PERIOSTIN COMPOUNDS FOR THE TREATMENT OF HAEMATOLOGICAL COMPLICATIONS

Abstract

The present invention pertains to periostin compounds for use in the prevention and treatment of haematological complications, such as adverse events from therapy or haematological diseases. In context of the present invention a therapeutic was developed for enhancing haematopoiesis in patients and to support haematopoietic stem cell (HSC) transplantation (HSCT) by administration of periostin compounds to patients or stem cell donors, or by contacting HSC directly with periostin compounds, for example ex vivo, to improve a transplant HSC preparation. The present invention provides periostin derived compounds such as polypeptides, peptides, nucleic acids, and other periostin-derived agents, that are used both in therapeutic applications and for improving haematopoiesis, for example in stem cell donor subjects or to treat HSC in vitro.

Claims

1. A periostin compound for use in the prevention or treatment of a haematological disorder in a subject, wherein the periostin compound is selected from a periostin protein, or a functional fragment or variant thereof, or a periostin nucleic acid encoding the periostin protein, or encoding the functional fragment or variant thereof.

2. The periostin compound for use according to claim 1, wherein the treatment comprises (i) the administration of the periostin compound to the subject suffering from the haematological disorder, or (ii) in-vitro or in-vivo treating a biological cell with the periostin compound and administering the so periostin compound-treated cell to the subject suffering from the haematological disorder.

3. The periostin compound for use according to claim 2, wherein the biological cell is an autologous or allogenic haematopoietic stem cell (HSC).

4. The periostin compound for use according to claim 3, wherein the allogenic HSC is derived from an umbilical cord blood sample or from a bone marrow sample or from a mobilized haematopoietic stem cell obtained from peripheral blood, or from a placenta.

5. The periostin compound for use according to claim 1, wherein the periostin is a human periostin isoform selected from isoform 1 to 7 (SEQ ID NO: 1 to 7).

6. The periostin compound for use according to claim 1, wherein the variant of the periostin protein comprises an amino acid sequence that is at least 80% identical to an amino acid sequence of a human periostin isoform shown in any one of SEQ ID NO: 1 to 7.

7. The periostin compound for use according to claim 1, wherein the haematological disorder is a haematological malignancy, a disease associated with a pathological haematopoietic stem cell (HSC) function, such as a disease associated with decreased/impaired haematopoiesis, or is a haematological adverse event caused by a primary treatment of the subject, for example with a primary therapeutic for another, for example non-haematological, disorder.

8. The periostin compound for use according to claim 7, wherein the haematological adverse event is caused by a treatment of the subject with an anti-cancer agent, such as a chemotherapeutic agent, or by a radiotherapy treatment, or by treatment with a vitamin-K antagonist, such as warfarin, fluindione, phenindione, acenocoumarol, dicoumarol, ethyl biscoumacetate, or phenprocoumon, preferably in a subject suffering from or suspected to develop thromboembolic complications.

9. A pharmaceutical composition for use in the prevention or treatment of a haematological disorder in a subject, comprising a periostin compound selected from a periostin protein, or a functional fragment or variant thereof, or a periostin nucleic acid encoding the periostin protein, or encoding the functional fragment or variant thereof, and a pharmaceutically acceptable carrier and/or excipient

10. An in-vitro method for preparing a stem cell transplant, the method comprising the steps of a. Providing a composition of stem cells, b. Contacting the composition of stem cells with a periostin compound as described in claim 1, c. Incubating the mixture of b for a sufficient amount of time to obtain a suitable stem cell transplant d. Optionally, culturing and/or purifying the stem cell transplant.

11. The method according to claim 10, wherein the stem cells are HSC, preferably HSC derived from an umbilical cord blood sample or from a bone marrow sample.

12. A method for improving haematopoietic stem cell function/activity, the method comprising contacting the HSC with a sufficient amount of a periostin compound as described in claim 1.

13. A non-therapeutic method for enhancing haematopoiesis in subject to improve or prepare a stem cell donation of the subject, the method comprising administering to the subject a periostin compound selected from a periostin protein, or a functional fragment or variant thereof, or a periostin nucleic acid encoding the periostin protein, or encoding the functional fragment or variant thereof.

14. The method according to claim 13, wherein the method is for improving the quality and/or quantity of a stem cell donation obtained from the subject.

15. A non-therapeutic method for obtaining stem cells from a subject, the method comprising the steps of a. administering to the subject a periostin compound selected from a periostin protein, or a functional fragment or variant thereof, or a periostin nucleic acid encoding the periostin protein, or encoding the functional fragment or variant thereof, b. harvesting stem cells from the subject.

Description

[0081] The present invention will now be further described in the following examples with reference to the accompanying figures and sequences, nevertheless, without being limited thereto. For the purposes of the present invention, all references as cited herein are incorporated by reference in their entireties. In the Figures:

[0082] FIG. 1: Warfarin compromises hematopoiesis and HSC function. A and B) Total number of leukocytes (P=0.008 and P=0.002; t-test, n=5) (A) and monocytes (P=0.007 and P=0.001; t-test, n=5) (B) in peripheral blood of control mice (black) or mice treated with 0.5 mg/kg (white) or 0.05 mg/kg (gray) warfarin 14 days after begin of treatment. C) Percentage of CD11b+ myeloid cells in peripheral blood (P=0.029; t-test, n=3). D) Number of colonies in methylcellulose from BM cells from vehicle (black)- or warfarin (white)-treated mice (P=0.02, t-test). E-F) Total number of leukocytes (P=0.029) 7 days after a one-time treatment with 200 mg/kg 5-fluorouracil (E) and Kaplan-Meier-style survival curve (F) of control mice (solid line) or mice treated with 0.5 mg/kg (dashed line) (P=0.0004; Log-rank test, n=8) or 0.05 mg/kg warfarin (dotted line) (P=0.016; Log-rank test, n=5) cotreated with 75 mg/kg 5-fluorouracil on days 0, 7, 14, 21. G-I) Competitive (P values as indicated; t-test, n=5) (G), serial (P values as indicated; t-test, n=5) (H) and limiting dilution (P=0.0006) (I) transplantation assays using CD45.1+ total donor bone marrow from control (solid line) mice or warfarin (dashed line)-treated mice. The limiting dilution assay (I) shows the percentage of non-engrafted mice (y-axis) in relation to the dose of transplanted bone marrow. The lines fitted by regression analysis (L-Calc) allow estimation of functional HSC frequency, which was 7.6 fold lower in warfarin-treated donors. (J) Number of colonies arising from Lin− bone marrow cells pretreated with vehicle or 2pLY1 warfarin for 7 days plated in methylcellulose in the absence of drugs. Colonies were scored on day 10 after plating.

[0083] FIG. 2: Warfarin impairs bone marrow stroma. A-B) Number of Nestin+ MSC (P=0.02; t-test, n=5) (A) or Co12.3 kb GFP+ osteoblastic cells (n.s.; t-test, n=4) (B) in crushed, collagenase-treated bones of Nestin-GFP− (A) or Co12.3 kb GFP+(B) control mice (black) or mice treated with warfarin (white) for 14 days. C) Trichrome stain of the distal femora of a representative control versus a warfarin-treated mouse. Note the dramatic reduction in cancellous bone in warfarin-treated animal compared to vehicle control. D-F) Ratios of bone volume/tissue volume (P<0.001; t-test, n=4-5) (D), osteoid volume/bone volume (P=0.046; t-test, n=4-5) (E) and osteoblast surface/bone surface (P=0.0008, t-test, n=4-5) (F) in control (black) versus warfarin-treated (white). G) Differentiation of primary murine stroma cells into adipocytes (top) or osteoblastic cells (bottom) in the presence of vehicle or warfarin or in the absence of differentiation factors (control). Oil Red 0 stains adipocytes red, while von Kossa stains osteoblastic cells brown. The scale bar depicts 100 μm. H) Number of cobblestones formed by Lin− cells under stroma which had been treated with vehicle (black) or warfarin (white) for 10 days (P=0.033; t-test).

