See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/47532416

Critical roles of Bcl11b in T-cell development and maintenance of T-cell identity. Immunol Rev

ARTICLE in IMMUNOLOGICAL REVIEWS · NOVEMBER 2010 Impact Factor: 10.12 · DOI: 10.1111/j.1600-065X.2010.00953.x · Source: PubMed

CITATIONS READS 37 37

3 AUTHORS, INCLUDING:

Peng Li Chinese Academy of Sciences

9 PUBLICATIONS 218 CITATIONS

SEE PROFILE

Available from: Peng Li Retrieved on: 14 December 2015 Pentao Liu Critical roles of Bcl11b in T-cell Peng Li development and maintenance of Shannon Burke T-cell identity

Authors’ address Summary: T-cell development primarily occurs in the thymus and 1 1 1 Pentao Liu , Peng Li , Shannon Burke involves in the interactions of many important transcription factors. Until 1 Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK. recently, no single has been identified to be essential for T-cell lineage commitment or maintenance of T-cell identity. Recent Correspondence to: studies have now identified the zinc finger transcription factor Bcl11b to Pentao Liu be essential for T-cell development and for maintenance of T-cell iden- Wellcome Trust Sanger Institute tity. Remarkably, T cells acquire NK cell properties upon Bcl11b deletion. Hinxton These reprogrammed cells have unique properties in proliferation, cyto- Cambridge, CB10 1HH kine dependency and killing target cells, and may therefore provide a UK new cell source for some cell-based therapies. Tel.: +44 1223496850 Fax: +44 1223496802 Keywords: Bcl11b, T cells, transcription factor, identity, natural killer, reprogramming, e-mail: [email protected] immunotherapy

Introduction The Bcl11 family has two members, Bcl11a and Bcl11b, both being Kruppel-like C2H2 type zinc finger transcription factors (1). Bcl11a was first discovered as a retroviral insertion site (Evi9) in myeloid leukemia tumors in the BXH-2 mouse (2, 3). Evi9 was later renamed as Bcl11a, since it was found ectopi- cally expressed in some B-cell caused by chromo- somal translocations (1). Bcl11a is required for normal lymphoid development. Germline deletion of Bcl11a causes neonatal lethality and an absence of B cells at the earliest B-cell development stages (4). Surprisingly, Bcl11a deficiency also causes abnormal T-cell development. Recipient mice of Bcl11a- deficient fetal liver cells develop T-cell leukemia, suggesting that Bcl11a might have important functions in T-cell progeni- tors, since it is expressed in these cells (4–6). Recent genome- wide association studies in have revealed association of the BCL11A with persistent fetal hemoglobin in the adult (7, 8). In the subsequent validation assays, knocking down BCL11A in human primary adult erythroid cells indeed Immunological Reviews 2010 leads to robust HbF expression (9). Further characterization of Vol. 238: 138–149 Printed in Singapore. All rights reserved Bcl11a mutant mice also uncovers the key role of Bcl11a in the fetal-to-adult expression switch of hemoglobin (9, 10). 2010 John Wiley & Sons A/S Immunological Reviews Recent studies demonstrate that Klf1 positively regulates 0105-2896 Bcl11a in this process (11, 12).

138 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

Bcl11b is the other member of the Bcl11 family in the mouse transcription has been identified for T-cell lineage commit- and human genomes. The Bcl11b locus was once named Rit1 ment and ⁄ or identity maintenance. (radiation-induced tumor suppressor 1), because homozygous analysis in thymocytes reveals that Bcl11b deletions and point were located to this locus in a is the most upregulated transcription factor at the transition genome-wide allelic loss analysis of c-ray induced mouse thy- from DN1 to DN2, suggesting the potential function of mic lymphomas (13–15). Therefore, Bcl11b was initially Bcl11b in early T-cell development and its possible connection identified as a tumor suppressor in T cells. with Notch signaling (6, 42). This review summarizes recent T-cell development takes place in the thymus. The earliest studies that identify Bcl11b as the key factor required for populations of progenitor thymocytes lack T-cell T-cell lineage commitment and for the maintenance of (TCR) coreceptors CD4 and CD8, and are therefore referred T-cell identity. as double negative (DN) cells (16). The DN population can be further subdivided by cell surface markers CD117 (c- ) Kit), CD44, and CD25 (17). The DN1 (CD44+CD25 ) pop- Bcl11b is specifically expressed in T cells ulation of progenitor thymocytes are thought to contain Bcl11b is highly expressed in mouse and human T lympho- multipotent progenitors (18–22). Although NK and myeloid cytes (43, 44). During evolution, a homolog of Bcl11b first potentials persist in DN2 cells (CD44+CD25+) (21, 22), the appears in cartilaginous fishes. In sea lampreys, a jawless ver- non-T-cell developmental potentials are lost in the DN3 tebrate, specific expression of a Bcl11b ortholog is detected in ) ) ) (CD44 CD25+) thymocytes. DN4 thymocytes (CD44 CD25 ) VLRA+ cells that are similar to T lymphocytes in vertebrates, have undergone b-selection after successful TCRb gene rear- but not in VLRB+ cells that are similar to B lymphocytes (45). rangement (23) and have already initiated the process of dif- In bony fish, the Bcl11b ortholog is expressed in the thymus ferentiating to the CD4+CD8+ double positive (DP) stage and positively regulates Ccr9 (46), which encodes the receptor (24, 25). Once thymocytes survive the positive and negative for ccl25, a novel chemokine. Both Ccr9 and Ccl25 are now selection processes, they migrate to the peripheral lymphoid shown to have important functions in the recruitment of pro- tissues where the interleukin-7 (IL-7) and the con- genitors to the thymus and early thymocyte development (47, stant interaction of T cells with self-peptide-major histo- 48). compatibility complex (MHC) play a critical role in T-cell Bcl11b expression has been characterized at the single cell homeostasis (26). level in a reporter knockin mouse where the Td-Tomato cassette T-cell development requires a complex interplay among key is targeted to the 3¢ untranslated region (UTR) of the Bcl11b transcription factors with Notch signaling playing a critical locus (5). In hematopoietic lineages, Bc11b expression is role (27). Notch triggers initiation of the T-cell program, and strictly T-cell restricted with transient low levels expression in is required to sustain the cells throughout the pro-T-cell stages some immature natural killer (NK) cells. In the mouse thy- (28–30). Loss of Notch signaling in DN1 cell converts them mus, it is expressed at high levels throughout T-cell develop- into DCs (31). Deletion of Notch1 in the thymus leads to accu- ment including almost all DN2-DN4 and DP thymocytes, mulation of B cells in the thymus, possibly by a cell-extrinsic CD4+ and CD8+ single positive T cells, cd T cells, and natural pathway (28, 31). In committed T cells, Notch signaling killer T (NKT) cells. However, early T-cell-lineage progenitors favors ab versus cd T-cell lineage (32, 33), influences CD4+ (ETP), defined as CD117++DN1 thymocytes (27), did not versus CD8+ lineage decisions (34) and affects T-helper devel- express Bcl11b (5). Indeed, Bcl11b expression is undetectable opment partly through Gata3 (35, 36). in all other thymocytes that express CD117, such as some Other transcription factors that are essential for various T- DN2 thymocytes (5), suggesting that Bcl11b might suppress cell subtypes are also characterized. For example, zinc finger expression of c-kit (CD117), which is often found in progeni- transcription factor Th-POK (Zbtb7b), which appears to be tors. repressed by Runx complexes (37), regulates CD4+ versus In peripheral mature T cells, Bcl11b is expressed in both CD8+ T-cell lineage commitment (38). T-bet (Tbx21), a CD4+ and CD8+ T cells with lower levels of Bcl11b expression T-box transcription factor, directs T-helper 1 (Th1) lineage in CD8+ T cells. Interestingly, some activated T cells ) commitment (39). Gata3 is necessary and sufficient for Th2 (CD44+CD62L ) express very low levels of Bcl11b (5), sug- cytokine gene expression in CD4+ T cells (40). Eomesoder- gesting that it might have functions in TCR signaling, perhaps min, another T-box transcription factor, controls effector normally suppressing some activation-associated thus CD8+ T-cell function (41). Despite these advances, no single inhibiting T-cell activation. The specific and dynamic expres-

