CD96 is a leukemic stem cell-specific marker
`in human acute myeloid leukemia
`
`Naoki Hosen*†‡, Christopher Y. Park*, Naoya Tatsumi§, Yusuke Oji§, Haruo Sugiyama§, Martin Gramatzki¶,
`Alan M. Krensky储, and Irving L. Weissman*‡
`
`*Departments of Pathology and Developmental Biology, Institute of Stem Cell Biology and Regenerative Medicine, Stanford Cancer Center, Stanford
`University School of Medicine, Stanford, CA 94305-5323; §Department of Functional Diagnostic Science, Osaka University School of Medicine, Osaka
`565-0871, Japan; ¶Division of Stem Cell Transplantation and Immunotherapy, Second Department of Medicine, University of Kiel, Kiel 24105, Germany;
`and 储Department of Pediatrics, Stanford University School of Medicine, Stanford, CA 94305
`
`Contributed by Irving L. Weissman, May 10, 2007 (sent for review March 20, 2007)
`
`Permanent cure of acute myeloid leukemia (AML) by chemotherapy
`alone remains elusive for most patients because of the inability to
`effectively eradicate leukemic stem cells (LSCs), the self-renewing
`component of the leukemia. To develop therapies that effectively
`target LSC, one potential strategy is to identify cell surface markers
`that can distinguish LSC from normal hematopoietic stem cells
`(HSCs). In this study, we employ a signal sequence trap strategy to
`isolate cell surface molecules expressed on human AML-LSC and
`find that CD96, which is a member of the Ig gene superfamily, is a
`promising candidate as an LSC-specific antigen. FACS analysis
`demonstrates that CD96 is expressed on the majority of
`CD34ⴙCD38ⴚ AML cells in many cases (74.0 ⴞ 25.3% in 19 of 29
`cases), whereas only a few (4.9 ⴞ 1.6%) cells in the normal
`HSC-enriched population (LinⴚCD34ⴙCD38ⴚCD90ⴙ) expressed
`CD96 weakly. To examine whether CD96ⴙ AML cells are enriched
`for LSC activity, we separated AML cells into CD96ⴙ and CD96ⴚ
`fractions and transplanted them into irradiated newborn Rag2ⴚ/ⴚ
`␥c
`ⴚ/ⴚ mice. In four of five samples, only CD96ⴙ cells showed
`significant levels of engraftment in bone marrow of the recipient
`mice. These results demonstrate that CD96 is a cell surface marker
`present on many AML-LSC and may serve as an LSC-specific
`therapeutic target.
`
`hematopoietic stem cell
`
`The cancer stem cell hypothesis holds that cancers are composed
`
`of a subset of cells that have the unique ability to transplant
`disease as well as self-renew (1–4). This concept may radically alter
`approaches to cancer therapies. We have proposed a multistep
`model of leukemogenesis in which protooncogenic events short of
`activating the self-renewal process occur in the self-renewing he-
`matopoietic stem cell (HSC) population (5). Therefore, the clonal
`progression of preleukemias likely occurs in a succession of HSC
`subclones until augmented or poorly regulated self-renewal path-
`ways are activated, leading to the emergence of final stage leukemic
`stem cells (LSCs) usually at the level of a downstream progenitor
`(5). We previously showed that in human acute myeloid leukemia
`(AML) characterized by t(8;21),
`the LSCs reside in the
`CD34⫹CD90⫺CD38loLin⫺ fraction of progenitors, whereas pre-
`LSC, detected during hematological remission, are present in the
`CD34⫹CD90⫹CD38loLin⫺ fraction (6), which is typically enriched
`for HSC (7). These observations suggest that cell surface antigens
`that are expressed on progenitor cells, but not on normal HSC, are
`possibly expressed on AML-LSC. Nevertheless, cell surface mark-
`ers that can distinguish AML-LSC from normal HSC are still
`lacking, except for the IL-3R␣ chain (CD123) (8).
`One promising strategy for therapeutic targeting of LSC is
`immunotherapy with monoclonal antibodies (9). In fact, differ-
`ent antibodies or immunoconjugates,
`including CD20 (10),
`CD33 (11), and CD52 (12), have been successfully used for the
`treatment of hematologic malignancies. Previous studies have
`pointed to antibody-dependent cell-mediated cytotoxicity
`(ADCC) and FcR-mediated phagocytosis by the macrophage
`
`lineage as important mechanisms of activity for these antibody
`therapies (13). Therefore, when a cell surface antigen specific for
`LSC is identified, depleting antibodies against it can be poten-
`tially developed as therapeutics. In addition, cytotoxic sub-
`stances such as chemotherapeutics, radioimmunoconjugates, or
`toxins can be coupled to enhance efficacy.
`To identify cell surface markers selectively expressed on
`AML-LSC, we applied a signal sequence trap PCR method (14)
`to highly purified CD34⫹CD38⫺ AML cells, a fraction enriched
`in AML-LSC activity (15). We find that CD96 is selectively
`expressed in AML-LSC. CD96, which is also known as Tactile,
`is a member of the Ig gene superfamily and was first identified
`as a gene expressed in activated T cells (16). Furthermore, we
`demonstrate that CD96⫹, but not CD96⫺, AML cells can engraft
`immunodeficient mice (17). Taken together, CD96 is a marker
`for AML-LSC and therefore a potential target for LSC-specific
`therapy.