[0084] FIG. 3: Warfarin reversibly impairs the HSC-supportive function of bone marrow stroma via a pAKT-dependent mechanism in HSC. A) Number of Lin− pAKT+ cells in the bone marrow of control mice or mice treated with warfarin 14 days after initiation of treatment (P=0.037, t-test, n=4).). B) Total number of hematopoietic cells two days after plating of Lincells from vehicle- or warfarin-treated mice on vehicle- or warfarin-treated stroma (P=0.009, t-test, n=3). C) Percentage of Lin− cells in the S/G2/M phase of the cell cycle 7 days after plating of Lin− cells from vehicle- or warfarin-treated mice on vehicle (gray)- or warfarin (white)-treated stroma (P=0.02, t-test, n=3). D) Percentage of Lin− pAKT+ cells 3 days after plating of Lin− cells from vehicle- or warfarin-treated mice on vehicle (gray)- or warfarin (white)-treated stroma (P=0.02, t-test, n=3). E) Total number of hematopoietic cells four days after plating of Lin− cells from vehicle- or warfarin-treated mice on vehicle- or warfarin-treated stroma in the presence of the AKT inhibitor MK-2206 (5 μM) (P=0.02, t-test, n=5). F-G) Total number of hematopoietic cells (P=0.0009; t-test, n=3) (F) and percentage of myeloid (P=0.001) and Lin− cells (P=0.0001) (G) grown in conditioned medium harvested from stroma cells grown in vehicle (black)- or warfarin (white)-containing medium 3 days after plating of Lin− cells. H-I) Percentage of total donor Actin DsRed+ leukocytes (P=0.025; t-test, n=10) (H) and percentage of donor CD11b+ Actin DsRed+ myeloid cells (P=0.023; t-test, n=8) (I) in recipients, which had been intrafemorally cotransplanted with untreated Actin DsRed+ Lin− and vehicle (black)- or warfarin (white)pretreated stroma cells 6-8 weeks previously. The stroma had, originally, been obtained from warfarin-treated mice and expanded in the continuous presence of warfarin in vitro.

[0085] FIG. 4: Warfarin impairs hematopoiesis via periostin. A) Co-immunoprecipitation of concentrated conditioned medium (CM) from vehicle- or warfarin-treated stroma cells with an anti-Gla antibody. The Western blot was performed with an antibody to periostin (90 kDa). B) Total number of hematopoietic cells in an in vitro co-culture of vehicle (black)- or warfarin (white)-treated stroma cells and hematopoietic cells from murine bone marrow, to which 2 μg/ml periostin (lighter gray) had been added. The increase of the number of hematopoietic cells (P=0.02, t-test, n=6) 5 days after addition of periostin is significant. C) Ratio of CD45.1/CD45.2 four weeks after competitive transplantation of CD45.1+ total donor bone marrow from vehicle (black)- or periostin (white)-treated mice, mixed with CD45.2+ competitor cells and transplanted into CD45.2+ recipient mice (P=0.02; t-test, n=10). D) Abundance of annexin+ DAPI+ LKS cells in mice treated with vehicle (black) or warfarin (white) and periostin (lighter gray) relative to vehicle-treated mice (P=0.02; t-test, n=10). E) Percentage of CD11b+ myeloid cells in the peripheral blood of mice treated with control (black)- or warfarin (white) plus BSA (dark gray) or periostin (light gray). The increase of the percentage of myeloid cells (P=0.007, t-test, n=5) after a 4 day treatment with periostin is significant.

[0086] FIG. 5: Warfarin impairs hematopoiesis via the periostin/integrin β3 axis. A) Number of LKS cells expressing integrin β3 in mice treated with vehicle (black) or warfarin (white) for 4 weeks and periostin (lighter gray) (P=0.0013; t-test, n=10). B) Protein lysates of 293T cells, transfected with an integrin β3-overexpressing construct and grown in conditioned medium from vehicle- or warfarin-treated stroma cells, were subjected to co-immunoprecipitation with an anti-integrin β3 antibody, and coimmunoprecipitation of periostin was assessed by immunoblotting, as indicated. Anti-IgG-isotype antibody was used as control in the co-immunoprecipitation. The input (left) shows the presence of integrin β3 (92 kDa) and periostin (90 kDa), and GAPDH (38 kDa) was used to control for equal loading. C) Adhesion of 100,000 Lin− Actin− DsRed+ cells to vehicle- or warfarin-treated stroma cells in the presence or absence of periostin (P=0.03; t-test).

[0087] FIG. 6: Vitamin K antagonism leads to reduction of human leukocytes and engraftment of human HSC. A) Percentage of human CD45+ leukocytes in the peripheral blood of vehicle (black)- or warfarin (white)-treated NSG mice 6 weeks after transplantation with 1.3×10.sup.5 human CD34+ HSC (P=0.015, t-test, n=9). B) Percentage of human CD45+ leukocytes in the peripheral blood of NSG mice 4 weeks after intrafemoral co-transplantation of 0.8×10.sup.5 human CD34+ HSC and 5×10.sup.5 vehicle (black)- or warfarin (white)-treated human autologous stroma cells (P=0.046, t-test, n=6). C-D) Total number of leukocytes (P=0.0007, t-test) (C) and monocytes (P=0.0042, t-test) (D) in the peripheral blood of control patients (black; n=145) and patients on warfarin (white; n=282). E-F) Percentage of basophils (P=0.007, t-test) (E) and eosinophils (P=0.0027, t-test) (F) of all leukocytes in the peripheral blood of control patients (black; n=145) and patients on warfarin (white; n=282). G-H) Total number of leukocytes (P=0.029, t-test) (G) and monocytes (P=0.046, t-test) (H) in the peripheral blood of control patients (black; n=145) and patients on fluindione (white; n=89), another vitamin K-antagonist similar to warfarin.

[0088] FIG. 7: A-B) Use of vitamin K antagonists (VKA) in men and women aged 70-79 years (n=5,464,258) (A) or 50-60 years (n=11,452,848) (B) with and without a diagnosis of myelodysplastic syndrome (MDS) in 2015. Data source: French health-care databases.

[0089] FIG. 8: A-B) Total number of hematopoietic cells in an in vitro co-culture of (A) K562 (P=0.05; t-test, n=3) and (B) BCR-ABl1+ Baf3 (P=0.07; t-test, n=3) leukemic cells on wildtype (black)-versus periostin deficient (white) bone marrow stroma cells. (C) Percentage of Mac-1 (CD11b)+ BCR-ABL1+ leukemic cells in the peripheral blood of wildtype (black) or periostin-deficient (white) recipient mice intravenously transplanted with transduced bone marrow (P=0.02; t-test, n=10).

