Overexpression of HES-1 is not sufficient to impose T-cell differentiation on human hematopoietic stem cells

Studies with Notch-1-deficient and transgenic mice have shown that signaling through the Notch-1 transmembrane receptor is necessary and sufficient for T-cell commitment.^1-3  In humans, we showed that retroviral overexpression of the intracellular domain of Notch-1 (ICN) in cord-blood CD34⁺ cells results in inhibition of B-cell development both in vitro and in vivo. Instead, cells are forced to differentiate along the T/NK pathway, resulting in ectopic development of T cells in the bone marrow (BM) of mice injected with ICN-transduced human hematopoietic stem cells (HSCs).⁴ 

A technique to direct HSCs toward the T-cell lineage could be of great therapeutic value to develop strategies to enhance T-cell development from transplanted donor HSCs after myeloablative therapy. Manipulation of HSCs with ICN is not an option however, since constitutive Notch-1 expression eventually leads to the development of T-cell neoplasms,⁵  and mutations in the NOTCH1 gene have been involved in most cases of human T-cell acute lymphoblastic leukemia (T-ALL).⁶  Understanding the cellular events leading to T-cell commitment after ICN overexpression might lead to the identification of an effector molecule that does not cause tumor development when overexpressed in HSCs.

An important target gene of Notch-1 signaling is the basic helix-loop-helix transcription factor HES1,⁷  which is expressed in thymocytes and thymic stroma and is essential for normal T-cell development.^8,9  HES-1 is up-regulated in hematopoietic precursors after ICN overexpression¹⁰  and after activation of endogenous Notch-1 signaling by binding with the Delta-like-1 ligand.^11,12  Here, we investigated whether HES-1 is the mediator of the effects on human hematopoietic differentiation seen previously with ICN overexpression.

Human HES-1 cDNA was cloned in the retroviral vector pLZRS-IRES-EGFP (LIE)¹³  upstream of the internal ribosomal entry site (IRES). ICN-LZRS was prepared by subcloning ICN from the ICN-MSCV construct used before.⁴  Infectious retrovirus was produced by transfection of Phoenix-A cells.¹⁴  Expression of HES-1 protein was verified by Western blotting on nuclear extracts of HES-1-transduced HEK-T cells using a polyclonal rabbit anti-HES-1 antibody.

CD34⁺ cells were isolated from cord-blood mononuclear cells by positive selection with EasySep magnetic beads (StemCell Technologies, Vancouver, BC, Canada), and CD34⁺ Lin^- cells were sorted after staining with anti-human CD34-APC, CD3-FITC, CD14-FITC, CD19-FITC, or CD19-PE monoclonal antibodies (Becton and Dickinson Immunocytometry Systems, San Jose, CA) and CD56-FITC (Serotec, Oxford, United Kingdom). The purity was always more than 98%. Retroviral transduction and coculture on MS-5 stromal cells were performed as described.⁴  Labeling of cells with the indicated antibodies and flow cytometric analysis were performed as described.⁴ 

Absolute cell numbers were calculated from the total viable cell number as counted with the hemocytometer, the frequency of enhanced green fluorescence protein⁺ (EGFP⁺) cells and the frequency of the indicated subpopulations as measured by flow cytometry. Because transduction efficiencies were different for LIE, HES-1, and ICN, these numbers were multiplied by a factor to normalize to 1 × 10³ EGFP⁺ cells at the start of all cultures.

The paired Student t test was used with a significance level of .05.

Cord-blood CD34⁺ cells were transduced with comparable efficiencies by the HES-1 and ICN retrovirus (17.0% ± 5.9% and 19.6% ± 7.5% after 1 day, respectively). The transduction efficiency for the control virus LIE was higher, that is, 47.2 % ± 12.4%. As anticipated, B-cell development in MS-5 cocultures was dramatically blocked by ICN overexpression (Figure 1A). In the cultures initiated with HES-1-transduced cells, B-cell differentiation was not completely suppressed but yet significantly reduced (Figure 1A). This reduction was correlated with the level of HES-1 expression and was not caused by a block at an immature B-cell stage (data not shown). These results indicate that the up-regulation of HES-1 in response to Notch-1 signaling might aid to the inhibition of B-cell development but is not the main mediator of this effect.

