eclipse TS100 inverted microscope on day 14. Results are expressed as the number of colonies per 100 cells plated. A two-tailed unpaired Student’s t test performed with EXCEL software revealed that there was no statistically significant difference in BFU-E colony formation by normal peripheral blood HSC before or after AG490 treatment (P ⫽ 0.67), whereas normal bone marrow and cord blood derived HSC appeared to be more sensitive to the effects of AG490 with regard to erythroid colony-forming potential (P ⫽ 0.054), as was the mixed-colony-forming potential of normal samples (P ⫽ 0.063). There were no statistically significant differences in the proportions of HSC-derived CFU-GM, CFU-G, CFU-Mega, or CFU-M before or after AG490 treatment. (B) Effect of JAK2 inhibition on PV HSC differentiation potential in vitro. Hematopoietic progenitor assays were performed on HSC, CMP, or MEP derived from PV bone marrow or peripheral blood samples (n ⫽ 9) that were FACS-sorted onto methylcellulose supplemented with or without AG490 in addition to cytokines. Colonies including CFU-Mega, CFU-GM, CFU-G, CFU-M, BFU-E, and CFU-E as well as CFU-Mix were scored with the aid of a Nikon Eclipse TS100 inverted microscope on day 14. Results are expressed as the number of colonies per 100 cells plated. A two-tailed unpaired Student’s t test performed with EXCEL software revealed that there was a statistically significant difference in the number CFU-Mix derived from PV HSC before and after AG490 treatment (n ⫽ 0.027). Although there was a trend toward a reduction in PV HSC-derived BFU-E after the addition of AG490, it was not statistically significantly different (P ⫽ 0.17) from untreated controls. There was no reduction in other colony types derived from HSC or other progenitor populations before or after AG490 treatment. (C) Qualitative assessment of JAK2 inhibition in normal versus PV HSC. The effect of JAK2 inhibition with AG490 (50 ␮M) on normal versus PV HSC in vitro differentiation potential was assessed. Representative phase-contrast photomicrographs were obtained with a Nikon Eclipse TS100 microscope and SPOT software. (Magnification: ⫻100.) Normal (Upper) or PV (Lower) HSC were FACS-sorted onto methylcellulose supplemented with or without AG490 (50 ␮M) in addition to cytokines. (D) Analysis of JAK2 V617F expression by normal versus PV HSC colonies before and after JAK2 inhibition. Sequencing analysis of JAK2 V617F mutation (JAK2⫹) expression was performed on HSC colonies derived from three normal individuals (Normal 1–3) versus four patients with PV (PV1– 4) before and after in vitro treatment with AG490 (50 ␮M), a JAK2 inhibitor. This analysis revealed that the proportion of HSC colonies harboring the JAK2 V617F mutation (JAK2⫹) as opposed to those without the JAK2 V617F mutation (JAK2⫺) varies among PV patients and that the JAK2 V617F mutation persists despite AG490 treatment in the HSC of three of four PV patients. Results are expressed as the number of JAK2⫺ and JAK2⫹ colonies of the total number of colonies derived from HSC from each individual before and after in vitro AG490 treatment. (E) PV HSC and their progeny harbor the JAK2 V617F mutation. Sequencing analysis revealed that PV samples from patients with the JAK2 V617F mutation in their peripheral blood mononuclear cells harbored the same G 3 T mutation (black arrow) at nucleotide 1849 of JAK2 (red) in their HSC but that their HSC-derived colonies had different sensitivities to in vitro JAK2 inhibition with AG490 (50 ␮M).