[0090] And in the sequences:

TABLE-US-00001 Human Periostin Isoform 1, also known as OSF-2OS (Uniprot Identifier: Q15063-1) SEQ ID NO: 1   10   20   30   40   50 MIPFLPMFSL LLLLIVNPIN ANNNYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYT TKIITKVVEP KIKVIEGSLQ PIIKTEGPTL        710        720        730        740        750 TKVKIEGEPE FRLIKEGETI TEVIHGEPII KKYTKIIDGV PVEITEKETR        760        770        780        790        800 EERIITGPEI KYTRISTGGG ETEETLKKLL QEEVTKVTKF IEGGDGHLFE        810        820        830 DEEIKRLLQG DTPVRKLQAN KKVQGSRERL REGRSQ Human Periostin Isoform 2, also known as OSF-2p1 (Uniprot Identifier: Q15063-2) SEQ ID NO: 2 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRES TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYK PIIKKYTKII DGVPVEITEK ETREERIITG        710        720        730        740        750 PEIKYTRIST GGGETEETLK KLLQEEVTKV TKFIEGGDGH LFEDEEIKRL        760        770 LQGDTPVRKL QANKKVQGSR RRLREGRSQ Human Periostin Isoform 3 (Uniprot Identifier: Q15063-3) SEQ ID NO: 3 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYR PTLTKVKIEG EPEFRLIKEG ETITEVIHGE        710        720        730        740        750 PIIKKYTKII DGVPVEITEK ETREERIITG PEIKYTRIST GGGETEETLK        760        770        780 KLLQEDTPVR KLQANKKVQG SRRRLREGRS  Q Human Periostin Isoform 4 (Uniprot Identifier: Q15063-4) SEQ ID NO: 4 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGTVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYK PIIKKYTKII DGVPVEITEK ETREERIITG        710        720        730        740        750 PEIKYTPIST GGGETEETLK KLLQEDTPVR KLQANKYMQG SRRRLREGRS Q Human Periostin Isoform 5 (Uniprot Identifier: Q15063-5) SEQ ID NO: 5 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYR PTLTKVKIEG EPEFRLIKEG ETITEVIHGE        710        720        730        740        750 PIIKKYTKII DGVPVEITEK ETREERIITG PEIKYTRIST GGGETEETLK        760        770        780        790        800 KLLQEEVTKV TKFIEGGDGH LFEDEEIKRL LQGDTPVFKL QANKKVQGSR RRLREGRSQ Human Periostin Isoform 6 (Uniprot Identifier: Q15063-6) SEQ ID NO: 6 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVKGST FKEIPVTVYS PEIKYTRIST GGGETEETLK KLLQEEVTKV        710        720        730        740 TKFIEGGDGH LFEDEEIKRL LQGDTPVRKL QANKKVQGSR RRLREGRSQ Human Periostin Isoform 7 (Uniprot Identifier: Q15063-7) SEQ ID NO: 7 10 20 30 40 50 MIPFLPMFSL LLLLIVNPIN ANNHYDKILA HSRIRGRDQG PNVCALQQIL         60         70         80         90        100 GTKKKYFSTC KNWYKKSICG QKTTVLYECC PGYMRMEGMK GCPAVLPIDH        110        120        130        140        150 VYGTLGIVGA TTTQRYSDAS KLREEIEGKG SFTYFAPSNE AWDNLDSDIR        160        170        180        190        200 RGLESNVNVE LLNALHSHMI NKRMLTKDLK NGMIIPSMYN NLGLFINHYP        210        220        230        240        250 NGVVTVNCAR IIHGNQIATN GVVHVIDRVL TQIGTSIQDF IEAEDDLSSF        260        270        280        290        300 RAAAITSDIL EALGRDGHFT LFAPTNEAFE KLPRGVLERI MGDKVASEAL        310        320        330        340        350 MKYHILNTLQ CSESIMGGAV FETLEGNTIE IGCDGDSITV NGIKMVNKKD        360        370        380        390        400 IVTNNGVIHL IDQVLIPDSA KQVIELAGKQ QTTFTDLVAQ LGLASALRPD        410        420        430        440        450 GEYTLLAPVN NAFSDDTLSM DQRLLKLILQ NHILKVKVGL NELYNGQILE        460        470        480        490        500 TIGGKQLRVF VYRTAVCIEN SCMEKGSKQG RNGAIHIFRE IIKPAEKSLH        510        520        530        540        550 EKLKQDKRFS TFLSLLEAAD LKELLTQPGD WTLFVPTNDA FKGMTSEEKE        560        570        580        590        600 ILIRDKNALQ NIILYHLTPG VFIGKGFEPG VTNILKTTQG SKIFLKEVND        610        620        630        640        650 TLLVNELKSK ESDIMTTNGV IHVVDKLLYP ADTPVGNDQL LEILNKLIKY        660        670        680        690        700 IQIKFVRGST FKEIPVTVYS PEIKYTRIST GGGETEETLK KLLQEDTPVR        710        720 KLQANKKVQG SRRRLREGRS Q Primer sequences SEQ ID NO: 8, 9 indicates data missing or illegible when filed

EXAMPLES

Example 1: Warfarin Compromises Hematopoiesis and HSC Function

[0091] In order to test whether warfarin may compromise hematopoiesis via an inhibition of coagulation factors II, VII, IX and X, the inventors treated wildtype mice with warfarin. This led to a modest, but significant increase of the international normalized ratio (INR), a derived measure of the prothrombin time, only at a dose of 0.5 mg/kg (P=0.009) compared to sham-operated control mice. However, when enumerating leukocytes in peripheral blood 14 days after the initiation of treatment, the inventors revealed that the absolute number of leukocytes (P=0.008; FIG. 1A) and monocytes (P=0.007; FIG. 1B) was reduced at the higher dose of warfarin (which prolongs the INR in mice), as well as at the lower dose of warfarin (FIGS. 1A and 1B) which is equivalent to the dose used in humans (approximately 0.05-0.07 mg/kg to achieve an INR between 2-3). The decrease in leukocytes and monocytes appeared as early as three days after initiation of treatment. The percentage of CD11b+ myeloid cells (P=0.003; FIG. 1C) was also decreased. Oral versus subcutaneous administration of warfarin similarly reduced leukocyte and monocyte counts, but subcutaneous treatment was associated with fewer lethal hemorrhagic complications. The lymphocyte count was also decreased, albeit to a lesser extent than myeloid cells. The absolute count and percentage of Lin− c-Kit+ Sca-1+(LKS), LKS CD150+CD48-(SLAM) at 2 and 4 weeks and the absolute count and percentage of myeloid progenitor cells in the bone marrow, however, did not significantly differ between controls and warfarin-treated mice. A colony-forming assay revealed that the colony-forming ability of HSPC from warfarin-treated compared to control mice was significantly impaired (P=0.02; FIG. 1D). In response to a one-time dose of 5-fluorouracil-induced ‘stress’ leukocyte counts were lower in warfarin-treated compared to control mice 7 days after treatment (P=0.03; FIG. 1E) and survival was similarly shortened in mice treated with warfarin at the higher (P=0.0004, long dashes, Log-rank test; FIG. 1F) or the lower dose (P=0.02, dotted line, Log-rank test; FIG. 1F) compared to control mice. Transplantation-based assays to test HSC function in control-versus warfarin-treated mice, revealed impaired function of whole bone marrow cells from a warfarin-treated compared to a control BMM in competitive (P=0.04 after 16 weeks; FIG. 1G), serial (P=0.004 after tertiary transplantation; FIG. 1H) and limiting dilution assays. In fact, the limiting dilution assay revealed a 7.6 fold reduction of functional HSC in warfarin-treated compared to control mice (P=0.0006; FIG. 1I). Colony formation assays of Lin− cells, previously exposed to vehicle or warfarin in vitro for 7 days, in order to rule out a direct toxic effect of warfarin on hematopoietic cells, did not yield significantly different numbers of colonies (FIG. 1J), and culturing of Lin− hematopoietic cells in warfarin-containing media had no direct effect on phenotypical Lin− or LKS cells, the percentage of annexin V+ dead cells, their cycling or myeloid cells. The possibility that warfarin may alter the homing of HSPC in a transplant setting could be excluded, as homing of Lin− and LKS cells to a warfarin-treated BMM was not compromised. In addition, as confirmation of the specificity of the effects of deficient γ carboxylation, administration of vitamin K1 to previously warfarin-treated mice ‘rescued’ the decreased percentage of myeloid cells (P=0.04). This suggested that the observed detrimental effect of warfarin on hematopoiesis was indeed due to deficient γ carboxylation of vitamin K-dependent factors. Collectively, these results indicate, that warfarin reduces myeloid cells in peripheral blood and impairs the ability of HSPC to adequately respond to hematopoietic ‘stress’. This detrimental effect on hematopoiesis seemed to occur indirectly, possibly via detrimental effects on bone marrow stroma.