`
`Results
`Identification of CD96 as an AML-LSC-Specific Marker. To identify
`cell-surface molecules expressed on AML-LSC, we used the signal
`sequence trap PCR screening method, which allows for the detec-
`tion of mRNAs that contain signal sequences by virtue of their
`ability to redirect a constitutively active mutant of c-mpl to the cell
`membrane, and thereby induce IL-3-independent growth of Ba/F3
`cells (14) (Fig. 1A). We initially applied this method to amplified
`cDNA from 3.6 ⫻ 104 FACS-purified CD34⫹CD38⫺ AML cells
`(FAB classification: M2). Completion of the screen resulted in the
`identification of 33 independent genes [supporting information (SI)
`Table 2; 31 known, 2 unknown]. We then used an unamplified
`cDNA library prepared from 3 ⫻ 107 FACS-purified
`CD34⫹CD38⫺ cells from another AML (M2) sample and identified
`41 genes (SI Table 2; 36 known, 5 unknown). Candidate AML-LSC
`marker genes were selected from these two data sets by excluding
`those genes that would not be expected to effectively discriminate
`AML-LSC from HSC, such as CD45. The expression levels of the
`
`Author contributions: N.H. and I.L.W. designed research; N.H., C.Y.P., N.T., Y.O., and H.S.
`performed research; M.G. and A.M.K. contributed new reagents/analytic tools; N.H. ana-
`lyzed data; and N.H., C.Y.P., M.G., A.M.K., and I.L.W. wrote the paper.
`
`Conflict of interest statement: I.L.W. was a member of the scientific advisory board of
`Amgen and owns significant Amgen stock. I.L.W. is also a cofounder and director of Stem
`Cells, Inc., and cofounded Cellerant, Inc. N.H. and I.L.W. have applied for a U.S. patent
`entitled ‘‘Identification of Cell Surface Marker for Acute Myeloid Leukemia Stem Cells’’
`through the Stanford University Office of Technology and Licensing.
`
`Abbreviations: AML, acute myeloid leukemia; HSC, hematopoietic stem cell; LSC, leukemic
`stem cell; NK, natural killer.
`†Present address: Department of Cancer Stem Cell Biology, Osaka University Graduate
`School of Medicine, Osaka 565-0871, Japan.
`‡To whom correspondence may be addressed. E-mail: hnaoki@imed3.med.osaka-u.ac.jp or
`irv@stanford.edu.
`
`This article contains supporting information online at www.pnas.org/cgi/content/full/
`0704271104/DC1.
`
`© 2007 by The National Academy of Sciences of the USA
`
`11008 –11013 兩 PNAS 兩
`
`June 26, 2007 兩 vol. 104 兩 no. 26
`
`www.pnas.org兾cgi兾doi兾10.1073兾pnas.0704271104
`
`Sanquin EX2004
`Forty Seven v. Stichting Sanquin Bloedvoorziening
`IPR2016-01529
`
`

`
`Isotype
`
`105 36
`104
`
`103
`
`102
`0
`
`0 102 103 104 105
`105 7.0
`104
`
`103
`
`102
`0
`
`CD96
`
`G8.5
`
`TH-111
`105 30.5
`0.9
`104
`
`1.6
`
`103
`
`105 32.9
`104
`
`103
`
`102
`0
`
`7.8
`0 102 103 104 105
`105 6.5
`0.3
`104
`
`102
`0
`
`5.7
`0 102 103 104 105
`5 7.0
`10
`4
`10
`
`103
`
`102
`0
`
`3
`10
`
`2
`10
`0
`
`CD34+
`CD38-
`Lin-
`
`CD34+
`CD38+
`Lin-
`
`0 102 103 104 105
`
`A
`
`105
`
`104
`
`103
`
`102
`0
`
`CD38
`
`Exp.2
`
`AML(M2)
`
`105
`104
`103
`102
`0
`
`CD38
`
`103
`
`104
`
`105
`
`102
`0
`CD34
`3X107 CD34+CD38- AML cells
`
`
`A
`
`Exp.1
`
`AML(M2)
`
`105
`104
`103
`102
`0
`
`CD38
`
`103
`
`104
`
`105
`
`102
`0
`CD34
`3.6X104 CD34+CD38- AML cells
`
`GGG
`CCC
`
`mRNA
`
`cDNA
`
`Amplification
`RsaI digestion
`
`BstI linker ligation
`
`CD34
`
`BstI linker
`ligation
`
`CD34-
`Lin-
`
`0 102 103 104 105
`105 0.1
`104
`
`19.5
`0 102 103 104 105
`105 0.1
`0.2
`104
`
`17.1
`0 102 103 104 105
`5 0.1
`0.1
`10
`4
`10
`
`103
`
`102
`0
`
`5.7
`0 102 103 104 105
`
`3
`10
`
`2
`10
`0
`
`8.4
`0 102 103 104 105
`
`1.9
`0 102 103 104 105
`
`103
`
`102
`0
`
`CD90
`
`-
`
`CD34
`-
`Lin
`
`+
`+
`-
`
`CD34
`CD38
`CD90
`-
`Lin
`
`+
`
`--
`
`CD34
`CD38
`CD90
`-
`Lin
`
`+
`-
`+
`
`CD34
`CD38
`CD90
`-
`Lin
`
`CD96
`
`35
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`B
`
` CD96 cells
`Percentages of
`
`+
`
`MEDICALSCIENCES
`
`Fig. 2. CD96 expression in normal BM cells. (A) Lin⫺ BM cells were separated
`into subpopulations according to the expression of CD34 and CD38 and then
`analyzed for CD90 and CD96 expression, with the anti-CD96 clones, G8.5 or
`TH-111. Numbers represent the percentages of cells in the gated populations.
