Introduction

Allogeneic hematopoietic stem cell transplantation (HSCT) is a key treatment for a large number of acquired and inherited diseases of the hematopoietic system. Initially restricted to patients with HLA-identical siblings due to risks of T-cell mediated graft rejection and graft versus-host disease (GvHD), it is now widely used also for patients without HLA-identical donors thanks to the use of antithymocyte globulin (ATG), changes in graft handling (T-cell depletion, CD34+ hematopopoietic stem, and progenitor cell (HSPCs) selection) and, recently, the use of post-transplant cyclophosphamide (PTC) [1,2,3]. Although these modifications have greatly improved HSCT outcomes, T-cell immunodeficiency following transplantation remains a major obstacle in an HLA-mismatched setting and hampers the more widespread use of this approach.

Whereas innate immunity recovers quickly, reconstitution of adoptive immunity is a long process. In the case of T-cell replete haplo-identical HSCT, donor T cells present in the graft survive and expand, but this expansion concerns mainly a CD8+ memory T-cell population and leads to a contraction and skewing of the TCR repertoire, thus producing a partially ineffective immune response [4]. In CD34+-selected haplo-identical transplantation, this first wave of homeostatically expanded T cells is virtually absent. Consequently, in both settings, reconstitution of a functional, fully diverse and naive peripheral T-cell pool relies on de novo production of naive T cells from grafted CD34+ HSPCs [4,5,6]. Even in young patients (i.e., most of those transplanted for severe combined immunodeficiency (SCID)), donor-derived T-cells only appear in the blood after 3–6 months” [7, 8]. A period of 6–12 months is required to achieve CD4 + cell counts that provide protective immunity. More complete restoration of the overall T-cell compartment (i.e. naive T-cells exhibiting a polyclonal T-cell receptor (TCR) repertoire) is an even longer process and may require up to 2 years in adults [9,10,11].

In primary immunodeficiency patients transplanted in an HLA-incompatible setting the profound immunodeficiency following such a procedure leads to severe, opportunistic viral, bacterial, and fungal infections accounting for ~40% of the mortality. The importance of a functional T-cell compartment early after HSCT is further highlighted by the fact, that T-cell-depleted HSCT is associated with an increased risk of graft rejection and, in patients with malignant hematologic disorders, relapse [12].

Gene therapy constitutes a valid alternative treatment for SCID-X1 patients without HLA-identical donor [13]. Indeed, the risk of vector-related leukemia has been greatly reduced by the introduction of new generations of safer vectors. Furthermore this treatment has been restricted to patients without HLA-identical donors (either matched related or matched unrelated). A restrospective analysis of haplo-identical transplantion versus gene therapy treatment for immune reconstitution of SCID-X1 patients showed that T-cell reconstitution was significantly faster in the gene therapy group (14 patients) as compared to the haplo-identical group (13 patients) [8]. However, this overall positive output veils some disparities (Fig. 1a). T-cell reconstitution was very slow and incomplete in some patients from both groups. All of them shared one common feature: the presence of active infection at time of treatment (disseminated BCGitis, adenovirus infection…) (Fig. 1b).

Fig. 1
figure 1

Immune reconstitution in SCID-X1 patients after gene therapy and In vitro generation of human T-lymphoid progenitors (HTLPs). a Reconstitution of naïve CD4 T cells over time in patients from trial #1 (upper graph) and trial #2 (bottom graph). b Correlation of CD3 reconstitution over time in the absence (upper graph) or presence of infection (bottom graph). Data include patients of both trials (indicated by SCID#1 and SCID#2). c Descriptive scheme of human T-cell development. d Experimental protocol for the 7-day generation of CD7+ T-cell precursors (=HTLPs) from CD34+ hematopoietic stem and precursor cells (HSPCs) by co- culture with immobilized notch ligand DL4 and a cocktail of cytokines

Full size image

Many well-established pre-transplant parameters influence immune reconstitution after HSCT: diagnosis, age of the recipient, degree of HLA mismatches, conditioning regimen, type of graft (origin and manipulation), but also the infectious status of donor and recipient. Immune reconstitution is further impacted by clinical complications including leukemia relapse, infections and GvHD, or treatments [14,15,16,17]. However, in the small comparative study published by our group [8] as well as in the study conducted by Clave and colleagues [12], GvHD did not significantly impact the immunological outcome in young patients.