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 139 Liu et al Æ Bcl11b has essential functions in T cells

sion of Bcl11b in T cells supports the hypothesis that it has ligand Delta-like 4, stem cell factor, Flt3 ligand (Flt3L), the important roles in T-cell development and function. culture media also contains 10 ng ⁄ ml IL-7 (54). Surprisingly, decreasing IL-7 in the culture media from 10.0 ng ⁄ ml to 1.0 ng ⁄ ml leads to the development of these differentiation- Essential functions of Bcl11b in T-cell development arrested wildtype DN2 cells to T cells. Changes of Bcl11b Loss-of-function studies in the mouse demonstrates that expression levels are associated with the IL-7 concentration Bcl11b is required for the early T-cell development and the switch as well as the differentiation transition. Similar obser- survival of T lymphocytes (44). The Bcl11b homozygous vations are also made in Bcl11b-deficient DN2 thymocytes of mutant knockout mice die in the first few days after birth the adult thymus that lower IL-7 concentrations is beneficial (44), likely due to neurological or other uncharacterized for T-cell development on OP9-DL1 stromal cells (56). Once defects (49, 50), and thymocyte development is blocked at Bcl11b is reintroduced to the differentiation-arrested DN2 the DN2-DN3 stage without obvious defects in other hemato- cells, T-cell development resumes regardless of IL-7 con- ) ) poietic lineages (44). In Bcl11 ⁄ thymocytes, Vb to Db rear- centration (54). These results thus suggest that in normal rangements is impaired which could contribute to the lack of T-cell development, Bcl11b could be the key transcription expression of pre-TCR complex, which in turn leads to the factor that terminates non-T-lineage potential in T-cell profound apoptosis in the thymocytes. Apoptosis is not progenitors in order for them to become committed to the believed to be the main reason for the failure of T-cell devel- T-cell lineage. opment upon loss of Bcl11b, because inactivation of is not Further analysis in the adult thymus demonstrates that acute sufficient to restore further T-cell development of these deletion of Bcl11b in T-cell progenitors (DN1 thymocytes) mutant thymocytes, even though some immature single-posi- stops T-cell development at the DN2 stages, consistent with tive (ISP) T cells are detected (51). The exact cause of T-cell developmental block at the DN2 stage observed in the Bcl11b- defects in the Bcl11b mutant mice thus remained unresolved. deficient fetal thymocytes (56). Detailed characterization of Nevertheless, the importance of Bcl11b in T-cell development these Bcl11b-deficient DN2 thymocytes shows that they already is clear, and is further demonstrated by Bcl11b haploinsuffi- express many T-cell-associated transcription factors such as ciency in thymocyte development as the heterozygous mutant Gata3 and Tcf7 but still retain expression of genes such as mice have roughly half the number of thymocytes compared Tal1, Bcl11a, Sfpi1, Erg, and Flt3 that are normally turned off to the wildtype littermates and are much more prone to lym- during or after the DN2 stage. Moreover, these DN2 cells have phomagenesis (52). In c-irradiated mice, loss of Bcl11b is tremendous proliferation potentials, suggesting that Bcl11b proposed to promote clonal expansion and differentiation suppresses the stemness in the early T-cell progenitors. The arrest of thymocytes (53). proliferation potential of these cells however is not consistent The Bcl11b-deficient DN2 thymocytes were recently further with Bcl11b suppressing some cell cycle regulators such as characterized and found to be able to proliferate extensively in p57 (57). Nevertheless, these new results indicate that loss of T-cell culture conditions (54). These Bcl11b-deficient thymo- further T-cell differentiation potential upon Bcl11b deletion cytes, however, are not found expanded in vivo, possibly due make these T-cell progenitors a potential target for further to limited niche space for extra progenitors in the thymic cor- accumulation of genetic mutations which eventually leads to tex, although wildtype DN2 thymocytes are characterized by tumorigenesis (5, 54, 56). Indeed, even loss of one allele of active proliferation (55). Another possibility is that their pro- Bcl11b makes some thymocytes lose differentiation potential, liferation in vitro is artificially enhanced by the culture condi- and these thymocytes are considered to resemble stem tion with high levels of IL-7 (10 ng ⁄ ml). Bcl11b-deficient cells (53). Interestingly, however, it was puzzling that no DN2 thymocytes do not develop into T cells, rather they retain tumor development has ever been reported in mice trans- the differentiation potential to several lineages including NK planted with Bcl11b homozygous knockout mutant fetal liver cells, dendritic cells, and macrophages but not B cells, similar stem cells. This might now be explained by the presence of to the early T-cell progenitors (21, 22). ITNKs in these recipient mice as described below. Similar cells which are differentiation-arrested yet prolifer- The Bcl11b-deficient T-cell precursors in the adult thymus ate rapidly could also be obtained by culturing wildtype DN2 appear to initiate T-cell specification in a manner similar to cells in stromal cell-free culture conditions (54). Failure wildtype controls, as indicated by the upregulation of Cd3e, of these cells to differentiate to T cells is not due to TCR Cd3g, Ptcra, and Rag1 between the DN1 and DN2a stages. rearrangement defects. Besides immobilized Notch receptor The inability of the cells to differentiate further along the

140 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

T lineage may thus be explained not by a failure to initiate thymocytes efficiently develop to late stage T cells on OP9- T-cell program activation but by a failure to repress alternative DL1 stromal cells (59). Once OHT is added to the culture and lineages, even under conditions of Notch and cytokine signal- thus Bcl11b is deleted, no cells expressing T-cell markers, such ing (56). However, it remains possible that some Bcl11b-defi- as CD3or TCRb, are found in the culture even after a couple of cient DN2 cells are differentiated further down to DN2b or weeks. Surprisingly, OP9-DL stromal cells in the culture dishes DN3 transiently, but the differentiation status is not main- are killed. Indeed, some cells express NK cell surface markers tained. Consequently, the further differentiated thymocytes such as NK1.1 and NKp46 (60). These results demonstrate could undergo apoptosis. Indeed, in the Bcl1b-deficient fetal again that Bcl11b-deficient DN1–2 thymocytes have either thymus, many thymocytes undergo apoptosis. Interestingly, failed to acquire, or acquired but subsequently lost, T-cell besides the developmental block, the Bcl11b-deficient DN2 features due to failure of maintenance and that they are unable thymocytes also express genes such as Id2, Il2rb, and Nfil3 to suppress non-T lineages such as NK cells. IL-2 or IL-15 that normally promote NK cell development, indicating that are known to promote NK cell development. Sup- Bcl11b suppresses NK development from the progenitor cells, plying IL-2 or IL-15 in the culture media greatly promotes besides promoting the T-cell lineage (56). proliferation and ⁄ or differentiation of Bcl11b-deficient DN1–2 thymocytes cells along NK cell lineage. Indeed, acute loss of Bcl11b leads to an increased IL-2 production in T cells in gene Bcl11b is required for both early T-cell development expression microarray analysis (5). and T-cell identity maintenance of committed T cells DN3 thymocytes are committed T cells and have lost poten- The critical role of Bcl11b in all T-cell compartments is exten- tials to differentiate to other non-T-cell lineages (21, 22). sively studied in our laboratory using CreERT2; Bcl11bflox ⁄ flox Similar to Bcl11b-deficient DN1 and DN2 thymocytes, the (flox ⁄ flox) mice, in which the tamoxifen (OHT)-inducible Cre mutant DN3 cells are also reprogrammed to NKp46+ NK-like recombinase is expressed from the ubiquitously expressed cells (5). The reprogramming process appears to be an intrin- Rosa26 locus (58). Bcl11b could be conveniently deleted in cul- sic property of the mutant thymocytes as NK-like cells are tured cells from this mouse by adding OHT to the culture readily produced in B or myeloid cell culture media. These media (Fig. 1). Alternatively, tamoxifen is directly adminis- NK-like cells retains V(D)J recombination at the TCRb locus, trated to the CreERT2; Bcl11bflox ⁄ flox mice to activate Cre recom- despite having no TCRb or CD3 detected on their cell surface, binase and delete Bcl11b. This mouse thus allows examination thus genetically confirming the T-cell origin of these ) of Bcl11b function in each and every compartment NKp46+CD3 cells from DN3 thymocytes. We named the (Fig. 1). In culture media containing IL-7 and Flt3L, wildtype NK-like cells that are derived from Bcl11b-deficient T cells as or Bcl11b conditional knockout (untreated) DN1 or DN2 induced-T-to-NK (ITNK) cells.