`Representative data from three different normal BM samples are shown. (B)
`Summary of the analysis of CD96 expression in normal Lin⫺ BM cells (n ⫽ 3).
`Error bars show the SD.
`
`In 19 of 29 samples (65.5%), the percentage of CD96-positive
`cells in the CD34⫹CD38⫺ AML-LSC-enriched fraction (15) was
`significantly higher than in normal human BM CD34⫹CD38⫺
`cells (74.0 ⫾ 25.3% vs. 12.2 ⫾ 2.7%) (Fig. 3 A–C and Table 1).
`CD96 is expressed almost exclusively in the CD90⫺ subset (Fig.
`3 A and B and SI Fig. 6). In the remaining 10 samples (34.5%),
`the frequency of CD96 expression in the CD34⫹CD38⫺ blasts
`was not increased compared to normal BM CD34⫹CD38⫺ cells
`(Table 1). To examine whether CD34⫹CD38⫺CD96⫹ blasts
`express lineage markers, we analyzed three AML samples for the
`expression of CD96, in addition to numerous lineage markers
`(CD2, CD3, CD4, CD8, CD10, CD11b, CD14, CD19, CD20,
`CD56, and Glycophorin-A), as well as CD34, CD38, and CD90
`(Fig. 3D). In all three samples, ⬎85% of CD34⫹CD38⫺CD96⫹
`AML blasts did not express lineage markers.
`Together these results indicate that CD96 is frequently ex-
`pressed on CD34⫹CD38⫺Lin⫺CD90⫺ AML blasts, which are
`enriched for LSC activity and exclude HSC. When the AML
`samples were grouped according to their FAB category, CD96
`expression in the AML-LSC population was more frequent in
`M2 compared to M0/M1 or M4/M5 samples (Fig. 3C).
`
`CD96 Is Expressed on Functional AML-LSC. To determine whether
`CD96 is expressed on functional AML-LSC, we used FACS to
`fractionate primary human AML specimens into CD96⫹ and
`CD96⫺ populations, and we transplanted similar numbers of
`cells from each fraction into sublethally irradiated (100 cGy)
`
`TPOR(-TM domain)
`
`pMX-SST
`
` AML1
`34+38-
`
` AML2
`34+38-
`
` AML3
`34+38-
`
`1
` Normal
`Lin-34+38-
`
`600
`500
`400
`300
`200
`100
`0
`
`levels of CD96 mRNA
` Relative expression
`
`B
`
`SST screening of CD34⫹CD38⫺ AML-LSCs. (A) The scheme of the SST
`Fig. 1.
`screening of CD34⫹CD38⫺ AML-LSCs. Full-length cDNA, either digested into
`pieces (experiment 1, Left) or undigested (experiment 2, Right), was ligated
`with a BstXI linker and subcloned into pMX-SST vector. TPOR, thrombopoietin
`receptor (a constitutively active mutant); TM, transmembrane. (B) CD96 mRNA
`expression level in CD34⫹CD38⫺ cells from three different AML (M2) samples
`and normal CD34⫹CD38⫺Lin⫺ cells. The expression levels are shown relative to
`those in normal CD34⫹CD38⫺Lin⫺ BM cells.
`
`35 selected genes were measured by quantitative RT-PCR in
`CD34⫹CD38⫺ cells isolated from three different AML (M2)
`samples, as well as normal bone marrow (BM)-derived HSC/
`progenitors (Lin⫺CD34⫹CD38⫺). Among the genes tested, only
`CD96 was found to be expressed at much higher levels in all three
`different CD34⫹CD38⫺ AML blasts compared to normal
`Lin⫺CD34⫹CD38⫺ BM cells. CD96 mRNA expression levels in
`CD34⫹CD38⫺ blasts from three different AML samples were 270,
`570, and 202 times higher than in normal Lin⫺CD34⫹CD38⫺ BM
`cells (Fig. 1B).
`
`CD96 Is Not Expressed by the Majority of Cells in the Normal HSC-
`Enriched Population. Two CD96 monoclonal antibodies, clone G8.5
`(16) and clone TH-111 (18), were used to analyze CD96 expression
`in normal human adult HSCs and progenitor cells by flow cytom-
`etry. BM cells from three independent healthy adult donors were
`stained with CD96, CD34, CD38, CD90, and lineage markers and
`then analyzed by flow cytometry. Only 4.9 ⫾ 1.6% of cells in the
`HSC-enriched population (Lin⫺CD34⫹CD38⫺CD90⫹) (7, 19) ex-
`pressed CD96 weakly (Fig. 2 A and B). In other non-HSC popu-
`lations, such as Lin⫺CD34⫹CD38⫺CD90⫺, Lin⫺CD34⫹CD38⫹,
`and Lin⫺CD34⫺, CD96 was detected on 18.2 ⫾ 8.3%, 25.0 ⫾ 3.5%,
`and 16.6 ⫾ 12.1% of the cells, respectively (Fig. 2 A and B).
`Colony-forming assays were performed by using FACS-purified
`Lin⫺CD34⫹CD38⫺CD90⫺CD96⫹ as well as Lin⫺CD34⫹CD38⫺
`CD90⫺CD96⫺ BM cells. Both fractions form GEMM, GM, and
`indicating that Lin ⫺CD34⫹
`BFU-E colonies (SI Fig. 5),
`CD38⫺CD90⫺CD96⫹ BM cells contain progenitor cells of myeloid/
`erythroid lineages.
`
`CD96 Is Frequently Expressed in the AML-LSC Population. Next, we
`examined CD96 expression in 29 primary human AML samples.