Whereas overexpression of ICN drives T-cell development, as shown by the massive generation of CD7⁺cyCD3⁺ T/NK progenitors under B-cell conditions in MS-5 cocultures, HES-1 overexpression did not mimic this effect (Figure 1A). Furthermore, HES-1, unlike ICN, did not promote NK-cell development in MS-5 cocultures with specific cytokines for NK-cell development (Figure 1B). Therefore, the Notch-1 signal instructing the CD34⁺ cell to commit to the T/NK lineage is probably not or not solely mediated by HES-1.

Our results are in line with those of Kawamata et al,¹⁵  who also noticed a partial block in B-cell development but no extrathymic T-cell development in the BM of mice transplanted with HES-1- or HES-5-transduced mouse BM Lin^- precursors. Also, Maillard et al¹⁶  communicated that overexpression of HES-1 in Notch-1-deficient murine precursors cannot rescue T-cell development in fetal thymus organ culture (FTOC). Combined with our data, it is clear that overexpression of HES-1 is not sufficient to impose T-cell differentiation on either murine or human HSCs.

Culturing CD34⁺ cells on MS-5 in the presence of SCF, FL3, TPO, IL-2, IL-7, and IL-15 (Mix 6) allows their expansion and differentiation toward monocytes. We showed earlier that ICN expression blocks both HSC expansion and differentiation toward monocytes.⁴  This was confirmed in the current experiments (Figure 2). HES-1-transduced CD34⁺ cells, on the other hand, generated comparable frequencies of monocytes as the control (Figure 2A). However, a significantly higher percentage of HES-1-transduced CD34⁺ cells remained CD34⁺, and absolute numbers of CD34⁺ cells were not reduced compared to the control (Figure 2B). The frequency of CD34⁺ cells also correlated with the level of HES-1 expression (data not shown). Although recent studies indicate that HES-1 maintains stem cells,^17,18  phenotypic analysis of the CD34⁺ population from our MS-5 cultures did not point to a specific increase of cells with a stem-cell-like phenotype (ie, CD34^brightCD133⁺) (data not shown).

Our experiments indicate that HES-1 is not the sole mediator of T-cell commitment induced by ICN overexpression. Based on the data from Kawamata et al, HES-5 also is not the downstream effector molecule of ICN.¹⁵  Deltex-1, which is also up-regulated after ICN overexpression (Deftos et al¹⁰  and own observation) cannot be the downstream mediator either, since forced expression of Deltex-1 in hematopoietic progenitors results in B-cell development at the expense of T-cell development.¹⁹  More target genes of Notch signaling are being discovered, however. HERP1 and HERP2, which are structurally very similar to HES, are both up-regulated by Notch signaling. They can form heterodimers with HES proteins that bind with greater affinity to the promoter of downstream genes and have a stronger repression activity than the respective homodimers.²⁰  This might explain why overexpression of HES-1 alone has little effect. Possibly one or more other transcription factors that are up-regulated by ICN overexpression must be overexpressed together with HES-1. HERP1 (Hey1) and HERP2 (Hey2) knock-out mice have been established recently,²¹  so it will be interesting to find out whether T-cell development is influenced in these mice. Another possible scenario is that ICN drives cells to the T/NK pathway by up-regulating T/NK-committed genes. For example, the pre-TCR- α chain (PTCRA) gene has been shown to be a direct target gene of Notch-1 in T cells.^10,22  Furthermore, several studies point to the existence of CBF1, suppressor of hairless Lag-1 (CSL)-independent Notch signaling pathways.²³  Although the CSL pathway is active in ICN-transduced cord blood CD34⁺ cells, as shown by the up-regulation of HES-1, it is possible that the effects of ICN on differentiation are mediated by CSL-independent mechanisms.

We thank the staff of the Bloedtransfusiecentrum Oost-Vlaanderen for the supply of umbilical cord blood samples and Marie-José De Bosscher for lymphoprepping and freezing them. We are grateful to Dr T. Sudo (Toray Industries Inc, Kamakura, Japan) for his kind gift of the anti-HES-1 antibody, to Nancy De Cabooter and Els Demecheleer for DNA sequencing, to Christian De Boever for help with artwork, and to Caroline Collier for animal care.