Beside our own observations (Fig. 1a, b), the deleterious impact of bacterial and viral infections on immune reconstitution has been observed by several groups in murine models [18,19,20] as well as by clinical observations [21]. Some of these studies show a significant association between infections (viral and bacterial) and poor thymic output quantified by T-cell receptor excision circles (TREC) in the blood [4, 22].

As mentioned above, delayed T-cell reconstitution represents a major obstacle to the widespread use of HLA-mismatched HSCT. In order to shorten post-transplant immunodeficiency and thus improve clinical outcome, we chose to implement a new treatment strategy consisting of co- transplantion of human T-lymphoid progenitors (HTLPs) together with non-manipulated HSPCs. These progenitors are able to seed the thymus and generate a wave of mature and polyclonal T-cells significantly faster than it is usually observed [11, 13]. Herein we are going to describe this approach as well as its potential applications for both HLA-mismatched HSCT and gene therapy.

Ex vivo generation of human T-lymphoid progenitors

T-cell generation proceeds through three main stages: the first one consists in the production of lymphoid progenitors able to leave the bone marrow (where they are generated from HSCs) and seed the thymus; the second one is T-cell commitment accompanied by the loss of other lineage potentials and first T-cell rearrangements at the δ, γ, and β loci leading to the production of either γδ TCR+ T cells (<5% of circulating mature T cells) or immature single positive cells that (stage 3) further rearrange TCR α locus and undergo positive and negative selection to ultimately give rise to functional, non autoreactive αβ TCR+ T cells (95% of circulating T cells) (Fig. 1c). Of note, the first step of thymic T-cell differentiation takes half of the time required for the whole thymopoietic process. Finding a way to produce quickly ex vivo large numbers of HTLPs appeared thus as a promising approach to fasten T-cell reconstitution after HSCT.

The early stage of T-cell development is dependant on Notch signaling in mouse and human, in particular Delta-like ligand 4 (DL-4) [23,24,25], as well as on key cytokines implicated in the survival and proliferation of thymocytes [26,27,28]. Based on these findings, we implemented a feeder cell-free culture system relying on a recombinant fusion protein composed of the extracellular domain of human DL-4 and the Fc part of human IgG2 (Fig. 1d). The use of this modification allowed immobilisation of DL-4 on the culture surface and thus eliminated the necessity of using a stromal cell line similar to a previous study implementing DL-1 for the expansion of CB HSCPs [29].

Our DL4-Fc culture system allowed the generation large numbers of CD34+/-CD7+CD5-icCD3+ HTLPs from cord blood CD34+ HSPCs within 7 days (an average of 2.5 HTLPs per CD34+ HSPC,) [30]. HTLPs displayed the gene expression profile of early thymic precursors as demonstrated by significant levels of transcripts of PTA, IL7RA, RAG1, and BCL11B. T-cell differentiation in limiting dilution conditions on OP9/DL-1 cells revealed a significant increase in in vitro T-cell potential (from 1/350 at day 0 to 1/12 at day 7). Once transplanted into both, irradiated adult or non-irradiated neonate NOD/SCID/γc(-/-) (NSG) recipients, these T-cell precursors seeded the thymus and generated mature, polyclonal, and functional T cells [30]. Co-transplantation of HTLPs and untreated CD34+ HSCPs was used to mimick future clinical applications. In this setting reconstitution of the T-cell compartment from cultured HTLPs was robust and rapid, with other hematopoietic lineages being produced from the non-manipulated HSCPs.