`
`Hosen et al.
`
`PNAS 兩
`
`June 26, 2007 兩 vol. 104 兩 no. 26 兩 11009
`
`Sanquin EX2004
`Forty Seven v. Stichting Sanquin Bloedvoorziening
`IPR2016-01529
`
`

`
`CD34+
`CD38-
`
`CD96
`0.4
`
`103
`
`56.2
`104
`105
`0.9
`
`Pt 16
`
`73.1
`
`16.5
`
`Isotype
`0.1
`
`105 1.9
`104
`
`103
`
`102
`6.9
`0
`
`0.2
`104
`105
`
`103
`
`0 102
`
`103
`
`104
`
`105
`
`105 0.6
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`B
`
`CD38
`
`CD34
`
`CD34+
`CD38 -
`
`CD96
`
`0.1
`
`0 102
`
`103
`
`95.3
`104
`105
`0.3
`
`Pt 7
`
`4.4
`
`78.7
`
`Isotype
`105 0.1
`104
`
`103
`
`102
`13.9
`0
`
`0.2
`104
`105
`
`103
`
`0 10 2 10 3 10 4 10 5
`
`5
`10
`
`4
`10
`
`3
`10
`
`2
`10
`0
`
`5
`10
`
`105
`
`104
`
`103
`
`102
`0
`
`CD38
`
`A
`
`0 102
`105 1.3
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`CD90
`
`0.2
`104
`105
`
`103
`
`0 102
`0.2
`
`0 102
`105 1.4
`104
`
`103
`
`102
`0
`
`0 102
`105 0.3
`104
`
`103
`
`55.7
`104
`105
`0.1
`
`CD34+
`CD38+
`
`CD34-
`
`103
`
`102
`0
`
`0.1
`105
`
`103
`
`104
`
`0 102
`
`103
`
`104
`
`6.9
`105
`
`CD34
`
`0 102
`105 0.1
`104
`
`103
`
`102
`0
`
`0 102
`105 0.5
`104
`
`103
`
`102
`0
`
`CD34 +
`CD38 +
`
`4
`10
`
`3
`10
`
`2
`10
`0
`
`0 102
`5 0.6
`
`10
`
`103
`
`96.9
`104
`105
`0.6
`
`103
`
`0.3
`104
`105
`0.3
`
`CD34 -
`
`4
`10
`
`3
`10
`
`2
`10
`0
`
`0 102
`
`103
`
`87.6
`104
`105
`
`0 102
`CD96
`
`D
`
`Pt 4
`
`Pt 6
`
`105
`
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`0 102
`
`103
`
`104
`
`105
`
`Isotype
`
`CD96
`
`105
`
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`0 102
`
`103
`
`104
`
`105
`
`105
`
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`88
`104
`105
`
`0 102
`
`103
`
`88.2
`104
`105
`
`103
`
`0.7
`104
`105
`
`103
`
`0 102
`CD96
`
`CD90
`
`C
`
`100
`
`80
`
`60
`
`40
`
`86.7
`104
`105
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`105
`
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`Lin
`
`0 102
`
`103
`
`104
`
`105
`
`105
`
`104
`
`103
`
`102
`0
`
`Pt 9
`
`CD38
`
`CD34
`
`CD96
`
`20
`
`0
`
` in CD34+CD38- AML cells
`Percentages of CD96+ cells
`
`FAB
`classification
`
`M0/1
`
`M2
`
`M3
`
`M4/5
`
`NBM
`
`CD96 expression in AML. (A and B) Expression profiles of CD34, CD38, CD90, and CD96 in AML samples. Cells were separated into subpopulations
`Fig. 3.
`according to the expression of CD34 and CD38 and then analyzed for CD90 and CD96 expression. Representative data from AML samples that contain high (A)
`and intermediate (B) percentages of CD96⫹ cells in the CD34⫹CD38⫺ population are shown. (C) Percentages of CD96⫹ cells in the AML CD34⫹CD38⫺ population.
`Each bar represents a single AML sample. Pts 10, 15, 16, and 24 are not included in this graph because of the lack of information about FAB classification. The
`frequencies of CD96⫹ cells in normal BM CD34⫹CD38⫺ cells are shown for comparison. The dotted line represents the average in normal BM CD34⫹CD38⫺ cells.
`(D) Expression of lineage markers and CD96 in CD34⫹CD38⫺ AML blasts. Numbers represent the percentage of CD96⫹Lin⫺ cells in CD34⫹CD38⫺ AML blasts.
`
`newborn Rag2⫺/⫺ ␥c
`⫺/⫺ mice via the facial vein (SI Table 3). This
`highly immunodeficient mouse strain lacks B, T, and natural
`killer (NK) cells (20), and newborn Rag2⫺/⫺ ␥c
`⫺/⫺ mice support
`efficient engraftment of human AML (C.Y.P., R. Majeti, and
`I.L.W., manuscript in preparation). Transplanted mice were
`killed at 6 to 10 weeks after transplantation and analyzed for
`engraftment of human leukemia cells in BM. The results of a
`typical experiment are shown in Fig. 4A, with CD96⫹ AML cells
`uniquely showing engraftment of human CD45 (hCD45)⫹ cells.
`We confirmed that engrafted hCD45⫹ cells were human myeloid
`leukemia blasts by measuring human CD13/CD14/CD33 expres-
`sion (Fig. 4A) and/or evaluating Wright–Giemsa-stained cyto-
`spin preparations of the peripheral blood or BM cells (Fig. 4B
`and SI Fig. 7).