(CFU-GM), burst-forming unit (BFU)-erythroid (BFU-E), and endogenous erythroid colonies derived from a patient with essential thrombocythemia. In three patients with MPDs, the JAK2 mutation was identified in CD34⫹ progenitor cells, with the mutant form predominating in a mixed clonality pattern with wild-type JAK2 (34). However, in preliminary studies, JAK2 V7617F was not detected in B or T lymphocyte fractions, allowing the possibility that the target population was not multipotent but myeloidcommitted (13, 15, 34). In this study, direct sequencing analysis of the JAK2 V617F mutation in primitive hematopoietic progenitors, including HSC and their progeny, revealed PV to be a HSC disease, wherein the JAK2 V617F mutation alters the differentiation potential of HSC. Our previous progenitor profiling studies involving chronic myelogenous leukemia, a clonal disorder wherein the initial BCR-ABL translocation occurs at the level of HSC, showed that as the disease progresses from a chronic MPD to blast crisis, there was an expansion of the GMP pool (23). Progression was attributed, in part, to aberrant acquisition of self-renewal potential as a result of ␤-catenin activation within the committed GMP population, as well as increased proliferative capacity secondary to BCR-ABL amplification (23). This study demonstrates that progenitor profiling could provide critical insights into the stage of hematopoiesis at which key events involved in progression of MPDs, such as chronic myelogenous leukemia, occur. In this study, phenotypic and functional progenitor profiling together with targeted JAK2 sequencing analysis revealed five previously unrecognized findings: (i) in PV, there is an increase in the number of cells with a HSC phenotype, expansion of the CMP pool, and emergence of an IL-3 receptor ␣-high progenitor population with disease progression; (ii) there is a quantitative and qualitative alteration in differentiation potential of PV HSC; (iii) cells with a HSC phenotype, from the majority of PV samples tested, harbored the JAK2 V617F mutation which (iv) was transmitted in a clonal manner to more committed progenitors; and (v) the multilineage differentiation potential of PV HSC was more susceptible to inhibition with a JAK2 antagonist, AG490, than normal HSC. Many scientific and clinical reports have demonstrated that human HSC, regardless of the tissue of origin, share the same phenotype (CD34⫹CD38⫺CD90⫹Lin⫺) and can be purified by FACS (19–22). In this study, we observed that the number of cells with a HSC phenotype was increased in the peripheral blood of patients with JAK2 V617F-positive PV compared with normal samples and that the in vitro differentiation capacity of PV was profoundly altered. These findings suggest that, in addition to enhanced proliferative capacity, one of the primary defects in PV may be altered differentiation potential at the stem cell level. However, additional mutations are likely responsible for propagation of JAK2 V617F-positive clones and expansion of the CMP compartment, such as changes in survival, self-renewal, or replicative capacity that may hasten the progression of PV and evolution to acute leukemia as has been demonstrated with other MPDs (23, 28–30, 35–38). We identified an IL-3 receptor ␣-high progenitor population that was unique to PV patients. Prior reports showed that PV marrow BFU-E, CFU-GM, and CFU-megakaryocyte (CFU-Mega) exhibit marked hypersensitivity to recombinant IL-3. The mechanism was not found to caused by enhanced binding of recombinant IL-3 to its receptor (3, 39). The growth factors to which hematopoietic progenitors are hypersensitive in MPDs (e.g., IL-3, granulocyte兾 macrophage colony stimulating factor, erythropoietin, stem cell factor, thrombopoietin, and insulin-like growth factor-1) all employ JAK2 for signaling, and therefore, constitutive activation of the JAK2 signaling pathway may partly explain this observation (18, 40). Functional analysis of IL-3 receptor signaling in the IL-3 receptor ␣-high population should be undertaken to evaluate its 6228 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0601462103

relevance to increased JAK2 tyrosine kinase activity, and more broadly, to the pathogenesis of PV and other MPDs. We used targeted JAK2 mutation screening of cells with a HSC phenotype and committed progenitors to confirm the clonal HSC origin of PV. The JAK2 V617F mutation was identified in HSC as well as their progeny including CMP, GMP, MEP, and the IL-3 receptor ␣-high population. JAK2 V617F appears to be propagated without diminution in mutant allele frequency from HSC to more committed hematopoietic progenitors. Analysis of a larger cohort of samples will be required to assess whether JAK2 mutant allele frequency increases preferentially in more terminally differentiated progenitor populations and whether this contributes to the proliferation of particular lineages in different MPDs. Although inhibition of JAK2 with AG490 decreased the aberrant erythroid potential of PV HSC, it did not eradicate all JAK2 V617F-positive colonies and also exhibited inhibitory effects on the erythroid potential of normal HSC but to a lesser extent than PV HSC. These data suggest that a tyrosine kinase inhibitor with more specificity for the JAK2 V617F protein will likely be required to preferentially target JAK2 mutation-positive clones and to avoid inhibitory effects on normal erythropoiesis. Materials and Methods Samples. Peripheral blood and bone marrow samples (n ⫽ 16) were