Fig. 1. Reprogramming mouse T cells to ITNKs upon Bcl11b deletion. ITNKs have been produced using three approaches. The CreERT2; Bcl11bflox ⁄ flox mice are treated with tamoxifen to delete Bcl11b. ITNKs are found in peripheral blood, the spleen, and the thymus of the treated mice in a couple of weeks. Alternatively, Bcl11b is also deleted in vitro in thymocytes. Whole thymocytes from the CreERT2; Bcl11bflox ⁄ flox mice are treated with tamoxifen, which are then sorted into different subsets and cultured on stromal cells for ITNK production. DP thymocytes sorted from these whole thymocytes ) ) ) ) are also transferred into Rag2 ⁄ Il2rg ⁄ mice by i.v. injection. ITNKs are detected in these recipient mice a couple weeks after the injection.

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 141 Liu et al Æ Bcl11b has essential functions in T cells

Besides NKp46, these in vitro culture-derived ITNK cells express other genes such as NKG2A ⁄ C ⁄ E, TRAIL, perforin and ITNKs are efficiently produced in vivo from T cells interferon-c, which are highly expressed in NK cells, but not Loss of T-cell identity and acquisition of NK cell properties some other key NK function genes, such as members of the upon Bcl11b deletion are not simply in vitro culture artifacts. Ly49 family or FasL. Morphologically, ITNK cells resemble Once flox ⁄ flox mice are injected with tamoxifen to delete NK cells rather than their parental DN3 thymocytes. They have Bcl11b, ITNKs, which include NKp46+CD3+ cells in the spleen ) larger cytoplasm and more granules, showing evidence of and both NKp46+CD3+ and NKp46+CD3 cells in the thy- high synthesis activity with abundant endoplasmic mus, are readily found (Fig. 1). These in vivo-derived ITNKs are reticulum. Interestingly, although ITNK cells kill foreign cells not CD1d-restricted NKT cells, which also do not express such as OP9-DL1 stromal cells and NIH3T3 cells very potently, NKp46. The decrease in the percentage and the absolute num- they do not attack cells from the same mouse, demonstrating ber of CD1d-restricted NKT cells upon Bcl11b deletion actually their self-recognition ability (5). indicates that Bcl11b has an important role in NKT cells. Fur- Besides morphology, the global gene expression profile of thermore, similar to ITNKs from cultured thymocytes, the ITNKs is much more similar to that of normal LAK cells (lym- in vivo reprogrammed ITNKs (both NKp46+CD3+ and the ) phokine-activated killer cells), composed of >90% NK cells, NKp46+CD3 ) also retain much higher proliferation than NK than that of their parental DN3 thymocytes (5). Expression of cells and could readily be expanded in NK cell culture condi- many T-lineage genes such as Notch1, Est1, Hes1, Gata3, tion, even though some of them are TCRb positive. Impor- Dtx1, and Tcf1 is decreased, while expression of some key NK tantly, these tamoxifen-injected mice did not develop any cell genes such as Id2, Zfp105, and E4bp4 is greatly abnormalities in up to 2 years after tamoxifen injection, sug- upregulated. Zbtb32 [Rog (Repressor of GATA)], which plays gesting that ITNKs spare normal cells and that Bcl11b-deficient important roles in negatively regulating T-cell activation (61), T cells (ITNKs) are not transformed. Although ITNKs from is highly expressed in ITNKs. Cell cycle regulator Cdkn1c mature T cells often retain TCRb and CD3 on the cell surface, (p57KIP2) is almost undetectable in DN3 thymocytes, but its TCR signaling appears to be compromised to some extent (5). expression is drastically increased in ITNKs, consistent with Reprogramming from T cells to ITNKs upon loss of Bcl11b the previous notion that it is a Bcl11b target gene candidate is a cell-autonomous process in vivo. ITNKs are also found in ) ) ) ) (57). the Rag2 ⁄ Il2rg ⁄ recipient mice that are injected with ) ) ) ) Similar to the tremendous proliferation potential of Bcl11b-deficent DP thymocytes (5) (Fig. 1). Rag2 ⁄ Il2rg ⁄ Bcl11b-deficient DN1–2 thymocytes, the DN3 thymocyte- mice have no B, T, or the vast majority of NK cells (62). derived ITNKs also show extensive in vitro expansion poten- ITNKs are readily detected in the spleen, bone marrow, and tial. From one mutant DN3 thymocyte, up to 0.5 million peripheral blood of the recipient mice. Many of the ITNKs stromal-killing cells could be obtained in 2–3 weeks in the express TCRb, CD3, and CD8. They also express Ly49 family presence of IL-2 in the media (5). Without IL-2, about genes, NKG2A ⁄ C ⁄ E, TRAIL, perforin, and interferon-c.The 50 000 ITNK cells were obtained from a single DN3 total number of ITNK cells does not vary dramatically for at thymocyte under the same culture conditions. These ITNKs least 3 months, perhaps representing a homeostasis of ITNKs are nevertheless able to kill stromal cells, demonstrating that in the in vivo microenvironment or due to the space limitation exogenous IL-2 or IL-15 is not essential for the generation in the in vivo niche. Importantly, neither recipient mice with and function of ITNKs in vitro. In the same single cell assay, ITNKs derived from flox ⁄ flox mice nor tamoxifen-treated flox ⁄ it is found that essentially each and every Bcl11b-deficient flox mice themselves show any noticeable autoimmune DN3 thymocytes is reprogrammed to ITNKs (5). It remains diseases or tumors, again indicating that ITNK cells do to be determined whether this is also true for other T cells. not indiscriminately kill normal host cells or do not have Nevertheless, ITNKs are efficiently produced from Bcl11b- malignant transformation. deficient DP and CD8+ T cells in cell culture. These ITNKs, Acute loss of Bcl11b in T cells leads to failure of T-cell line- in contrast to those reprogrammed from DN1–3 thymo- age commitment and of T-cell identity maintenance. Constitu- cytes, retained TCRb and CD3 on the cell surface. Similar to tive expression of Cre recombinase in T cells, however, leads the DN3-derived ITNKs in vitro, they express NKG2A ⁄ C ⁄ E, to somewhat different phenotypes. For example, deletion of TRAIL, perforin and interferon-c but not some other key Bcl11b in early DP thymocytes using CD4-Cre caused expres- NK function genes, such as members of the Ly49 family sion of some of the genes found in mature SP T cells such as (5). Zbtb7b (Th-POK) and Runx3 (63). But no NK cell gene