`Transplantation experiments were performed by using five dif-
`ferent AML samples (Fig. 4D and SI Table 3). For four of five cases,
`cells were separated into CD96⫹ and CD96⫺ populations regardless
`of the expression of CD34 or CD38 (Pts 5, 11, 14, 26). For one
`patient (Pt 3), cells were separated into CD34⫹CD38⫺CD96⫹ and
`CD34⫹CD38⫺CD96⫺ fractions. In four of five samples (Pts 3, 5, 11,
`26), only CD96⫹ AML cells showed significant levels of engraft-
`ment in the BM of recipient mice, whereas CD96⫺ AML cells did
`not engraft. In the case of Pt 14, high levels of engraftment were
`observed with both CD96⫹ and CD96⫺ AML cells. It should be
`noted that for one specimen (Pt 26), the enrichment of LSC activity
`in the CD96⫹ AML fraction was observed despite the low per-
`centage of CD34⫹ cells in this fraction. Finally, we analyzed CD96
`expression on engrafted hCD45⫹ cells from BM of mice trans-
`planted with CD96⫹ AML cells. As shown in Fig. 4C, transplan-
`tation of purified CD96⫹ cells resulted in engraftment of both
`CD96⫹ and CD96⫺ cells, recapitulating the heterogeneity of CD96
`
`expression in the primary AML specimen. Collectively, these
`results demonstrate that CD96 is expressed on functional LSC in
`human AML.
`
`Discussion
`CD96 has been reported to be expressed in T and NK cells, but
`not in B cells, granulocytes, monocytes, or RBCs (16, 18). In
`nonhematopoietic tissue, CD96 is expressed in the convoluted
`tubular epithelium of the kidney, the mucosal epithelium of the
`small and large intestines, and the vascular endothelium (18). It
`was also reported that CD96 is expressed in ⬇30% of human
`AML samples, regardless of disease subtype, by examining whole
`BM cells from AML patients with immunohistochemical-
`staining (18); its expression was rare in the FAB-M5 subtype. In
`this study, we demonstrate that CD96 is frequently expressed in
`the CD34⫹CD38⫺ LSC population in ⬎60% of the human
`primary AML samples examined. In some of the AML samples
`(e.g., Pt 16) (Fig. 3B), CD96 is highly expressed on the CD34⫹
`blasts, but at a lower frequency in CD34⫺ blasts. Such samples
`may have been scored as negative for CD96 when unsorted
`leukemia cells were examined.
`The xenotransplantation experiments reveal that CD96 is
`expressed on AML-LSC in all of the cases examined, and that in
`four of five cases, CD96⫹AML cells are highly enriched for LSC
`activity compared to CD96⫺ AML blasts. Furthermore, CD96 is
`not detectable on the majority of cells in the normal adult BM
`HSC-enriched population (7). Collectively, these results indicate
`that AML-LSC can be distinguished from normal HSC by the
`presence of CD96 expression. This finding suggests that CD96
`may prove to be an excellent target for antibody therapy against
`LSC because hematopoietic progenitors are regenerated rapidly
`
`11010 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0704271104
`
`Hosen et al.
`
`Sanquin EX2004
`Forty Seven v. Stichting Sanquin Bloedvoorziening
`IPR2016-01529
`
`

`
`Table 1. Patient characteristics
`
`Patient
`no.
`
`FAB
`classification
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`29
`
`M5
`M2
`M1
`M0/1
`M4Eo
`M2
`M2
`M2
`M2
`NA
`M3
`M2
`M2
`M2
`NA
`NA
`M5a
`M4Eo
`M2
`M5
`M2
`M4/5a
`M1
`NA
`M1
`M4
`M0/1
`M5a
`
`M2
`
`Cytogenetic abnormality
`
`FLT3 mutation
`Inv(9)(p11q13)
`Complex*
`NA
`Inv(16), FLT3D835
`t(8;21), t(6;17)
`Normal karyotype
`NA
`NA
`FLT3ITD⫹
`t(15;17) ⫹ complex†
`NA
`NA
`t(8;21)
`NA
`6p⫹
`Normal karyotype
`del(16)(p13.1p13.3)
`NA
`NA
`NA
`13⫹
`Normal karyotype
`5q⫺
`NA
`NA
`NA
`MLL translocation(FISH) 0.6
`FLT3 mutation
`NA
`
`CD96⫹/CD34⫹
`CD38⫺, %
`
`CD96⫹/CD34⫹
`CD38⫹, %
`
`CD96⫹/
`CD34⫺, %
`
`CD96⫹/
`Lin⫺CD34⫹CD38⫺, %
`
`99.6
`
`94.9
`
`91.3
`
`20.7
`0
`18.3
`
`99.5
`99.2
`99.2
`97.8
`97.3
`97.4
`95.4
`93.4
`88.2
`79.8
`72.7
`65.9
`59.9
`56.2
`51.1
`43.3
`42.6
`41.9
`22.1
`16.8
`15.1
`6.2
`3.8
`3.1
`2.8
`1.5
`1.1
`1.2
`
`0
`
`99.2
`95.1
`99.8
`99.8
`96.3
`98.6
`96.9
`98.7
`93.4
`94.5
`91.6
`93.9
`51.6
`55.7
`37.8
`59.5
`86.2
`53.7
`21.9
`26.2
`14.2
`9.2
`4.0
`4.1
`1.1
`2.6
`3.1
`3.4
`
`24.5
`
`28.7
`51.8
`91.0
`32.9
`57.4
`69.9
`87.6
`33.6
`56.5
`3.8
`88.4
`63.4
`59.8
`6.9
`63.2
`17.0
`40.0
`17.0
`5.8
`3.1
`6.8
`34.7
`16.4
`20.2
`11.2
`16.1
`66.0
`
`2.5
`
`NA, not available.