donated by patients with PV. Normal bone marrow (n ⫽ 14) or cord blood (n ⫽ 2) and peripheral blood samples (n ⫽ 4) were provided by healthy volunteers. Samples were obtained with informed consent according to Stanford University Institutional Review Boardapproved protocols. Normal bone marrow and cord blood samples were also purchased from All Cells (Berkeley, CA).

Huma HSC and Myeloid Progenitor Flow-Cytometric Analysis and Cell Sorting. Mononuclear fractions were extracted from peripheral

blood or bone marrow after Ficoll density centrifugation according to standard methods (23, 24). Samples were analyzed fresh or subsequent to rapid thawing of samples previously frozen in 90% FBS and 10% DMSO in liquid nitrogen. In some cases, CD34⫹ cells were enriched from mononuclear fractions with the aid of immunomagnetic beads (CD34⫹ Progenitor Isolation Kit; Miltenyi Biotec, Auburn, CA). Before HSC and myeloid progenitor FACS analysis and sorting, CD34-enriched cell populations or mononuclear cells were stained with lineage marker-specific phycoerythrin-Cy5-conjugated antibodies including CD2 (RPA-2.10), CD11b (ICRF44), CD20 (2H7), CD56 (B159), and GPA (GA-R2) from Becton Dickinson Pharmingen; CD3 (S4.1), CD4 (S3.5), CD7 (CD7-6B7), CD8 (3B5), CD10 (5-1B4), CD14 (TUK4), and CD19 (SJ25-C1) from Caltag (South San Francisco, CA); APC-conjugated anti-CD34 (HPCA-2; Becton Dickinson Pharmingen); and biotinylated anti-CD38 (HIT2; Caltag) in addition to phycoerythrin-conjugated anti-IL-3 receptor ␣ (9F5; Becton Dickinson Pharmingen) and FITC-conjugated anti-CD45RA (MEM56; Caltag) followed by staining with streptavidin-Alexa Fluor 594 (Invitrogen) to visualize CD38-biotinstained cells and resuspension in propidium iodide to exclude dead cells. The methodology was the same for HSC staining except that a phycoerythrin-conjugated anti-human CD90 antibody (Becton Dickinson Pharmingen) was used instead of an anti-IL-3 receptor ␣ antibody. Unstained samples and isotype controls were included to assess background fluorescence. After staining, cells were analyzed and sorted by using a modified FACSVantage (Becton Dickinson Immunocytometry Systems) equipped with a 599-nm dye laser and a 488-nm argon laser. Double-sorted HSC were identified as CD34⫹CD38⫺CD90⫹ and lineage-negative. CMP were identified based on CD34⫹CD38⫹IL-3R␣⫹CD45RA⫺Lin⫺ staining, and their progeny including GMP were CD34⫹CD38⫹IL3R␣⫹CD45RA⫹Lin⫺, whereas MEP were identified based on CD34⫹CD38⫹IL-3R␣⫺CD45RA⫺Lin⫺ staining (16, 17, 23, 24). Jamieson et al.