142 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

expression is reported. In DP thymocytes, Bcl11b deletion mature T cells are readily converted to ITNKs. However, it is using CD4-Cre causes defects in initiation of positive selection possible that a de-differentiation step may still exist upon which includes impaired proximal TCR signaling, attenuated Bcl11b loss in T cells. The process could be transient or take a extracellular signal-regulated kinase phosphorylation and relatively longer time for some T-cell subsets. For example, IT- calcium flux, and increased susceptibility to spontaneous NKs from mature T cells may eventually become more similar apoptosis (64). If Bcl11b is deleted using dLck (Lck distal) pro- to ITNKs from DN1 thymocytes over time. moter–iCre, which is primarily expressed in peripheral CD8+ Bcl11b is also called CTIP2, because it was independently iso- T cells, both antigen-specific clonal expansion and cytolytic lated for its interaction with all members of the chicken ovalbu- effector function of CD8+ T lymphocytes are impaired (65). min upstream transcription factor (COUP-TF) Besides loss-of-function studies in the mouse, Bcl11b is also subfamily of orphan nuclear receptors (77). COUP-TFs usually implicated in human T-cell leukemia. A specific cryptic trans- mediate transcriptional repression by recruiting nuclear recep- location, t(5;14)(q35.1;32.2), present in some pediatric tor co-repressor (NCoR) and ⁄ or silencing mediator for retinoid T-ALL patients and T-ALL cell lines activates expression and thyroid (SMRT) to the template (78, of HOX11L2 by juxtaposition with strong T-cell enhancer 79). COUP-TF family members play important roles in devel- elements at the 3¢ of the BCL11B locus (43, 66, 67). Addition- opment (80–82). As a transcription factor, Bcl11b is also func- ally, a novel chromosomal aberration, inv(14)(q11.2q32.31), tionally associated with the nucleosome remodeling and was reported in T-ALL samples. In this inversion, the 5¢ part histone deacetylase (NuRD) complex to repress targeted pro- of BCL11B, including exons 1–3, was joined to the TRDD3 moters. Further analysis shows that both metastasis-associated segment of the TCRd locus. Consequently, in-frame transcripts MTA1 and MTA2 interact directly with BCL11B (83). with truncated BCL11B and TCRd constant region (TRDC) Interestingly, MTA1 and MTA2 are also Bcl11a-interacting pro- were highly expressed in screened T-ALLs samples but not in teins in erythrocytes (9). In HEK293 cells, Bcl11b recruits sir- normal T cells (68). Interestingly, although BCL11B is consid- tuin 1 (SIRT1), a trichostatin-insensitive, nicotinamide- ered as a , high levels of BCL11B mRNA sensitive class III histone deacetylase, to the promoter region of expression is found in the majority of T-ALL and T-cell leuke- a reporter gene template (84, 85). At the HIV-1 promoter mia ⁄ cell lines (67, 68). Suppression of BCL11B by region, Bcl11b recruits histone deacetylase 1 (HDAC1) and RNA interference causes apoptosis of these tumor cells, which HDAC2 to promote local histone H3 deacetylation (86). is possibly caused by the decrease of an anti-apoptotic protein, Indeed, deletion of Bcl11b in T cells in the first 24 h leads to gene BCL-xL, and involvement of the mitochondrial apoptotic path- expression changes associated with chromatin remodeling (5). way (69–71), whereas normal mature T cells are unaffected In human CD4+ T lymphocytes, BCL11B can also act as a (70). Therefore, BCL11B could be an attractive therapeutic transcription activator. A decrease in the level of endogenous RNAi target for treating T-cell malignancies. BCL11B reduces the level of expression of IL-2, while over- expression of BCL11B augments its expression. Further ChIP- assays show BCL11B promotes IL-2 gene expression by bind- Bcl11b in T-cell development and identity maintenance: ing and activating the IL-2 promoter following activation downstream target genes and upstream regulators through TCR (87). Additionally, BCL11B can also activate IL-2 Master regulators that are able to promote a specific cell lineage expression by enhancing NF-jB (nuclear factor jB) activity in and that are required to maintain lineage identity have been the context of TCR ⁄ CD28-triggered T-cell activation (87, 88). identified for several hematopoietic cell lineages. For example, The precise in vivo binding site of Bcl11b is still elusive. ectopically expressing Cebpa in pro-B and pro-T cells trans- Besides IL-2, Bcl11b can apparently regulate cell cycle by sup- forms them into macrophages at a frequency of around 60% pressing p21 and p57 expression. CDKN1A ⁄ p21(WAF1) (72, 73). Fibroblast cells expressing MyoD are converted to (p21), a cyclin-dependent kinase inhibitor, is a major cell myogenic colonies at 25–50% efficiency (74). Additionally, cycle regulator of the response to DNA damage, senescence, loss of Pax5 in B cells enables de-differentiation of B cells to and tumor suppression. Bcl11b (CTIP2) is a constitutive p21 become multi-potent progenitors (75). Bcl11b currently is the gene suppressor that cooperates with SUV39H1 and histone only transcription factor known for T-cell identity mainte- methylation to silence the p21 gene transcription (89). nance. However, unlike de-differentiation in B cells upon loss Expression of another cyclin-dependent kinase inhibitor, of Pax5 (76), deletion of Bcl11b in T cells does not appear to CDKN1C (p57KIP2), is also shown to be repressed by Bcl11b have obvious de-differentiation steps, because both pro-T and in association with the NuRD complex (57). Indeed, ITNKs

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 143 Liu et al Æ Bcl11b has essential functions in T cells

express high levels of these cell cycle inhibitors (5), a fact ing (101). The fact that ITNKs instead of T cells grow out apparently representing a paradox because ITNKs have enor- from Bcl11b-deficient DN1 and DN2 thymocytes on either mous proliferation potential in vitro. Perhaps the effects of cell OP9 or OP9-DL1 stromal cells demonstrates that loss of T-cell cycle regulation by Bcl11b are more obvious on ITNKs in vivo. development potential and acquisition of NK cell properties Loss of Bcl11b in T cells leads to expression of many genes upon Bcl11b deletion is Notch signaling independent and sug- important for NK cell development or function. An increase of gests that Bcl11b could act downstream of Notch signaling in expression of E4bp4, Il2rb, IL-2, and granzyme B (GZMB)is T cells. ChIP analysis using CSL antibodies confirms that the detected in microarray analysis 48 h after Bcl11b deletion (5). canonical Notch signaling directly regulates Bcl11b at the Il2rb, which encodes CD122 (a receptor for IL-2 and IL-15), is transcription level (5). Although Bcl11b acts downstream of considered as a NK cell surface marker and is required for NK Notch signaling, it could also be regulated by other factors or cell proliferation and differentiation (90). The helix-loop- signaling pathways in T-cell development and identity main- helix transcription factor Id2, which antagonizes the basic tenance, since loss of Bcl11b in T cells does not have the same helix-lool-helix E proteins E2A and HEB, is essential for full phenotype as loss of canonical Notch signaling. Therefore, NK lineage development, since the Id2-knockout mice exhibit Bcl11b is likely regulated by additional factors in T cells. As a severe peripheral NK cell deficiency (91, 92). Conversely, described earlier, IL-7 signaling is proposed to affect Bcl11b forced expression of Id2 or Id3 is able to redirect pro-T cells (54). In the stromal-free culture media with an immobilized to NK cell differentiation (93, 94). Acting in a cell-intrinsic Notch ligand, higher concentration of IL-7 blocks DN2 cells manner ‘downstream’ of the IL-15 receptor and through Id2, further differentiation into T cells. This block can be overcome the basic (bZIP) transcription factor E4bp4 by either lowering the IL-7 level or by forcibly expressing (Nfil3) is indispensable for production of NK cell lineage, Bcl11b. Peripheral T cells require IL-7 signaling for survival because overexpression of E4bp4 promotes NK cell generation and homeostasis, but ITNKs from DN3 thymocytes or mature from hematopoietic progenitor cells while lack of E4bp4 T cells, which no longer express Bcl11b, apparently do not impairs NK cell development and NK cell-medicated cytotox- require IL-7 to proliferate in vitro. Therefore, it is possible that icity (95, 96). Additionally, deletion of Pu.1, CBFb,orEts-1 IL-7 signaling interacts with Bcl11b in T cells. How IL-7 con- adversely affects NK cells, but this is likely due to their impor- trols Bcl11b activities is not clear from the current data (102). tant roles in the lymphoid lineages (97–99). Recently, a NK- specific zinc finger transcription factor, Zfp105, was shown to promote differentiation from hematopoietic stem cells to the ITNKs: a new cell source for cancer immunotherapy NK lineage (100). These key NK cell genes are all overexpres- applications sed in ITNKs or in Bcl11b-deficient thymocytes detected either The immune system is alert to the presence of tumors, so it in reverse transcriptase polymerase chain reaction or in micro- functions as an extrinsic tumor suppressor in naive hosts. Can- array analysis (5, 54, 56). However, it remains to be deter- cer immunosurveillance in principle could be exploited for mined whether these genes are the direct transcription use in cancer therapies. In particular, there have been intensive downstream targets of Bcl11b. To comprehensively identify efforts made by many groups in the area of cellular therapy to the potential target genes regulated by Bcl11b, ChIP-Seq not only identify how best to harness the immune response should be performed using Bcl11b antibodies in wildtype and for use against tumors but also why most immunotherapy Bcl11b-deficient T cells. The ChIP-Seq predictions should be fails to control tumors (103, 104). validated by expression analysis and functional confirmation. Two general approaches are taken in cancer immunotherapy. Which signaling pathways regulate Bcl11b? Notch signaling First, many vaccine formulations have been trialed to stimulate has critical roles at several stages of T-cell development and the endogenous anti-tumor T-cell response. Recent approaches normally suppresses NK lineage from the T ⁄ NK progenitors. have included DNA and viral vaccines, dendritic cells pulsed Consequently, OP9-DL1 stromal cells suppress the growth of with tumor peptide (105), and tumor cells engineered to NK cells from bone marrow cells or thymocytes if exogenous express immunostimulatory factors such as cytokines (106). IL-2 or IL-15 is not added in the culture. Previous studies pro- However, the overall response rates of patients given cancer pose that Bcl11b is regulated by Notch signaling during T-cell vaccines has been estimated to be as low as 3.3% (103), and specification and commitment based on gene expression (6). many are only as good as IL-2 therapy alone (107). Recently, CG6530, the Drosophila ortholog of Bcl11 genes was An attractive alternative strategy to cancer vaccines is adop- identified as a direct downstream target gene of Notch signal- tive cell transfer (ACT), the use of T cells or tumor-infiltrating