`*46,XX,add(5)(q31),der(12)t(12;15)(q22⬃24;q11.2),der(12)add(12)(p13)?del(12)(q12), add(15)(q11.2)关cp3兴/46,XX关cp17兴
`†46,XX, add(7)(q22), t(15,17)(q22:q12), add(x)(q22), add(3)(p11), add(4) (q21), der(9)add(9)(p13) del(9) (q?), t(11;20)(q13:q13)
`
`MEDICALSCIENCES
`
`from HSC. This therapy can be optimized by the development of
`CD96 antibodies that can induce cytotoxicity, such as ADCC,
`augmented macrophage phagocytosis, or complement-
`dependent cytotoxicity (21), although the expression of CD96 in
`T and NK cells, histiocytes, and nonhematopoietic cells has to be
`considered (18). In addition, ex vivo purging or FACS selection
`of stem cells with CD96 antibodies may be an avenue to pursue
`in autologous transplantation for AML patients.
`An important observation made in our studies is that CD96⫹
`AML cells are enriched in LSC activity even in a sample that
`contained a low percentage of CD96⫹ cells within the
`CD34⫹CD38⫺ population (Pt 26). These results indicate that the
`enrichment of LSC in CD96⫹ AML is not simply a reflection of
`the enrichment of CD34⫹ AML cells in the CD96⫹ population.
`This finding suggests that CD96 may play a functional role in
`LSC biology. CD96 expressed on NK cells has been shown to
`bind to the polio virus receptor (CD155) and thereby mediate
`NK–target cell
`interactions such as those between NK and
`cancer cells (22). In addition, CD155 has been implicated as a
`human homologue of a protochordate histocompatibility gene
`(23). CD96 expressed on AML-LSC may also interact with its
`ligand on other cells in the BM, possibly niche cells for AML-
`LSC, and in these cells binding to the ligand cannot result in
`killer function, but may play a role in leukemia properties. More
`information may be obtained by examining the expression of
`CD155 in BM nonhematopoietic cells, such as osteoblasts or
`endothelial cells.
`Although CD96 may play a functional role in AML-LSC
`biology, it may also have implications for the cell of origin for
`
`AML-LSC. One possibility is that AML-LSC may arise from the
`Lin⫺CD96⫹CD34⫹CD38⫺ CD90⫺ population, ordinarily a non-
`self-renewing multipotent progenitor (R. Majeti, C.Y.P., and
`I.L.W., unpublished data); at a stage in leukemic progression in
`AML, this population may have gained the property of self-
`renewal. The lack of CD90 expression on most AML
`CD34⫹CD38⫺ cells (6, 24) suggests that the final leukemogenic
`event occurs in the CD90⫺ population, which is downstream of
`the CD90⫹ HSC population (7). In support of this hypothesis, we
`previously reported that AML1/ETO was expressed in
`Lin⫺CD34⫹CD38⫺CD90⫹ cells in remission, although CD90
`expression was absent in the leukemia cells harboring the
`AML1/ETO translocation (6).
`In summary, we provide evidence that CD96 is expressed on
`the majority of Lin⫺CD34⫹CD38⫺CD90⫺ AML blasts, the same
`population previously characterized as containing AML-LSC
`activity (6, 15, 25). Furthermore, we demonstrate that AML-
`LSC activity is highly enriched in the CD96⫹ AML blast fraction.
`Because the majority of cells in the normal HSC population do
`not express CD96, these results suggest that CD96 is a marker of
`AML-LSC and a candidate therapeutic target.
`
`Materials and Methods
`Leukemia and Normal BM cells. Diagnostic AML BM cells were
`obtained after informed consent and with the approval of the
`research ethics committee at Stanford University or Osaka
`University. Normal BM cells from healthy volunteers were
`purchased from All Cells (Emeryville, CA).
`
`Hosen et al.
`
`PNAS 兩
`
`June 26, 2007 兩 vol. 104 兩 no. 26 兩 11011
`
`Sanquin EX2004
`Forty Seven v. Stichting Sanquin Bloedvoorziening
`IPR2016-01529
`
`

`
`31.6
`0 102
`103
`
`104
`
`105
`
`105
`
`104
`
`103
`
`102
`0
`
`A
`
`CD38
`
`CD34
`
`C
`
`105
`
`104
`
`B
`
`93.8
`
`105
`
`104
`
`103
`
`102
`0
`
`0 102
`hCD3/19
`
`103
`
`104
`
`105
`
`BM chimerism
`
`hCD13/14/33
`
`12.9
`
`0 102
`
`103
`
`104
`
`105
`
`105
`
`104
`
`103
`
`102
`0
`
`105
`
`104
`
`103
`
`102
`0
`
`0 102
`mCD45.2
`
`103
`
`104
`
`105
`
`CD96 +
`
`CD96 -
`
`hCD45
`
`100cGy
`
`105
`
`104
`
`103
`
`0.8
`
`99.2
`
`0 102
`
`103
`
`104
`
`105
`
`102
`0
`
`CD38
`
`CD96
`
`100cGy
`
`Rag2 -/-γ c-/-
`new born mice
`
`hCD45+ cells
`Isotype
`CD96
`
`D
`
` Pt 3
`(CD34+CD38-)
`0.8
`99.2
`
`Pt 5
`72.7
`
`25.4
`
`Pt 11
`92.8
`
`6.6
`
`Pt 14
`26
`
`71.8
`
`Pt 26
`13.4
`86.1
`
`105
`
`104
`
`105
`
`104
`
`105
`
`104
`
`105
`
`104
`
`105
`
`104
`
`105
`
`104
`
`CD34
`
`0 102
`CD96
`100
`
`103
`
`102
`0
`
`52.8
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`103
`
`102
`0
`
`hCD45
`
`CD96
`
`103
`
`102
`0
`
`103
`
`102
`0
`
`103
`
`102
`0
`
`103
`
`102
`0
`
`103
`
`102
`0
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`0 102
`
`103
`
`104
`
`105
`
`100
`
`10
`
`1
`
`0.1
`
`100
`
`10
`
`1
`
`- +
`
`0.1
`
`- +
`
`100
`
`100
`
`10
`
`1
`
`0.1
`
`10
`
`1
`
`0.1
`
`- +
`
`- +
`
`10
`
`1
`
` cells
`
`+
`
`Percentages of hCD45
`
`0.1
`
`CD96
`
`- +
`
`CD96⫹ AML cells are enriched in LSC activity. (A) Representative results of transplantation of CD34⫹CD38⫺CD96⫹ or CD96⫺ AML cells from Pt 3. (Left)
`Fig. 4.