Example 2: Warfarin Impairs Bone Marrow Stroma

[0092] Patients on longterm treatment with warfarin experience bone loss (Rezaieyazdi et al., 2009) and in men taking warfarin the risk of fracture is increased (Gage et al., 2006). In conjunction with our data on the detrimental effects of warfarin on hematopoiesis and previous publications on the support of HSPC by mesenchymal stem cells (MSC) (Mendez-Ferrer et al., 2010) the inventors enumerated Nestin-GFP+ cells, which label MSC, and osteoblastic cells in Nestin-GFP or Co12.3 kb GFP mice, in which the expression of green fluorescent protein (GFP) is driven by the Nestin (Mendez-Ferrer et al., 2010) or Co12.3 kb promoter (Kalajzic et al., 2002), respectively. This revealed a significant reduction of Nestin+ MSC (P=0.047; FIG. 2A), but no reduction of Co12.3 kb+ osteoblastic cells (FIG. 2B) in mice.

[0093] Histomorphometric analysis of distal femora of warfarin-treated mice revealed a significant reduction in cancellous bone mass (FIG. 2C), as shown by a decreased ratio of bone volume/tissue volume (BV/TV P<0.001; FIG. 2D), reduced trabecular number and thickness and increased trabecular separation (FIG. 2C) in warfarin-treated mice compared to control mice. Osteoid volume (OV/BV, P=0.046; FIG. 2E), osteoblast surface (Ob.S/BS, P=0.0008; FIG. 2F) and osteoclast surface (Oc.S/BS P=0.01) were significantly increased upon treatment with warfarin, as well as the number of osteoblasts (N.Ob/B.Pm, P=0.0001). Osteoclasts were not directly affected by warfarin treatment, as shown by the absence of changes in osteoclast differentiation or number (N.Oc/B.Pm). Taken together these data suggest that warfarin treatment reduces bone formation and osteoblast function (as shown by increased osteoid volume) and increases bone resorption, as shown by the significant increase in osteoclast surface per bone surface.

[0094] Hypothesizing that warfarin impairs MSC differentiation, the inventors demonstrated that the differentiation of murine stroma cells to adipocytes and osteoblasts was significantly compromised in presence of warfarin (FIG. 2G). Consistently, the relative expression of osteoblastic genes such as Runt-related transcription factor 2 (Runx2), alkaline phosphatase (Alpl), Col1 (Col1a1), osterix (Osx), osteopontin (Spp1), osteocalcin (Bglap) and others involved in osteoblastic differentiation and function were significantly reduced in primary calvaria cells treated with warfarin compared to vehicle further supporting the hypothesis that warfarin impairs osteoblast function, as suggested by our histomorphometry studies. There was a modest, but significant increase of apoptosis of stroma cells in vitro (P=0.05) after treatment with warfarin.

[0095] In order to test the support of HSPC by bone marrow stroma, the inventors performed a cobblestone-formation assay, in which untreated Lin− hematopoietic cells were plated on previously vehicle- or warfarin-treated stroma cells. This revealed a significantly decreased number of cobblestone-forming areas, when the stroma had previously been treated with warfarin (P=0.03; FIG. 2H). Further, coculture of Lin− hematopoietic cells with stroma cells in the presence of warfarin revealed a modest increase of annexin V+ dead hematopoietic cells compared to vehicle treatment (P=0.04). Consistently, in the bone marrow of mice treated with warfarin the inventors observed a modest increase of annexin V+ DAPI+ dead Lin− cells after two weeks (P=0.038) and of annexin V+ DAPI+ dead LKS cells after 6 weeks (P=0.03).

[0096] In summary, these data suggest that warfarin reduces the number of MSC, as well as their differentiation capacity to adipocytes and osteoblastic cells. Warfarin impairs osteoblastic function, as well as the HSC-supportive ability of bone marrow stroma cells, leading to compromised function of HSPC.

Example 3: Rebounding of Warfarin-Exposed HSPC in an Untreated Microenvironment

[0097] Although a mere reduction of HSC number in warfarin-treated mice did not seem to account for impaired hematopoietic reconstitution upon transplantation (FIG. 1G-I), the inventors assessed whether transplantation of equal numbers of CD45.1+ LKS cells from control or warfarin-treated mice would recapitulate the impaired reconstitution of hematopoietic cells after transplantation of whole bone marrow. Transplantation of equal numbers of CD45.1+ LKS cells from control or warfarin-treated mice into CD45.2+ recipients led to a higher percentage of CD45.1+ donor cells in the peripheral blood of recipients one month (P=0.03), but not two months (P=0.65) after transplantation, if the transplanted bone marrow was derived from warfarin-treated donors. This phenomenon may possibly be due to exhaustion of HSC, which may have entered the cell cycle upon transplantation into an untreated BMM. A higher number of transplanted CD45.1+ LKS cells (and a lower number of competitor cells) achieved higher donor chimerism, but the difference in engraftment between control versus warfarin-treated LKS disappeared and at 16 weeks post transplantation reconstitution by warfarin-treated LKS cells was lower than in the case of control LKS cells, suggesting that functional differences may become more apparent, the greater the ‘stress’ of transplantation. Further, hypothesizing that an altered cell cycle may account for HSC impairment in a warfarin-versus an untreated BMM, the inventors showed that the percentage of LKS cells in the Go phase of the cell cycle was significantly increased in mice treated with warfarin compared to vehicle (P=0.005). Consistent with a role of pAKT in cell cycling (Liu et al., 2014) the inventors observed a significant reduction of pAKT+ Lin− cells in the bone marrow of warfarin-treated compared to control mice (P=0.04, FIG. 3A). However, transplantation of Lin− cells from a warfarin-treated BMM into an untreated BMM led to an increase of the percentage of Lin− cells in the S/G2/M phases of the cell cycle (P=0.04). Testing the possibility of ‘recovery’ of warfarin-treated HSPC upon exposure to ‘fresh’ or untreated stroma as shown in vivo the inventors found an increased number of hematopoietic cells (P=0.009; FIG. 3B) and an increase of the percentage of Lin− cells in the S/G2/M phases of the cell cycle (P=0.02; FIG. 3C), when warfarin-treated Lin− cells were plated on untreated compared to warfarin-treated stroma. Consistently, the number of pAKT+ Lin− cells increased upon plating of previously warfarin-exposed Lin− cells on untreated stroma (P=0.02; FIG. 3D). After adding the AKT-inhibitor MK-2206 to the cocultures, the inventors could confirm that the temporary ‘recovery’ of warfarin-exposed Lincells after plating on untreated stroma was AKT-dependent (P=0.02; FIG. 3E), as treatment with MK-2206 prevented the temporary ‘recovery’ of hematopoietic cells.