(CD34⫹CD38⫺CD90⫹Lin⫺), CMP (CD34⫹CD38⫹IL-3R␣⫹CD45RA⫺Lin⫺), GMP (CD34⫹CD38⫹IL-3R␣⫹CD45RA⫹Lin⫺), and MEP (CD34⫹CD38⫹IL-3R␣⫺CD45RA⫺Lin⫺) were sorted with the aid of a FACSVantage directly onto 35-mm plates containing complete methylcellulose (GF ⫹ H4435; StemCell Technologies, Vancouver) according to the manufacturer’s specifications, with or without a 50 ␮M concentration of the JAK2 inhibitor AG490 (Tyrphostin B42; Calbiochem) (23–26). Colonies were incubated in a 37°C 7% CO2 humidified incubator and scored on day 14 as CFU-Mix (mixed colonies composed of granulocytes, erythrocytes, megakaryocytes, and macrophages), BFU-E or CFUerythroid (CFU-E), CFU-granulocyte (CFU-G), CFU-macrophage (CFU-M), CFU-Mega, or CFU-GM (24). Phase-contrast photomicrographs of colonies were obtained on day 14 with a Nikon phase-contrast inverted microscope at ⫻100 magnification with the aid of SPOT software. JAK2 Mutation Screening. Mononuclear cells. JAK2 V617F mutation

genotyping was performed on peripheral blood or bone marrow mononuclear cells derived from patients with PV, as well as normal peripheral blood, bone marrow, and cord blood. Red blood cells were lysed, and DNA was extracted with the QIAamp DNA Blood Mini kit according to the manufacturer’s directions (Qiagen, Valencia, CA) and then stored at ⫺80°C until amplification-based testing. Extracted DNA was prepared for JAK2 mutation analysis by LightCycler (Roche Applied Science, Indianapolis) methodology (see section B of Supporting Methods, which is published as supporting information on the PNAS web site). HSC- and progenitor-targeted JAK2 mutation analysis. Targeted JAK2 V617F sequencing analysis was performed on cDNA derived from FACS-sorted HSC, CMP, GMP, and MEP from normal peripheral blood, bone marrow, or cord blood versus PV peripheral blood or bone marrow. In some experiments, methylcellulose-containing colonies (whole-plate) derived from individual progenitor populations were resuspended in 1 ml of TRIzol (Invitrogen), RNA was extracted, and cDNA was made and sequenced for the JAK2 V617F mutation (see section C of Supporting Methods for information on PCR and JAK2 primers). Clonal JAK2 mutation analysis. To further investigate whether JAK2 V617F occurred as a clonal mutation, sequencing analysis of JAK2 was performed on individual colonies derived from HSC, CMP, GMP, and MEP populations with or without in vitro inhibition of 1. Prchal, J. F. & Axelrad, A. A. (1974) N. Engl. J. Med. 290, 1382–1384. 2. Zanjani, E. D., Lutton, J. D., Hoffman, R. & Wasserman, L. R. (1977) J. Clin. Invest. 59, 841–848. 3. Dai, C. H., Krantz, S. B., Dessypris, E. N., Means, R. T., Jr., Horn, S. T. & Gilbert, H. S. (1992) Blood 80, 891–899. 4. Dai, C. H., Krantz, S. B., Koury, S. T. & Kollar, K. (1994) Br. J. Haematol. 88, 497–505. 5. Correa, P. N., Eskinazi, D. & Axelrad, A. A. (1994) Blood 83, 99–112. 6. Axelrad, A. A., Eskinazi, D., Correa, P. N. & Amato, D. (2000) Blood 96, 3310–3321. 7. Spivak, J. L. (2004) Semin. Hematol. 41, 1–5. 8. Adamson, J. W., Fialkow, P. J., Murphy, S., Prchal, J. F. & Steinmann, L. (1976) N. Engl. J. Med. 295, 913–916. 9. Prchal, J. F., Adamson, J. W., Steinmann, L. & Fialkow, P. J. (1976) J. Cell. Physiol. 89, 489–492. 10. Fialkow, P. J., Faguet, G. B., Jacobson, R. J., Vaidya, K. & Murphy, S. (1981) Blood 518, 916–919. 11. Gilliland, D. G., Blanchard, K. L., Levy, J., Perrin, S. & Bunn, H. F. (1991) Proc. Natl. Acad. Sci. USA 88, 6848–6852. 12. Cashman, J. D., Eaves, C. J. & Eaves, A. C. (1988) J. Clin. Invest. 81, 87–91. 13. Baxter, E. J., Scott, L. M., Campbell, P. J., East, C., Fourouclas, N., Swanton, S., Vassiliou, G. S., Bench, A. J., Boyd, E. M., Curtin, N., et al. (2005) Lancet 365, 1054–1061. 14. Levine, R. L., Wadleigh, M., Cools, J., Ebert, B. L., Wernig, G., Huntly, B. J., Boggon, T. J., Wlodarska, I., Clark, J. J., Moore, S., Adelsperger, J., et al. (2005) Cancer Cell 7, 387–397. 15. James, C., Ugo, V., Le Couedic, J. P., Staerk, J., Delhommeau, F., Lacout, C., Garcon, L., Raslova, H., Berger, R., Bennaceur-Griscelli, A., et al. (2005) Nature 434, 1144–1148. 16. Kralovics, R., Passamonti, F., Buser, A. S., Teo, S. S., Tiedt, R., Passweg, J. R., Tichelli, A., Cazzola, M. & Skoda, R. C. (2005) N. Engl. J. Med. 352, 1779–1790. 17. Steensma, D. P., Dewald, G. W., Lasho, T. L., Powell, H. L., McClure, R. F., Levine, R. L., Gilliland, D. G. & Tefferi, A. (2005) Blood 106, 1207–1209. 18. Kaushansky, K. (2005) Blood 105, 4187–4190. 19. Weissman, I. (2005) J. Am. Med. Assoc. 294, 1359–1366. 20. Negrin, R. S., Atkinson, K., Leemhuis, T., Hanania, E., Juttner, C., Tierney, K., Hu, W. W.,