`Expression profiles of CD34, CD38, and CD96 and the sorting gates. CD96⫹ or CD96⫺ AML cells were transplanted into sublethally irradiated newborn Rag2⫺/⫺
`⫺/⫺ mice. (Right) Analyses of human (hCD45⫹) versus mouse (mouse CD452⫹) chimerism in the bone marrow (BM) at 6 weeks after transplant. (B) Wright–Giemsa
`␥c
`stain of FACS-sorted hCD45⫹ BM cells from an engrafted mouse transplanted with CD96⫹ AML cells from Pt 3 (X630), showing that engrafted hCD45⫹ cells have
`myeloblastic morphology. (C) CD96 expression on engrafted hCD45⫹ cells in the BM of mice transplanted with CD96⫹ AML cells from Pt 3. (D) (Upper) Expression
`profiles of CD34 and CD96 and the sorting gates for CD96⫹ or CD96⫺ populations for each AML sample. Note that the CD34⫹CD38⫺ population is pregated in
`⫺/⫺ mice. Analyses of human engraftment are shown as the
`the case of Pt 3. CD96⫹ or CD96⫺ AML cells were transplanted into irradiated newborn Rag2⫺/⫺ ␥c
`percentage of hCD45⫹cells in the BM at 6 to 10 weeks after transplantation. Each dot corresponds to an individual mouse recipient.
`
`Cell Separation, Immunophenotyping, and Sorting. BM mononu-
`clear cells were isolated by Ficoll (GE Healthcare, Piscataway,
`NJ) gradient centrifugation and cryopreserved for later use.
`CD34-positive normal BM cells were enriched by using the
`MACS (magnetically activated cell sorting) CD34 isolation kit
`(Miltenyi Biotec, Bergisch Gladbach, Germany). Single-cell
`suspensions were washed with PBS containing 2% FCS, incu-
`bated with 10% human AB serum for 20 min to prevent
`nonspecific antibody binding, and stained with either of two
`CD96 antibodies (G8.5; provided by A.M.K.) (16) or TH-111
`(provided by M.G.) (18) or mouse IgG isotype control (eBio-
`science, San Diego, CA) for 30 min on ice. Cells were then
`washed and incubated with PE-conjugated anti-mouse IgG
`polyclonal antibody (eBioscience). Finally, cells were stained
`with Cy5-PE-conjugated lineage markers (CD2, CD3, CD4,
`CD8, CD10, CD11b, CD14, CD19, CD20, CD56, and Glycoph-
`orin-A) (Caltag, Carlsbad, CA), APC-conjugated CD34 (BD
`Bioscience, San Jose, CA), biotin-conjugated CD38 (Invitrogen,
`Carlsbad, CA), and FITC-conjugated CD90 (BD Bioscience) for
`30 min on ice, followed by the incubation with streptoavidin-
`Cy7PE (eBioscience). Cells were resuspended in 1 ␮g/ml pro-
`pidium iodide. Analysis and cell sorting were performed on a
`FACS Aria (Becton Dickinson, San Jose, CA). Data analysis was
`done with Flow Jo software (Tree Star, Ashland, OR).
`
`FACS-purified, and total RNA was extracted by using the
`TRIzol reagent (Invitrogen). In the first experiment, cDNA was
`synthesized by using a PCR cDNA synthesis kit (SMART;
`Clontech, Palo Alto, CA) and oligo dT primer. cDNA was
`amplified by PCR and then digested with RsaI restriction
`enzyme and linked with a BstXI adaptor. cDNA fragments
`ranging from 0.5 to 2.0 kb were selected by electrophoresis on an
`agarose gel and subcloned into pMX-SST vector (a kind gift
`from Toshio Kitamura, Tokyo University, Tokyo, Japan). In the
`second experiment, full-length cDNA libraries were generated
`from FACS-purified CD34⫹CD38⫺ AML cells by using a cDNA
`library synthesis kit (Invitrogen) and random hexamers. cDNAs
`larger than 1 kb were selected by electrophoresis on an agarose
`gel and subcloned into pMX-SST vector. The SST-REX library
`was subjected to screening by transduction into BaF3 cells as
`previously described.