Mobilized peripheral blood (mPB) is currently the main source of HSPCs in allogenic HSCT as such adult HSPCs are available in large quantity and exhibit several advantages over cord blood grafts in the clinical setting. However, the lymphoid potential of adult HSCPs, especially T-lymphoid potential, is diminished as compared to the one of CB [31]. We recently demonstrated the capacity to generate HTLPs from adult mPB CD34+ cells in DL-4 culture conditions in a very similar way as for cord blood cells [32]. Within 7 days, adult HSPCs produced CD7+ HTLPs expressing T lineage master genes (TCF7, IL7Ra and BCL11B, GATA3, and CD3E) and exhibited an in vitro T-cell differentiation potential similar to the one of CB-derived HTLPs (around 1/19). HTLPs derived from adult HSPCs expressed chemokine receptors implicated in thymus homing (CCR7, CCR9, and CXCR4) and efficiently produced polyclonal T cells upon transplantation in NOD/SCID/γc−/− (NSG) mice. However, due to intrinsic differences in terms of T-cell potential, survival, and proliferation, the yields of HTLPs recovered after 7 days were lower (around 0.45 HTLP/CD34+ HSPC). The lower efficiency of adult HSPCs to produce HTLPs in DL-4 culture conditions was accompanied by a higher rate of apoptosis and a lower rate of proliferation. To overcome this hurdle, we tested several molecules aiming at either increasing Notch signaling or enhancing T-cell differentiation and HTLPs survival and proliferation. The addition of a human fragment of fibronectin and Tumor Necrosis Factor alpha (TNFα) improved the yields of production of adult HTLPs by 10 fold (patents WO2016/055396, WO2018/146297). Our improved results demonstrate that, like CB HSPCs, adult HSPCs provide an effective and available source of in vitro cultured HTLPs in the context of future clinical applications directed to shorten T-cell recovery after HSCT.

Cell therapy based on HTLPs

Based on these preclinical data, the entire platform, including all reagents, DL-4, medium, serum, cytokines, was translated into a GMP grade process. Of note, results obtained in GMP conditions in the clinical runs were identical to the ones obtained following the research grade protocol. Subsequently, we carefully monitored the composition of the day-7-cellular product, with a special focus on cellular contaminants. Apart from HTLPs, DL-4 cultures contain no or very rare T-cells (undetectable upon analysis of TCR rearrangements by robust and highly sensitive techniques used to detect minimal residual disease in hematologic malignancies, coll. with Hematology Lab directed by E. McIntyre, Necker Hospital) and mostly myeloid precursors (around 12%).

As a third step, we tested the toxicity of the DL-4 protein and the day 7-cellular product of DL-4 cultures by injection of a 10 fold higher dose than used in any future clinical trial into the mice.

Of note, quantification of residual DL-4 at the end of the culture performed by a home-made ELISA using antibodies specific for either the DL-4 extracellular domain or the Fc part of DL-4 yielded negative results in all experiments but one the dose of which was defined as reference for injection into mice. We did not observe any toxicity or tumorigenicity upon injection of the DL-4 protein and the cellular product in mice. Furthermore, we also analyzed the caryotype of the T-cell precursors and performed CGH array, both of which were normal (data not shown).

Based on these results, we are currently implementing two phase 1/2 clinical trials using human in vitro generated HTLPs, with the financial support of the French Ministry of Health and the sponsorship of Assistance-Publique Hopitaux de Paris. The first clinical trial will include 12 SCID patients undergoing haplo-identical HSCT. In this trial, patients will receive a first transplant of CD34+-selected cells, followed 8 days later by the injection of HTLPs generated from the residual CD34+ fraction at doses (0.1 up to 1.5 × 10e6 CD7 + cells/kg, one single dose of HTLPs per patient). The protocol has been approved by the French Drug Agency and is open to inclusion. The primary objective will be to assess the procedure’s safety: dose-limiting toxicity (DLT), including grade III-IV graft versus host disease (GVHD) and grade III or higher CTCAE adverse events (AEs). Secondary objectives will include: graft failure, the presence of naïve CD4+ T cells at 6 months, kinetics of T-cell reconstitution, incidence of infections, relapse rate, and overall survival.

The second trial will include 10 adult patients treated by double cord blood HSCT for leukemia. One cord blood will be injected without manipulation. The other one will be exposed 7 days to DL-4 culture conditions before infusion into the patient.

Both trials will allow not only evaluation of the procedure’s safety, but also the ability of adult versus CB-derived T-cell precursors to accelerate immune reconstitution in a different clinical context.

The first clinical trial for SCID patients will start in Q2 2019, followed by the second one in Q1 2020.