144 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

lymphocytes in cell-based cancer immunotherapy (27). In a the inhibitory receptors on NK cells and HLA on the tumor landmark study, the transfer of autologous LAK cells, a mix of cell surface (119). IL-2-stimulated cytotoxic cells, together with high-dose IL-2, Despite the wealth of experience acquired with cell-based to patients with metastatic solid tumors gave objective immunotherapy, several important hurdles remain before the responses in almost 50% of cases and increasing survival full clinical potential of these strategies can be realized. Adop- (108, 109). ACT strategies can be significantly improved in tive NK cell therapy demands that large numbers of cells with the context of patient lymphodepletion regimes, which are a well defined phenotype and high purity need to be gener- believed to eliminate immunosuppressive T-regulatory cells ated. The isolation, culture, and expansion of clinical NK cell and afford ‘space’ for transferred cells by removing endoge- products has been hampered by their relatively low represen- nous cells which compete for survival factors (110). In addi- tation in the blood and the fact that their in vitro proliferation tion, current approaches have begun enhancing the specificity potential in IL-2 and IL-15 is limited (120), despite of recent of transferred cells by genetically modifying lymphocytes advances (117, 121). in vitro by introducing high-affinity tumor-specific TCR genes Another important consideration is that the transferred cells (111) or enhancing other lymphocyte functions (112). must persist long enough to have a substantial tumor-killing In T-cell-based therapy, direct immunological pressure effect (122). To support transferred T and NK cells, exoge- from T cells on tumor cells (immunoediting) can lead to nous cytokines, typically IL-2, are used. Even with IL-2, trans- outgrowth of tumor clones which express low or no HLA ferred NK cells and autologous T cells are generally only antigens (113–115) or that have lost the targeted antigen detectable for around 2 weeks (116, 123). Moreover, clinical (116). These evading tumor cells that have low or no HLA use of IL-2 is usually associated with severe side effects (109), antigens nevertheless are ideal targets for the use of NK cells and it also supports the expansion of T-regulatory cells that in cellular cancer immunotherapy (117). NK cells detect and may counteract effector cells (124–126). kill tumor variants that lacked major histocompatibility com- As described above, the unique properties of ITNKs can plex class I expression (118). In some settings, the use of NK make them an attractive cell source for cancer immunotherapy cells in cancer immunotherapy is more effective with (Fig. 2). A large number of ITNKs could readily be obtained alloreactive NK cells as this exploits the mismatch between from either thymocytes or from mature peripheral T cells. The

Fig. 2. A potential platform for the production and application of human ITNKs. Autologous T cells from PBMCs or endogenous tumor-infiltrat- ing T cells from the patient or allogeneic thymocytes are isolated and cultured in the in vitro reprogramming conditions. Several alternative approaches to modify the T cells, including Bcl11b inactivation, conditioning in optimized cocktails of growth factors, and insertion of tumor-specific Tg- TCR ⁄ chimeric receptors in this reprogramming platform. Then, ITNKs are validated for their killing and self-tolerance capacities. Selected ITNKs are expanded and infused into patients where these ITNKs encounter and kill tumor cells. Reduction in the tumor volume by appropriate preparative regimes, including radiotherapy or chemotherapy and the co-administration of cytokine adjuncts may potentiate ITNK efficacy.

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 145 Liu et al Æ Bcl11b has essential functions in T cells

Bcl11b-deficient thymocytes exhibit enormous proliferation Mouse and human Bcl11b proteins are highly similar. Knock- potentials (5, 54, 56). From one DN3 thymocytes, we are able ing down Bcl11b in human T-cell leukemia cells shows that to obtain a half million ITNKs. Reprogramming of T cells to Bcl11b is required for their survival. We expect that inactivat- ITNKs occurs even in the absence of exogenous cytokines. ing human BCL11B in T cells would lead to production of ITNKs (in vivo reprogrammed) also appear to be much more ITNKs. It will be interesting to test the functions of human potent killers for certain tumor cells than regular LAK cells in ITNKs in tumor killing ability and other biological properties, the in vitro killing assay, and these cells are able to prevent relevant to cancer immunotherapy. Are human ITNKs able to melanoma metastasis in the mouse. ITNK killing appears to be kill human tumors in vitro and in immune-compromised mice? MHC-I-independent. However, they clearly can distinguish Is this killing HLA-independent? If human ITNKs behave normal cells from tumor cells, since mice transplanted with similarly to their mouse counterparts, then ITNKs may indeed ITNKs and flox ⁄ flox mice treated with OHT to delete Bcl11b do provide a unique cancer immunotherapy cell source (Fig. 2). not develop any obvious abnormalities related to ITNK cells. Before ITNKs can be used for any immunotherapy, however, several issues must be addressed. First, we are not entirely cer- Concluding remarks tain about the persistence of ITNKs in the mouse. It is clear that Bcl11b is now identified to be the critical transcription factor ITNKs are able to survive in the mouse for at least 3 months. for T-cell development and for maintenance of T-cell identity. Importantly, these cells are not transformed as mice harboring T-cell lineage commitment at DN2 to DN3 stages is lost once ITNKs do not develop tumors. Do ITNKs reprogrammed from Bcl11b is gone, which leads to expression of genes of alterna- different T-cell subsets have different in vivo half-life? Second, tive lineages in the T-cell progenitors. Deletion of Bcl11b in are these ITNKs similar in terms of tumor cell killing potential? committed or mature T cells causes loss or decreased expres- Since ITNKs derived from CD8+ T cells still retain the TCR sion of T-cell genes with concomitant expression of genes complex on the cell surface and TCR signaling in these ITNKs usually associated with NK cells. These NK-like cells (ITNKs) still appears to be functional, can the TCR signaling still be uti- reprogrammed from T cells are remarkable in that they have lized for targeting ITNKs to tumor cells expressing specific enormous proliferation potential in vitro, have a killing ability antigens? In this case, engineering a tumor-specific chimeric for tumor cells that is major histocompatibility complex-inde- receptor or transgenic TCR may enhance ITNK killing potency. pendent, and yet do not kill normal cells. Therefore, ITNKs Alternatively, this could be achieved by obtaining ITNKs from may provide a new cell source for cell-based therapies. Identi- T cells enriched for tumor-specific CD8+ cells following their fication of downstream target genes and upstream regulators isolation from tumor-infiltrating lymphocyte preparations. of Bcl11b will enable better understanding of T-cell develop- Third, does BCL11B act in an analogous way in human T cells? ment and maintenance of T-cell identity in the future.