Jamieson et al.

JAK2 with AG490. Individual colonies were plucked and resuspended in 200 ␮l of RLT buffer supplemented with ␤-mercaptoethanol (RNeasy kit; Qiagen) and frozen immediately at ⫺80°C. Samples were thawed and RNA was extracted, followed by cDNA preparation and PCR amplification with JAK2-specific primers (see section C of Supporting Methods). Mutation scanning and DNA sequencing. Mutation analysis of the JAK2 cDNA PCR product was conducted by using fluorescent denaturing high-performance liquid chromatography (DHPLC) technology and Surveyor mismatch cleavage analysis, both with the WAVE-HS System (Transgenomic). Aliquots of PCR product (3–15 ␮l) from all samples were scanned for mutations by DHPLC, confirmed by Surveyor mismatch cleavage, and identified with bidirectional sequence analysis on an ABI 3100 sequencer using BigDye V3.1 terminator chemistry (Applied Biosystems). In addition, for semiquantitative determination of mutant and normal allele frequencies, relative peak areas of denaturing high performance liquid chromatography elution profiles, and Surveyor mismatch cleavage products were determined after normalization and comparison with reference controls by using the WAVE NAVIGATOR software. Statistical Analysis. Standard deviation, standard error of the mean,

and numbers of HSC and progenitors per 105 mononuclear cells were measured by using FLOJO (Treestar, San Carlos, CA) and Microsoft EXCEL software. Two-tailed Student’s t test (EXCEL) was used to analyze statistical differences in the number of different progenitors between PV patient samples and normal controls. We thank Lenn Fechter and Drs. Steven Coutre´, Stanley Schrier, Caroline Berube, and Lawrence Leung (all of Stanford University) for providing samples from patients with PV; Drs. William Maloney, Derrick Rossi, and David Bryder (all of Stanford University) for their assistance in obtaining and processing normal age-matched bone marrow samples; Libuse Jerabek for excellent laboratory management; the Stanford University FACS facility for expert assistance; members of the Stanford University Division of Hematology for their support; and the patients who made this research possible. This work was supported by a Stanford Center for Clinical Immunology Yu–Bechmann fellowship for Genomics and Oncology and an Aplastic Anemia and Myelodysplastic Syndrome International Foundation Investigator Award (both to C.H.M.J.); the Walter and Beth Weissman Fund; the Smith Family Fund; National Institutes of Health Grants CA086065 and CA86017 (to I.L.W.), K23 HL04409 (to J.G.), and 2P01CA49605 (to C.J. and J.L.Z.); and a de Villiers grant from the Leukemia Society (to I.L.W.). Johnston, L. J., Shizurn, J. A., Stockerl-Goldstein, K. E., et al. (1992) Biol. Blood Marrow Transplant. 