Gene therapy based on HTLPs

Our DL-4 culture system provides a unique plateform not only for cellular but also for gene therapy applications. We also explored the possibility to genetically modify HTLPs by combining gene transduction by SIN-retroviral (SIN-RV) and lentiviral (LV) vectors with DL-4 based culture. Several conditions were tested, transduction followed by DL-4 culture, DL-4 culture followed by transduction and a combination of both using either a VSVG SIN-retroviral vector and a BaEv –pseudotyped LV expressing either IL2RG gene or mCherry. The presence of gene-corrected HTLPs was evaluated at day 4 and day 7 of culture by the frequency and number of mCherry+ HTLPs at day 4 and day 7 and the quantification of T-cell potential in limiting dilution assay as described elsewhere [27]. Healthy donor HSPCs and SCID-X1 patients’ HSPCs were used, the results were compared to the ones obtained in classical gene transduction protocols. As a reminder, SCID-X1 patients exhibit a severe T-cell defect due to an early-stage blockade of T-cell differentiation caused by a loss-of-function mutation in the γC gene. SCID-X1 γC deficiency can be safely and efficiently corrected by the ex vivo transduction of HSPCs, as demonstrated in patients included in gene therapy protocols [33]. For both types of vectors (SIN-RV and BaEv-LV) and samples (healthy or SCID-X1 HSPCs), the combination of transduction and DL-4 culture conditions led to the appearance of a population of gene-corrected HTLPs from day 4 on (>30% [34]). Furthermore, we demonstrated that upon co-culture on OP9/DL1 cells, CD4+CD8+ double positive cells and TCR+ T cells appeared earlier and in higher numbers when gene transduction was combined to the DL-4 plateform. Our data indicated the correction of γC deficiency in BM SCID-X1 CD34+ HSPCs under both conventional and DL-4 culture conditions. However, T-cell differentiation was faster and more efficient under the DL-4 condition.

The possibility to obtain gene-corrected HTLPs able to accelerate T-cell reconstitution following gene therapy in infected SCID patients, but also for other clinical indications, needs to be further evaluated. We therefor plan to confirm both our in vitro and in vivo results in order to develop a GMP compatible protocol combining DL-4 culture with gene therapy.

A gene therapy protocol for infected SCID X1 patients including both, the infection of gene-corrected HSPCs and gene-corrected HTLPs, remains the ultimate goal. This proof of concept will be of key importance for the extension of the DL-4 protocol to other clinical applications.

Methods

Search strategy and selection criteria The first three references correspond to three major breakthroughs in HLA-partially incompatible HSCT. We searched PubMed between 1 Jan 2007, and 31 August 2018, with the terms « allogeneic hematopoietic stem cell transplantation », « haplo-identical HSCT » in combination with « thymus », « thymopoiesis », « immune reconstitution », « viral complications », « T-cell receptor excision circles », « SCID », and « lymphoid progenitors », « T-cell development » in combination with « Notch », « Delta-like-4 ligand » and « Delta-like-1 ligand ». We restricted our search to English publications. We selected reports from the past 5 years but did not exclude important and highly cited older publications. We searched the reference lists of articles identified by this search strategy and selected the 33 that we judged relevant. Review articles are also cited to provide more detail.

Conclusions

We have demonstrated the possibility to generate large amounts of HTLPs exhibiting the phenotypic and molecular signatures of their thymic counterparts from any source of CD34+ HSPCs.

Ex vivo generated HTLPs were able to colonize the thymus of NSG mice and produce polyclonal mature T cells in vitro and in vivo. The whole plateform has been moved to Necker GMP facility to implement two clinical trials including SCID and leukemic patients undergoing allogeneic transplantation in different settings. If we successfully shorten the severe immunodeficiency period following transplantation, the benefit for the patients and public health will be immense:

First of all, the reduction of incidence and severity of opportunistic infections (e.g., adenovirus, cytomegalovirus, enterovirus, and fungi infections) would significantly shorten the length of hospitalization for each patient. As a consequence the use of partially incompatible HSC donors could be extended to all patients requiring this procedure (even in transplantation units that are not familiar with this high-risk procedure), with a significant decrease in mortality and morbidity.

Apart from pure cell-therapy approaches, the combination of DL-4 culture with gene transduction protocols will open new perspectives in terms of gene therapy applications.