References

1. Satterwhite E, et al. The BCL11 gene fam- 6. Tydell CC, David-Fung ES, Moore JE, Rowen 10. Sankaran VG, et al. Developmental and spe- ily: involvement of BCL11A in lymphoid L, Taghon T, Rothenberg EV. Molecular dis- cies-divergent globin switching are driven malignancies. Blood 2001;98:3413–3420. section of prethymic progenitor entry into by BCL11A. Nature 2009;460:1093–1097. 2. Nakamura T, Largaespada DA, Shaughnessy the T lymphocyte developmental pathway. J 11. Borg J, et al. Haploinsufficiency for the ery- JD Jr, Jenkins NA, Copeland NG. Coopera- Immunol 2007;179:421–438. throid transcription factor KLF1 causes tive activation of Hoxa and Pbx1-related 7. Menzel S, et al. A QTL influencing F cell hereditary persistence of fetal hemoglobin. genes in murine myeloid leukaemias. Nat production maps to a gene encoding a zinc- Nat Genet 2010;42:801–805. Genet 1996;12:149–153. finger protein on 2p15. Nat 12. Zhou D, Liu K, Sun CW, Pawlik KM, Tow- 3. Nakamura T, et al. Evi9 encodes a novel Genet 2007;39:1197–1199. nes TM. KLF1 regulates BCL11A expression zinc finger protein that physically interacts 8. Uda M, et al. Genome-wide association and gamma- to beta-globin gene switching. with BCL6, a known human B-cell proto- study shows BCL11A associated with Nat Genet 2010;42:742–744. oncogene product. Mol Cell Biol persistent fetal hemoglobin and ameliora- 13. Wakabayashi Y, et al. Homozygous dele- 2000;20:3178–3186. tion of the phenotype of beta-thalassemia. tions and point mutations of the 4. Liu P, et al. Bcl11a is essential for normal Proc Natl Acad Sci USA 2008;105:1620– Rit1 ⁄ Bcl11b gene in gamma-ray induced lymphoid development. Nat Immunol 1625. mouse thymic lymphomas. Biochem Bio- 2003;4:525–532. 9. Sankaran VG, et al. Human fetal hemoglo- phys Res Commun 2003;301:598–603. 5. Li P, et al. Reprogramming of T cells to nat- bin expression is regulated by the develop- 14. Matsumoto Y, et al. Allelic loss analysis of ural killer-like cells upon Bcl11b deletion. mental stage-specific repressor BCL11A. gamma-ray-induced mouse thymic lym- Science 2010;329:85–89. Science 2008;322:1839–1842. phomas: two candidate tumor suppressor

146 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

gene loci on 12 and 16. function by the Notch pathway. Annu Rev lymphocytes. Nat Immunol 2003;4:533– Oncogene 1998;16:2747–2754. Immunol 2005;23:945–974. 539. 15. Shinbo T, et al. Allelic loss mapping and 30. Rothenberg EV. Negotiation of the T lineage 45. Guo P, et al. Dual nature of the adaptive physical delineation of a region harboring a fate decision by transcription-factor inter- immune system in lampreys. Nature putative thymic lymphoma suppressor gene play and microenvironmental signals. 2009;459:796–801. on mouse chromosome 12. Oncogene 2007;26:690–702. 46. Bajoghli B, et al. Evolution of genetic net- 1999;18:4131–4136. 31. Feyerabend TB, et al. Deletion of Notch1 works underlying the emergence of thymo- 16. Anderson G, Moore NC, Owen JJ, Jenkinson converts pro-T cells to dendritic cells and poiesis in vertebrates. Cell 2009;138:186– EJ. Cellular interactions in thymocyte devel- promotes thymic B cells by cell-extrinsic 197. opment. Annu Rev Immunol 1996;14:73– and cell-intrinsic mechanisms. Immunity 47. Wurbel MA, et al. The chemokine TECK is 99. 2009;30:67–79. expressed by thymic and intestinal epithelial 17. Godfrey DI, Kennedy J, Suda T, Zlotnik A. A 32. Washburn T, et al. Notch activity influences cells and attracts double- and single-positive developmental pathway involving four phe- the alphabeta versus gammadelta T cell line- thymocytes expressing the TECK receptor notypically and functionally distinct subsets age decision. Cell 1997;88:833–843. CCR9. Eur J Immunol 2000;30:262–271. of CD3)CD4)CD8) triple-negative adult 33. Wolfer A, Wilson A, Nemir M, MacDonald 48. Schwarz BA, Sambandam A, Maillard I, Har- mouse thymocytes defined by CD44 and HR, Radtke F. Inactivation of Notch1 man BC, Love PE, Bhandoola A. Selective CD25 expression. J Immunol 1993; impairs VDJbeta rearrangement and allows thymus settling regulated by cytokine and 150:4244–4252. pre-TCR-independent survival of early alpha chemokine receptors. J Immunol 18. Allman D, et al. Thymopoiesis independent beta lineage thymocytes. Immunity 2007;178:2008–2017. of common lymphoid progenitors. Nat 2002;16:869–879. 49. Arlotta P, Molyneaux BJ, Chen J, Inoue J, Immunol 2003;4:168–174. 34. Fowlkes BJ, Robey EA. A reassessment of the Kominami R, Macklis JD. Neuronal subtype- 19. Ikawa T, Kawamoto H, Fujimoto S, Katsura effect of activated Notch1 on CD4 and CD8 specific genes that control corticospinal Y. Commitment of common T ⁄ natural killer T cell development. J Immunol 2002; motor neuron development in vivo. Neuron (NK) progenitors to unipotent T and NK 169:1817–1821. 2005;45:207–221. progenitors in the murine fetal thymus 35. Amsen D, Antov A, Flavell RA. The different 50. Arlotta P, Molyneaux BJ, Jabaudon D, revealed by a single progenitor assay. J Exp faces of Notch in T-helper-cell differentia- Yoshida Y, Macklis JD. Ctip2 controls the Med 1999;190:1617–1626. tion. Nat Rev Immunol 2009;9:116–124. differentiation of medium spiny neurons 20. Michie AM, et al. Clonal characterization of 36. Ho IC, Tai TS, Pai SY. GATA3 and the T-cell and the establishment of the cellular archi- a bipotent T cell and NK cell progenitor in lineage: essential functions before and after tecture of the striatum. J Neurosci the mouse fetal thymus. J Immunol T-helper-2-cell differentiation. Nat Rev 2008;28:622–632. 2000;164:1730–1733. Immunol 2009;9:125–135. 51. Okazuka K, et al. p53 prevents maturation 21. Wada H, et al. Adult T-cell progenitors 37. Setoguchi R, et al. Repression of the tran- of T cell development to the immature retain myeloid potential. Nature scription factor Th-POK by Runx complexes CD4)CD8+ stage in Bcl11b) ⁄ ) mice. 2008;452:768–772. in cytotoxic T cell development. Science Biochem Biophys Res Commun 2005;328: 22. Bell JJ, Bhandoola A. The earliest thymic 2008;319:822–825. 545–549. progenitors for T cells possess myeloid line- 38. He X, et al. The zinc finger transcription 52. Kamimura K, et al. Haploinsufficiency of age potential. Nature 2008;452:764–767. factor Th-POK regulates CD4 versus CD8 T- Bcl11b for suppression of lymphomagenesis 23. Dudley EC, Petrie HT, Shah LM, Owen MJ, cell lineage commitment. Nature and thymocyte development. Biochem Bio- Hayday AC. T cell receptor beta chain gene 2005;433:826–833. phys Res Commun 2007;355:538–542. rearrangement and selection during thymo- 39. Szabo SJ, Kim ST, Costa GL, Zhang X, 53. Go R, et al. Bcl11b heterozygosity promotes cyte development in adult mice. Immunity Fathman CG, Glimcher LH. A novel tran- clonal expansion and differentiation arrest 1994;1:83–93. scription factor, T-bet, directs Th1 lineage of thymocytes in gamma-irradiated mice. 24. Petrie HT, Hugo P, Scollay R, Shortman K. commitment. Cell 2000;100:655–669. Cancer Sci 2010;101:1347–1353. Lineage relationships and developmental 40. Zheng W, Flavell RA. The transcription 54. Ikawa T, et al. An essential developmental kinetics of immature thymocytes: CD3, factor GATA-3 is necessary and sufficient for checkpoint for production of the T cell line- CD4, and CD8 acquisition in vivo and in vitro. Th2 cytokine gene expression in CD4 T age. Science 2010;329:93–96. J Exp Med 1990;172:1583–1588. cells. Cell 1997;89:587–596. 55. Ciofani M, Zuniga-Pflucker JC. The thymus 25. Nikolic-Zugic J, Moore MW, Bevan MJ. 41. Pearce EL, et al. Control of effector CD8+ T as an inductive site for T lymphopoiesis. Characterization of the subset of immature cell function by the transcription factor Annu Rev Cell Dev Biol 2007;23:463–493. thymocytes which can undergo rapid in vitro . Science 2003;302:1041– 56. Li L, Leid M, Rothenberg EV. An early T cell differentiation. Eur J Immunol 1989; 1043. lineage commitment checkpoint dependent 19:649–653. 42. David-Fung ES, et al. Transcription factor on the transcription factor Bcl11b. Science 26. Takada K, Jameson SC. Naive T cell homeo- expression dynamics of early T-lymphocyte 2010;329:89–93. stasis: from awareness of space to a sense of specification and commitment. Dev Biol 57. Topark-Ngarm A, et al. CTIP2 associates place. Nat Rev Immunol 2009;9:823–832. 2009;325:444–467. with the NuRD complex on the promoter of 27. Rothenberg EV, Moore JE, Yui MA. Launch- 43. Bernard OA, et al. A new recurrent and p57KIP2, a newly identified CTIP2 target ing the T-cell-lineage developmental pro- specific cryptic translocation, gene. J Biol Chem 2006;281:32272– gramme. Nat Rev Immunol 2008;8:9–21. t(5;14)(q35;q32), is associated with 32283. 28. Radtke F, et al. Deficient T cell fate specifica- expression of the Hox11L2 gene in T acute 58. Hameyer D, et al. Toxicity of ligand-depen- tion in mice with an induced inactivation of lymphoblastic leukemia. Leukemia dent Cre recombinases and generation of a Notch1. Immunity 1999;10:547–558. 2001;15:1495–1504. conditional Cre deleter mouse allowing 29. Maillard I, Fang T, Pear WS. Regulation of 44. Wakabayashi Y, et al. Bcl11b is required for mosaic recombination in peripheral tissues. lymphoid development, differentiation, and differentiation and survival of alphabeta T Physiol Genomics 2007;31:32–41.