[0098] Taken together, this suggested that the detrimental effect of warfarin on HSPC may be due to an induction of a quiescent state via a reduction of pAKT. This may be followed by an increase in cycling status upon replating in vitro or transplantation into an untreated environment via an increase of pAKT, but this effect decreases at later time points, likely due to HSC exhaustion.

Example 4: Warfarin Impairs the HSC-Supportive Function of Bone Marrow Stroma

[0099] Hypothesizing that HSC-supportive cytokines may be reduced in a warfarin-treated BMM, the inventors cultured untreated Lin− bone marrow cells in conditioned medium which had been harvested from stroma cells grown in vehicle- or warfarin-containing medium. This revealed a decrease of the total number of hematopoietic cells (P=0.0009; FIG. 3F), as well as a decrease of the myeloid (P=0.001; FIG. 3G) and Lin− cell (P=0.0001; FIG. 3G) fractions in those cultures grown in conditioned medium from warfarin-treated stroma cells. In order to test, which proteins known to support HSC function may be decreased in warfarin—compared to vehicle-treated stroma cells, the inventors performed a focused gene expression analysis. This revealed significant decreases of osteopontin, runt-related transcription factor 2 (Runx 2), as well as chemokine (C-C motif) ligand 3 (Ccl3), angiopoietin (Angpt1), insulin-like growth factor 1 (Igf1), Notch1 and Wnt7b, while stem cell factor (SCF; Kitl) was unchanged and there was a trend towards increased expression of C-X-C motif chemokine 12 (CxCl12). Indeed, granulocyte-colony-stimulating factor (G-CSF)-mediated mobilization of LKS and LKS SLAM cells into peripheral blood was impaired in warfarin-treated mice (P=0.042 and P=0.026, respectively).

[0100] In order to test the support of HSPC by stroma directly and in vivo, the inventors intrafemorally cotransplanted vehicle- or warfarin-treated stroma cells with untreated Lin− bone marrow cells. This led to decreased engraftment of total donor bone marrow (P=0.025; FIG. 3H) and donor-derived myeloid cells (P=0.023; FIG. 3I) with persistence and modest expansion of the injected stroma cells at low levels in mice intrafemorally cotransplanted with warfarin—compared to vehicle-treated stroma 14 days after transplantation. Analysis of the identity of the intrafemorally transplanted stroma cells revealed that 60% of these were F4/80+ macrophages at the time of transplantation. Similarly, 14 days after transplantation 80% of the engrafted stroma cells were constituted by F4/80+ macrophages, regardless of whether they had previously been treated with vehicle or warfarin. Macrophages in the bone marrow of primary mice treated with warfarin were not reduced compared to controls. Coculture of untreated Lin− cells on warfarin-treated, sorted F4/80+ macrophages increased the percentage of Annexin V+ apoptotic leukocytes compared to coculture on vehicle-treated macrophages (P=0.034), and protein analysis of sorted F4/80+ macrophages revealed that macrophages are a major source of the vitamin K-dependent protein periostin. However, coimmunoprecipitation of periostin in conditioned medium from vehicle-versus warfarin-treated macrophages and subsequent protein analysis revealed no significant reduction of total periostin levels. Taken together, these in vitro and in vivo data suggest that warfarin-treated stroma is detrimental for hematopoiesis via reduction of functional HSC-supportive proteins leading to a reduction of the reconstitution potential of HSC. This effect may be at least partially mediated by periostin produced by F4/80+ macrophages.

Example 5: Warfarin Impairs Hematopoiesis Via Periostin

[0101] Several proteins produced by stromal cells in the BMM, such as osteocalcin, protein Z, Gas 6, matrix gla protein (MGP) and periostin, which is known to be expressed in MSC (Khurana et al., 2016) and osteoblastic cells (Horiuchi et al., 1999; Khurana et al., 2016), require vitamin K for their γ carboxylation (Coutu et al., 2008) and function. Specifically, as a form of posttranslational modification the vitamin-K-dependent enzyme γ glutamylcarboxylase modifies glutamic residues to γ-carboxyglutamic acid (Gla) (Coutu et al., 2008). Therefore, the inventors hypothesized that vitamin K-dependent factors in the BMM and particularly periostin may be responsible for the observed effects on HSPC. Periostin is a secreted extracellular matrix protein, which is expressed by MSC (Coutu et al., 2008; Khurana et al., 2016), osteocytes, periosteal osteoblasts (Bonnet et al., 2012) and bone marrow macrophages, and whose function depends on γ carboxylation. In fact, periostin was shown to be the most abundant Gla-containing protein secreted by MSC (Coutu et al., 2008) and to regulate HSC function (Khurana et al., 2016), while periostin-deficient mice are characterized by anemia, myelomonocytosis and lymphopenia (Khurana et al., 2016). Treatment of periostin knockout mice with warfarin did not lead to a decrease of total leukocytes or monocytes compared to control periostin knockout mice—contrary to wildtype mice treated with warfarin (FIGS. 1A and 1B). In addition, repetitive doses of 75 mg/kg 5-fluorouracil to warfarin-treated periostin knockout mice did not lead to lower leukocyte counts or decreased survival compared to control periostin knockout mice, as previously observed in wildtype mice (FIGS. 1E and 1F). This suggested that periostin and, possibly, the 7-carboxylation of periostin plays a major role in the decrease of myeloid cells after warfarin treatment. Deficiency of other γ-carboxylated proteins in the BMM such as matrix gla protein (El-Maadawy et al., 2003) or osteocalcin (Ducy et al., 1996) to date has not been associated with a hematological phenotype.

[0102] In order to test whether the 7-carboxylation of periostin was decreased by warfarin, the inventors performed a coimmunoprecipitation experiment using an anti-gla antibody on conditioned medium from stroma cells treated with vehicle or warfarin (FIG. 3F). Indeed, no γ-carboxylated periostin could be detected in the conditioned medium of warfarin-treated stroma cells (FIG. 4A). Hypothesizing that the lack of (γ-carboxylated) periostin in warfarin-treated mice or in cocultures was the cause of the impairment of hematopoiesis, the inventors added recombinant periostin to cocultures of stroma cells treated with vehicle or warfarin and Lin− bone marrow cells. This restored or ‘rescued’ the reduced absolute number of hematopoietic cells (P=0.02; FIG. 4B) and led to a strong trend towards restoration of the percentage of LKS cells in warfarin-treated cultures (P=0.056). Additionally, competitive transplantation of total bone marrow from a CD45.1+ donor mouse which had been treated with vehicle or periostin for four days into CD45.2+ recipients led to increased engraftment of CD45.1+ periostin-treated bone marrow four weeks after transplantation (P=0.02; FIG. 4C) but not at later time points (data not shown), suggesting that periostin may augment HSC function short-term. However, treatment of donor mice with vitamin K prior to competitive transplantation did not lead to increased engraftment of donor cells. Competitive transplantation of previously periostin- or vitamin K-treated bone marrow cells also did not lead to an increase in the percentage of LKS cells, an increase in the cell cycle of transplanted LKS cells or an increase in the number of colonies in methylcellulose compared to vehicle. Treatment of mice with periostin and warfarin (or vehicle), however, significantly reduced annexin V+ apoptotic LKS cells compared to mice treated with warfarin (or vehicle) alone (P=0.02; FIG. 4D). Further, in vivo administration of periostin to mice treated with warfarin ‘rescued’ the percentage of myeloid cells (P=0.007; FIG. 4E).