`
`Quantitative RT-PCR. Quantitative RT-PCR analysis was per-
`formed by using SYBR Green on an ABI 7700 real-time PCR
`machine (Applied Biosystems, Foster City, CA) according to the
`manufacturer’s protocols. The expression level of each gene was
`normalized to the ␤-actin expression level. PCR conditions and
`primer sequences are available on request.
`
`Signal Sequence Trap Cloning. SST-REX cloning (14) was per-
`formed with slight modifications. CD34⫹CD38⫺ AML cells were
`
`Colony Assay. Methylcellulose culture assays were performed in
`Methocult H4334 (Stem Cell Technologies, Vancouver, BC,
`
`11012 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0704271104
`
`Hosen et al.
`
`Sanquin EX2004
`Forty Seven v. Stichting Sanquin Bloedvoorziening
`IPR2016-01529
`
`

`
`MEDICALSCIENCES
`
`Canada) according to the manufacturer’s instructions. Colony
`numbers were enumerated and scored on day 14 of culture.
`
`ⴚ/ⴚ Mice. FACS-sorted
`Transplantation into Newborn Rag2ⴚ/ⴚ ␥c
`AML cells were transplanted into newborn Rag2⫺/⫺ ␥c
`⫺/⫺ mice
`within 72 h after birth. Mice were irradiated with 100 cGy 4 to
`24 h before transplantation and injected with sorted cells via the
`anterior facial vein. Mice were killed at 6 to 10 weeks of age to
`evaluate BM engraftment. Human chimerism was examined by
`staining with human CD45 antibody. Cells were stained with
`APC or Cy7PE-conjugated human CD45 (BD Bioscience),
`Alexa488-conjugated mouse CD45.2 (AL1–4A2), and Pacific
`blue or Cy5PE-conjugated anti-mouse Ter119 (eBioscience) and
`analyzed on a FACS Aria. In addition, PE-conjugated human
`CD13, CD14, and CD33 (BD Bioscience) and Cy5PE-
`
`conjugated human CD3 and CD19 (Caltag) were used for
`lineage analysis of engrafted human cells. When hCD45-positive
`cells were detected, their morphology was examined by Wright–
`Giemsa staining of a cytospin preparation.
`
`We thank Toshio Kitamura (Tokyo University, Japan) for kind gifts of
`pMX-SST vector and Plat-E virus producer cell line; H. Ueno for
`technical advice; R. Majeti and B. Tan for critical reading of manuscript
`and great discussion; L. Jerabek for excellent laboratory management; C.
`Muscat for antibody preparation; L. Hidalgo, D. Escoto, and J. Dollaga
`for animal care; and all the members of I.L.W.’s laboratory for great
`discussions and technical help. This work was supported by the National
`Institutes of Health Grants CA55209 and CA86017 (to I.L.W.), the
`Walter and Beth Weissman Foundation, the Smith Family Fund, the
`Floren Fund, the Japanese Society of Promotion of Science Fellowship
`(N.H.), the Yamada Memorial Foundation, and the Mitsubishi Pharma
`Research Foundation.
`
`1. Reya T, Morrison SJ, Clarke MF, Weissman IL (2001) Nature 414:105–111.
`2. Passegue E, Jamieson CMH, Allies LE, Weissman IL (2003) Proc Natl Acad Sci
`USA 100:11842–11849.
`3. Warner JK, Wang JCY, Hope KJ, Jin L, Dick JE (2004) Oncogene 23:7164–
`7177.
`4. Huntly BJ, Gilliland DG (2005) Nat Rev Cancer 5:311–321.
`5. Weissman IL (2005) J Am Med Assoc 294:1359–1366.
`6. Miyamoto T, Weissman IL, Akashi K (2000) Proc Natl Acad Sci USA
`97:7521–7526.
`7. Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B (1992) Proc
`Natl Acad Sci USA 89:2804–2808.
`8. Jordan CT, Upchurch D, Szilvassy SJ, Guzman ML, Howard DS, Pettigrew AL,
`Meyerrose T, Rossi R, Grimes B, Rizzieri DA, et al. (2000) Leukemia
`14:1777–1784.
`9. Jordan CT, Guzman ML, Noble M (2006) N Engl J Med 355:1253–1261.
`10. Maloney DG, Liles TM, Czerwinski DK, Waldichuk C, Rosenberg J, Grillo-
`Lopez A, Levy R (1994) Blood 84:2457–2466.
`11. Sievers EL, Larson RA, Stadtmauer EA, Estey E, Lowenberg B, Dombret H,
`Karanes C, Theobald M, Bennett JM, Sherman ML, et al. (2001) J Clin Oncol
`19:3244–3254.
`12. Keating MJ, Flinn I, Jain V, Binet JL, Hillmen P, Byrd J, Albitar M, Brettman
`L, Santabarbara P, Wacker B, et al. (2002) Blood 99:3554–3561.
`
`13. Smith MR (2003) Oncogene 22:7359–7368.
`14. Kojima T, Kitamura T (1999) Nat Biotechnol 17:487–490.
`15. Bonnet D, Dick JE (1997) Nat Med 3:730–737.
`16. Wang PL, O’Farrell S, Clayberger C, Krensky AM (1992) J Immunol 148:2600–
`2608.
`17. Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, Caceres-Cortes J,
`Minden M, Paterson B, Caligiuri MA, Dick JE (1994) Nature 367:645–648.
`18. Gramatzki M, Ludwig WD, Burger R, Moos P, Rohwer P, Grunert C, Sendler
`A, Kalden JR, Andreesen R, Henschke F, et al. (1998