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 147 Liu et al Æ Bcl11b has essential functions in T cells

59. Schmitt TM, Zuniga-Pflucker JC. Induction ven apoptosis. J Proteome Res 2010;9: 84. Senawong T, et al. Involvement of the his- of T cell development from hematopoietic 3799–3811. tone deacetylase SIRT1 in chicken ovalbu- progenitor cells by delta-like-1 in vitro. 72. Xie H, Ye M, Feng R, Graf T. Stepwise min upstream promoter transcription factor Immunity 2002;17:749–756. reprogramming of B cells into macrophages. (COUP-TF)-interacting protein 2-mediated 60. Walzer T, et al. Identification, activation, Cell 2004;117:663–676. transcriptional repression. J Biol Chem and selective in vivo ablation of mouse NK 73. Laiosa CV, Stadtfeld M, Xie H, de Andres- 2003;278:43041–43050. cells via NKp46. Proc Natl Acad Sci USA Aguayo L, Graf T. Reprogramming of com- 85. Senawong T, Peterson VJ, Leid M. BCL11A- 2007;104:3384–3389. mitted T cell progenitors to macrophages dependent recruitment of SIRT1 to a pro- 61. Miaw SC, Choi A, Yu E, Kishikawa H, Ho and dendritic cells by C ⁄ EBP alpha and PU.1 moter template in mammalian cells results IC. ROG, repressor of GATA, regulates the transcription factors. Immunity 2006; in histone deacetylation and transcriptional expression of cytokine genes. Immunity 25:731–744. repression. Arch Biochem Biophys 2000;12:323–333. 74. Davis RL, Weintraub H, Lassar AB. Expres- 2005;434:316–325. 62. Colucci F, Soudais C, Rosmaraki E, Vanes L, sion of a single transfected cDNA converts 86. Marban C, et al. Recruitment of chromatin- Tybulewicz VL, Di Santo JP. Dissecting NK fibroblasts to myoblasts. Cell 1987;51:987– modifying enzymes by CTIP2 promotes cell development using a novel alymphoid 1000. HIV-1 transcriptional silencing. EMBO J mouse model: investigating the role of the 75. Mikkola I, Heavey B, Horcher M, Busslinger 2007;26:412–423. c-abl proto-oncogene in murine NK cell M. Reversion of commitment upon 87. Cismasiu VB, et al. BCL11B participates in differentiation. J Immunol 1999;162: loss of Pax5 expression. Science 2002; the activation of IL2 gene expression in 2761–2765. 297:110–113. CD4+ T lymphocytes. Blood 2006;108: 63. Kastner P, et al. Bcl11b represses a 76. Cobaleda C, Jochum W, Busslinger M. 2695–2702. mature T-cell gene expression program Conversion of mature B cells into T cells by 88. Cismasiu VB, et al. BCL11B enhances in immature CD4(+)CD8(+) thymo- dedifferentiation to uncommitted progeni- TCR ⁄ CD28-triggered NF-kappaB activation cytes. Eur J Immunol 2010;40:2143– tors. Nature 2007;449:473–477. through up-regulation of Cot kinase gene 2154. 77. Avram D, Fields A, Pretty On Top K, expression in T-lymphocytes. Biochem J 64. Albu DI, et al. BCL11B is required for posi- Nevrivy DJ, Ishmael JE, Leid M. Isolation 2009;417:457–466. tive selection and survival of double-posi- of a novel family of C(2)H(2) zinc 89. Cherrier T, et al. p21(WAF1) gene promoter tive thymocytes. J Exp Med 2007;204: finger proteins implicated in transcrip- is epigenetically silenced by CTIP2 and 3003–3015. tional repression mediated by chicken SUV39H1. Oncogene 2009;28:3380–3389. 65. Zhang S, et al. Antigen-specific clonal ovalbumin upstream promoter transcrip- 90. Waldmann TA, Tagaya Y. The multifaceted expansion and cytolytic effector function of tion factor (COUP-TF) orphan nuclear regulation of interleukin-15 expression and CD8+ T lymphocytes depend on the tran- receptors. J Biol Chem 2000;275:10315– the role of this cytokine in NK cell differen- scription factor Bcl11b. J Exp Med 10322. tiation and host response to intracellular 2010;207:1687–1699. 78. Wang LH, Tsai SY, Cook RG, Beattie WG, pathogens. Annu Rev Immunol 1999;17: 66. MacLeod RA, Nagel S, Kaufmann M, Janssen Tsai MJ, O’Malley BW. COUP transcription 19–49. JW, Drexler HG. Activation of HOX11L2 by factor is a member of the steroid receptor 91. Yokota Y, et al. Development of peripheral juxtaposition with 3¢-BCL11B in an acute superfamily. Nature 1989;340:163–166. lymphoid organs and natural killer cells lymphoblastic leukemia cell line (HPB-ALL) 79. Shibata H, Nawaz Z, Tsai SY, O’Malley BW, depends on the helix-loop-helix inhibitor with t(5;14)(q35;q32.2). Genes Chromo- Tsai MJ. Gene silencing by chicken ovalbu- Id2. Nature 1999;397:702–706. somes Cancer 2003;37:84–91. min upstream promoter-transcription factor 92. Ikawa T, Fujimoto S, Kawamoto H, Katsura 67. Nagel S, Kaufmann M, Drexler HG, MacLe- I (COUP-TFI) is mediated by transcriptional Y, Yokota Y. Commitment to natural killer od RA. The cardiac gene NKX2- corepressors, -corepressor cells requires the helix-loop-helix inhibitor 5 is deregulated by juxtaposition with (N-CoR) and silencing mediator for retinoic Id2. Proc Natl Acad Sci USA 2001;98:5164– BCL11B in pediatric T-ALL cell lines via a acid receptor and 5169. novel t(5;14)(q35.1;q32.2). Cancer Res (SMRT). Mol Endocrinol 1997;11:714– 93. Blom B, et al. Disruption of alpha beta but 2003;63:5329–5334. 724. not of gamma delta T cell development by 68. Przybylski GK, et al. Disruption of 80. van der Wees J, et al. Developmental overexpression of the helix-loop-helix pro- the BCL11B gene through inv(14) expression and differential regulation by tein Id3 in committed T cell progenitors. (q11.2q32.31) results in the expression of retinoic acid of Xenopus COUP-TF-A and EMBO J 1999;18:2793–2802. BCL11B-TRDC fusion transcripts and is COUP-TF-B. Mech Dev 1996;54:173–184. 94. Fujimoto S, Ikawa T, Kina T, Yokota Y. associated with the absence of wild-type 81. Mlodzik M, Hiromi Y, Weber U, Goodman Forced expression of Id2 in fetal thymic T BCL11B transcripts in T-ALL. Leukemia CS, Rubin GM. The Drosophila seven-up cell progenitors allows some of their prog- 2005;19:201–208. gene, a member of the steroid receptor gene eny to adopt NK cell fate. Int Immunol 69. Kamimura K, Mishima Y, Obata M, Endo T, superfamily, controls photoreceptor cell 2007;19:1175–1182. Aoyagi Y, Kominami R. Lack of Bcl11b fates. Cell 1990;60:211–224. 95. Gascoyne DM, et al. The basic leucine zip- tumor suppressor results in vulnerability to 82. You LR, Lin FJ, Lee CT, DeMayo FJ, Tsai MJ, per transcription factor E4BP4 is essential DNA replication stress and damages. Onco- Tsai SY. Suppression of Notch signalling by for development. Nat gene 2007;26:5840–5850. the COUP-TFII transcription factor regulates Immunol 2009;10:1118–1124. 70. Grabarczyk P, et al. Inhibition of BCL11B vein identity. Nature 2005;435:98–104. 96. Kamizono S, et al. Nfil3 ⁄ E4bp4 is required expression leads to apoptosis of malignant 83. Cismasiu VB, Adamo K, Gecewicz J, Duque for the development and maturation of NK but not normal mature T cells. Oncogene J, Lin Q, Avram D. BCL11B functionally cells in vivo. J Exp Med 2009;206:2977– 2007;26:3797–3810. associates with the NuRD complex in T lym- 2986. 71. Karanam NK, et al. Proteome analysis phocytes to repress targeted promoter. 97. Colucci F, Samson SI, DeKoter RP, reveals new mechanisms of Bcl11b-loss dri- Oncogene 2005;24:6753–6764. Lantz O, Singh H, Di Santo JP. Differential