Example 6: Warfarin Impairs Hematopoiesis Via Periostin/Integrin β3 Signaling Axis

[0103] In vivo administration of periostin to mice treated with warfarin also ‘rescued’ the number of LKS cells positive for integrin β3, whose binding to its ligand periostin (Gillan et al., 2002) is known to mediate HSC support (P=0.0013; FIG. 5A) (Khurana and Verfaillie, 2013). The increased percentage of integrin β3+ LKS cells was not due to increased expression of integrin β3+ on LKS cells, as the median fluorescence intensity remained unchanged. An altered physical interaction between periostin and integrin β3 was evident in a coimmunoprecipitation experiment, in which periostin in the conditioned medium of stroma cells exposed to vehicle, but not warfarin (FIG. 4A), efficiently bound integrin β3 overexpressed on 293T cells (FIG. 5B). Quantification showed this to be significant (P=0.02). In order to test, whether the periostin/integrin β3-axis mediates the adhesion of HSPC to stroma, the inventors performed an adhesion assay plating untreated Lin− cells on vehicle- or warfarin-treated stroma cells in the presence or absence of periostin. Indeed, addition of periostin to the cocultures restored or augmented the adhesion of Lin− cells to stroma in warfarin-, but more prominently in vehicle-treated conditions (P=0.03; FIG. 5C). Due to the role of pAKT, which lies downstream of integrin β3 (Hynes, 2002), in periostin-mediated cell survival (Bao et al., 2004) the inventors tested the effect of inhibition of integrin β3 by cilengitide on pAKT in Lin− cells after coculture on BM stroma cells. Similar to the reduction of pAKT+ Lin− cells in mice treated with warfarin (FIG. 3A), inhibition of integrin β3 led to a decrease of pAKT in Lin− cells (P=0.03).

[0104] Overall, these data suggest that impairment of HSPC and myeloid cells after warfarin treatment is at least partially mediated by decreased binding of carboxylated periostin to integrin β3 on HSPC leading to decreased support of HSC, likely decreased pAKT signaling and decreased self-renewal of HSC.

Example 7: Vitamin K Antagonism Leads to Reduction of Human Leukocytes and Engraftment of Human HSC

[0105] Vitamin K antagonists (VKA) are drugs widely used in several conditions in an effort to reduce thromboembolic complications. A possible limitation of its use due to detrimental hematopoietic effects would have extensive consequences. Therefore, in order to test a potential effect of warfarin on human hematopoietic cells, the inventors transplanted untreated human CD34+ cells into vehicle- or warfarin-treated NOD SCID interleukin-2 receptor γ knockout (NSG) mice and demonstrated reduced engraftment of human CD45+ leukocytes in warfarin—compared to vehicle-treated NSG mice (P=0.015; FIG. 6A). Further, intrafemoral cotransplantation of untreated human CD34+ cells with (in vitro) vehicle- or warfarin-treated human stroma cells demonstrated decreased engraftment of human CD45+ leukocytes, when the hematopoietic cells were cotransplanted with warfarin-treated stroma (P=0.046; FIG. 6B), similar to FIGS. 3H and 31, where this effect seems to have been due to decreased 7-carboxylation of periostin by warfarin-treated macrophages. 6 months after transplantation the transplanted human CD45+ cells were detectable at low levels in the injected and the contralateral femora, but only trends towards lower huCD45+ engraftment in recipients of warfarin-treated stroma were observed. Consistently, although still in the normal reference range, the absolute number of leukocytes (P=0.0007; FIG. 6C) and monocytes (P=0.004; FIG. 6D) and the percentage of basophils (P=0.007; FIG. 6E) and eosinophils (P=0.0027; FIG. 6F) in 282 patients on warfarin, but without hematologic disease, were significantly decreased compared to 145 control patients. Similarly, the absolute number of leukocytes (P=0.029; FIG. 6G) and monocytes (P=0.046; FIG. 6H) was also reduced in 89 patients on fluindione, another VKA, compared to controls. Platelets were not reduced in warfarin- or fluindione-treated patients.

Example 8: VKA Use is More Frequent in People with Versus without a Diagnosis of MDS

[0106] We hypothesized that the use of VKA may be associated with an increased risk of myelodysplastic syndrome (MDS), a clonal hematopoietic stem cell disorder, which—in mice—may also arise due to impairment of the BMM (Raajimakers et al., 2010; Zambetti et al., 2016).

[0107] In a population of men and women aged between 70 and 79 years in 2015 (n=5,464,258) VKA use was more frequent in people with versus without a diagnosis of MDS, both overall (14.66% versus 5.76%, P<0.001; FIG. 7A) and in the various age and sex strata (P<0.001; FIG. 7A). The odds of VKA use was significantly higher in patients with versus without a diagnosis of MDS (adjusted odds ratio: 2.49; 95% confidence interval: 2.31 2.68). As somatic mutations associated with clonal hematopoiesis and the risk of MDS increase with age (Jaiswal et al., 2014) and VKA use may trigger the development of these mutations or MDS risk, the inventors investigated VKA use in patients aged 50-60 (n=11,452, 848). This revealed that VKA use was also more frequent in 50-60 year old patients with versus without a diagnosis of MDS (7.1% versus 0.85%, P<0.001; FIG. 7B). The odds of VKA use in this age group was significantly higher in patients with versus without a diagnosis of MDS and higher than in patients aged 70-79 years (adjusted odds ratio: 8.2; 95% confidence interval: 6.64-10.13). Performance of a myeloid sequencing panel in 5 control and 4 patients on VKA, however, did not reveal any mutations associated with clonal hematopoiesis.

[0108] In summary, these data confirm the inventor's data in mice and suggest that treatment with two different vitamin K antagonists reduces the frequency of certain human leukocyte populations, albeit within the normal reference range, and impairs human HSC. Further, the odds of VKA use is increased in patients with MDS.

Example 9: Periostin in Haematological Malignancies

[0109] In order to test if periostin may play a role in haematological malignancies, the inventors co-cultured K562 or Baf3 p210 cells, both of which are positive for the oncogene BCR-ABL1, which is found in chronic myeloid leukaemia (CML) and B-cell acute lymphoblastic leukaemia (B-ALL), on wildtype versus periostin-deficient bone marrow stromal cells. This revealed increased proliferation of the leukemic cells when co-cultured on periostin-deficient compared to wildtype bone marrow stromal cells (FIGS. 8A-B). To further check if periostin deficiency has an effect on the progression of leukaemia in vivo, the inventors induced CML in wildtype versus periostin deficient mice via the transplantation of wildtype bone marrow cells transduced with retrovirus expressing the BCR-ABL1 oncogene. It was observed that periostin deficiency in recipient mice accelerated disease progression, as tested by the increased percentage of leukemic cells (GFP+(BCR-ABL1+) CD11b+) in peripheral blood (FIG. 8C). Conversely, this suggested that administration of periostin may decrease the progression of myeloid leukaemias.

Methods

Mice

[0110] 7 to 10 week old C57BL6/N (CD45.2) or SJL (CD45.1; B6.SJL-Ptprca Pepcb/BoyJ), Collar 2.3 kb-GFP (kind gift from D. Rowe), Nestin-GFP (Mignone et al., 2004) and periostin knockout mice (kind gift from Juerg Huelsken) were used for these experiments. NOD SCID interleukin (IL)-2 receptor γ deficient (NSG) mice were purchased from Charles River Laboratories (Sulzfeld, Germany) and bred in our facility. All animal studies were approved by the local German government (Regierungspräsidium Darmstadt) in Hessen, Germany, and by the Institutional Animal Care and Use Committee (IACUC) of Boston University.