6, 262–271. 21. Baum, C. M., Weissman, I. L., Tsukamoto, A. S., Buckle, A. M. & Peault, B. (1992) Proc. Natl. Acad. Sci. USA 89, 2804–2808. 22. Peault, B., Weissman, I. L., Buckle, A. M., Tsukamoto, A. & Baum, C. (1993) Nouv. Rev. Fr. Hematol. 35, 91–93. 23. Jamieson, C. H., Ailles, L. E., Dylla, S. J., Muijtjens, M., Jones, C., Zehnder, J. L., Gotlib, J., Li, K., Manz, M. G., Keating, A., et al. (2004) N. Engl. J. Med. 351, 657–667. 24. Manz, M. G., Miyamoto, T., Akashi, K. & Weissman, I. L. (2002) Proc. Natl. Acad. Sci. USA 99, 11872–11877. 25. Ugo, V., Marzac, C., Teyssandier, I., Larbret, F., Lecluse, Y., Debili, N., Vainchenker, W. & Casadevall, N. (2004) Exp. Hematol. (Charlottesville, Va) 32, 179–187. 26. Faderl, S., Harris, D., Van, Q., Kantarjian, H. M., Talpaz, M. & Estrov, Z. (2003) Blood 102, 630–637. 27. Reya, T., Morrison, S. J., Clarke, M. F. & Weissman, I. L. (2001) Nature 414, 105–111. 28. Jamieson, C. H. M., Weissman, I. L. & Passegue´, E. (2004) Cancer Cell 6, 531–533. 29. Jamieson, C. H. M., Passegue´, E. & Weissman, I. L. (2004) in Stem Cells in the Nervous System: Functional and Clinical Implications, ed. Gage, R. (Springer, Berlin), pp. 157–182. 30. Passegue´, E., Jamieson, C. H. M., Ailles, L. E. & Weissman, I. L. (2003) Proc. Natl. Acad. Sci. USA 100, 11842–11849. 31. Miyamoto, T., Weissman, I. L. & Akashi, K. (2000) Proc. Natl. Acad. Sci. USA 97, 7521–7526. 32. Morrison, S. J. & Weissman, I. L. (1994) Immunity 1, 661–673. 33. Akashi, K., Traver, D., Miyamoto, T. & Weissman, I. L. (2000) Nature 404, 193–197. 34. Lasho, T. L., Mesa, R., Gilliland, D. G. & Tefferi, A. (2005) Br. J. Haematol. 130, 797–799. 35. Bonnet, D. & Dick, J. E. (1997) Nat. Med. 3, 730–737. 36. Huntly, B. J., Shigematsu, H., Deguchi, K., Lee, B. H., Mizuno, S., Duclos, N., Rowan, R., Amaral, S., Curley, D., Williams, I. R., et al. (2004) Cancer Cell 6, 587–596. 37. Passegue´, E., Wagner, E. F. & Weissman, I. L. (2004) Cell 119, 431–443. 38. Xie, S., Lin, H., Sun, T. & Arlinghaus, R. B. (2002) Oncogene 21, 7137–7146. 39. Kobayashi, S., Teramura, M., Hoshino, S., Motoji, T., Oshimi, K. & Mizoguchi, H. (1993) Br. J. Haematol. 83, 539–544. 40. Yamaoka, K., Saharinen, P., Pesu, M., Holt, V. E., III, Silvennoinen, O. & O’Shea, J. J. (2004) Genome Biol. 5, 253–258.

PNAS 兩 April 18, 2006 兩 vol. 103 兩 no. 16 兩 6229

DEVELOPMENTAL BIOLOGY

Hematopoietic Progenitor Assays. Normal and PV HSC