148 2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 Liu et al Æ Bcl11b has essential functions in T cells

requirement for the transcription factor PU.1 with HLA-A2-positive advanced melanoma. Proc Natl Acad Sci USA 2002;99:16168– in the generation of natural killer cells versus J Clin Oncol 2008;26:2292–2298. 16173. B and T cells. Blood 2001;97:2625–2632. 108. Rosenberg SA, et al. Observations on the 117. Burke S, Lakshmikanth T, Colucci F, Carbone 98. Barton K, et al. The Ets-1 transcription fac- systemic administration of autologous lym- E. New views on natural killer cell-based tor is required for the development of natu- phokine-activated killer cells and recombi- immunotherapy for melanoma treatment. ral killer cells in mice. Immunity nant interleukin-2 to patients with Trends Immunol 2010;119:1251–1263. 1998;9:555–563. metastatic cancer. N Engl J Med 118. Karre K. Natural killer cell recognition of 99. Guo Y, Maillard I, Chakraborti S, Rothen- 1985;313:1485–1492. missing self. Nat Immunol 2008;9:477– berg EV, Speck NA. Core binding factors are 109. Atkins MB, et al. High-dose recombinant 480. necessary for natural killer cell development interleukin 2 therapy for patients with met- 119. Ruggeri L, et al. Effectiveness of donor nat- and cooperate with Notch signaling during astatic melanoma: analysis of 270 patients ural killer cell alloreactivity in mismatched T-cell specification. Blood 2008;112:480– treated between 1985 and 1993. J Clin hematopoietic transplants. Science 492. Oncol 1999;17:2105–2116. 2002;295:2097–2100. 100. Chambers SM, et al. Hematopoietic finger- 110. Gattinoni L, et al. Removal of homeostatic 120. Carlens S, et al. A new method for in vitro prints: an expression database of stem cells cytokine sinks by lymphodepletion expansion of cytotoxic human CD3-CD56+ and their progeny. Cell Stem Cell 2007;1: enhances the efficacy of adoptively trans- natural killer cells. Hum Immunol 578–591. ferred tumor-specific CD8+ T cells. J Exp 2001;62:1092–1098. 101. Krejci A, Bernard F, Housden BE, Collins S, Med 2005;202:907–912. 121. Alici E, et al. Autologous antitumor activity Bray SJ. Direct response to Notch activation: 111. Rosenberg SA, Restifo NP, Yang JC, Morgan by NK cells expanded from myeloma signaling crosstalk and incoherent logic. RA, Dudley ME. Adoptive cell transfer: a patients using GMP-compliant components. Science Signal 2009;2:ra1. clinical path to effective cancer immuno- Blood 2008;111:3155–3162. 102. Nie L, et al. Regulation of lymphocyte therapy. Nat Rev Cancer 2008;8:299–308. 122. Robbins PF, et al. Cutting edge: persistence development by cell-type-specific interpre- 112. Jena B, Dotti G, Cooper LJ. Redirecting of transferred lymphocyte clonotypes corre- tation of notch signals. Mol Cell Biol T-cell specificity by introducing a tumor- lates with cancer regression in patients 2008;28:2078–2090. specific chimeric antigen receptor. Blood receiving cell transfer therapy. J Immunol 103. Rosenberg SA, Yang JC, Restifo NP. Cancer 2010;116:1035–1044. 2004;173:7125–7130. immunotherapy: moving beyond current 113. Chang CC, Ferrone S. Immune selective 123. Miller JS, et al. Successful adoptive transfer vaccines. Nat Med 2004;10:909–915. pressure and HLA class I antigen defects in and in vivo expansion of human haploidenti- 104. Zitvogel L, Tesniere A, Kroemer G. Cancer malignant lesions. Cancer Immunol cal NK cells in patients with cancer. Blood despite immunosurveillance: immunoselec- Immunother 2007;56:227–236. 2005;105:3051–30587. tion and immunosubversion. Nat Rev 114. Dunn GP, Old LJ, Schreiber RD. The three Es 124. Ghiringhelli F, et al. CD4+CD25+ Immunol 2006;6:715–727. of cancer immunoediting. Annu Rev Immu- regulatory T cells inhibit natural killer cell 105. Ueno H, et al. Harnessing human dendritic nol 2004;22:329–360. functions in a transforming growth factor- cell subsets for medicine. Immunol Rev 115. Smyth MJ, Dunn GP, Schreiber RD. Cancer beta-dependent manner. J Exp Med 2005; 2010;234:199–212. immunosurveillance and immunoediting: 202:1075–1085. 106. Schreiber S, et al. Immunotherapy of meta- the roles of immunity in suppressing tumor 125. Zhang H, et al. Lymphopenia and interleu- static malignant melanoma by a vaccine development and shaping tumor immuno- kin-2 therapy alter homeostasis of consisting of autologous interleukin 2- genicity. Adv Immunol 2006;90:1–50. CD4+CD25+ regulatory T cells. Nat Med transfected cancer cells: outcome of a phase 116. Yee C, et al. Adoptive T cell therapy using 2005;11:1238–1243. I study. Hum Gene Ther 1999;10:983–993. antigen-specific CD8+ T cell clones for the 126. Ahmadzadeh M, Rosenberg SA. IL-2 admin- 107. Sosman JA, et al. Three phase II cytokine treatment of patients with metastatic istration increases CD4+ CD25(hi) Foxp3+ working group trials of gp100 (210M) pep- melanoma: in vivo persistence, migration, regulatory T cells in cancer patients. Blood tide plus high-dose interleukin-2 in patients and antitumor effect of transferred T cells. 2006;107:2409–2414.

2010 John Wiley & Sons A/S • Immunological Reviews 238/2010 149