[0111] In vivo and in vitro drug treatment 0.5 mg/kg/d or 0.05 mg/kg/d of warfarin, resuspended in phosphate-buffered saline was administered to mice via subcutaneously implanted osmotic minipumps (ALZET minipumps, Cupertino, Calif.). The osmotic minipumps were changed after 14 days (if the experiment went beyond 14 days of treatment). Sham operated (skin incision and suture) mice served as controls. Warfarin treatment continued for 10-14 days. In the xenotransplantation experiments the mice were treated with warfarin via the drinking water (or normal drinking water) at a dose of 0.72 mg/100 ml for a total of 3 days a week (Pfeilschifter et al., 2011).

[0112] In the mobilization experiments vehicle- or warfarin-treated C57/B16 mice were treated with 200 μg/kg G-CSF daily for 4 days, sacrificed on day 5 and peripheral blood analyzed for LKS and LKS SLAM cells by flow cytometry.

[0113] 5-fluorouracil was intraperitoneally injected every week at a dose of 75 mg/kg for a total of 4 doses or as a one time dose of 200 mg/kg. Recombinant mouse periostin (R&D Systems, Minneapolis, Minn., cat. no. 2955-F2), produced in the Sf21 (baculovirus)-derived insect cell line, was administered intravenously at a dose of 2-4 μg per day and vitamin K was given at a dose of 15 mg/kg by oral gavage for four consecutive days.

[0114] Concentrations of drugs for in vitro use were 211 M for warfarin, 2 μg/ml for periostin, 511 M for the AKT inhibitor MK-2206 (Selleck Chemicals, Houston, Tex.) and cilengitide (Selleck Chemicals, Houston, Tex., Cat. No. #S7077) was used at 0.05 μM. Phosphate buffered saline served as the vehicle control.

Bone Marrow Transplantation and In Vivo Assays

[0115] SJL (CD45.1) donor mice were treated with vehicle or warfarin, euthanized and long bones were flushed, followed by RBC lysis (Life Technologies, Darmstadt, Germany). For competitive transplantation 2×10.sup.6 total BM cells from SJL mice were co-transplanted with 1×10.sup.6 CD45.2+BM competitor cells into lethally irradiated (900 cGy) C57BL/6N (CD45.2+) recipient mice. For serial transplantation 2×10.sup.6 total CD45.1+BM cells were transplanted into lethally irradiated C57BL/6N recipient mice. Cell engraftment and chimerism were assessed by flow cytometry of peripheral blood leukocytes for CD45.1 and CD45.2 after 4, 8, 12 and 16 weeks post transplantation. In the limiting dilution experiment we treated primary CD45.1+ donor mice with vehicle or warfarin for 14 days and, consequently, transplanted 5×10.sup.5, 12.5×10.sup.4, 6×10.sup.4 and 1.5×10.sup.4 pooled BM cells into 5 CD45.2+ recipients per group adding 5×10.sup.5 CD45.2+ supporter cells. In indicated experiments we transplanted 5,000 or 10,000 sorted CD45.1+ LKS cells from mice treated with vehicle or warfarin plus 1×10.sup.6 CD45.2+ total BM supporter cells into CD45.2+ recipient mice.

[0116] In the homing assay, we intravenously transplanted 9×10.sup.6 Actin DsRed+ whole BM cells into C57BL/6N mice treated with vehicle or warfarin for 14 days. 18 hours later, we analyzed BM and spleen cells of recipient mice for the presence of DsRed+ LKS cells.

[0117] For the xenotransplantation experiments, we intravenously injected 1.3×10.sup.5 human CD34+ cells, obtained via magnetic separation by magnetic beads (Miltenyi Biotech, Bergisch Gladbach, Germany) from BM filters used after harvest of non-mobilized HSPC from healthy allogeneic donors, into irradiated (250 cGy) NSG mice, which had previously been treated with vehicle or warfarin for 14 days.

[0118] For the intrafemoral co-transplantation experiments 0.8×10.sup.5 human CD34+ cells or 1×10.sup.5 Lin− cells from Actin-DsRed reporter mice were mixed with 5×10.sup.5 human or murine stroma cells which had been expanded in vitro and pretreated with vehicle or 2 μM warfarin for 14 days. The human stromal cells had been obtained via flushing of discarded bones from the orthopedic operating room, as approved of by the local ethics committee.

Analysis of Mice

[0119] We assessed the complete blood count of vehicle or warfarin-treated mice using a complete blood count analyzer (Scil Vet ABC, Gurnee, Ill.). Peripheral blood and bone marrow samples were stained with antibodies for flow cytometry, which was performed on a BD LSR Fortessa (BD Biosciences, Heidelberg, Germany). The lineage antibody cocktail contained antibodies to B220, CD5, Ter119, F4/80 and CD11b. Engraftment of human cells in NSG mice was assessed using anti-human CD45 (BD Biosciences, San José, Calif.).

[0120] For cell cycle analysis cells were permeabilized and fixed with the cytoperm/cytofix kit (BD Biosciences, San José, Calif.), followed by staining with an antibody to Ki67-PE (Biolegend, San Diego, Calif.) overnight at 4 C. Consequently, cells were washed and resuspended in PBS with DAPI (1 ng/ml). For detection of apoptosis cells were first stained for surface markers, washed with Annexin V buffer (Life Technologies, Darmstadt, Germany), stained with Annexin V-PE (Life Technologies, Darmstadt, Germany), washed with Annexin V binding buffer, resuspended in PBS with DAPI and analyzed.

[0121] The PE-labelled anti-pAKT antibody (Cell signaling, #5315, Danvers, Mass.) was used for intracellular flow cytometry after fixation and permeabilization of the cells.

Bone Histomorphometry

[0122] The femora were isolated and first fixed in 4% paraformaldehyde at 4C overnight and then in 70% ethanol for an additional 5 days. Fixed bones were dehydrated in graded ethanol, then embedded in methyl methacrylate without demineralization. Undecalcified 5 μm thick longitudinal sections were obtained using a microtome (RM2255, Leica Biosystem, IL, USA). The sections were stained with Goldner Trichrome and at least two consecutive sections per specimen were examined for measurements of cellular parameters. A standard histomorphometric analysis of the femur methaphysis was performed using the Osteomeasure analysis system (Osteometrics Inc, Decatur, Ga., USA). Measurements were performed 200 μm below the distal growth plate. The observer was blinded to the experimental group at the time of measurement. The cellular parameters were calculated and expressed according to the standardized nomenclature (Dempster et al., 2013).

[0123] In Vitro Assays

[0124] In the co-culture and cobblestone assays 20,000 Lin− cells were plated on stroma cells, which had been grown and expanded from vehicle- or warfarin-treated mice in medium (aMEM medium supplemented with 20% fetal bovine serum, 1% penicillin/streptomycin, 1% L-Glutamine) containing vehicle or warfarin (2 μM). Conditioned medium was harvested from stroma from 5 mice grown in the above medium containing vehicle or warfarin for 7 days. Consequently, 100,000 Lin− cells were grown for 2-3 days in the conditioned medium and analyzed for cell counts by flow cytometry. Where indicated, the conditioned medium was concentrated using Microsep Advance Centrifugal Devices 30K Omega (Pall Life Sciences, Portsmouth, UK).

[0125] For all coculture assays hematopoietic and stroma cells were derived from the same mouse strain (C57/B16). Stroma cells were used after a maximum of two passages and when they were 70-80% confluent. Stroma cells were derived by crushing bones from mice, plating them in α-MEM medium containing 20% fetal calf serum, penicillin/strptomycin and by removing non-adherent hematopoietic cells the following day, as described (Mukherjee et al., 2008).

[0126] For the methylcellulose colony assays, we plated 10,000 total BM cells from vehicle- or warfarin-treated mice in methylcellulose (M3434, Stem Cell Technologies, Vancouver, Canada). We scored colonies after 10 days.

[0127] To test the direct effect of warfarin on hematopoietic cells 100,000 Lin− cells were cultured in 2 μM vehicle or warfarin in the absence of stroma cells for a total of 7 days. For the in vitro rescue experiment with periostin stroma cells were grown in warfarin for 14 days, before vehicle or periostin were added.

[0128] For the macrophage coculture assay F4/80+ macrophages were sorted from the bone marrow of control or warfarin-treated mice, plated and cocultured with 20,000-30,000 Lin− BM cells as from the following day.

Differentiation Assays

[0129] Bones from mice were crushed and plated and hematopoietic cells were removed the following day. As described (Mukherjee et al., 2008), after reaching confluence the stroma cells were cultured in α-MEM medium containing 10 mM β-glycerolphosphate, 10 nM dexamethasone and 50 μg/ml ascorbic acid for differentiation into osteoblastic cells (Mukherjee et al., 2008) or 500 nM insulin and 100 nM dexamethasone for differentiation into adipocytes. Differentiation was tested after 14-20 days by staining for von Kossa in the case of osteoblastic cells and Oil Red 0 in the case of adipocytes following standard staining protocols. Osteoclasts were purified from bone marrow macrophages, following standard protocols. Briefly, bone marrow cells were isolated from the femora and tibiae of 6-8 week old C57/B16 animals and cultured overnight at 37C in α-MEM medium containing 10% fetal bovine serum and 1% antibiotic/antimycotic. The following day non-adherent cells were collected and separated by centrifugation on 50% Ficoll-Paque. Bone marrow-derived macrophages were seeded at 20,000 cells/well and cultured for 3 days in the presence of macrophage colony-stimulating factor (M-CSF) (50 ng/ml). On day 4 cells were also treated with receptor activator of nuclear factor kappa-B ligand (RANKL) (50 ng/ml) for an additional 4-5 days prior to tartrate-resistant acid phosphatase (TRAP) staining. Cells were treated with vehicle or warfarin (2 μM) for the entire culture period. For TRAP staining cells were fixed in 10% formalin for 10 minutes, permeabilized in acetone:ethanol (50:50) for 1 minute and then stained with TRAP solution (1 mg/ml of naphtol AS-MX in 0.1 M sodium acetate, 0.05M sodium tartrate and 0.6 mg/ml violet blue salt) at 37 C for 10-15 minutes.

Cloning of the Integrin β3 Construct

[0130] Integrin β3 was amplified by PCR from 3T3 fibroblasts using the forward and the reverse primers, ATATATATGAATTCATGCGAGCGCAGTGG (SEQ ID NO: 8) and TATATAGAATTCTTAAGTCCCCCGGTAGGT (SEQ ID NO: 9), respectively. Both primers contained an EcoR1 restriction site, which was used to clone the 2.3 kb integrin β3 fragment into the MSCV IRES GFP vector. Integrin β3 expression was checked by flow cytometry (BD Biosciences, San José, Calif.) and Western Blotting (Abcam, Cambridge, UK, and Cell Signaling, Danvers, Mass.), as described (Krause et al., 2013).

Quantitative PCR

Quantitative PCR was Performed Using Standard Protocols.

Measurement of the INR

[0131] A Coaguchek XS (Roche Diagnostics, Mannheim, Germany) and Coaguchek XS PT test strips were used to determine the INR of vehicle- and warfarin-treated mice.

[0132] Western Blotting and co-immunoprecipitation HEK293T cells were transfected with the MSCV integrin β3 IRES GFP plasmid using calcium phosphate transfection. Conditioned medium from vehicle or warfarin-treated stroma cells was added 48 h post transfection for 6 hours followed by protein isolation using RIPA buffer (150 mM sodium chloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, sodium dodecyl sulfate, 50 mM Tris, pH 8.0). Magnetic beads (Dynabeads Protein G) were used for protein immunoprecipitation (IP). Protein lysates were first incubated in 2 μg of antibody for 3 h at 4 C and then incubated with magnetic beads overnight at 4 C. After washing three times in ice cold lysis buffer immunoprecipitated proteins were mixed with 5× Laemmli elution buffer, heated to 95 C for 4 min and eluted. Proteins were separated using SDS-PAGE (4-12% gradient) and blotted onto nitrocellulose membranes. Membranes were incubated overnight in the respective primary antibodies at 4 C and then in secondary antibodies for 1 h at room temperature.

[0133] For the anti-gla coimmunoprecipitation experiments we used 4 ng of anti-Gla antibody (BioMedica Diagnostics Inc., Windsor, Canada). In the periostin-overexpression experiments HEK293T cells were transfected with a Myc-DDK-tagged periostin-expressing plasmid (Origene #MR210633). 12 hours after transfection, the media was changed to fresh media containing warfarin (2 μM and 50 μM) or vitamin K1 (10 ng/ml) and left for 48 hours. Cell lysates were then prepared using RIPA lysis buffer.

Isolation of Human CD34+ and Stroma Cells and Patient Data

[0134] Stroma and hematopoietic cells (purified for CD34+) were taken from the iliac crest of healthy donors of allogeneic bone marrow.

[0135] For FIGS. 6C-6H: Data of patients on VKA were obtained from the French health care databases. Patients with hematological disease, inflammation or sepsis were excluded. Patients admitted to the hospital for dermatological problems without hematological disease or sepsis acted as controls. The control patients did not have clots and were not treated with VKA or any other anticoagulant.

[0136] Sequencing of patients Sequencing of 50 to 60 year old patients on warfarin or control patients was performed using the TrueSight Myeloid Sequencing Panel (Illumina, San Diego, Calif.).

[0137] The control group was composed of patients with no past history of clots, no VKA treatment and no genetic predisposition to thrombosis, while the study group consisted of patients with long-term VKA treatment for thrombosis. None of these patients had a known hematological disorder.

[0138] Epidemiological study The national healthcare administrative databases in France were used to explore a potential association between VKA use and a diagnosis of MDS. The French National Health Insurance Information System (SNIIRAM) collects all individualized and anonymous healthcare claims including drug usage and severe and long-term conditions listed in the International Classification of Diseases, 10th edition (ICD-10). Information from the SNIIRAM database was cross-referenced with the French hospital discharge database (PMSI), which provides discharge diagnoses (ICD-10 codes) for all patients.

[0139] The study population consisted of French men and women aged 70-79 years (n=5,464,258) or 50-60 years (n=11,452,848) in 2015. Individuals were considered VKA users if they had at least one VKA claim in 2015, and MDS diagnosis was identified from ICD-10 codes (D46) allocated to hospital stays and/or long-term illness diagnosis in 2015 (n=5,840 cases). In France, fluindione, warfarin and acenocoumarol are available as VKA.

Statistical Methods

[0140] Chi-squared test and logistic regression adjusted for age and sex were used to compare the odds of VKA use in individuals with versus without a diagnosis of MDS, overall and by age and sex strata.

[0141] Differences in survival were assessed by Kaplan-Meier non-parametric estimates (Log-rank test) and between groups by student's t-test. The data were presented as mean±s.d. We used L-Calc software (Stemcell Technologies, Vancouver, Canada) to calculate HSC frequency by Poisson statistics. P values <0.05 were accepted as significant.